Compare commits

..

48 Commits

Author SHA1 Message Date
40827d5371 T14.2 (D27): the offline sign-in window follows the server
/api/auth/me returns session_expires_at (the session token exp, UTC). The browser
keeps it and never extends it: the fixed 12 hours and the sliding reset are gone.
Offline before that time the app opens as normal. Offline after it the app opens in
read-and-queue mode: no redirect, one mode line, changes are queued and held until
a page load gets a good answer from the server. wpCacheAuth refreshes the cached
user without moving the end time.

D26 and D27 recorded as Accepted. tests/session_expiry_check.py (22 checks) fails
against the old 12 hour rule.
2026-10-17 14:18:53 -07:00
73f0b5e69d T14.1 (D26): a sign-in problem never destroys a queued change
A 401 holds the queue: nothing is marked dead, no try is counted, and only
wp-auth-ready releases it. A 403 marks only that change dead, with a reason.
qWrite reports failure; a change the device cannot store stays in memory and the
badge says so. Each change has an owner and is sent only under that person.

tests/outbox_check.py (26 checks) fails against the old 401 rule.
2026-10-17 14:00:52 -07:00
512f0b6b1d Wave 14: do not widen B5; badge states belong to D26 and BL-061 2026-10-17 13:27:37 -07:00
ae649b5ed7 Wave 14 spec (D26, D27, B5, BL-061) and the T13.1 gate; decisions 2026-10-17 2026-10-17 13:27:28 -07:00
4fc6e8a01b T13.11: session limit defaults to 8 hours (D25) 2026-10-17 13:25:58 -07:00
2c199f6f3f T13.6: confirm the auto-disable audit row in the Admin Console (D18) 2026-10-17 13:15:09 -07:00
c3519d3e53 T13.6: record the live Okta sync run and the final verification (D18) 2026-10-17 12:39:02 -07:00
1a3468effd T13.10: DPoP for the Okta Management API calls, auto-detected (D24) 2026-10-16 17:10:19 -07:00
83e77ec622 T13.9 follow-up 3: read pretty-printed and hand-copied JWKs, report the line and column of a syntax error (D23) 2026-10-16 16:56:11 -07:00
c75450a000 T13.9 follow-up 2 fix: keep quote marks that belong to brace-less JWK fields (D23) 2026-10-16 16:24:21 -07:00
5f8ebab5b2 T13.9 follow-up 2: name PEM keys, forgive missing braces and double encoding, give a length hint (D23) 2026-10-16 16:24:05 -07:00
5cf6355a01 T13.9 follow-up: forgive paste damage in OKTA_SYNC_PRIVATE_KEY and name the failing step (D23) 2026-10-16 16:04:52 -07:00
d137d690f9 T13.9: sign in to the Okta Management API as an OAuth service app, private_key_jwt (D23) 2026-10-16 15:29:32 -07:00
ad1b8273c0 T13.8: run the stack in Docker on a PC, local compose override and steps (D22); log BL-070 2026-10-16 15:08:01 -07:00
bcef0e8145 T13.7: dry run for the Okta sync, OKTA_SYNC_DRY_RUN (D22) 2026-10-16 14:53:05 -07:00
2d4e96d668 T13.6: record full-suite result on the release tree (D18) 2026-10-16 14:17:39 -07:00
493ead725d Release: wave 11 + T13.2-T13.5 (D18 Okta sync), T13.1 held back until T14.1; mobile roadmap docs (BL-062..BL-069) 2026-10-16 14:03:33 -07:00
2abd55c745 T13.4 follow-up: make the sync's log lines visible under gunicorn (D18)
Nothing configures Python logging and gunicorn only configures its own loggers,
so INFO lines from wpsuite.* were dropped: with the production launch command a
healthy start printed no Okta lines, and an operator could not tell the sync was
off from on-and-quiet. okta_sync's logger now has its own stderr handler, only
when the root logger has none. Verified with real gunicorn, 2 workers, with and
without a token. Other modules' output is unchanged.
2026-10-16 13:30:47 -07:00
bbfa403117 D18 docs: admin instructions for accounts the Okta sync disabled (T13.6)
New DEPLOYMENT.md section: how to see what the sync did, how to re-enable a
person (fix Okta first, then Enable in the User Directory), who may do it
(admin; super user only for exclusive-scope accounts), the host CLI when no
admin can sign in, and how to turn the sync off if it disabled the wrong people.
2026-10-16 12:05:03 -07:00
253886923c T13.6 (part): test T13.1's sliding session, and record what it found
T13.1 shipped with its done-when list unchecked and no test of its own. Adds
tests/session_idle_check.py (28 checks): continuous activity survives 30 minutes
but not the ceiling, an idle session gets 401, the cookie is not rewritten per
request, a pre-D18 token still gets a ceiling via iat, identity carries forward,
both variables are documented. Mutation-tested (no throttle, no ceiling, a fresh
ceiling for legacy tokens each fail it).

Notes in wave-13.md: /api/auth/* is exempt from the refresh, and the wave exit
criteria say AUTH_SESSION_HOURS 2 while T13.1 and the code say 8.
2026-10-16 11:41:29 -07:00
763ab8d98d T13.5: verify an auto-disable reads clearly in the admin console (D18)
No UI change. The existing Activity log card already renders a sync disable
legibly: system:okta_sync / user deprovisioned by sync / user / <name> /
okta_status: SUSPENDED|MISSING. tests/okta_sync_ui_check.py runs the real sync
on a throwaway database and reads the real page at 1440px and 390px (28/28):
rows, search, filter, no sideways scroll, no JS errors, and the disabled
account refused on its next request. Does not depend on CR-019's usage card.
2026-10-16 11:36:38 -07:00
f30989d4b8 T13.4: run the Okta sync on a schedule (D18)
A daemon timer thread starts from a new lifespan hook in app.py and calls the
T13.3 sync every OKTA_SYNC_INTERVAL_SECONDS (default 900, minimum 300). With no
OKTA_API_TOKEN it starts nothing and the startup log says so.

The api runs 2 gunicorn workers, so each starts its own timer. A single atomic
conditional UPDATE on one app_settings row (key okta_sync) is the gate: exactly
one worker wins each cycle, on SQLite and Postgres, with no lock held. The row
survives a restart, so a restart inside the interval starts no second run. A
fetch is now capped at 120 s in total, so a run is always shorter than the
minimum interval and cannot overlap the next. No migration needed.

docker-compose.yml passes OKTA_API_TOKEN, OKTA_ORG_URL,
OKTA_SYNC_PROFILE_ATTRIBUTE and OKTA_SYNC_INTERVAL_SECONDS to the api service.

tests/okta_sync_schedule_check.py: 38 checks including an 8-worker race on a
file database (exactly one winner, 16 rounds) and fail-closed through the
scheduled path. okta_auth_check.py still 22/22.
2026-10-16 11:34:13 -07:00
103e4d40ee T13.3: Okta deprovisioning sync job, fail closed (D18)
run_sync() compares the Okta user list (T13.2) to active local accounts and
disables the ones Okta no longer vouches for. One AuditLog row per change,
actor system:okta_sync, action user_deprovisioned_by_sync.

Fail closed: any exception from the fetch, an empty or untrusted answer, or a
failed save changes nothing and writes nothing. The fetch runs before the
database is read, so a failed fetch has no path to a change. It only ever
disables; an already-disabled account is untouched and gets no new row.

Decided with Matt 2026-10-16: disable on missing, SUSPENDED or DEPROVISIONED.
PASSWORD_EXPIRED, LOCKED_OUT and RECOVERY do not disable (the job never
re-enables, so a literal "not ACTIVE" would lock people out for good).
Added a guard: no match between any active local account and any Okta user is
refused as a wrong identity attribute, not read as a mass offboarding.

tests/okta_sync_job_check.py: 71 checks; removing the fail-closed branch and the
guard fails 28 of them. BL-069 logs the missing per-run cap.
2026-10-16 11:31:39 -07:00
2f7de6e33b T13.2: Okta Management API client (D18)
New server/okta_sync.py: lists every Okta user and status with its own
read-only credential (OKTA_API_TOKEN, SSWS), separate from the OIDC client.
Follows pagination. Every failure raises OktaSyncError; an error is never an
empty list, because T13.3 reads an empty list as "nobody is active".

Also refuses: a next-page link to another host (token never leaves the org),
a redirect, a chain over 50 pages (a cut-off list looks like missing users),
a bad item among good ones (never skipped), and an empty list.

tests/okta_sync_check.py: 52 checks, self-contained (httpx MockTransport).
Not wired into app.py yet. T13.3 uses it; T13.4 starts it on a schedule.
2026-10-16 11:28:57 -07:00
d42ae0c65b Merge pull request 'feat/wave-12-bulk-users' (#4) from feat/wave-12-bulk-users into main
Reviewed-on: #4
2026-10-06 22:41:38 +00:00
9fc2d307bb Merge pull request 'feat/wave-16-on-main' (#3) from feat/wave-16-on-main into main
Reviewed-on: #3
2026-10-06 22:40:22 +00:00
e50a05a1e0 T12.6: wave 12 closed - browser_check passes after the re-point
Matt reran tests/browser_check.py on the re-pointed probe: all
passed. Every T12 done-when and every wave 12 exit criterion is now
checked in wave-12.md.
2026-10-06 14:05:43 -07:00
912cf75027 T12.6: browser results, screenshots, browser_check re-pointed
Matt's run (Edge/Chrome, one at a time): bulk_users_ui_check 19/19,
console_dialogs_check 16/16, mobile_check 24/24, a11y_check 22/22.

browser_check 67/71: its four failures asserted column POSITIONS
(9 columns; r.cells[3]/[4] for Permissions/Project role). CR-020's
selection column comes first, so those read Username and Email. The
probe now expects 10 columns with the selection column first, finds
the cells by header name, and checks your own row has no selection
box. Verified on the real page in the DOM harness; rerun to close.

After-screenshots committed in docs/reference/baseline/wave-12/.
BL-061 logged: the pre-existing sync badge covers a control for ~2 s
after load (pointer-events:none, so never blocks a click); belongs
with wave 14's badge work.
2026-10-06 11:18:39 -07:00
d04a0ff414 T12.6: wave 12 verification record, browser probe, BL-060
- tests/bulk_users_ui_check.py (new): selection, keyboard Space on a
  row box, toolbar role change, super-user CSV with good and bad lines,
  add-to-project with a role, typed DELETE, all end to end against a
  real server; 390px bar layout; writes after-screenshots to
  docs/reference/baseline/wave-12/. Needs a browser: to be run on
  Matt's machine.
- tests/cdp.py: --do-not-de-elevate and the rc=0 grace, the same
  change as wave 16 T16.4 (identical content, merges cleanly), so the
  probes start from an elevated shell on this branch too.
- Verification record in wave-12.md: bulk_users_check 61/61 (main
  7/18), single-user routes identical before/after on 16 calls, DOM
  harness 48/48, jsdom against a real server 13/13, smoke 27/27,
  seed_demo OK. Server-checkable done-when boxes ticked; the browser,
  screenshot and full-suite boxes stay open until run on Edge.
- BL-060 logged: a super user's refusal reason promises a project
  access change neither the dialog nor the bulk route allows. Needs
  Nick; not fixed here (scope).
- file-map.md: wave 12's two new checks.
2026-10-06 11:08:03 -07:00
1c4d1bbccb T12.5: CR-020 bulk delete with typed confirmation
'Delete selected…' in the bulk bar. The dialog lists every affected
username (scrolling inside the dialog for a long list), says the
delete cannot be undone and offers disable instead, and sends nothing
until the word DELETE is typed in capitals, whatever the count
(confirmed 2026-10-06). Wrong words keep the dialog open with the
reason; cancel sends nothing.

Your own account cannot be selected; if it ever reached the request,
the server refuses that row and the report names it with the reason.

The bulk delete is a prompt-style dialog, so it is covered in
tests/console_dialogs_check.py (its docstring asks for exactly this):
new section 4 drives it in the browser with native dialogs poisoned.
DOM harness: 48/48.
2026-10-06 11:05:01 -07:00
74d3bc4af7 T12.4: CR-020 CSV upload (usernames only)
POST /api/auth/bulk-users/resolve reads a list of usernames (first
column; comma, semicolon or tab; header skipped) and answers line by
line, the CR-005 validate-and-report pattern. A line is rejected, with
its number and reason, when it names nobody the caller can see ('User
not found', as everywhere else), an account they may not change (the
single route's own reason), their own account, or a username already
listed (duplicate of line N). An email address resolves only when
exactly one account has it. Nothing is changed by this route.

The console's 'Select from CSV' turns the accepted accounts into the
ordinary selection (filter reset so every selected row is visible) and
lists every rejected line. The action, confirm and per-row report are
the checkbox ones, through the same T12.1 routes, which re-check each
row: a CSV cannot reach an account a checkbox could not.
Usernames only, action in the toolbar (decided 2026-10-06). Files over
1 MB are refused before upload; FileReader fallback for browsers
without Blob.text().

bulk_users_check.py: 61/61. DOM harness: 40/40.
2026-10-06 11:04:01 -07:00
681048197a T12.3: CR-020 bulk-action toolbar
Role change, enable/disable, and project add/remove on the selected
rows, wired to the T12.1 endpoints. Delete is T12.5.

- The toolbar offers only what the caller may do: grantable roles,
  managed projects (archived last, marked), grantable per-project roles.
- Every action confirms through wpConfirmDialog naming the exact count
  (and for projects, the project and role) before anything is sent.
- The result lists every failed row with the server's reason, plus the
  changed / already-set counts; a partial failure is a warning, an
  all-failed batch an alert, a server error says nothing was changed.
  Never a blanket success. Announced through the live-region banner.
- The directory reloads afterwards, which clears the selection.

DOM harness (jsdom, real page and scripts, mocked API): 30/30 with
T12.2.
2026-10-06 11:02:09 -07:00
cf46b3ce42 T12.2: CR-020 row selection in the User Directory
- A checkbox per row the caller may change; select-all covers only the
  rows the current filter and search show, and is indeterminate on a
  partial selection.
- Your own row and rows you cannot change show a disabled box whose
  label gives the reason.
- Changing the filter or the search clears the selection and says so
  in a live region (decided 2026-10-06), so no hidden row stays
  selected.
- The bulk bar appears at one or more selected rows and is hidden (out
  of the tab order) at zero. Actions arrive in T12.3.
- Selection is updated in place, not by re-rendering, so focus stays on
  the checkbox just used. One delegated listener; no ids in inline
  handlers.
- The email-width rule now targets td.ell, not nth-child(3), which the
  new column would have shifted.

Checked in a DOM harness (jsdom) against the real page and scripts:
14/14; against main the selection checks fail. Browser probe at
390px/1440px follows in T12.6.
2026-10-06 11:01:02 -07:00
0e79e17a57 T12.1: CR-020 bulk user endpoints
POST /api/auth/bulk-users/{active,role,projects,delete} take a list of
user ids (max 500) and run each row through the same guards and audit
as the single-user route: the rules now live once, in guard_target and
apply_user_{active,role,delete,projects}, which the single-user routes
call too.

- A refused row is reported with its reason; the rest still run; the
  response counts ok / unchanged / failed. Never a blanket success.
- The caller's own account is refused only on its own row.
- One AuditLog row per changed account, same action and shape as the
  single route. A row already in the requested state is 'unchanged'
  and writes no audit row.
- Project assignment is add/remove to ONE project with a per-project
  role (confirmed 2026-10-06); roles on the user's other projects are
  carried over as stored.
- Routes sit under /api/auth/bulk-users/ because /api/auth/users/bulk/
  would match {user_id}='bulk'.

New tests/bulk_users_check.py (no browser): 47/47. Against main: the
7 single-route checks pass, every bulk check fails.
Single-user routes compared before/after on 16 calls (200/400/403/404):
identical responses, audit rows and memberships. Smoke test 27/27.
2026-10-06 10:57:48 -07:00
286e629782 wave 12: plan for CR-020 on a branch from main; details confirmed
Brings docs/waves/wave-12.md and decisions-2026-09-17.md across from
feat/waves-11-13 (planning docs only, no wave 11 or 13 code), so the
PR into main is wave 12 alone.

Confirmed 2026-10-06 by Matt Mabrey: typed DELETE for bulk delete;
project assignment is add/remove per project; changing the filter
clears the selection; the CSV holds usernames only.
2026-10-06 10:53:29 -07:00
da1325c8a1 T11.7: full verification pass - wave 11 closed out 2026-09-24 16:40:03 -07:00
4f3a94bd06 T11.6 follow-up: fix autosave_check.py regression from the retired tracker
Found running the full tests/*_check.py suite as part of a deployability
check on feat/waves-11-13: autosave_check.py's "section 7" asserted that
calling trackStepDwell() still wrote an event into wp-usage.js's
localStorage log (wp_suite_analytics_v1). That assertion predates CR-019
and was checking that the autosave guard hadn't accidentally clobbered
the pre-existing analytics dwell listener - a real concern at the time.

T11.6 (previous commit b2a083a) retired that listener on purpose: track()
is now a documented no-op and wp-usage.js is gone, so the localStorage key
is never written to any more. The test was still asserting the old
behavior, so it now correctly failed (0 -> 0) - this was a gap in T11.6's
own verification (only tests/baseline_shots.py was run for it at the
time; the full suite wasn't).

Fix: rewrote section 7 to check what's actually still true - autosave's
own registration (WPAutosave / sopDraftId), and that the retired
trackStepDwell() is a harmless no-op (exists, doesn't throw) rather than
missing or broken - instead of asserting it still records anywhere.

Verified: tests/autosave_check.py now passes 33/33 (was 33/34, exit 1).
Also ran the complete tests/*_check.py suite (39 scripts) as a
deployability check for the whole feat/waves-11-13 branch:
  - 38/39 exit 0 after this fix
  - token_check.py exits 2 when run with no arguments by design (it needs
    --out or --compare) - a harness-invocation detail, not a test failure;
    already confirmed non-regression in wave 10
  - generalinfo_check.py has one known pre-existing failure (a raw colour
    literal check against html/wp-creation-styles.css, a file untouched by
    any wave 11/13 work) - already diagnosed as pre-existing in wave 10,
    unrelated to CR-019/D18
Also re-verified the usage_events migration (T11.1) round-trips cleanly
on SQLite (upgrade/downgrade/upgrade) and renders clean, valid DDL for
Postgres via `alembic upgrade ... --sql` (TIMESTAMP WITH TIME ZONE / JSON,
no SQLite-only syntax) - no BL-027-class defect.
Backend smoke test (27/27) and seed_demo.py still pass.
2026-09-24 16:38:55 -07:00
2842ec996c Add session notes for 2026-09-23 (CR-019 / D18 work) 2026-09-23 15:06:02 -07:00
b2a083ac7c T11.6: retire the per-browser Usage report (CR-019)
Removes the old "Usage logs" card from admin.html/admin.js (the
usage-admin panel + loadUsage(), D5/T7.10) now that T11.5's real,
server-side Activity & usage card exists and reads real per-user data
instead of per-browser localStorage.

wp-usage.js is deleted outright, along with its three <script> includes
(admin.html, wp-creation-index.html, work-package-suite.html) - it had
no reader left once the panel above it was removed (the "download the
full event log" button lived only in that panel), and per the original
CR-019 decision record it was never reliably tied to a real identity,
so it was never a candidate data source for the new report either.

Its two call sites (wp-creation-app.js, work-package-suite-app.js) keep
a local track() function as a documented no-op rather than having each
of their ~45 individual track('event', ...) call sites deleted one at a
time - that would be a much larger, riskier diff for the same outcome
(no data is recorded either way), and each call site still marks what
was worth recording if usage analytics are ever rebuilt server-side.
work-package-suite-app.js's dwell-timer plumbing (_stepEnter /
trackStepDwell), which only ever fed track(), was left in place for the
same reason: inert, not broken.

Also removes tests/usage_check.py, which tested exactly the retired
feature, and updates its line in docs/reference/file-map.md to point at
the 2026-09-17 decision record instead.

Verified:
  - grep across the whole repo for WPUsage / wp-usage.js / usage-admin /
    usage_check: no live references remain, only explanatory comments
    and planning docs (decisions-2026-09-17.md, wave-11.md) that
    describe the removal itself
  - node --check on all three touched .js files: no syntax errors
  - full backend smoke test (27/27) and seed_demo.py still pass
  - tests/baseline_shots.py --pages admin,creator,sop at 390px/1440px,
    run locally: all three pages render with no new JS errors (the one
    "beforeunload" log line on sop/creator at 1440px is pre-existing
    harness noise from wp-autosave.js's unsaved-work guard, unrelated
    to this change) and no horizontal overflow; refreshed baseline
    screenshots committed alongside this change
2026-09-23 14:46:32 -07:00
0b5ab59518 T11.5: verify Activity & usage card at 390px/1440px 2026-09-23 14:40:30 -07:00
a6c3fbfe50 T11.5: admin console Activity & usage card (CR-019)
New card in admin.html/admin.js, above the old per-browser Usage logs
card (which T11.6 retires next). Filters (date range, project, user,
tool) drive GET /api/usage/summary; two export buttons call
GET /api/usage/export (raw / sanitized) and save the CSV via a Blob,
same download pattern wp-usage.js already uses.

Access: the card lives inside admin.html, already gated admin-only
client-side by gateByRole() (unchanged) - a non-admin never sees the
card. The underlying API is gated server-side by require_user_manager
regardless (admin or project_super_user on >=1 project), independent
of and stricter than the client-side gate, so a non-admin request is
refused even if someone reached the endpoint directly.

Accessibility (C1): every control is a real <input>/<select>/<button>,
keyboard-operable. Status/export banners use the existing '.banner' /
'*-banner' id convention, which console-util.js's MutationObserver
already turns into aria-live (role=status, or role=alert on a '.bad'
banner) - no new announcement plumbing needed. No new CSS: reuses
.toolbar/.banner/.card/.row/.kv/.users, so nothing here adds a second
token source (the token rule).

Verified so far:
  - node --check html/admin.js: no syntax errors
  - every id admin.js's new code references exists in admin.html
    (scripted diff against the full getElementById/id= sets)
  - live-server check: GET /api/usage/summary with the exact
    (possibly-empty) query string _activityFilters() builds returns
    the {active_users, per_user_last_active, by_tool, event_count}
    shape renderActivity() expects; project_id filter narrows
    correctly; GET /api/usage/export?sanitize=true returns
    text/csv with the expected header row
  - full smoke test + seed_demo.py still pass

NOT yet verified: rendering at 390px/1440px with before/after
screenshots (CLAUDE.md verification step). This sandbox has no
headless-capable browser (no chromium/msedge on PATH) and the
playwright/chromium download is blocked by this environment's
network allowlist, so tests/cdp.py's harness can't run here. Deferred
to T11.7, same as wave 10's browser checks — flagging rather than
skipping silently.
2026-09-23 12:31:55 -07:00
2de76d52e6 T11.4: usage export endpoint (raw + sanitized CSV)
CR-019 / wave 11. Adds GET /api/usage/export, reusing _usage_query()'s
scoped filters (from/to/project_id/username/tool) and the same
require_user_manager gate as /api/usage/summary. Two modes:

- raw (default): real usernames, for internal admin use.
- sanitize=true: usernames replaced with an HMAC-SHA256 pseudonym
  (keyed with auth.SECRET_KEY, 16 hex chars, 'u_' prefix) so the file
  can be fed into PowerBI or another external reporting tool without
  carrying real identities. HMAC chosen over a plain hash since the
  username space is small enough to brute-force a bare digest.

Both modes emit at, username, project_id, tool, event as columns and
deliberately omit the detail JSON column in both modes to avoid an
identity leak riding along inside free-form detail data. Response is
returned with a Content-Disposition: attachment header and a filename
that encodes mode + date.

Verified locally against a throwaway SQLite DB with two seeded users
and four seeded UsageEvent rows:
  - raw export contains the real usernames and matches the summary
    endpoint's event_count for the same session state
  - sanitized export contains no real username or email anywhere in
    the file body, across two independently-issued export calls
  - the same real user maps to the same pseudonym both within one
    export and across the two separate export calls
  - raw and sanitized rows line up 1:1 on at/tool/event for the same
    filter set
  - the tool= filter narrows the export the same way it narrows the
    summary
  - a plain project_user is refused with 403; an unauthenticated
    request is refused with 401
  - full smoke test (27/27) and seed_demo.py both still pass
2026-09-23 12:25:05 -07:00
8e863ae7d0 T11.3: usage aggregation endpoint with filters (CR-019)
GET /api/usage/summary: active-user counts by day/week/month, per-user
last-active, per-tool breakdown. Filters (from/to/project_id/username/
tool) combine. Gated by require_user_manager - same boundary as the User
Directory. A project_super_user is scoped to events tied to projects
they manage plus their own activity (managed_project_ids), never another
user's suite-wide activity outside that; an app admin sees everything.

_usage_query() factored out so T11.4's export can never disagree with
what this endpoint counted - same filtered row set, not two derivations.

Verified: admin sees all seeded events; a project_super_user scoped to
one of two projects correctly sees only that project's events plus their
own account-wide activity, and specifically does NOT see the admin's
other-project or no-project activity; date/tool/project filters each
narrow results correctly and combine.
2026-09-23 12:19:14 -07:00
6cde6e3f60 T13.1: idle timeout + absolute ceiling (D18)
Sessions now slide on activity (AUTH_IDLE_MINUTES, default 30) capped by
a hard ceiling from original sign-in (AUTH_SESSION_HOURS, meaning changed,
default 12 -> proposed 8). login_at carried across reissues so the ceiling
survives refreshes; pre-D18 tokens with no login_at fall back to iat.
Refresh is throttled (~IDLE_MINUTES/3) so the cookie isn't rewritten on
every request. Wired into auth_gate (server/app.py) - no DB hit, reads
only the already-validated claims.

Verified: 7 unit-level checks (fresh-token expiry, past-ceiling refusal,
throttling, mid-session extension, legacy-token fallback both live and
expired, idle cutoff itself) all pass, plus the full 27-check smoke
suite still passes end to end through the new middleware path.
2026-09-23 11:40:59 -07:00
358469531c D18: revise T13.1 to idle timeout + absolute ceiling
Matt asked whether idle time would be a better fit than a flat session
length. It is, but idle-alone weakens D18's own purpose - a continuously
active session would never force a fresh Okta recheck on its own. Decided:
both. AUTH_IDLE_MINUTES (new, default 30) slides the session on activity;
AUTH_SESSION_HOURS (existing var, meaning changes to an absolute ceiling,
default 12 -> proposed 8) caps it regardless of activity.

Docs only in this commit - implementation is T13.1, next.
2026-09-23 11:38:37 -07:00
850b78972b T11.2: capture usage events (CR-019)
POST /api/usage/ping writes one page_open UsageEvent per authenticated
page load, identity from the session (get_current_user), never from the
client. Wired from exactly one place - auth-guard.js's proceed(), after
wp-auth-ready - so this can't drift into six separate per-page copies.

okta_callback now also writes one login event per sign-in.

Verified locally: unauthenticated ping -> 401; a real fake-Okta sign-in
writes exactly one login row and one page_open row, no duplicates.
2026-09-23 11:24:49 -07:00
0652fa732d T11.1: UsageEvent model + migration (CR-019)
New usage_events table, separate from audit_log - see the model docstring
for why. Verified: applies and downgrades cleanly on SQLite, and the
postgresql-dialect --sql render has no risky defaults (the BL-027 class
of defect). No app wiring yet - that's T11.2.
2026-09-23 11:23:14 -07:00
75ac930d0c waves 11-13: planning docs for CR-019, CR-020, D18
CR-019 - usage/activity metrics (admin console), wave 11
CR-020 - bulk editing of users, wave 12
D18    - Okta/AD deprovisioning detection and auto-disable, wave 13

Raised by Matt Mabrey 2026-09-17. Decision record and task breakdowns
only in this commit - no feature code yet.
2026-09-18 09:21:29 -07:00
63 changed files with 7632 additions and 465 deletions

10
.env.local.example Normal file
View File

@@ -0,0 +1,10 @@
# Copy to .env in the repository root for a local Docker run (T13.8).
# .env is ignored by git. Use values that are not used anywhere else.
POSTGRES_USER=wpsuite
POSTGRES_PASSWORD=local-only-change-me
POSTGRES_DB=wpsuite
# Generate: python -c "import secrets; print(secrets.token_urlsafe(48))"
AUTH_SECRET_KEY=local-only-change-me-to-a-long-random-string
# Rehearse the Okta sync first. Add OKTA_API_TOKEN and OKTA_ORG_URL only when you
# want the local stack to call the real Okta, and keep this on until you read the log.
OKTA_SYNC_DRY_RUN=true

View File

@@ -24,7 +24,7 @@ In the stack's **Environment variables** section, add:
| Name | Value | Notes |
|------|-------|-------|
| `AUTH_SECRET_KEY` | a long random string | **Required.** Signs the login session cookies. |
| `AUTH_SESSION_HOURS` | `12` | *Optional.* Hours a login lasts before re-auth (defaults to 12). |
| `AUTH_SESSION_HOURS` | `8` | *Optional.* Hours a login lasts before re-auth (defaults to 8). |
Generate the secret on the host with:

View File

@@ -404,6 +404,148 @@ Directory page — no shell access needed.
unreachable for everyone, including admins, until Okta is restored — the same
posture the abandoned LDAPS design took, carried forward deliberately (D16).
### Accounts the Okta sync disabled (D18)
**Who this is for:** app admins, and project super users who manage accounts.
When Okta credentials are set (`OKTA_API_TOKEN`, or the service-app pair below), the `api` container asks Okta every 15 minutes
(`OKTA_SYNC_INTERVAL_SECONDS`) who is still active. It **disables** a local
account when the person is missing from Okta's list, or when Okta shows the
status SUSPENDED or DEPROVISIONED. A password expiry, a lockout, or a recovery
in progress does not disable anyone. With no token set, the sync is off and the
startup log says `Okta sync DISABLED`.
**Two ways to sign in to Okta.** Set one of them.
- *API token (SSWS):* set `OKTA_API_TOKEN` to the token string. It has about 40
letters, digits, hyphens and underscores, and no quote marks.
- *Service app with a private key (`private_key_jwt`):* set `OKTA_SYNC_CLIENT_ID`
(the app's client ID, it starts with `0oa`) and `OKTA_SYNC_PRIVATE_KEY` (the
private key as a JWK). The key starts with `{` and has fields such as `kty`,
`kid` and `d`. In Portainer, encode it as base64 first, because Portainer can
change the quote marks in JSON. In Windows PowerShell, save the key to a file
`key.json` and run:
[Convert]::ToBase64String([IO.File]::ReadAllBytes("key.json"))
Paste the one-line result as the value. Delete `key.json` afterward. The key is
a secret: never put it in Git, a ticket, or a chat.
The Okta admin must set up the app as follows: the app type is API Services; the
scope `okta.users.read` is granted on the Okta API Scopes tab; an admin role
that can read users (for example Read-only Administrator) is assigned to the app.
DPoP can be on or off. By default (`OKTA_SYNC_DPOP=auto`) the sync asks for a plain
token first, and if Okta answers `invalid_dpop_proof` it uses DPoP. Set
`OKTA_SYNC_DPOP` to `on` or `off` only to force one way. If both
`OKTA_API_TOKEN` and the service-app pair are set, the token is used.
The startup line in the log says which credential is in use. A failed sign-in
shows the HTTP status and the Okta error code, for example
`Okta answered HTTP 400 to the token request, error invalid_scope`. Common codes:
`invalid_client` (wrong client ID or key, or the key is not registered on the app),
`invalid_scope` (the scope is not granted), `access_denied` (no admin role assigned), `invalid_dpop_proof` (DPoP was used and Okta
rejected the proof: send the log lines to the developer), `use_dpop_nonce` (Okta kept
asking for a new nonce).
**First deploy: use the dry run.** Set `OKTA_SYNC_DRY_RUN=true` in `.env` (or the
Portainer stack environment) with the token, then run `docker compose up -d --build api`.
The sync runs as normal but changes no account and writes no audit row. Read the
result:
docker compose logs api | grep "Okta sync"
The summary line shows how many Okta users and active local accounts it checked,
how many matched, and how many it would disable. A line `DRY RUN would disable ...`
names each account. If the match count is close to your number of real users and
the list holds only people who have left, remove `OKTA_SYNC_DRY_RUN` (or set it
to `false`) and run `docker compose up -d api` again to go live. If the match
count is low, the Okta attribute is wrong. Set `OKTA_SYNC_PROFILE_ATTRIBUTE`
and run the dry run again.
The sync only ever disables. **It never re-enables an account.** A person comes
back only when an admin enables the account by hand.
**How to see what the sync did**
1. Open the **Admin Console** (`admin.html`). Find the **Activity log** card.
2. Click **Refresh**. Search for `okta_sync`, or set the type to **User accounts**.
3. A row from the sync reads: Who `system:okta_sync`, Action `user deprovisioned
by sync`, Item the username, Detail `okta_status: SUSPENDED` (or `MISSING` if
the person was not in Okta's list at all).
Only app admins see these rows. They belong to no project.
**How to re-enable a person**
1. **Fix the person in Okta first.** Ask the Okta admin to make the person ACTIVE
again (or to restore them, for a rehire). If you skip this step, the next sync
cycle (up to 15 minutes later) disables the account again.
2. Open the **User Directory** (`users.html`).
3. Find the person. A disabled account shows the **Enable** button on its row.
Click it.
4. The person can sign in on their next attempt. The app writes a `user enabled`
row to the Activity log under your name.
**Who may do step 3**
* An **app admin** may enable any account.
* A **project super user** may enable an account only if every project the person
belongs to is a project that super user administers. If the person is also on a
project that the super user does not administer, the row is read-only and shows
the reason. An app admin must make the change.
* Nobody can change their own account.
**If no admin can sign in** (for example, a bad sync run disabled every admin),
use the shell on the host. Replace `alice` with the username:
```bash
docker compose exec api python -m server.manage_users list
docker compose exec api python -m server.manage_users enable alice
```
**If the sync disabled the wrong people**
1. Turn the sync off. Remove `OKTA_API_TOKEN` from `.env` and restart the api
(`docker compose up -d api`). The sync starts only when the token is set.
2. Find the cause before you turn it on again. The usual cause is that the Okta
profile attribute does not match `users.username`. See
`OKTA_SYNC_PROFILE_ATTRIBUTE` in `server/.env.example`.
3. Enable the people who were disabled by mistake (steps above).
4. Set the token again and restart.
If the sync cannot reach Okta, or Okta gives an answer the sync cannot trust, the
sync changes **no** account in that cycle and writes a line to the api log:
`docker compose logs api | grep "Okta sync"`.
## Run the stack on your own PC (T13.8)
Use this to check a build in Docker before you deploy it. It needs Docker Desktop.
Do not use it on the host.
1. Copy `.env.local.example` to `.env` in the repository root. Change the two
`CHANGE` values: the database password and `AUTH_SECRET_KEY`.
2. Run these commands in the repository root:
docker compose -f docker-compose.yml -f docker-compose.local.yml up -d --build
3. Open `http://localhost:8080/api/health`. The answer is `{"ok":true}`.
4. Read the api log: `docker compose logs api`. The log shows each database
migration, then one `Okta sync` line. With no `OKTA_API_TOKEN`, the line says
`DISABLED`. That is correct.
5. To stop it and keep the data, run
`docker compose -f docker-compose.yml -f docker-compose.local.yml stop`.
To stop it and delete the data, replace `stop` with `down -v`.
What this proves: the images build, all migrations run on Postgres, and the app
starts. What it does not prove: sign-in. Okta sends the browser back to the
redirect URI that is set for the Okta app, and that is the production host name.
To rehearse the Okta sync against your real Okta, add `OKTA_API_TOKEN` and
`OKTA_ORG_URL` to `.env` and keep `OKTA_SYNC_DRY_RUN=true`. The local database has
no real users, so every real Okta user shows as a non-match. Use it only to check
that the token works and that the log lines appear.
The override file `docker-compose.local.yml` explains each change that it makes.
## Permissions roles
`User.role` is the **permissions** role; `User.project_role` is the person's **job

38
docker-compose.local.yml Normal file
View File

@@ -0,0 +1,38 @@
# Local override (T13.8). Use it only on your own PC, never on the host.
#
# docker compose -f docker-compose.yml -f docker-compose.local.yml up -d --build
#
# What it changes, and why:
# 1. It publishes the web server on http://localhost:8080 (loopback only). On the
# host a reverse proxy reaches the web server over the `proxy` network and no
# port is published.
# 2. It replaces the external `proxy` network with an ordinary bridge that this
# stack owns. A `proxy` network that already exists on a PC may have no gateway
# (internal), and Docker does not publish ports through such a network.
# 3. It makes the web server answer for any host name, so http://localhost:8080
# works. The shipped nginx config answers only for wp.controls.dev, and
# without this every /api request from localhost gets a 404. The change is
# made when the container starts: it removes the stock default.conf, which
# answers to "localhost" first, and marks the app server as the default. The
# shipped config file is not edited.
# 4. It turns the backup service off. Its scripts are checked out with CRLF line
# endings on Windows and cannot start (BL-070). To start it anyway, add
# --profile backup.
#
# Sign-in needs a real Okta redirect URI, so this stack proves the build, the
# database migrations on Postgres, and the Okta sync. It does not prove sign-in.
services:
webserver:
ports:
- "127.0.0.1:8080:80"
command:
- sh
- -c
- rm -f /etc/nginx/conf.d/default.conf && sed -i 's/listen 80;/listen 80 default_server;/' /etc/nginx/conf.d/wp-suite.conf && exec nginx -g 'daemon off;'
backup:
profiles: ["backup"]
networks:
proxy: !override
name: wp_local_proxy
driver: bridge

View File

@@ -31,7 +31,7 @@ services:
# REQUIRED — compose fails fast if it's unset, and the API refuses to
# start in production without it (see server/auth.py).
AUTH_SECRET_KEY: ${AUTH_SECRET_KEY:?set AUTH_SECRET_KEY in .env (see server/.env.example)}
AUTH_SESSION_HOURS: ${AUTH_SESSION_HOURS:-12}
AUTH_SESSION_HOURS: ${AUTH_SESSION_HOURS:-8}
# Okta OIDC — the only sign-in path (D15/D16). Not marked required the
# way AUTH_SECRET_KEY is: the API starts without these, it just refuses
# every sign-in and says so in the startup log (server/okta_auth.py
@@ -49,6 +49,22 @@ services:
# here instead, so an operator who leaves .env's copy commented out gets
# the real default, not a broken one.
OKTA_IDENTITY_CLAIM: ${OKTA_IDENTITY_CLAIM:-preferred_username}
# Okta deprovisioning sync (D18, wave 13). OFF until OKTA_API_TOKEN is set; the
# startup log says which. The token is a read-only Okta API token, NOT the OIDC
# client secret above. Empty values here mean "unset" to the code, so the
# ${VAR:-} form is safe for all four (unlike OKTA_IDENTITY_CLAIM above, whose
# code default only applies when the variable is absent).
# Reaching Okta needs outbound https. The `outbound` network the api service
# already joins is an ordinary bridge with a gateway, so no new network is
# needed; the host firewall must allow https to the Okta org.
OKTA_API_TOKEN: ${OKTA_API_TOKEN:-}
OKTA_ORG_URL: ${OKTA_ORG_URL:-}
OKTA_SYNC_CLIENT_ID: ${OKTA_SYNC_CLIENT_ID:-}
OKTA_SYNC_PRIVATE_KEY: ${OKTA_SYNC_PRIVATE_KEY:-}
OKTA_SYNC_DPOP: ${OKTA_SYNC_DPOP:-}
OKTA_SYNC_PROFILE_ATTRIBUTE: ${OKTA_SYNC_PROFILE_ATTRIBUTE:-}
OKTA_SYNC_INTERVAL_SECONDS: ${OKTA_SYNC_INTERVAL_SECONDS:-}
OKTA_SYNC_DRY_RUN: ${OKTA_SYNC_DRY_RUN:-}
# Optional — SMTP password for WP-assignment emails. Email is off by
# default and enabled from the Admin console; this is the only email
# secret and it is never stored in the DB. Leave unset until configured.

Binary file not shown.

Before

Width:  |  Height:  |  Size: 382 KiB

After

Width:  |  Height:  |  Size: 307 KiB

Binary file not shown.

Before

Width:  |  Height:  |  Size: 315 KiB

After

Width:  |  Height:  |  Size: 260 KiB

Binary file not shown.

Before

Width:  |  Height:  |  Size: 193 KiB

After

Width:  |  Height:  |  Size: 169 KiB

Binary file not shown.

Before

Width:  |  Height:  |  Size: 135 KiB

After

Width:  |  Height:  |  Size: 125 KiB

Binary file not shown.

Before

Width:  |  Height:  |  Size: 78 KiB

After

Width:  |  Height:  |  Size: 66 KiB

Binary file not shown.

Before

Width:  |  Height:  |  Size: 41 KiB

After

Width:  |  Height:  |  Size: 34 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 128 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 112 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 97 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 78 KiB

View File

@@ -283,7 +283,9 @@ python tests/triage_check.py # A6 - the sidebar answers the stand-up
python tests/qa_gate_check.py # CR-014/D2/D9/D10 - QA gate + capture sink 41 checks
python tests/files_check.py # CR-007/D8 - drawings upload + real offline 36 checks
python tests/sticky_bar_check.py # B6 - save reachable on every wizard step 12 checks
python tests/usage_check.py # D5 - one analytics core, admin report 15 checks
# tests/usage_check.py (D5) removed at T11.6 - it tested the per-browser
# analytics core and admin report, both retired in favour of CR-019's
# server-side Activity & usage card (see docs/waves/decisions-2026-09-17.md).
python tests/creator_dialogs_check.py # S1 creator - 0 natives, errors at fields 20 checks
```
@@ -316,6 +318,17 @@ python tests/critical_reopen_check.py # BL-021 - on-hold mail reaches PM + CM
python tests/console_dialogs_check.py # BL-024 - consoles/launcher: 21 natives -> 0 17 checks
```
Wave 12 (CR-020, bulk user editing) adds these:
```bash
python tests/bulk_users_check.py # CR-020 - bulk routes, guards, audit, CSV; NO browser 61 checks
python tests/bulk_users_ui_check.py # CR-020 - selection, toolbar, CSV, typed delete at 1440/390px
```
`console_dialogs_check.py` gained section 4 (the bulk delete prompt). `tests/cdp.py`
passes `--do-not-de-elevate` on Windows so the probes start from an Administrator
shell (same change as wave 16, T16.4).
**Three probes were re-pointed at `T7.1`.** `sections_check.py` 5b drove the live
cross-frame toggle hand-off, which no longer exists; it now pins the surviving path and
the fact that an UNSAVED toggle correctly does not travel - a behaviour change, recorded

View File

@@ -662,3 +662,182 @@ deliberately deferred.
worth confirming rather than assuming).
- **Suggested wave or follow-up:** next housekeeping pass — needs Nick on
whether to delete, archive, or rewrite it as historical record.
### BL-060 — A super user is told they can change access the server then refuses
- **Found during:** T12.6 (CR-020 verification, the CSV path)
- **Where:** `server/app.py` `manage_user_problem` (the reason text), `get_user_projects` / `set_user_projects` (the rule)
- **What:** for an account that is also on a project the super user does not
administer, the refusal reason ends "You can still change their access and
role on your own projects." Neither the project-access dialog's routes nor
the CR-020 bulk project route allow that: all of them call
`require_manage_user`, which refuses the account. The bulk route mirrors the
single route on purpose (CR-020: same rules), so the words promise something
neither path does. Either the text or the rule is wrong; which one is a
product call.
- **Why not now:** changing either is out of CR-020's scope (the wave must
not widen what a super user can do), and it needs a decision.
- **Suggested wave or follow-up:** needs Nick: should a super user be able to
add or remove a shared account on their own projects? Then fix the text or
the rule, not both.
### BL-061 — The sync badge covers controls while it shows
- **Found during:** T12.6 (the wave 12 screenshots)
- **Where:** `html/project-data.js` `renderSyncBadge` (fixed bottom-right, `z-index:9998`)
- **What:** for about 2 seconds after each sync the "Everything sent to the
project" badge sits over whatever is in the bottom-right corner: a Delete
button on the User Directory at 1440px, the bulk bar's "Set permissions" at
390px. It is `pointer-events:none`, so clicks pass through, but the control is
hidden while it shows, and on the console pages there is no outbox work to
report at all.
- **Why not now:** `project-data.js` is shared by every page and owned by the
outbox work in wave 14 (T14.1 to T14.3 change this badge's states anyway).
- **Suggested wave or follow-up:** wave 14, alongside T14.1's badge changes:
do not show the badge on pages with nothing to sync, and keep it clear of
page controls.
<<<<<<< ours
### BL-069 - The sync has no cap on one run and no last-admin guard
- **Found during:** T13.3 (D18)
- **Where:** `server/okta_sync.py` `run_sync()`
- **What:** a run disables every account the rules select, however many. If the
identity attribute is partly wrong (it matches some accounts and not others),
one run can disable most of the company, including every admin. The
zero-match guard only stops the case where nothing matches. Disabled accounts
cannot sign in, so re-enabling needs an admin who is still active, or the
`manage_users.py` command line on the host.
- **Why not now:** D18 does not ask for it, and the right limit is a product
call: a count, a share of accounts, a rule that the last active admin never
disables, or all three.
- **Suggested wave or follow-up:** before the first live run, or in T13.6 if
Matt decides. Needs Matt.
=======
## Future build: mobile app and scaling
Logged 2026-10-12 from the meeting with Nick. The answers are in
`docs/waves/decisions-2026-10-12.md` (Q2 to Q14). The path from here to a
mobile version outside the VPN, across projects, is in
`docs/waves/roadmap-mobile-scaling.md`. These are features for a future
build. They have no wave and no `CR-`/`D-` id yet. Give each one an id when a
wave is planned. Items marked "to confirm" have an open question listed in the
decision file ("Still open"). Do not build them until it is answered.
Follow-up answers from Matt on 2026-10-12 are folded into the entries below.
Server work is needed for BL-062 to BL-066 (CLAUDE.md, "Frontend and backend
boundary"). Do not use `localStorage` as the store of record.
### BL-062 - Offline Field View for crews
- **From:** Q2, Q3
- **What:** Field View opens offline for up to **7 days** after the last sign-in
(today: 12 hours). Only crews in Field View work offline. The creator has no
offline mode. When the app opens online again and the account is disabled,
wipe that account's local data.
- **Why not now:** new scope. It needs an id, and it follows the outbox work in
wave 14 (T14.1 to T14.3).
- **Suggested wave or follow-up:** future mobile wave. Check it against the D18
session rules (idle timeout, absolute ceiling): a 7 day offline window and a
session ceiling must agree on what "signed in" means.
### BL-063 - Offline edit conflicts: merge, notify, log
- **From:** Q4
- **What:** When an offline edit meets a newer server version, the app merges
the two, shows a notification to the crew, writes a log entry, and updates the
owner. If both sides changed the same field, **the field value wins**, and the
app keeps a **history of changes**. This replaces the earlier default (keep
both, crew chooses).
- **To confirm:** who "the owner" is, how long the history is kept, and who can
read it. See the decision file, "Still open" item 2.
- **Suggested wave or follow-up:** same wave as BL-062.
### BL-064 - Field photo limits
- **From:** Q5
- **What:** a limit of **5 MB per photo**. This equals the D8 limit that the
server already enforces for drawings (`FILE_MAX_BYTES`), so photos can use the
same upload path and the same D19 download headers.
- **To confirm:** the count per package, and whether the app resizes a photo
before upload (phone photos are often 3 MB to 8 MB). Decision file, "Still
open" item 1.
- **Suggested wave or follow-up:** same wave as BL-062.
### BL-065 - Batch print and merged PDF
- **From:** Q6, Q7, Q8, Q9
- **What:** print many packages in one PDF.
- Limit: **25 packages on phones**. On desktop the user can scale the filter
up to 50 or 100. Open: who sets the larger value (each user, or an admin
setting). Decision file, "Still open" item 5.
- Drawings go **after the form**, in **upload order**, **scaled to fit**.
- SVG drawings are **left out**. A page names each file that was left out.
- Every page shows the **package number, the print date, and page X of Y**.
- **Related:** BL-000d (merge or list attachments on export) asked the same
question for CR-008. This entry answers the layout part of it.
- **Suggested wave or follow-up:** future print wave.
### BL-066 - Keep drawings when a package is deleted
- **From:** Q13
- **What:** deleting a package no longer deletes its drawings. Keep them for
data retention. **The storage quota is freed** when the package is deleted.
Storage will move to cloud storage in a future deployment. This replaces the
earlier default (delete with the package).
- **Today:** drawing bytes live in the Postgres `wp_files` table (CR-007, D8).
The 2 GB per project limit counts them. Kept files must not count toward that
limit, so the count has to change.
- **To confirm:** retention period, who can delete for good, who can recover.
See the decision file, "Still open" item 3.
- **Suggested wave or follow-up:** roadmap Phase 3 (object storage). It affects
D19 (safe download) paths too: recovered files must keep the attachment and
sandbox headers.
### BL-067 - Ready for QA progress weight is 0.85
- **From:** Q12
- **What:** change the progress weight for Ready for QA from 0 to **0.85**.
Today a package falls from 75% to 0% when it reaches Ready for QA.
- **Why not now:** it changes a shipped number on dashboards and roll-ups. Use
its own item id and re-run the roll-up checks.
- **Suggested wave or follow-up:** future wave. Small task.
### BL-068 - One cost-code list, kept in the SOP
- **From:** Q14
- **What:** the creator has 38 cost codes, the SOP has 39, and 13 differ. Keep
**one list, in the SOP**, and make the creator read it.
- **Blocked on:** the correct codes. The **PM** supplies them (decided
2026-10-12). The date is not set. This is a data question and not a code
question.
- **Related:** BL-019 (closed) made a stored code that has left the list stay
as an option. Keep that behavior when the list changes.
- **Suggested wave or follow-up:** future wave, after the codes arrive.
>>>>>>> theirs
### BL-070 - Shell scripts get CRLF line endings on Windows, and the backup container cannot start
- **Found during:** T13.8 (local Docker run)
- **Where:** `scripts/*.sh`, no `.gitattributes` in the repository
- **What:** Git for Windows checks out `scripts/entrypoint.sh` with CRLF line endings.
The container then fails with `exec /app/entrypoint.sh: no such file or directory`
and restarts in a loop. The files are stored with LF in Git, so the host (Linux) is
not affected. Any PC build is.
- **Why not now:** T13.8 only makes the local run work. The override turns the backup
service off. A `.gitattributes` rule (`*.sh text eol=lf`) is a repository-wide change.
- **Suggested wave or follow-up:** next housekeeping pass. Add `*.sh text eol=lf` and
renormalize.
### BL-071 - smoketest.py cannot sign in against http://localhost
- **Found during:** T13.6 (final verification)
- **Where:** `server/smoketest.py` `seed_session_cookie()`
- **What:** the script puts the session cookie in a Python cookie jar for the domain `localhost`. The jar treats
a host without a dot as `localhost.local`, so it does not send the cookie, and every call after sign-in returns 401.
With `http://127.0.0.1:PORT` the same script passes 27 of 27.
- **Why not now:** it is a test-script quirk and not part of D18. The documented usage lines show `http://localhost:8000`.
- **Suggested wave or follow-up:** next housekeeping pass: use the domain `localhost.local` for a dotless host,
or change the usage lines to `127.0.0.1`.

View File

@@ -0,0 +1,97 @@
# Context handoff: Work Package Suite
> Update 2026-10-16: waves 11 and 13 (T13.2 to T13.5) are on the release branch `release/waves-11-13-okta-sync`, with T13.1 held back. The branch and wave status tables below describe 2026-10-12. See `docs/waves/wave-13.md` for the current state.
Date: 2026-10-12
Purpose: Give a new chat the facts about waves 11, 12, 13, and 16.
This file is a handoff note. It is not a spec. The spec files stay the source of truth (`IMPLEMENTATION.md`, `docs/waves/*`, `CLAUDE.md`).
It was checked against `git` on 2026-10-12. Earlier chat summaries about this repo are out of date. Use this file.
## 1. Branch state
| Branch | State |
|---|---|
| `main` | At `d42ae0c` (2026-10-06). Contains wave 12 (CR-020, bulk user editing) and wave 16 (D19, D20, D21), both merged by PR. |
| `feat/waves-11-13` | Local and `origin` are the same commit, `da1325c`. It is pushed. It is not merged. It is 13 commits ahead of `main` and 19 behind. |
| `notes/cr019-usage-metrics-d18-deprovisioning` | Holds the same wave 11 work plus the session notes. |
Correction to an earlier statement: the Gitea push is no longer blocked. The branch is on `origin`.
Merge rule from `wave-12.md`: T13.1 (idle timeout) must not reach `main` before T14.1. Do not merge `feat/waves-11-13` into `main` as a whole without a decision from the user. Wave 11 and T13.1 are on the same branch.
Working tree on `main` (2026-10-12) has two untracked files: `WorkPackageSuite_Consolidated_Change_Request_R2.docx` and `docs/waves/decisions-2026-08-25.md`. Do not delete them.
## 2. What is on `feat/waves-11-13`
| Commit | Task | Item |
|---|---|---|
| `75ac930` | Planning docs for CR-019, CR-020, D18 | CR-019, CR-020, D18 |
| `0652fa7` | T11.1 UsageEvent model and migration | CR-019 |
| `850b789` | T11.2 capture usage events | CR-019 |
| `3584695`, `6cde6e3` | T13.1 idle timeout with an absolute ceiling | D18 |
| `8e863ae` | T11.3 aggregation endpoint with filters | CR-019 |
| `2de76d5` | T11.4 usage export (raw and sanitized CSV) | CR-019 |
| `a6c3fbf`, `0b5ab59` | T11.5 admin console "Activity & usage" card, checked at 390px and 1440px | CR-019 |
| `b2a083a` | T11.6 retire the per-browser Usage report | CR-019 |
| `2842ec9` | Session notes for 2026-09-23 | none |
| `4f3a94b` | T11.6 follow-up: fix `tests/autosave_check.py` | CR-019 |
| `da1325c` | T11.7 full verification pass. Wave 11 closed on 2026-09-24. | CR-019 |
Notes:
- `html/wp-usage.js` and `tests/usage_check.py` were deleted in T11.6. The `track()` functions in `wp-creation-app.js` and `work-package-suite-app.js` are now documented no-ops. About 45 call sites stay.
- T11.7 found a real regression that T11.6 caused. `tests/autosave_check.py` still expected the retired tracker to write to `localStorage`. The test was fixed. It now passes 33 of 33. The cause was that T11.6 ran only `tests/baseline_shots.py`, not the full suite. Run the full suite for any frontend task.
## 3. T11.7 verification results (2026-09-24)
- Alembic: upgrade, downgrade, upgrade on SQLite passed. The SQL render for Postgres was clean (this catches SQLite-only DDL, the "BL-027 class" of defect).
- Test scripts: 38 of 39 passed after the autosave fix.
- Two known failures that are not regressions:
- `token_check.py` exits 2. This is a CLI usage detail.
- `generalinfo_check.py` finds an old color-literal gap in `wp-creation-styles.css`.
## 4. Wave status
| Wave | Items | Status |
|---|---|---|
| 10 | Okta OIDC sign-in (D15, D16) | Done. In production. |
| 11 | CR-019 usage and activity metrics | Done on `feat/waves-11-13`. Not on `main`. |
| 12 | CR-020 bulk user editing | Done. Merged to `main` on 2026-10-06 (T12.1 to T12.6). |
| 13 | D18 Okta/AD deprovisioning sync | T13.1 done on `feat/waves-11-13`. T13.2 and later are not started. T13.3 must fail closed. Do not relax that. |
| 14, 15 | Not checked in this handoff | T14.1 must land before T13.1 reaches `main`. Read the wave files before you plan. |
| 16 | D19 safe downloads, D20 one SOP per project, D21 session decode fails closed | Done. Merged to `main` on 2026-10-06. |
Wave 16 decisions (`docs/waves/decisions-2026-10-02.md`): SEC-06 (session revocation) waits for the AWS work. Other SEC findings stay as backlog drafts in `docs/reference/security-review/backlog-drafts.md`.
## 5. Rules that apply here (`CLAUDE.md`)
- Every change traces to an item ID (`CR-`, `F`, `S`, `A`, `B`, `C`, `D`, task `T<wave>.<n>`). New scope gets a new ID.
- One task per PR. Put the task ID and item IDs in the commit message and PR title.
- Do not fix other problems that you see. Log them in `docs/waves/backlog.md`.
- Do not reorder waves. Do not start a wave before the previous wave is merged, unless the task says it is independent.
- Server work is needed for CR-004, CR-018, CR-007, CR-011, CR-014, CR-013, and B4. Do not use `localStorage` for these.
- Do not change: Actual Hours in Closeout (CR-017), localization (A7), uppercase card headers in `console.css` (A5), the logged-override path `confirmEarlyRelease()` (A1), or hidden-not-deleted fields (CR-002, CR-016).
- Token rule: one place defines color, spacing, and type. Do not add raw hex values to page stylesheets.
- Accessibility (C1): use real `<button>`, `<a>`, or input elements. Add `aria-live` for status changes. Keep focus visible. Text contrast is 4.5:1 (3:1 for large text).
- Frontend checks: run `uvicorn server.app:app` on a throwaway SQLite database. Check 390px and 1440px. Capture before and after screenshots. Run the smoke test.
- If a done-when check cannot be run, do not mark the task complete. Say which check failed and why.
## 6. Environment on the user's PC
- Python was installed with `winget`. Tests use the repo `.venv`.
- Browser tests (`tests/cdp.py` and about 40 `*_check.py` scripts) need Edge or Chrome. `cdp.py` starts from an elevated shell (wave 16 note). The Linux sandbox cannot run them.
- See `C:\Users\m.mabrey\GIT\Field Value Driver\sde-meeting-toolkit\docs\context\environment-notes.md` for tool problems (PowerShell `$` removal, long git paths, the 180 second command limit).
## 7. Open items
1. Decide how `feat/waves-11-13` reaches `main`. The T13.1 and T14.1 order is the constraint.
2. Plan the next wave. T13.2 and later need an explicit instruction.
3. Future build list added 2026-10-12: `BL-062` to `BL-068` in `docs/waves/backlog.md` (mobile app, offline Field View, batch print, and related items). Answers are in `docs/waves/decisions-2026-10-12.md` (follow-ups from Matt included). The path to a mobile version outside the VPN is in `docs/waves/roadmap-mobile-scaling.md` (phases 0 to 5; confirmed: all users in the Okta group, multiple customers, company-managed devices; nine open questions in its section 7). These files are uncommitted edits on `main`.
4. Work from the other chat (the SDE Meeting Toolkit) does not touch this repo. See the toolkit `docs/context/` folder.
## 8. Related files
- `IMPLEMENTATION.md` (sections 2 and 4)
- `docs/waves/wave-11.md`, `wave-12.md`, `wave-13.md` (on `feat/waves-11-13`), `wave-16.md`
- `docs/waves/decisions-2026-09-17.md` (CR-019, CR-020, D18), `decisions-2026-10-02.md` (D19, D20, D21)
- `docs/waves/notes-2026-09-23.md` (on `feat/waves-11-13`)
- `docs/waves/backlog.md`

View File

@@ -0,0 +1,355 @@
# Decisions — September 17, 2026
Three items. Like `D11`, `D15`, `D16` and `D17`, these are new scope raised after
R2 closed out (see `docs/reference/completion.md`, T9.7), not a reopening of
anything already decided there.
`CR-019` and `CR-020` get `CR` ids because they are field/product-facing feature
requests — the same kind of thing `CR-001`-`CR-018` were — not internal
engineering calls made mid-build. `D18` gets a `D` id because it is exactly that:
an internal security/architecture call, the same category as `D15` (Okta vs.
LDAPS), not a feature a user asked for.
Requested/raised by Matt Mabrey, 2026-09-17.
---
## CR-019 — Usage and activity metrics (admin console)
- **Area:** Admin Console / Monitoring
- **Priority:** proposed High — the current admin console has no reliable way to
answer "who is using this and how much," which is the same visibility gap the
original UX review found in `B4` (numbers that look authoritative but are not).
- **Source:** Matt Mabrey, 2026-09-17.
### Why this is not already built
`D5`/`T7.10` already shipped a usage-analytics feature — `html/wp-usage.js`, with
a report at `admin.js:666-699` — but it reads `localStorage`, which is scoped to
one browser. The report's own empty state says exactly this: *"No usage recorded
in this browser yet."* It cannot show who across the team is active or what they
use, because each person's activity exists only on their own machine. This is the
same class of defect `B4` named for the pipeline strip — a number that looks
authoritative but is not — just never generalized to this feature. `CR-019` is
therefore new server-side work, not a UI addition on top of what exists.
Two things already in the schema are relevant and reused rather than duplicated:
`User.last_login_at` (one timestamp, no history) and `AuditLog` (append-only,
already records business mutations — WP created, status changed, role changed —
per user, with a timestamp). Neither captures navigation or feature-open events,
which is the actual gap.
### Intent
Give admins a real, server-backed picture of who is using the suite, what parts
of it they use, and how active they are — replacing the per-browser report as
the thing anyone actually looks at.
### Decisions, made 2026-09-17
1. **Retention: indefinite.** No automatic purge of usage-event data. (Separate
from `AuditLog`'s own retention, which this does not change.)
2. **Scope: suite-wide, filterable.** Not per-project by default; the console
provides filters (date range, project, user, tool/page) rather than scoping
the data itself.
3. **Export: raw view shows real identities; export supports sanitization.**
The admin console's own tables and the CSV export both default to real
usernames — this is an internal audit tool, not a public one. But the export
also offers a "sanitize" toggle that replaces the actor with a **stable
pseudonymous id** (a per-user hash, consistent across rows and across
export runs) rather than dropping the identity field outright — so an
external system (Power BI or similar) can still group and trend "by user"
without ever receiving a real name. Flagged here as the recommended
approach rather than the only one Matt confirmed in so many words: if a
fully-anonymous (no stable id at all) export turns out to be what's
actually wanted, that is a one-line change to the same feature, raise it
in the PR rather than treating it as blocking.
4. **The old per-browser report is retired**, not kept alongside the new one.
Once `CR-019` ships, `admin.js`'s existing `usage-admin` panel (reading
`WPUsage.load(...)` per browser) is removed rather than left next to the
real report, where it would show a smaller, misleading number for whoever
happens to have it open. `html/wp-usage.js` and its two call sites
(`html/work-package-suite-app.js`'s wizard dwell-tracking, and the
creator's equivalent) are a separate question — the *recording* code can
stay or go independent of the *admin report* being retired, since dwell
events were never reliably tied to a real identity anyway. Default to
removing both unless a task finds a reason to keep the recorder; log that
reason rather than deciding it here.
### Acceptance criteria
- A new admin-only tab in `admin.html` (same role gate as the User Directory)
shows: active users over a selectable date range (day/week/month), each
user's last-active timestamp, and a breakdown of which tools/pages get
opened and how often.
- Server-side event capture, keyed to the authenticated session (real identity,
not a browser-local guess) — a new table, not an extension of `AuditLog`,
since page-open/navigation events are not business mutations and mixing them
in would make `AuditLog` noisy for its existing, narrower purpose.
- The console's filters cover date range, project, user, and tool/page, and
combine (e.g., "user X, last 30 days, field view only").
- CSV export from the console, in both raw (real usernames) and sanitized
(stable pseudonymous id per user) modes.
- Retention is indefinite; nothing in this item purges data.
- View-only, refresh-on-load. No alerting — matches how the rest of the admin
console works today; if that changes later it is new scope, not a rider on
this item.
- The old per-browser "Usage" report and its admin-console panel are removed
in the same wave, not left running alongside the new one.
- Accessible per `CLAUDE.md`'s standing `C1` rules (this is a new component,
not a legacy one — it ships accessible or it is not done, same as
everything else built since wave 7).
### Frontend/backend boundary
This needs server work, the same way `CR-004`/`CR-018`/`B4` did: a real table,
a capture endpoint, an aggregation endpoint, and an export endpoint. If a task
under this item is being built by writing to `localStorage`, it is rebuilding
the exact defect this item exists to replace — stop and say so, per
`CLAUDE.md`.
### Scheduling
New wave. Wave 10 (Okta) is merged, so this does not wait on anything.
Task breakdown: `docs/waves/wave-11.md`.
---
## CR-020 — Bulk editing of users in the admin console
- **Area:** Admin Console / User Directory
- **Priority:** proposed High — three of the four actions below touch access
control (role, project assignment, active/disabled) and the fourth is a hard
delete; getting the guardrails right matters more than getting it built fast.
- **Source:** Matt Mabrey, 2026-09-17.
### Why this is not already built
Every user-editing action in `html/users.js` today is one row, one action:
a role `<select>` per row, an activate/disable button per row, a
per-user project-membership checklist opened one user at a time, and a
per-row delete. There is no row selection in the User Directory table at all.
Server-side, every corresponding endpoint
(`/api/auth/users/{id}/active`, `/role`, `/project-role`, `/projects`, and
`DELETE /api/auth/users/{id}`) takes exactly one `user_id`. Bulk editing is new
UI (selection) and, for most actions, either a loop over the existing
single-user endpoints or new endpoints that accept a list — the task decides
which, per user count and transaction-safety needs.
**Deletion is already a hard delete today** (`delete_user`, `server/app.py:1121`
— `db.delete(u)`, not a deactivate). Bulk delete inherits that: it is not this
item's job to invent a soft-delete pattern that does not exist for the single
case, but the confirmation step around it has to be sized for the fact that a
bad multi-select now removes more than one account at once, permanently.
### Decisions, made 2026-09-17
1. **All four bulk actions are in scope:** role change, activate/disable,
project assignment (add to / remove from a project, including the
project-role), and delete.
2. **Selection works two ways:** checkboxes (select-all and individual) in the
existing User Directory table, respecting whatever filter is already
applied (role, active/disabled, project) — and a CSV upload, for a one-off
bulk operation against an external list (e.g., an offboarding list that
didn't originate in this app). Both are in scope, not a choice between them.
### Acceptance criteria
- The User Directory table gains row checkboxes and a select-all that respects
the current filter; a bulk-action toolbar appears once at least one row is
selected.
- CSV upload as an alternative to checkbox selection: a list of usernames plus
the action to apply. Validates every row, reports rejected ones by row (bad
username, user not found, actor lacks permission over that user) rather than
silently skipping them — the same validate-and-report pattern `CR-005`
established for list uploads.
- Every existing single-user guardrail carries forward unchanged: an actor
cannot include their own account in a bulk action that would disable, demote,
or delete it; a super user's bulk action is scoped to only the users and
projects `require_see_user`/`require_manage_user` already let them touch
today (a super user cannot use a bulk action to reach a user or project
outside what they manage, even via CSV); `grantable_roles` still gates which
roles an actor may assign in bulk, the same as one at a time.
- Every affected row is written to `AuditLog` individually, exactly as the
single-user endpoints do today (one `role_changed` / `user_deleted` / etc.
row per user) — a bulk action is many audited changes, not one opaque batch
entry, so per-user history stays intact and readable in isolation.
- Confirmation before applying, using the `wp-dialog` kit (`T7.9`), not a
native `confirm()`. The dialog names exactly how many users are affected and,
for delete specifically, lists the affected usernames before committing.
- Partial failure is reported, not hidden: if some rows in a batch fail (scope,
already-deleted, bad CSV row), the action applies to what it can and states
exactly which rows failed and why. It never reports success on a batch that
partly failed.
- Accessible per `C1`: real controls, keyboard-operable selection and bulk-
action toolbar, `aria-live` announcing the result.
### Recommended, not yet confirmed — raise in the PR if this is wrong
- **Bulk delete gets an extra confirmation step beyond naming the count** —
proposed as typing a confirmation phrase (e.g. the word `DELETE`) regardless
of how many rows are selected, since one bad multi-select now removes more
than one account, permanently, with no soft-delete to fall back on. This is
a recommendation, not a confirmed requirement — Matt has not signed off on
the exact mechanism.
- **Project assignment is add/remove, not replace-the-whole-list** — a bulk
"add these users to project X" or "remove these users from project X"
action, rather than a bulk action that overwrites a user's entire project
list. Proposed because add/remove is less likely to clobber project
memberships the actor didn't intend to touch; a replace-the-whole-list
version is a materially different (and riskier) feature if that turns out to
be what's actually wanted.
### Confirmed 2026-10-06 (Matt Mabrey, at the start of wave 12)
Both recommendations above are confirmed as written, and the two details
`wave-12.md` left to the build are decided:
1. **Bulk delete:** the dialog lists the affected usernames and the request is
not sent until the word `DELETE` is typed, whatever the count.
2. **Project assignment is add/remove** to or from one project at a time, with
the per-project permissions role (`ProjectMember.role`) set on add. It never
replaces a user's whole project list.
3. **Changing the table filter clears the selection** (T12.2), and the toolbar
says so, so no row hidden by the filter stays selected.
4. **The CSV holds usernames only** (T12.4). The action and its settings are
chosen in the toolbar and apply to every valid row of the file.
### Frontend/backend boundary
Selection state (which rows are checked) is fine as client-side UI state — it
is not persisted data. Everything the bulk action actually does (role,
active/disabled, project membership, delete) already requires server work
today for the single-user case, and bulk does not change that: no new
localStorage-derived state, no client-side aggregation of what changed.
### Scheduling
New wave. Independent of wave 11 (`CR-019`) — the two do not touch the same
code and can build in either order or in parallel. Task breakdown:
`docs/waves/wave-12.md`.
---
## D18 — Detecting and acting on Okta/AD deprovisioning
- **Raised by:** Matt Mabrey, 2026-09-17, in response to a direct question about
what happens when someone's AD account is removed and Okta subsequently drops
them.
- **Amends:** nothing decided, closes a gap `D15`/`D16` left unaddressed. Those
items designed how identity flows *into* this app (Okta authenticates, this
app JIT-provisions and owns roles); neither addressed what happens when an
identity is withdrawn. This app has never had any deprovisioning signal —
push or pull — since Okta went live.
### What was found
Traced through `server/auth.py` and `server/okta_auth.py`: this app has no
connection back to Okta after initial sign-in. Consequences, confirmed against
the actual code:
1. A local `users` row is never touched by anything Okta-side. `is_active`
stays `True` indefinitely unless an admin manually disables or deletes the
account through the User Directory. There is no flag distinguishing a
terminated employee's account from a current one.
2. `get_current_user` (`server/auth.py:237`) re-checks `is_active` and
`token_version` on **every request** — so a manual disable takes effect
immediately, on the very next request. The gap is not enforcement, it's
detection: nothing tells an admin to go flip that switch.
3. A session already live when someone is deprovisioned keeps working, fully,
for up to `AUTH_SESSION_HOURS` (12 hours today, `server/auth.py:63`),
because validity is checked against this app's own JWT and DB state only,
never against Okta.
4. A NEW session can't be established once Okta drops the account —
`okta_login`/`okta_callback` requires completing Okta's own sign-in, which
Okta itself refuses. The front door closes on its own; the back door (an
already-open session, and the stale local record) does not.
### Decisions, made 2026-09-17
1. **Build a scheduled sync against Okta's Management API**, not an Event Hook.
This app has, since `D15`, only ever reached out to Okta — never received
anything from it — and a polling design keeps that shape rather than
introducing a new inbound, internet-reachable endpoint with its own
signature-verification surface. Traded deliberately: this is
poll-interval-late rather than real-time, which is judged acceptable for an
HR/offboarding-driven event, not a to-the-second requirement.
2. **Revised 2026-09-23, in response to Matt's question about idle time
instead of a flat session length: sessions now slide on activity, with a
hard ceiling underneath.** A flat `AUTH_SESSION_HOURS` forces a re-check
with Okta on a fixed schedule regardless of activity; a pure idle timer
with no ceiling does the opposite — a continuously-active session would
never force a fresh Okta check on its own, which is a worse fit for the
exact threat this item exists to address (someone still clicking around
after being deprovisioned). Decided: **both**.
- `AUTH_IDLE_MINUTES` (new, default **30**): a session with no request
for this long stops being valid. Implemented as a sliding JWT expiry —
the token is reissued with a fresh `exp` on activity, throttled so the
cookie isn't rewritten on literally every request.
- `AUTH_SESSION_HOURS` (existing var, meaning changes to an **absolute
ceiling**): no session survives past this many hours from the original
sign-in, no matter how continuously active it is. Default changing from
12 to a proposed **8** — flagged as a recommendation, not confirmed.
- Both defaults, and the mechanism itself, should be sanity-checked
against the tenant's actual Okta SSO session policy — if Okta's own
session silently outlives either number, re-authentication here is
likely a fast redirect, not a real re-login screen, so these numbers
cost less than they look like they do. Confirm before treating either
as final.
3. **The sync only ever disables an account — it never re-enables one.** A
rehire showing active in Okta again does not automatically restore access;
an admin re-enabling the account is a deliberate act, consistent with
`D16`'s posture that this app never auto-grants access on its own initiative.
4. **Fail closed on the side of INACTION, not disablement.** This is the
opposite failure direction from `D16`'s login-time posture ("if Okta is
unreachable, the app is unreachable for everyone"). Here, an Okta API
error, timeout, empty response, or anything the sync can't confidently
parse must result in **no change to any account** that cycle, plus a
logged failure. A sync job that treats "couldn't reach Okta" as "nobody is
active" is a far worse outcome than a missed cycle — it would lock out the
entire org on an Okta API hiccup. This is the single most important
acceptance criterion in this item.
5. **Every auto-disable is audited individually**, same as every other
account-state change in this app: an `AuditLog` row per user, with an actor
value that's clearly the sync job and not a person (e.g.
`system:okta_sync`), so it reads correctly in the User Directory's history
and is never confused with an admin's own action.
6. **New credential required:** a read-scoped Okta API token (or an Okta
service-app OAuth2 client), separate from the `OKTA_CLIENT_ID`/
`OKTA_CLIENT_SECRET` pair used for sign-in. This needs provisioning by
whoever administers the Okta tenant — the same dependency that gated the
original OIDC rollout (D15's "security scoping reply").
### Recommended, not yet confirmed — raise in the PR if this is wrong
- **Sync interval:** proposed every 15 minutes. Frequent enough that the
detection gap is small, infrequent enough not to hammer Okta's API or need
special rate-limit handling. Not confirmed with IT/security.
- **Where the job runs:** proposed as an in-process background task inside the
existing `api` container (it already has `outbound` network access to reach
Okta, and already holds the Okta client config) rather than a new sidecar
container. The `backup` container (`docker-compose.yml`) is the existing
precedent for a scheduled-interval container in this stack, if isolation
from the `api` process is preferred instead — a reasonable alternative, not
the recommendation.
- **Admin visibility:** at minimum, an auto-disable is a normal, readable
`AuditLog` entry (visible whever admin already reviews audit history, and
naturally covered once `CR-019`'s activity view exists). Whether it should
also trigger an email/notification to admins is a genuine open question —
proposed as a fast-follow rather than blocking this item, since `D10`
already established the pattern for admin-controlled email toggles this
could reuse.
### Frontend/backend boundary
Entirely server-side and infrastructure. No new `localStorage` state, no
frontend surface beyond what already reads `is_active` and `AuditLog` today
(the User Directory, and eventually `CR-019`'s activity view).
### Scheduling
New wave, independent of wave 11 and wave 12 in the sense that nothing here is
blocked by them — but note it touches the same `is_active`/account-state
surface `CR-020`'s bulk actions touch in `server/app.py`. Not a hard
dependency; sequence commits to avoid an avoidable merge conflict, per
`CLAUDE.md`'s "one task per PR" spirit. Task breakdown: `docs/waves/wave-13.md`.

View File

@@ -0,0 +1,51 @@
# Decisions: October 12, 2026 (meeting with Nick)
Topic: mobile app (offline Field View), batch print, and related scaling questions.
Recorded by Matt Mabrey from the meeting with Nick, 2026-10-12. Follow-up answers from Matt are in the section "Follow-up answers".
This file records the answers. It does not schedule any work. The features that come from these answers are logged in `docs/waves/backlog.md` as `BL-062` to `BL-068` for a future build. The path to a mobile version that works outside the VPN is in `docs/waves/roadmap-mobile-scaling.md`. None of these have a wave or a `CR-` or `D-` id yet. Assign the ids when a wave is planned.
Question numbers (`Q2` to `Q14`) come from the decision table that Matt received. `Q1`, `Q10`, and `Q11` were not in the table. They are not recorded here.
## Answers
| Q | Decision needed | Default that was proposed | Answer |
|---|---|---|---|
| Q2 | Days the app may open offline after the last sign-in (today: 12 hours) | 7 days. Wipe a disabled account's data on the next online open. | **7 days.** The default wipe rule stands. |
| Q3 | Must planners edit in the creator offline, or only crews in Field View? | Field View only | **Only crews in the field.** The creator needs no offline mode. |
| Q4 | An offline edit meets a newer server version. What does the crew see? | Keep both, show differences, crew chooses. Server wins if nobody acts. | **Merge with a notification. Log it. Update the owner.** If the same field changed on both sides, **the field value wins**, and the app keeps a history of changes. This replaces the default. |
| Q5 | Photo limit per photo and per package | Resize to 1280 px, 50 per package | **5 MB per photo.** |
| Q6 | Maximum packages in one batch print | 50, with a message to narrow the filter | **25 on phones.** On desktop the user can scale the filter up to 50 or 100. |
| Q7 | Where drawings go in the PDF, and at what size | After the form, upload order, scaled to fit | **After the form, upload order, scaled to fit.** Same as the default. |
| Q8 | SVG drawings in the merged PDF | Leave out, with a page that names the file | **Leave out, with a page that names the file.** Same as the default. |
| Q9 | What goes on every page | Package no., print date, page X of Y | **Package no., print date, page X of Y.** Same as the default. |
| Q12 | Progress weight for Ready for QA (today 0, so a package drops from 75% to 0%) | 0.85 | **0.85.** Same as the default. |
| Q13 | What happens to a package's drawings when the package is deleted | Delete with the package | **Keep them (data retention). The storage quota is freed.** Storage will move to cloud storage in a future deployment. This replaces the default. |
| Q14 | Which cost-code list is correct (creator has 38, SOP has 39, 13 differ) | One list in the SOP. Needs codes from Accounting or the PM. | **One list in the SOP.** The **PM** supplies the codes. |
## Follow-up answers (Matt, 2026-10-12)
These answer the six questions that were open after the first record.
1. **Q5:** the 5 MB limit is **per photo**. This matches the limit that already exists for drawings (D8, enforced in `server/app.py`: 5 MB a file, 2 GB per project).
2. **Q6:** 25 on phones. Desktop users can scale up (50 or 100).
3. **Q4:** on a conflict in the same field, **the field value wins**. The app keeps a history of changes.
4. **Q13:** the quota is freed. Cloud storage will hold the kept files in a future deployment.
5. **Q14:** the **PM** gives the codes.
6. **Scaling:** Matt asked for a path to a mobile version that scales to multiple projects, outside VPN access, for field agents on phones and tablets on site. That path is in `docs/waves/roadmap-mobile-scaling.md`.
7. **Who uses it:** all end users will eventually be in the Okta group.
8. **Scale:** expect multiple **customers**. The data model needs a customer (organization) level above a project. See Phase 4 of the roadmap.
9. **Devices:** company-managed.
## Still open
1. **Q5:** the count of photos per package, and whether the app resizes a photo before upload. Phone photos are often 3 MB to 8 MB. A 5 MB cap per photo with no resize will reject some normal photos.
2. **Q4:** who "the owner" is (package owner, project owner, or the person who made the server change). Also: how long the history of changes is kept, and who can read it.
3. **Q13:** how long a kept drawing is retained, who can delete it for good, and who can recover it. The quota is freed, so the retained bytes must not count toward the 2 GB project limit. Decide where they are counted instead.
4. **Q14:** the date the PM gives the codes.
5. **Q6:** who sets the larger desktop limit (each user, or an admin setting).
6. The questions in `roadmap-mobile-scaling.md`, section 7 (nine open). The first two (one person on more than one customer, and the isolation model) change the data design. Answer them first.
## Source
Meeting with Nick. Status column in the original table: "Open. Team builds to the default" for every row, as of October 12. The answers above came after that status.

View File

@@ -0,0 +1,115 @@
# Decisions: October 16, 2026
## D22: dry run for the Okta deprovisioning sync
**Asked by:** Matt Mabrey, 2026-10-16. **Extends:** D18 (wave 13). **Task:** T13.7.
**Why:** The sync disables a local account when Okta does not list the person.
If the Okta profile attribute is wrong for some accounts, the first live run can
disable people who have not left (see BL-069). The attribute `profile.login` is
the default for `preferred_username`. It is not confirmed against the tenant, and
the person who deploys has no access to the Okta admin console.
**Decision:** Add the environment variable `OKTA_SYNC_DRY_RUN`.
- Values `1`, `true`, `yes`, `on` (any case) turn it on. Any other value, or no
value, leaves the sync live. It is off by default, so a deploy that does not set
it behaves as before.
- In a dry run the sync does every step of a live run: the fetch, the match, the
zero-match guard and the fail-closed rules. It then changes no account and writes
no audit row.
- It logs one line for each account it would disable, and one summary line with the
number of Okta users, active local accounts, matches, and accounts that would be
disabled. Read them with `docker compose logs api | grep "Okta sync"`.
- The startup line shows the mode.
**Not decided here:** a limit on accounts disabled in one run, and a rule to keep the
last active admin (BL-069). The dry run lets a person see the result before the
first live run. It does not stop a bad live run.
**Rules from D18 that stay:** fail closed, only disable, one audit row for each
change, actor `system:okta_sync`.
## D23: sign in to the Okta Management API as an OAuth service app
**Asked by:** Matt Mabrey, 2026-10-16. **Extends:** D18, T13.2. **Task:** T13.9.
**Why:** The first deploy of the sync ended with `Okta answered HTTP 401`. The
credential that Okta security gave Matt is an API Services app that uses client
authentication `private_key_jwt`. It is a private key (a JWK with the fields `kty`,
`use` and `kid`), not an SSWS API token. The sync sent that value as an SSWS token,
and Okta refused it.
**Decision:** The sync accepts a second credential.
- `OKTA_SYNC_CLIENT_ID` and `OKTA_SYNC_PRIVATE_KEY` (a JWK as JSON, or as base64 of
that JSON). The sync signs a JWT (RS256 or ES256, from the key's `alg`; `kid` goes
in the header; `iss` and `sub` are the client ID; `aud` is `<org>/oauth2/v1/token`;
life 5 minutes; a new `jti` each time) and posts it to `<org>/oauth2/v1/token`
with `grant_type=client_credentials` and `scope=okta.users.read`.
- The access token is used as `Authorization: Bearer` on `/api/v1/users`. It is
requested again at the start of each run. It is not stored.
- `OKTA_API_TOKEN` still works. If both are set, the token wins. A deploy that sets
only a token is unchanged.
- A DPoP token (`token_type` not Bearer) is refused with a clear message.
**Rules from D18 that stay:** fail closed. Any problem with the key, the token
request or the user list raises an error, and the sync changes no account. The key,
the signed JWT and the access token are never logged and never put in an error
message. An error carries the HTTP status and the Okta error code only when that
code is a short lower-case word.
**Not decided here:** how often the key is replaced. A key is registered on the Okta
app and has no expiry that this code checks.
## D24: DPoP for the Okta Management API calls
**Asked by:** Matt Mabrey, 2026-10-16. **Extends:** D23. **Task:** T13.10.
**Why:** The first live token request with the service-app key ended with
`error invalid_dpop_proof`. The Okta app requires DPoP (RFC 9449). With DPoP, every
request carries a signed proof from a key that only the caller holds, so a stolen
access token alone does not work. The choice was to ask the Okta admin to switch DPoP
off, or to build it. Matt chose to build it.
**Decision:**
- `OKTA_SYNC_DPOP` is `auto` (default), `on` or `off`. In auto mode the sync asks for a
plain Bearer token first. If Okta answers `invalid_dpop_proof`, the sync asks again with
DPoP. So one build works whether the Okta app requires DPoP or not.
- Each run makes its own throwaway EC P-256 key. It is held in memory only, never
stored, logged or configured. The public half goes in each proof header.
- Each request gets its own proof (ES256, `typ` dpop+jwt): `htm`, `htu` without the
query, a one-time `jti`, `iat`, the server's latest `nonce`, and on an API call `ath`
(a hash of the access token). The same key is used for the token request and the API
calls. A new client assertion is built for every token attempt.
- A nonce challenge is answered once per request, at the token endpoint and at the API.
The token request is tried at most 4 times.
- A DPoP-bound token is sent as `Authorization: DPoP <token>`. If the setting is `off`
and Okta returns a DPoP-bound token, the sync stops with a clear message.
**Rules from D18 and D23 that stay:** fail closed; the key, the assertions, the proofs and
the access token are never logged or put in an error message; an error carries the HTTP
status and Okta's short error code only. The SSWS token path is unchanged.
**Not verified:** this has been tested only against a fake Okta that follows the RFC. The
first run against the real tenant is the real test.
## D25: the session limit is 8 hours for now
**Decided by:** Matt Mabrey, 2026-10-17. **Closes:** the 2-versus-8 question in `wave-13.md`. **Task:** T13.11.
**The conflict:** the wave 13 exit criteria said `AUTH_SESSION_HOURS` defaults to 2. T13.1's own text and
the T13.1 code say 8. The shipped default on `main` was 12.
**Decision:** The default is **8 hours**. This replaces the exit criterion of 2 hours.
- `AUTH_SESSION_HOURS` defaults to 8 in `server/auth.py`, `docker-compose.yml`, `server/.env.example` and
`DEPLOY-login-portal.md`. A value that is set still wins.
- Until T13.1 lands, 8 hours is one flat limit from the moment of sign-in, with no sliding. Once T13.1 lands, the
same variable is the absolute ceiling and an idle timer sits under it. The default does not change then.
- A person who signs in at the start of a shift is signed out 8 hours later, whatever they were doing. Signing in
again needs Okta to be reachable.
**Open, for wave 14:** how the limit works for a crew that is offline. `wave-14.md` sets that rule.
**Not decided here:** the idle time (T13.1 proposes 30 minutes). It is a proposed value, not a decision.

View File

@@ -0,0 +1,35 @@
# Decisions: October 17, 2026
Status key: **Proposed** means the text is a recommendation. It needs Matt's OK before any build starts.
**Accepted** means Matt approved it and the work may start.
Matt accepted D26 and D27 as written on 2026-10-17 (open questions 1, 2 and 3 in `wave-14.md`: hold the changes of
another person, the offline window ends with the session and then works in read-and-queue mode, and a full device
keeps the change in memory and tells the person). The idle time (question 4) was not decided. It belongs to `T13.1`.
## D26: a sign-in problem never destroys a queued change (Accepted)
**Why:** The outbox (`html/project-data.js`) marks any `4xx` answer except `429` as dead. A `401` means the session ended, and it
gets the same treatment. With a session limit (D25, 8 hours) or an idle timeout (T13.1), a crew that works offline past the limit
comes back online and every queued change is rejected for good. The outbox also hides a full device store, and it sends a queued
change under whoever is signed in, not under the person who made it.
**Proposal:**
- `401` holds the queue. It does not mark anything dead and does not count as a try. The queue sends after a good sign-in.
- `403` stays a permanent refusal, with a clear reason.
- A full device store is reported. The change is never dropped silently.
- Each queued change has an owner. It is sent only when that person is signed in.
**Tasks:** `T14.1`, `T14.3`. **Needs Matt:** open questions 1 and 3 in `wave-14.md`.
## D27: the offline sign-in window follows the server (Accepted)
**Why:** `auth-guard.js` allows the app to open offline for a fixed 12 hours after the last good check, and every page load resets that
time. The server session lasts 8 hours (D25). The two do not agree.
**Proposal:** `/api/auth/me` returns `session_expires_at`. The browser stores it and does not extend it. Offline, the app opens until
that time. After that time, offline, it works in read-and-queue mode and does not redirect. Online, a `401` goes to sign-in and the
queue is kept. The 7 day offline window (`BL-062`) is a later wave.
**Task:** `T14.2`. **Needs Matt:** open question 2 in `wave-14.md`.

View File

@@ -0,0 +1,85 @@
# Session notes — 2026-09-23
Working notes for the `feat/waves-11-13` line (CR-019, CR-020 reserved, D18),
written up before merge. Not a spec document — `wave-11.md`, `wave-13.md` and
`decisions-2026-09-17.md` are the source of truth for scope and acceptance
criteria. This is the "what actually happened building it" record.
## What shipped today
**CR-019 (wave 11) — usage/activity metrics, T11.1 through T11.6 complete:**
- `UsageEvent` table + migration (T11.1), a capture endpoint wired into
`auth-guard.js` so every protected page pings it once per load, plus a
`login` event at Okta sign-in (T11.2).
- `GET /api/usage/summary` (T11.3) — active users by day/week/month,
per-user last-active, per-tool breakdown, filterable by date/project/
user/tool, all combinable.
- `GET /api/usage/export` (T11.4) — raw and sanitized CSV. Sanitized mode
replaces the username with an HMAC-SHA256 pseudonym (keyed with
`auth.SECRET_KEY`), stable per user across rows and across separate
export calls, so an external tool (Power BI etc.) can still group by
user without ever seeing a real name.
- A new "Activity & usage" card in the admin console (T11.5): real filter
controls, the summary tables, both export buttons. Client-side gated
admin-only same as the rest of the console; the API underneath is
independently gated server-side regardless.
- Retired the old per-browser "Usage logs" panel and `wp-usage.js`
entirely (T11.6) — it had no reader left and was never a real data
source for the new report anyway. The scattered `track()` call sites in
the creator and wizard were left in place calling a documented no-op,
rather than deleting ~45 individual call sites for the same effect.
Only **T11.7 (final wave verification)** is left before wave 11 is fully
closed out — everything under it has already been verified per-task, so
this is a consolidation pass, not new work.
**D18 (wave 13) — Okta/AD deprovisioning, T13.1 only:**
- Session lifetime changed from one flat `AUTH_SESSION_HOURS` to a sliding
idle timeout (`AUTH_IDLE_MINUTES`, default 30) capped by a hard ceiling
from original sign-in (`AUTH_SESSION_HOURS`, default 8, meaning changed
from "session length" to "absolute ceiling"). Both defaults are flagged
in `.env.example` and `DEPLOYMENT.md` as proposed, not confirmed against
the tenant's actual Okta SSO policy.
- **T13.2 onward (the actual Okta Management API sync job) is paused** —
explicit call from Matt: no Okta API credential yet, come back to it
later. Not started, not blocked on anything code-side.
**CR-020 (wave 12, bulk user editing):** not started. Reserved, scoped in
`wave-12.md`, no code touched.
## Environment work (not itself a task, but load-bearing)
- Fixed a CRLF/LF mismatch that was making every tracked file look modified
to this session's git client (`core.autocrlf true`, repo-local, no file
content changed).
- This machine had no Python. Installed it via `winget` (`Python.Python.3.12`)
and set up a `.venv` in the repo with `server/requirements.txt` installed,
specifically so `tests/baseline_shots.py` could run locally — this
sandbox has no headless-capable browser and can't download one (network
allowlist), so the 390px/1440px screenshot verification CLAUDE.md asks
for had to run on Matt's own machine instead, using the browser already
installed there (Edge).
- Established a repeatable local verification loop for every task: throwaway
SQLite, fake-Okta sign-in, promote to admin, `seed_demo.py` +
`smoketest.py` (27/27 passing throughout), plus `baseline_shots.py` for
anything touching `html/`.
## Standing constraints, still in effect
- Everything stays local on this branch line. No `git push` at any point
today.
- One task per PR discipline was kept even though all of it landed on one
branch — each commit corresponds to exactly one task ID, in wave order,
each individually verified before the next started.
## Before merging
- Run T11.7 (full wave-11 verification pass) and record it in `wave-11.md`.
- Decide where this branch actually merges to — `feat/waves-11-13` has all
of today's commits already; this notes branch was cut from it so it can
fast-forward back in, or merge as its own PR if the notes should be
reviewed separately from the code.
- T13.2+ and all of wave 12 remain explicitly out of scope until Matt says
otherwise.

View File

@@ -0,0 +1,178 @@
# Roadmap: mobile version, multiple projects, access outside the VPN
Date: 2026-10-12
Status: plan for a future build. It is not a spec. No wave, `CR-`, or `D-` id exists for it yet.
Inputs: `docs/waves/decisions-2026-10-12.md` (Q2 to Q14), `BL-062` to `BL-068` in `docs/waves/backlog.md`, and the code as of `main` (`d42ae0c`).
Goal (from Matt, 2026-10-12): field agents use the app on phones and tablets on site, from outside the VPN, across more than one project.
Scope facts confirmed by Matt on 2026-10-12:
- All end users will be in the Okta group. There are no contractors without an Okta account.
- The product will scale to **multiple customers**, not only more projects for one customer.
- The devices will be **company-managed**.
## 1. Where the product is today
| Area | Fact today | Source |
|---|---|---|
| Mobile shell | A PWA already exists: `manifest.webmanifest` (standalone, "Field View" shortcut) and `sw.js` (caches the app shell and uploaded drawings). | `html/manifest.webmanifest`, `html/sw.js` |
| Offline writes | Writes queue in a sync outbox in `project-data.js`. Reads fall back to a `localStorage` cache. `/api/*` is never cached by the service worker. | `html/sw.js` header |
| Offline drawings | Drawings for the user's assigned packages are prefetched and cached (CR-007, D8). | `html/sw.js`, `html/field.js` |
| Files | The bytes live in the Postgres `wp_files` table. Limit: 5 MB a file, 2 GB per project. | `server/models.py` |
| Sign-in | Okta OIDC. A signed cookie (`HttpOnly`, `SameSite=Lax`, `Secure` over HTTPS). Idle timeout (D18, T13.1) is on `feat/waves-11-13`, not on `main`. | `server/auth.py`, `docs/waves/wave-13.md` |
| Projects | `Project` and `ProjectMember` exist. Access is by project role. There is no organization or customer level above the project. | `server/models.py` |
| Hosting | Docker Compose on an internal host. NGINX serves the site. An external proxy ends TLS. The database and backup containers have no internet route. | `docker-compose.yml`, `nginx/conf.d/wp-suite.conf` |
| Asset picker | Reads the Micron SQL Server catalog through `MICRON_DB_URL`. That database sits on the company network. | `server/assets_db.py` |
| Abuse limits | No rate limiting or login throttling was found in `server/`. CSRF origin checks and security headers (CSP, HSTS) exist. | `server/app.py`, `wp-suite.conf` |
| Security review | The 2026-10-01 review put session revocation (SEC-06) on hold until the AWS work. Wave 16 fixed SEC-01, SEC-02, SEC-07. | `docs/waves/decisions-2026-10-02.md` |
Result: the "mobile app" is mostly built as a PWA. The main work is not a new app. It is (a) finishing offline Field View, (b) making the service safe to put on the internet, and (c) removing the single-host limits.
## 2. Path in phases
Each phase has a gate. Do not start a phase before the gate of the one before it is met, unless this file says the phases are independent. Sizes are rough estimates (S under 1 week, M 1 to 3 weeks, L over 3 weeks). They are not commitments.
### Phase 0: Close the open decisions (S)
- Answer the questions in section 7 and the "Still open" list in `decisions-2026-10-12.md`.
- Assign `CR-`/`D-` ids and write the wave files.
- **Gate:** every item in `BL-062` to `BL-068` is free of "to confirm".
### Phase 1: Finish offline Field View, still inside the VPN (L)
Scope: `BL-062`, `BL-063`, `BL-064`.
1. Land the outbox work (wave 14, T14.1 to T14.3). `wave-12.md` says T13.1 must not reach `main` before T14.1.
2. Set the offline window to 7 days (Q2). Reconcile it with the D18 idle timeout and absolute ceiling. Define what "signed in" means offline. Decide what the 7 day window does when Okta says the account is disabled (wipe on the next online open).
3. Conflict handling (Q4): merge, field value wins on the same field, history of changes, a notification to the crew, a log entry, and an update to the owner.
4. Photos: capture on the device, 5 MB per photo (Q5). Add resize before upload if Matt chooses it. Store photos like drawings (server-side, with D19 download headers).
5. Add a phone and tablet test matrix (real devices): iOS Safari installed to the home screen, Android Chrome, one tablet.
6. Test the weak-link case: no signal, then a slow signal, then a sync with a conflict.
Known platform risks to test on the devices (verify; do not assume):
- iOS may delete site storage after a period of no use unless the PWA is installed to the home screen.
- iOS Safari has no Background Sync. The outbox then sends only while the app is open. A crew member must open the app to push changes.
- A lost device holds up to 7 days of project data. Web storage is not encrypted by the app. See Phase 2.
**Gate:** the crew completes a full shift offline on a test device, reconnects, and the server holds every change. The conflict case works as Q4 says. `tests/mobile_check.py` and the 390px/1440px checks pass.
### Phase 2: Make the service safe to reach from the internet (M to L)
Today the design assumes the firewall. Outside the VPN, the app becomes the perimeter. Do this before any real user connects from the internet.
| Control | Why |
|---|---|
| Rate limiting and login throttling on the API (none today) | Stops guessing and scraping. The toolkit (`sde-meeting-toolkit`) needed the same fix. |
| A WAF or reverse proxy with request limits in front of NGINX | Filters bad requests before they reach the API. |
| Real client IP handling (`X-Forwarded-For` from the trusted proxy only) | Limits and logs must use the real address, not the proxy. |
| Okta policy for the app: MFA, session length, device trust | Okta is the only sign-in path (D15, D16). The policy is the control. Ask the Okta admin. |
| Session revocation (SEC-06) | When a phone is lost, an admin must end that session now. This was put on hold for the AWS work. Bring it forward. |
| A rule for lost devices | Devices are company-managed (decided). Require a device PIN and disk encryption through the device policy. Remote wipe through the device manager. Pair it with session revocation (above), because a wipe cannot reach an offline device. Confirm with IT that the policy covers every phone and tablet that runs the app, and that a lost device is reported quickly. |
| Okta access rule | All end users are in the Okta group (decided). Gate the app on that group. Today the API creates the account on first sign-in; check that a person who is not in the group is refused, and that removal from the group ends access (this is what D18 and wave 13 do). Ask the Okta admin to add MFA and device-trust conditions for the outside path. |
| Cookie flags on every cookie (the OIDC state cookie uses `https_only=False` in `server/app.py`) | Must be `Secure` on the public site. |
| Review the 2026-10-01 findings that stayed in backlog drafts | Re-rank them for an internet-facing service. |
| Outside review (a penetration test) | Required before the customer-IP data is reachable from the internet. |
| Alerting on failed sign-ins and on odd upload sizes | Detect abuse early. |
**Gate:** the controls above are in place and an outside test finds no high findings open.
### Phase 3: Hosting that can leave the office network (L)
The AWS work is already planned (SEC-06 note). Proposed shape. The choice of services is open for Matt and the AWS owner.
- A public edge (load balancer, WAF, TLS) in front of the API and the static site.
- API containers on a managed container service, with a managed Postgres database.
- Files in object storage instead of the database (`BL-066`, Q13). Move the bytes out of `wp_files`. Keep the metadata in Postgres. Kept files from deleted packages go to a retention area with a lifecycle rule. The quota counts only live files.
- Backups and restore tests on the new stack (the current backup container writes to a host folder).
- A private link back to the Micron asset database, or a different design for the asset picker (the catalog is on the company network). Until this exists, the picker falls back to manual entry outside the VPN. Check that fallback works for a user who is online but cannot reach the database (the code treats "unconfigured" as a normal state, but not "unreachable").
- Infrastructure as code, so the environment can be rebuilt.
- Keep the internal deployment working until the cutover.
**Gate:** a full restore test passes, the migration runs on Postgres (render the SQL for the Postgres dialect first, BL-027), and a rollback plan is written and rehearsed.
### Phase 4: Many projects and many customers (L)
Matt confirmed multiple customers. A customer boundary is now part of the plan, not an option. The data model has `Project` and `ProjectMember` but nothing above a project, so this phase changes the base design. Plan for it early. Phase 3 choices (storage layout, database layout) must already allow it.
- Add an organization (customer) level above `Project`. Every project, user role, file, report, export, and search belongs to one organization. A user may belong to more than one organization (for example Prime Controls staff who work for several customers). Decide this rule first (section 7, question 1).
- Isolation: choose the model before any code. Options: one database with an organization id on every row and checks on every query; a separate schema per customer; or a separate database or stack per customer. Matt and the AWS owner decide. The choice sets cost, backup, and the strength of the separation that a customer can be shown.
- Files: one storage prefix per customer, with its own access rule and, if a customer asks, its own encryption key.
- Customer-specific data: the cost-code list, the asset catalog (today Micron only), and the SOP templates must belong to the organization or the project. `BL-068` (one cost-code list in the SOP) must not become one list for all customers.
- Okta: map Okta groups to organizations and projects, so the Okta admin controls who sees which customer.
- Reports and the admin console: super users see only the organizations they are given. Check CR-019 usage reports, the user directory, and the bulk user tools from wave 12 against this rule.
- Onboarding a customer: a runbook with the steps and a checklist.
- Project onboarding: create a project, load its SOP, cost codes (`BL-068`, one list in the SOP, codes from the PM), people, and roles. Use the bulk user tools from wave 12.
- Okta group to project membership mapping, so the Okta admin controls who joins a project.
- Per-project storage and photo quotas (2 GB per project exists today) and per-project usage reports (CR-019 on `feat/waves-11-13`).
- Performance work with a load test: many crews syncing at shift start. Check the outbox flush, the `wp_files` move, and the list endpoints on a large project.
- Data retention rules per project.
**Gate:** two real customers run side by side. A test user in one customer cannot reach any data of the other through any screen, report, export, search, file link, or API call. A test that tries to cross the boundary on purpose passes. The load test meets targets that Matt sets.
### Phase 5: Native wrapper (only if needed) (M)
Wrap the PWA in a native shell (for example Capacitor) only if Phase 1 testing shows a platform limit that the PWA cannot meet: storage eviction, no background sync, camera control, or a company rule for app-store distribution. A wrapper adds app-store review, signing, and update cost. Do not start it by default.
## 3. Order and dependencies
```
Phase 0 -> Phase 1 -> Phase 2 -> Phase 3 -> Phase 4
\_ Phase 5 (decide after Phase 1 device tests)
```
- Phase 2 may start in part while Phase 1 runs (rate limiting, Okta policy review). No crew uses the internet path before Phase 2 is done.
- Phase 3 (storage move) and `BL-066` (retained drawings) are the same work. Do them together.
- Phase 4 needs the Phase 2 and Phase 3 gates. Opening a second project to the internet first would repeat any gap twice.
## 4. Work that this path adds to the backlog
| Id | Item | Phase |
|---|---|---|
| BL-062 to BL-068 | Already logged | 1, 3, 4 |
| New | API rate limiting and login throttling | 2 |
| New | Real client IP handling behind the proxy | 2 |
| New | Session revocation (SEC-06) brought forward | 2 |
| New | Device policy check with IT (PIN, encryption, remote wipe) | 1, 2 |
| New | Okta group gate and MFA/device-trust conditions for the outside path | 2 |
| New | Object storage and lifecycle rules for files, one prefix per customer | 3 |
| New | Asset picker when the Micron database is unreachable | 3 |
| New | Organization (customer) level above `Project`, with the isolation model chosen first | 3, 4 |
| New | Per-customer cost codes, asset catalog, and SOP templates | 4 |
| New | Okta group to organization and project mapping | 4 |
| New | Cross-customer access tests (a test that tries to cross the boundary) | 4 |
| New | Customer onboarding runbook | 4 |
| New | Load test and capacity targets | 4 |
Log these in `docs/waves/backlog.md` when Matt agrees with the plan. This file does not add them, so the backlog stays free of unconfirmed scope.
## 5. Risks
1. **Customer data on lost devices.** Offline data for 7 days on a phone or tablet. Devices are company-managed, which lowers this risk a lot, but only if the device policy (PIN, disk encryption, remote wipe) is real and covers every device. An offline device cannot receive a wipe until it connects. Keep the 7 day window only if IT confirms the policy.
2. **Customer boundary added late.** The data model has no organization level. If Phases 1 to 3 are built without it, every later change touches every table and every query. Choose the isolation model in Phase 0 and design Phase 3 storage around it.
3. **Contract terms.** Customer drawings and asset data may have rules about where they are stored and who can reach them. With several customers, each contract may differ. Check before Phase 3 (cloud storage) and before Phase 2 (internet access).
4. **Okta.** The app has no password of its own. A weak Okta policy for the outside path is a weak app. The Okta admin owns that setting.
5. **Single database host.** The current database is one container with a host-folder backup. Phase 3 fixes this. Until then, do not add projects that need high availability.
6. **Sequence risk.** T13.1 (idle timeout) must wait for T14.1 (outbox). A short idle timeout can end a session while a crew is offline. Settle the rule in Phase 1.
## 6. What does not change
These decisions stay as recorded: Actual Hours stays (CR-017), localization stays (A7), uppercase card headers stay (A5), the logged-override path for predecessors stays (A1), removed fields are hidden and not deleted (CR-002, CR-016), and the token rule for CSS. Mobile work follows the accessibility rules (C1): real buttons, visible focus, `aria-live` for status, hover is not enough on tablets.
## 7. Questions that decide the plan
Answered on 2026-10-12:
- Who are the outside users? **All end users will be in the Okta group.** No separate sign-in design is needed.
- What does "multiple projects" mean? **Multiple customers.** See Phase 4.
- Which devices? **Company-managed.**
Answer these before Phase 0 ends. The first four are new and follow from the answers above.
1. **Can one person work for more than one customer?** For example, a Prime Controls planner on two customers' projects. This sets how the organization level and the Okta groups map to each other.
2. **Which isolation model?** One database with an organization id on every row, a schema per customer, or a database or stack per customer. Do any customers already require a separate stack in their contract?
3. **Who manages the Okta groups?** One group for all users, or one group per customer or project? Does the Okta admin accept that mapping?
4. **What does the device policy enforce today?** PIN, disk encryption, remote wipe, and a time limit for a device to check in. Ask IT.
5. **Connectivity on site.** No signal for days, or weak signal that comes and goes? This sets how long the offline window must really be, and how big the photo queue gets.
6. **AWS.** Is the AWS move approved? Who owns the account? Which region? Can the Micron database be reached from it?
7. **Contract.** Can customer drawings and asset data be stored in cloud storage and reached from the internet? Does each customer have its own terms?
8. **Distribution.** Is an installed PWA enough, or does the company need an app-store app? Company-managed devices can often install a web app from a device policy. Ask IT.
9. **Targets.** How many customers, projects, and users, and how many crews syncing at the same time, in the first year?

280
docs/waves/wave-11.md Normal file
View File

@@ -0,0 +1,280 @@
# Wave 11 — Usage and activity metrics
**Items:** `CR-019`
**Depends on:** wave 10 merged (it is; this wave does not wait on anything else)
**Decision record:** `docs/waves/decisions-2026-09-17.md`
Seven tasks, one concern each, in build order. Do not start a task whose
dependency is not merged. `CR-020` (bulk user editing) is reserved but not
scoped — it does not belong in this wave.
---
### T11.1 — CR-019: `usage_events` table + migration
- **Items:** `CR-019`
- **Depends on:** nothing (first task)
- **Blocks:** T11.2
- **Surface:** `server/`
- **Files:** `server/models.py`, `server/alembic/versions/`
**Do:** Add a `UsageEvent` model — append-only, same spirit as `AuditLog` but for
navigation/feature-open events rather than business mutations. Suggested shape:
`id`, `at` (indexed), `user_id` (or username — match whatever `AuditLog.actor`
does today for consistency), `project_id` (nullable — not every event is
project-scoped, e.g. opening the admin console), `tool` (e.g. `creator`,
`wizard`, `field_view`, `dashboard`, `admin`, `directory`), `event` (e.g.
`page_open`, `login`), `detail` (JSON, optional). Write the migration. Do not
touch `AuditLog` — this is a new table, not an extension of it (see the
decision record's reasoning).
**Do not:** fold this into `AuditLog`. They serve different questions and mixing
them makes the existing audit trail noisier for its existing readers.
**Done when:**
- [ ] `UsageEvent` exists with an indexed `at` column (this table will be scanned
by date range constantly)
- [ ] migration applies cleanly against both SQLite (dev) and Postgres (prod) —
render `alembic upgrade --sql` for postgresql and read it before calling
this done, per the class of defect `BL-027` logged
- [ ] no change to `AuditLog`'s shape or behavior
---
### T11.2 — CR-019: capture the events
- **Items:** `CR-019`
- **Depends on:** T11.1
- **Blocks:** T11.3
- **Surface:** `server/` + `html/`
- **Files:** `server/app.py` (new endpoint), `html/auth-guard.js`
**Problem:** Six pages need this and none should implement it separately — that
is exactly how `S4`'s "no global nav on two pages" and the four parallel token
systems (`S5`) happened. `auth-guard.js` is already loaded first, in the `<head>`,
on all six protected pages (`index.html`, `field.html`, `users.html`,
`wp-creation-index.html`, `work-package-suite.html`, `admin.html`) and already
knows the verified user once the `wp-auth-ready` event fires. That is the one
place this belongs.
**Do:** Add a small `POST /api/usage/ping`-style endpoint that writes one
`UsageEvent` row per call, keyed to the session (server trusts the session, not
anything the client claims about identity). Call it once from `auth-guard.js`
after `wp-auth-ready`, tagging `tool` from the page's own path. Also write a
`login` event at the point a session is actually established (reuse whatever
`okta_callback` already does at sign-in — do not add a second source of truth
for "did this person log in").
**Do not:** build a per-page capture call. If a page needs this and
`auth-guard.js` does not cover it, fix `auth-guard.js`, not the page.
**Done when:**
- [ ] one `page_open` event is recorded for a real sign-in on each of the six
pages, verified per page
- [ ] exactly one `login` event per Okta sign-in, not one per page load after
it
- [ ] the endpoint rejects a request with no valid session (this is server-
enforced identity, not client-reported)
- [ ] no page other than `auth-guard.js` calls this endpoint directly
---
### T11.3 — CR-019: aggregation endpoint with filters
- **Items:** `CR-019`
- **Depends on:** T11.2
- **Blocks:** T11.4, T11.5
- **Surface:** `server/`
- **Files:** `server/app.py`
**Do:** Build the read side: active-user counts by day/week/month, per-user
last-active timestamp (derived from `UsageEvent`, not `User.last_login_at`,
which only ever holds one value), and a per-tool usage breakdown. Accept query
filters: date range, project, user, tool — combinable, per the decision record.
This is server aggregation, the same principle `B4` established for the
pipeline strip: the browser asks for a number, the server computes it from real
rows, nothing is derived client-side from a partial cache.
**Done when:**
- [ ] active-user counts are correct against a seeded fixture with known dates
- [ ] filters combine correctly (verified: user + date range + tool together
narrows correctly, not just each alone)
- [ ] a project filter that matches nothing returns an empty result, not an
error or the unfiltered total
---
### T11.4 — CR-019: export, raw and sanitized
- **Items:** `CR-019`
- **Depends on:** T11.3
- **Blocks:** T11.5
- **Surface:** `server/`
- **Files:** `server/app.py`
**Do:** A CSV export endpoint over the same filtered query T11.3 exposes.
Two modes: raw (real usernames, the console's default) and sanitized. Sanitized
mode replaces the actor field with a stable pseudonymous id — a per-user hash,
consistent across rows in the same export and across separate export runs —
so an external system (Power BI or similar) can still group and trend "by
user" without ever receiving a real name. Do not simply drop the identity
column; that breaks per-user grouping downstream, which defeats the point of
an activity export.
**Done when:**
- [ ] raw export contains real usernames
- [ ] sanitized export never contains a real username or email anywhere in the
file, including in a `detail` blob if one is included
- [ ] the same real user maps to the same pseudonymous id within one export AND
across two separate export runs (a hash of something stable, not a
per-request random id)
- [ ] both modes otherwise contain identical rows for the same filter
---
### T11.5 — CR-019: admin console Activity tab
- **Items:** `CR-019`
- **Depends on:** T11.3, T11.4
- **Blocks:** T11.7
- **Surface:** `html/`
- **Files:** `html/admin.html`, `html/admin.js`
**Do:** New tab, same role gate as the User Directory. Filters (date range,
project, user, tool) driving the tables from T11.3; export buttons (raw and
sanitized) calling T11.4. Build accessible from the start per `C1` — this is a
new component, not a legacy one carrying an old defect forward: real
`<button>`/`<select>` controls, keyboard-reachable, `aria-live` on any
count that updates without a page reload, focus visible throughout.
**Done when:**
- [x] the tab is reachable only by an admin (the card lives inside admin.html,
already gated client-side by gateByRole(); the API underneath it is
independently gated server-side by require_user_manager regardless)
- [x] every filter is a real form control, keyboard-operable (date/select/text
inputs and a `<button>`, no click-div)
- [x] both export buttons produce the files T11.4 defines (verified against
the live endpoint in T11.4's own checks, and present/wired here)
- [x] works at 390px and 1440px — verified 2026-09-23 via
`tests/baseline_shots.py --pages admin` run locally on Windows (this
sandbox has no headless browser available; the script was run on the
user's machine instead, after installing Python via winget since it
wasn't present). Screenshots in `docs/reference/baseline/admin-390.png`
/ `admin-1440.png`. No JS errors, no horizontal overflow at either
width; the card rendered with real seeded data (events, by-tool,
per-user-last-active tables) confirming the filters and summary read
correctly, not just that the markup exists.
---
### T11.6 — CR-019: retire the per-browser Usage report
- **Items:** `CR-019`
- **Depends on:** T11.5
- **Blocks:** T11.7
- **Surface:** `html/`
- **Files:** `html/admin.js` (the `usage-admin` panel, `admin.js:666-699`),
`html/wp-usage.js` and its two call sites
**Do:** Remove the old per-browser `usage-admin` panel from `admin.js` now that
the real one exists, per the 2026-09-17 decision. Decide what happens to
`wp-usage.js`'s recording calls (wizard dwell-tracking, the creator's
equivalent): they were never reliably tied to a real identity, so they are not
a data source the new report can adopt. Default to removing the recorder too
unless it is still doing something useful on its own (re-read what it actually
records before deciding — do not assume from this file alone).
**Do not:** leave the old panel in place "just in case." Two activity reports
showing two different numbers is worse than one.
**Decision (2026-09-23):** `wp-usage.js` is removed, not kept — it had no
reader left once the admin panel above it was removed (the "download the
full event log" button lived only in that panel), and per the original
decision record it was never reliably tied to a real identity, so it was
never a candidate source for the new report either. The file itself and its
three `<script>` includes (`admin.html`, `wp-creation-index.html`,
`work-package-suite.html`) are gone. Its two call sites
(`wp-creation-app.js`, `work-package-suite-app.js`) keep a local `track()`
function as a documented no-op rather than having each of their ~45
individual `track('event', …)` call sites deleted one at a time — that
would be a far larger, riskier diff for the same outcome (no more data is
recorded either way), and it keeps each call site as a marker of what was
worth recording if usage analytics are ever rebuilt server-side. The
dwell-timer plumbing that only ever fed `track()` (`work-package-suite-app.js`'s
`_stepEnter`/`trackStepDwell`) was left in place for the same reason — it is
inert now, not broken, and touching it buys nothing.
Also removed: `tests/usage_check.py` (tested exactly the retired feature —
D5/T7.10's per-browser analytics core and admin report) and its line in
`docs/reference/file-map.md`, replaced with a note pointing at this decision.
**Done when:**
- [x] the old `usage-admin` panel and its markup are gone from `admin.html`/
`admin.js`
- [x] a decision on `wp-usage.js` itself is recorded (removed, or kept with a
stated reason) — not left ambiguous — see above
- [x] nothing else in the app references the removed code; grep confirms
(only remaining hits are this file, the 2026-09-17 decision record, and
the two explanatory code comments left at the retired call sites — all
prose, not live references)
---
### T11.7 — CR-019: verification
- **Items:** `CR-019`
- **Depends on:** T11.6
- **Blocks:** nothing
- **Surface:** `html/` + `server/`
- **Files:** as touched above
**Do:** Full verification per `CLAUDE.md`: run the app locally, exercise the
new tab at 390px and 1440px, before/after screenshots, run the existing smoke
test and `seed_demo.py`, run the full suite.
**Done when:**
- [x] all `CR-019` acceptance criteria in `decisions-2026-09-17.md` are met or
a failure is stated with a reason — all met
- [x] screenshots committed — `admin-390.png`/`admin-1440.png` (and
`creator`/`sop` at both widths, touched by T11.6's script-tag removal)
- [x] smoke test and `seed_demo.py` both still pass — 27/27, verified
2026-09-24 against the exact branch state about to be considered
deployable
- [x] full test suite passes — verified 2026-09-24: ran all 39
`tests/*_check.py` scripts plus the `usage_events` migration
round-trip. Result: 38/39 exit 0.
- `token_check.py` exits 2 with no arguments by design (needs `--out`
or `--compare`) — a CLI-invocation detail, not a test failure;
already confirmed non-regression in wave 10.
- `generalinfo_check.py` has one pre-existing failure (a raw colour
literal in `html/wp-creation-styles.css`, a file no wave 11/13 task
touched) — already diagnosed as pre-existing in wave 10.
- `autosave_check.py` **did** regress from T11.6 (it asserted the
retired analytics tracker still wrote to localStorage) — found by
this run, fixed same day (commit `4f3a94b`), now 33/33.
- Migration (T11.1): SQLite upgrade → downgrade → upgrade round-trip
clean; `alembic upgrade ... --sql` against the `postgresql` dialect
renders valid DDL (`TIMESTAMP WITH TIME ZONE`, `JSON`, no
SQLite-only syntax) — no `BL-027`-class defect.
---
## Wave 11 exit criteria
- [x] real, server-side activity data exists per user, indefinitely retained
- [x] the admin console shows it, filterable by date/project/user/tool
- [x] export works in both raw and sanitized form
- [x] the old per-browser report is gone, not duplicated
- [x] `CR-019` fully accounted for, no open acceptance criteria
Wave 11 is closed out as of 2026-09-24. `CR-020` (wave 12) remains reserved
and unstarted. `D18`/wave 13 (`T13.2` onward) remains paused pending an Okta
API credential, per standing instruction — `T13.1`'s session changes were
included in this same verification pass and did not regress anything above.

263
docs/waves/wave-12.md Normal file
View File

@@ -0,0 +1,263 @@
# Wave 12 — Bulk editing of users
**Items:** `CR-020`
**Depends on:** wave 10 merged (it is). Independent of wave 11 (`CR-019`) — no
shared files, may build in either order or in parallel.
**Decision record:** `docs/waves/decisions-2026-09-17.md` (open details
confirmed 2026-10-06, see "Confirmed 2026-10-06" in the CR-020 section)
**Branch (decided 2026-10-06):** `feat/wave-12-bulk-users`, cut from `main`
(`df20b8f`, production). Only this file and the decision record come across
from `feat/waves-11-13`, so the PR into `main` is wave 12 alone: no wave 11
code and no T13.1, which must not reach `main` before T14.1. Wave 16 is not
merged yet; it touches different functions, and whichever lands second rebases.
Six tasks, in build order.
---
### T12.1 — CR-020: bulk endpoints
- **Items:** `CR-020`
- **Depends on:** nothing (first task)
- **Blocks:** T12.2, T12.4
- **Surface:** `server/`
- **Files:** `server/app.py`
**Do:** Add bulk variants of the four existing single-user actions — role
change, active/disabled, project assignment (add/remove + project-role), and
delete. Each takes a list of `user_id`s plus the action's parameters and
applies `require_see_user`/`require_manage_user`/`grantable_roles` **per row**,
exactly as the single-user endpoint does today — a super user's bulk request
cannot reach further than their existing single-user requests can. Do not skip
the self-action guard: an actor's own account is rejected out of any batch
that would disable, demote, or delete it, same as today.
Each successful row writes its own `AuditLog` entry via `log_event`, same
action names the single endpoints already use. A row that fails is reported in
the response (user id, reason) and does not stop the rest of the batch from
being attempted.
**Do not:** invent a single opaque "bulk_action" audit entry in place of the
per-row entries. Do not build a soft-delete path for bulk delete that doesn't
exist for the single case — bulk delete stays a hard delete, matching
`delete_user` today.
**Done when:**
- [x] each of the four bulk actions is callable with a list of user ids and a
single set of parameters
- [x] a batch containing the actor's own account rejects only that row, not
the whole batch — verified for disable, demote, and delete
- [x] a super user's batch that includes a user/project outside what they
manage rejects only that row, with a stated reason
- [x] every successful row produces its own `AuditLog` entry, identical in
shape to what the single-user endpoint would have written
- [x] a batch with some failing rows still applies to the rows that succeed,
and the response lists exactly which rows failed and why
---
### T12.2 — CR-020: row selection in the User Directory table
- **Items:** `CR-020`
- **Depends on:** T12.1
- **Blocks:** T12.3
- **Surface:** `html/`
- **Files:** `html/users.js`, `html/users.html`
**Do:** Add a checkbox per row and a select-all control, respecting whatever
filter (`role`, `active`/`disabled`, project) is currently applied — select-all
selects the filtered set, not every user in the system regardless of what's
shown. A bulk-action toolbar appears once at least one row is checked and
disappears at zero.
**Done when:**
- [x] select-all selects exactly the rows currently visible under the active
filter, not the full unfiltered table
- [x] changing the filter while rows are selected does something sane and
visible (either clears the selection or keeps it explicit which rows are
still selected) — pick one and state it, don't leave it undefined
- [x] the toolbar is keyboard-reachable and only present when >=1 row is
selected
---
### T12.3 — CR-020: bulk-action toolbar
- **Items:** `CR-020`
- **Depends on:** T12.2
- **Blocks:** T12.5
- **Surface:** `html/`
- **Files:** `html/users.js`
**Do:** Wire the toolbar to T12.1's endpoints for role change, activate/
disable, and project assignment (add to / remove from a project + project-
role). Confirmation before applying uses the `wp-dialog` kit (`T7.9`) —
`wpConfirmDialog`, not a native `confirm()` — naming exactly how many users are
affected. On completion, report per-row results if anything failed (T12.1
already returns this) rather than a single success/failure toast that hides a
partial failure.
**Delete is built separately, in T12.5** — do not wire delete here.
**Done when:**
- [x] role change, activate/disable, and project assignment each work end to
end against a multi-row selection
- [x] the confirmation dialog names the exact affected count before anything is
sent
- [x] a batch with a partial failure shows which rows failed, not just an
undifferentiated error
- [x] `aria-live` announces the outcome
---
### T12.4 — CR-020: CSV upload path
- **Items:** `CR-020`
- **Depends on:** T12.1
- **Blocks:** T12.6
- **Surface:** `html/` + `server/`
- **Files:** `html/users.js`, `server/app.py`
**Do:** An upload accepting a list of usernames plus the action to apply,
following the validate-and-report pattern `CR-005` established: reject and
report bad rows (username not found, actor lacks permission over that user)
rather than silently dropping them. This is a second entry point onto the same
T12.1 endpoints, not a third implementation of the bulk logic.
**Done when:**
- [x] a CSV with a mix of valid and invalid usernames applies to the valid
rows and reports the invalid ones by row, with a reason
- [x] the same permission/self-action guards from T12.1 apply here — a CSV
cannot reach a user a checkbox-driven batch couldn't
- [x] duplicate usernames in one CSV are handled without double-applying or
erroring confusingly
---
### T12.5 — CR-020: bulk delete confirmation
- **Items:** `CR-020`
- **Depends on:** T12.3
- **Blocks:** T12.6
- **Surface:** `html/`
- **Files:** `html/users.js`
**Do:** Wire delete into the toolbar with a heavier confirmation than the other
three actions, per the decision record's recommendation: list the affected
usernames and require typing a confirmation phrase (e.g. `DELETE`) before the
request is sent, regardless of how many rows are selected. This is flagged in
the decision record as a recommendation Matt has not explicitly signed off on
— if the PR reviewer wants a lighter or heavier mechanism, that's the moment to
change it, not a reason to skip building a real confirmation now.
**Done when:**
- [x] the affected usernames are listed in the confirmation dialog before
delete is sent
- [x] the request is not sent until the confirmation phrase is typed correctly
- [x] the actor's own account, if somehow selected, is rejected with a clear
reason rather than silently included or silently dropped
---
### T12.6 — CR-020: verification
- **Items:** `CR-020`
- **Depends on:** T12.4, T12.5
- **Blocks:** nothing
- **Surface:** `html/` + `server/`
- **Files:** as touched above
**Do:** Full verification per `CLAUDE.md`: run locally, exercise bulk role
change, activate/disable, project assignment, CSV upload, and bulk delete at
390px and 1440px, before/after screenshots, smoke test, `seed_demo.py`, full
suite.
**Done when:**
- [x] all `CR-020` acceptance criteria in `decisions-2026-09-17.md` are met or
a failure is stated with a reason
- [x] screenshots committed
- [x] smoke test and `seed_demo.py` both still pass
- [x] full test suite passes
---
## Wave 12 exit criteria
- [x] all four bulk actions work against both a checkbox selection and a CSV
upload
- [x] every existing single-user guardrail (self-action, scope, grantable
roles) holds under bulk use
- [x] bulk delete requires typed confirmation and lists affected usernames
- [x] partial failures are always reported, never hidden behind a blanket
success
- [x] `CR-020` fully accounted for, no open acceptance criteria
## Verification record
Built 2026-10-06 on `feat/wave-12-bulk-users`
(from `main`, `df20b8f`), one commit per task. Run in the build sandbox (Linux,
Python 3.10, no browser).
| Check | Result |
|---|---|
| `tests/bulk_users_check.py` (new, no browser) | **61/61**: all four actions, own-row refusals, super-user scope, per-row audit in the single route's shape, partial failure, project add/remove keeping other projects and roles, CSV resolve, 403/401/422 gates |
| Same check against `main` | 7/18: the 7 single-route checks pass; every bulk check fails, because the routes do not exist there |
| Single-user routes before/after (16 calls: 200/400/403/404, as admin and super user) | **Identical** responses, audit rows and memberships against `main`. The routes now call the same `apply_*` functions as the bulk routes |
| Page logic, DOM harness (jsdom, the real `users.html` and scripts, mocked API) | **48/48**: selection, select-all over the filtered rows, filter clears selection, toolbar, confirm counts, per-row report, CSV, typed DELETE |
| Same flows, jsdom against a real server seeded by `browser_check.seed()` | **13/13**, including the CSV report texts the browser probe expects |
| `server/smoketest.py` | **27/27** |
| `server/seed_demo.py` | Seeds the demo project |
**Decisions applied (confirmed 2026-10-06):** typed `DELETE`; project assignment
is add/remove per project, with the per-project role (`ProjectMember.role`) set
on add; filter or search change clears the selection; CSV holds usernames only.
A row already in the requested state reports `unchanged` and writes no audit
row (nothing changed). An email address in the CSV resolves only when exactly
one account has it.
**Found and logged, not fixed:** BL-060, a super user's refusal reason promises
a project-access change that neither the dialog nor the bulk route allows.
**Browser probes, run 2026-10-06 on Matt's machine (Windows, Chrome 153 / Edge 154,
`.venv`), one at a time:**
| Probe | Result |
|---|---|
| `tests/bulk_users_ui_check.py` (new) | **19/19**: selection, Space on a row box, role change, super-user CSV (another job / not found / herself / duplicate, each by line), add-to-project with a role, typed DELETE, 390px bar fit, no JS errors |
| `tests/console_dialogs_check.py` | **16/16**, including the new section 4 (bulk delete prompt, natives poisoned) |
| `tests/mobile_check.py` | **24/24**, users page included |
| `tests/a11y_check.py` | **22/22**: users page 66 focusable elements, all with a visible ring; helper text 5.02:1 at worst |
| `tests/browser_check.py` | **67/71**, then re-pointed (below); **rerun after the re-point: all passed** (reported by Matt, 2026-10-06) |
**`browser_check.py` re-pointed (T12.6).** Its four failures were assertions
written against column POSITIONS: the manager table had 9 columns, and the
own-row checks read `r.cells[3]` (Permissions) and `r.cells[4]` (Project role).
CR-020's selection column is first, so the table has 10 and those indexes
landed on Username and Email. Not an app defect. The probe now expects 10
columns with the selection column first, finds the two cells by their header
name rather than their position (so the next column cannot break it the same
way), and adds one check: your own row has no selection box. Verified against
the real page in the DOM harness, then rerun in the browser: all passed.
**Screenshots** (after; the before-shots are `baseline/users-1440.png` and
`baseline/after-wave9/users-390.png`): `baseline/wave-12/users-1440.png`,
`users-1440-selected.png`, `users-390.png`, `users-390-selected.png`. Reviewed:
the bar, its four groups and the selected-row marking render as designed at
both widths; nothing scrolls sideways at 390px.
**Seen in the screenshots, not caused by wave 12, logged as BL-061:** the
pre-existing sync badge ("Everything sent to the project", `project-data.js`)
is drawn over a row's Delete button at 1440px and over "Set permissions" at
390px for about 2 seconds after a load. It is `pointer-events:none`, so it never
blocks a click, but it hides a control while it shows.
**Wave 12 closed 2026-10-06.** Every task done-when and every exit criterion
above is checked. Pushed as `feat/wave-12-bulk-users`; PR into `main` open.
Merge after wave 16 (`feat/wave-16-on-main`); both orders tested conflict-free.

486
docs/waves/wave-13.md Normal file
View File

@@ -0,0 +1,486 @@
# Wave 13 — Okta/AD deprovisioning sync
**Items:** `D18`
**Depends on:** wave 10 merged (it is). Not blocked by wave 11 or wave 12, but
shares `server/app.py` account-state surface with wave 12 (`CR-020`) — sequence
commits to avoid an avoidable conflict.
**Decision record:** `docs/waves/decisions-2026-09-17.md`
Six tasks, in build order. **T13.3's fail-closed behavior is the most
important done-when list in this wave — do not relax it to ship faster.**
---
### T13.1 — D18: idle timeout with an absolute ceiling
- **Items:** `D18`
- **Depends on:** nothing (first task, independent of the rest)
- **Blocks:** nothing
- **Surface:** `server/`
- **Files:** `server/auth.py`, `server/app.py` (`auth_gate` middleware),
`server/.env.example`, `DEPLOYMENT.md`
**Revised 2026-09-23** — originally just "shrink `AUTH_SESSION_HOURS`". Matt
asked whether an idle timeout would be a better fit than a flat session
length. It is, but not by itself — see the decision record's reasoning on why
idle-with-no-ceiling is actually a worse fit for this item's own threat model
than a flat expiry would have been. Build both.
**Do:**
- Add `login_at` to the JWT payload in `create_token()` — the original
sign-in time, distinct from `iat`, which becomes "when THIS token was
issued" once tokens start getting reissued. `login_at` never changes across
reissues; it's what the absolute ceiling is measured from.
- Add `AUTH_IDLE_MINUTES` (default 30). `AUTH_SESSION_HOURS` stays the name
for the absolute ceiling, default changing from 12 to a proposed 8 — update
its docstring/comment in `auth.py` and `.env.example`, since its MEANING is
changing (session length -> hard ceiling on top of a sliding idle window),
not just its value.
- In `auth_gate` (`server/app.py`), after confirming a request is
authenticated: if `now < login_at + AUTH_SESSION_HOURS` (the ceiling hasn't
passed), compute `new_exp = min(now + AUTH_IDLE_MINUTES, login_at +
AUTH_SESSION_HOURS)`. If `new_exp` is meaningfully later than the current
token's `exp` (throttle this — do not reissue on every single request, only
when enough time has passed to be worth a new cookie; a few minutes of
slack is fine), mint a refreshed token carrying forward `sub`/`username`/
`role`/`ver`/`login_at` unchanged, and set it on the response.
- If the ceiling HAS passed, do not refresh — let the existing token expire
on its own terms (it may already be invalid, or may tick over within the
idle window; either way, no new one is issued past the ceiling).
- The middleware should not need a DB round trip to do this — everything
needed (`sub`, `username`, `role`, `ver`, `login_at`) is already in the
validated claims. `get_current_user`'s existing per-request `is_active`/
`token_version` check is unaffected and still runs separately.
**Do not:** reissue the cookie on every request unconditionally — that's a
`Set-Cookie` header on every API call for no benefit over a coarser refresh.
Do not let the ceiling check silently vanish for tokens issued before this
change ships — a token with no `login_at` claim should fall back to treating
its own `iat` as `login_at`, not bypass the ceiling entirely.
**Done when:**
- [x] a session with continuous activity stays alive past 30 minutes but is
cut off at the `AUTH_SESSION_HOURS` ceiling regardless
*(2026-10-16, T13.6: a request every minute for 10 hours: alive well past 30 minutes, never extended past login + 8 h, cut off at the ceiling, original `login_at` kept.)*
- [x] a session with no activity for 30+ minutes is rejected on its next
request
*(2026-10-16, T13.6: an idle token gets 401 on a gated route; one inside the window gets 200.)*
- [x] the cookie is not rewritten on every single request — verify the
refresh is throttled, not unconditional
*(2026-10-16, T13.6: 20 rapid requests on a fresh session: 0 rewrites; 480 requests over 8 h: about one rewrite per 10 minutes; a refreshed cookie is not rewritten again on the next request. Mutation test: removing the throttle fails 3 checks.)*
- [x] a pre-existing token with no `login_at` claim (simulating a session
issued before this shipped) still gets a hard ceiling, via the `iat`
fallback
*(2026-10-16, T13.6: a token with iat 9 h ago and a still-valid old exp is not extended; one with iat 1 h ago is refreshed with `login_at` taken from `iat`; no `login_at` and no `iat` is not refreshed and does not crash. Mutation test: giving it a fresh ceiling fails.)*
- [x] `.env.example` and `DEPLOYMENT.md` explain both variables and flag both
defaults as proposed, not confirmed against Okta's own session policy
*(2026-10-16, T13.6: both variables are in both files, and both call the defaults proposed.)*
- [x] existing session/auth tests updated for the new mechanism, not just a
new number
*(2026-10-16, T13.6: there were none. T13.1 shipped with no test of its own. `tests/session_idle_check.py` is that test (T13.6 added it).)*
---
**Found while writing that test (T13.6, 2026-10-16):**
- Nothing under `/api/auth/` slides the session. `auth._EXEMPT_PREFIXES` covers
that whole prefix, so `auth_gate` never refreshes there (including
`/api/auth/me`). A page that only called `/api/auth/me` would not extend its
session. The spec says "after confirming a request is authenticated", and these
routes are not confirmed by the gate, so this is as written. Noted, not changed.
- The wave's exit criteria below say `AUTH_SESSION_HOURS` defaults to **2**. T13.1
(this section), `auth.py`, `.env.example` and `DEPLOYMENT.md` all say **8**. The
code and the T13.1 text agree, so the exit line looks like a leftover from
before the 2026-09-23 revision. Not changed here: it is a product number. Needs
Matt to say which one he means.
### T13.2 — D18: Okta Management API client
- **Items:** `D18`
- **Depends on:** nothing (independent of T13.1)
- **Blocks:** T13.3
- **Surface:** `server/`
- **Files:** new `server/okta_sync.py` (or extend `server/okta_auth.py` —
task's call), `server/.env.example`
**Do:** A small client for Okta's user-list endpoint, authenticated with a new,
separate credential (e.g. `OKTA_API_TOKEN`) — not the OIDC client secret used
for sign-in. Fetch the full user list (paginated per Okta's API) rather than
one-by-one lookups per local user; this app has ~20-odd accounts today, and a
list-and-diff is simpler and cheaper than N calls. Return each Okta user's
identity-claim value (matching `OKTA_IDENTITY_CLAIM`, already confirmed live
per wave 10) and status.
Follow the existing pattern for external credentials in this repo (`MICRON_DB_URL`,
`SMTP_PASSWORD`): env-only, never logged, never returned to the browser in an
error message.
**Do not:** reuse `OKTA_CLIENT_ID`/`OKTA_CLIENT_SECRET` for this. Sign-in and
the management API are different trust boundaries with different scopes;
conflating them means a compromise or rotation of one affects the other
unnecessarily.
**Done when:**
- [x] the client authenticates with its own credential, distinct from the OIDC
client
*(2026-10-16: `OKTA_API_TOKEN`, sent as `Authorization: SSWS`. `tests/okta_sync_check.py` sets a different OIDC client secret and asserts it is never sent.)*
- [x] it fetches the complete Okta user list, handling pagination
*(2026-10-16: follows `Link: rel="next"`, 200 per page, 2-page case checked. A next link to another host is refused, a chain longer than 50 pages raises instead of cutting the list off, redirects are never followed.)*
- [x] an auth failure or malformed response raises a clear, specific error
rather than returning an empty list indistinguishable from "everyone was
deprovisioned" — this distinction is what T13.3 depends on
*(2026-10-16: `OktaSyncError` for HTTP 401/403/404/429/5xx, redirects, network errors, timeouts, non-JSON, non-list, a bad item (never skipped), and an empty list. 31 failure checks. Mutation test: making the client return `[]` on an error fails 10 of them.)*
- [x] the credential is never logged or surfaced in any API response
*(2026-10-16: error messages carry the class name and HTTP status only, never the URL or a header. `describe()` has no token. Nothing is wired to an API response yet; T13.4 wires the startup log.)*
---
### T13.3 — D18: the sync job
- **Items:** `D18`
- **Depends on:** T13.2
- **Blocks:** T13.4
- **Surface:** `server/`
- **Files:** `server/okta_sync.py`, `server/app.py` (or wherever `log_event`
lives)
**Do:** Compare the Okta user list (T13.2) against local `users` rows. For any
local `is_active=True` user whose Okta identity is missing from the list, or
present with a non-active status, set `is_active=False` and write an
`AuditLog` row (`actor="system:okta_sync"`, action e.g.
`user_deprovisioned_by_sync`, detail naming the Okta status found). Never
touch a user already `is_active=False`. Never re-enable anyone.
**The fail-closed rule, non-negotiable:** if T13.2's client raises an error of
any kind (network, auth, malformed response, timeout), this task takes **no
action on any account** for that run and logs the failure clearly (server log,
at minimum). An error is never treated as "Okta returned zero active users."
Write a test that asserts this directly: feed the sync a failing client and
assert zero rows changed and zero `AuditLog` entries written.
**Done when:**
- [x] a user missing from Okta's list, or present but not active, is disabled
with a correctly-detailed audit row
*(2026-10-16: see "Decided 2026-10-16" below for which statuses count as "not active". Audit row: actor `system:okta_sync`, action `user_deprovisioned_by_sync`, `entity_type` user, `detail` `{okta_status, via}`; a missing user is recorded as `MISSING`.)*
- [x] a user already disabled is left alone (no duplicate audit row each run)
*(2026-10-16: a second run over the same data writes no new rows.)*
- [x] an active Okta user already active locally produces no audit row (only
changes are logged, not a clean bill of health every cycle)
*(2026-10-16: verified in `tests/okta_sync_job_check.py`.)*
- [x] **a simulated Okta API failure results in zero account changes and zero
audit rows** — this is the one check that must never be skipped or
weakened
*(2026-10-16: 10 stand-in failures (OktaSyncError, RuntimeError, ConnectionError, TimeoutError, ValueError, KeyError, empty list, None, two kinds of garbage), 4 through the real client (HTTP 401/403/429/500), the default path with no token, a failed commit, and a wrong identity attribute. Each: `ok` is False, zero accounts changed, zero audit rows. Mutation test: removing the fail-closed branch and the zero-match guard fails 28 of these.)*
- [x] the job never re-enables an account under any input
*(2026-10-16: a disabled account that is ACTIVE in Okta stays disabled, and `okta_sync.py` contains no assignment of `is_active` to True.)*
---
**Decided 2026-10-16 (Matt, in chat):** which Okta statuses disable an account.
The spec says "a non-active status". Read literally, that also catches
`PASSWORD_EXPIRED`, `LOCKED_OUT` and `RECOVERY`, which people leave on their
own, and this job never re-enables, so one AD password expiry would lock someone
out of the app until an admin stepped in. Decided: an account disables when it is
**missing** from Okta's list, or every Okta entry for it is **SUSPENDED** or
**DEPROVISIONED** (`DISABLING_STATUSES` in `server/okta_sync.py`). Okta's default
user list already leaves out DEPROVISIONED users, so those arrive as "missing".
**Added 2026-10-16, inside decision 4's "anything the sync can't confidently
parse":** if there are active local accounts and **none** of them matches any
Okta user, the run is refused and nothing changes. That pattern means the
identity attribute is wrong (`OKTA_SYNC_PROFILE_ATTRIBUTE`), not that the whole
company left. It is a guard, not a new feature. Raise it in the PR.
**Open, for the PR / Nick:** which Okta profile attribute equals the
`preferred_username` claim in this tenant. The client assumes `profile.login`,
Okta's default. Confirm before the first live run. A wrong value is caught by the
zero-match guard above only when it matches nobody at all; a value that matches
some accounts and not others would disable the rest. See `BL-069`.
### T13.4 — D18: run it on a schedule
- **Items:** `D18`
- **Depends on:** T13.3
- **Blocks:** T13.6
- **Surface:** `server/`, `docker-compose.yml`
- **Files:** `server/app.py` (startup hook) or a new sidecar per the `backup`
container's pattern — task's call, per the decision record's noted
alternative
**Do:** Wire T13.3 to run on an interval (proposed 15 minutes, env-overridable
— e.g. `OKTA_SYNC_INTERVAL_SECONDS`, matching `BACKUP_INTERVAL_SECONDS`'s
naming). Default choice is an in-process background task in the `api`
container; if a separate container is chosen instead, follow the `backup`
service's shape (own Dockerfile or script, `internal` network plus whatever
egress reaching Okta requires — check whether `outbound` as currently defined
is sufficient or Okta needs a distinct allowance).
**Done when:**
- [x] the job runs automatically on the configured interval without manual
invocation
*(2026-10-16: a daemon timer thread starts from the app's lifespan hook (`server/app.py`), checks once a minute, and ran the sync on its own in `tests/okta_sync_schedule_check.py`. With `OKTA_API_TOKEN` unset it starts nothing and the startup log says `Okta sync DISABLED`. Booted the real app both ways; the thread starts and ends with the app.)*
- [x] interval is env-configurable with a sane default
*(2026-10-16: `OKTA_SYNC_INTERVAL_SECONDS`, default 900. A non-number uses the default. A value under 300 is raised to 300. Passed through `docker-compose.yml`.)*
- [x] a container restart does not produce a duplicate/overlapping run, and a
slow cycle does not stack with the next one
*(2026-10-16: the api runs 2 gunicorn workers, so each starts its own timer. One row in `app_settings` (key `okta_sync`) is the gate: a worker claims a cycle with one atomic conditional UPDATE, so exactly one wins. Checked with 8 racing workers on a file database, 15 rounds plus the first-insert race: exactly 1 winner every time. The row outlives a restart, so a restart inside the interval starts nothing. A slow run cannot overlap the next: a fetch is capped at 120 s in total and the interval cannot go below 300 s. Inside one process the thread runs one cycle at a time. This needed no migration: `app_settings` already exists.)*
---
**Decided 2026-10-16 (task's call, per this task's "Do"):** in-process timer in the
`api` container, as D18 proposed, not a sidecar. The `backup` pattern was the
alternative; it would need its own image and would not remove the 2-worker
duplicate question, because the sync reads the application database anyway.
**Network check, as the task asked:** the `outbound` network the `api` service
already joins is an ordinary bridge with a gateway, so Okta is reachable with no
network change. The host firewall must allow https to the Okta org.
**One change to T13.2's code:** `fetch_users()` now also stops after 120 seconds in
total (`TOTAL_DEADLINE_SECONDS`), which is what lets the scheduler promise no
overlap. Tested in `tests/okta_sync_check.py` section 8.
**T13.4 follow-up, 2026-10-16:** the sync's INFO log lines did not appear in
`docker compose logs api`. Nothing in the app configures Python logging, and
gunicorn configures only its own loggers, so INFO from any `wpsuite.*` logger is
dropped. The T13.4 check booted the app in-process with logging switched on,
which hid it. With the exact Dockerfile launch (2 gunicorn workers), a healthy
start printed zero Okta lines. `okta_sync.py` now gives its own logger one stderr
handler (only when the root logger has none). Re-run with the real gunicorn
command: `configured` and `scheduled` (or `DISABLED`) now print once per worker.
No other module's output changed. App-wide logging stays a separate question
(BL-028 is the same family).
### T13.5 — D18: admin visibility
- **Items:** `D18`
- **Depends on:** T13.3
- **Blocks:** T13.6
- **Surface:** `html/`
- **Files:** `html/users.js` / `html/admin.js` (wherever audit history is
already surfaced)
**Do:** Confirm an auto-disable reads clearly wherever admins already look at
account history — the actor string (`system:okta_sync`) should be
self-explanatory in context, not require reading server logs to understand.
Do not build new UI beyond making sure the existing audit surface renders this
actor sensibly. Email/notification to admins on auto-disable is a noted
fast-follow (the decision record flags it as open, not required here) — do not
build it in this task; log it instead if it's tempting to add.
**Done when:**
- [x] an auto-disabled account's audit entry is visible and legible in the
existing admin UI without special-casing
*(2026-10-16: run on the PC against the real page, 1440px and 390px, `tests/okta_sync_ui_check.py` 28/28. The admin Activity log card reads: Who `system:okta_sync`, Action `user deprovisioned by sync`, Type `user`, Item the username, Detail `okta_status: SUSPENDED, via: okta_sync` (a person missing from Okta reads `okta_status: MISSING`). Search for `okta_sync` and the User accounts filter both find the rows. No sideways scroll, no JavaScript errors. The disabled account is refused on its very next request; an account still active in Okta is not. A project super user does not see these rows: they belong to no project, so only admins do, the same as other user-management events. No code in `html/` changed.)*
- [x] nothing here silently assumes `CR-019`'s activity view exists yet — this
must work standalone
*(2026-10-16: the check reads only the Activity log (AuditLog) card. The new Activity & usage card is on the page but is not read.)*
---
**No UI was built.** The existing card renders the new rows legibly as is. Two
things you might expect and did not get, both on purpose: no admin email on an
auto-disable (D18 calls it a fast-follow; not built, not logged as a new item
because D18 already names it), and no per-user history view in the User
Directory (it has none today, and the task says not to build one).
### T13.6 — D18: verification
- **Items:** `D18`
- **Depends on:** T13.4, T13.5
- **Blocks:** nothing
- **Surface:** `server/`
- **Files:** as touched above
**Do:** Full verification per `CLAUDE.md`. Beyond the usual run: specifically
re-run T13.3's fail-closed test in isolation and confirm it still passes after
T13.4's scheduling wrapper is in place — the scheduling layer must not
introduce a path that swallows the client's error and proceeds anyway.
**Done when:**
- [ ] all `D18` acceptance criteria in `decisions-2026-09-17.md` are met or a
failure is stated with a reason
- [x] the fail-closed behavior is verified end to end through the scheduled
wrapper, not just the bare sync function
- [x] smoke test and `seed_demo.py` both still pass
- [x] full test suite passes (48 of 50; the two failures are the known ones, see the result note below)
---
## Wave 13 exit criteria
- [x] `AUTH_SESSION_HOURS` defaults to 8, documented (D25 replaced the 2 hours in this line; T13.1's idle timer is still held back)
- [x] the sync job runs on a schedule and correctly disables accounts Okta no
longer shows as active
- [x] every auto-disable is individually audited with a clearly non-human actor
- [x] an Okta API failure of any kind changes zero accounts — verified, not
assumed
- [x] the sync never re-enables an account
- [ ] `D18` fully accounted for, no open acceptance criteria
## Release note (2026-10-16)
T13.1 (commit `6cde6e3`) is held back from `main` until T14.1 lands, per the merge
rule in `wave-12.md`. The release branch contains wave 11 and T13.2 to T13.5.
`tests/session_idle_check.py` skips (exit 0) when `auth.maybe_refresh_token` is absent.
Sessions keep the flat `AUTH_SESSION_HOURS` lifetime (default 12) until T13.1 lands.
T13.6 result (2026-10-16, merged release tree, Edge on the PC): 47 test scripts, 45 pass.
The two failures are not regressions. `token_check.py` exits 2 (CLI usage detail).
`generalinfo_check.py` finds an old color-literal gap in `wp-creation-styles.css`.
Alembic upgrade, downgrade, upgrade passed. Import check passed.
## T13.7: D22 dry run for the Okta sync
**Item:** D22 (`docs/waves/decisions-2026-10-16.md`). Extends T13.3 and T13.4.
**Done when:**
- [x] `OKTA_SYNC_DRY_RUN` on: the sync fetches and matches, logs the accounts it would disable, and changes no account and writes no audit row (`tests/okta_sync_dry_run_check.py`)
- [x] both guards (zero match, fail closed) still apply in a dry run
- [x] flag off or unset: behavior is unchanged (`okta_sync_job_check.py` still passes)
- [x] the startup line names the mode
- [x] `.env.example`, `docker-compose.yml` and `DEPLOYMENT.md` describe the flag and the first-deploy steps
- [x] mutation check: if the dry-run guard is removed, the new test fails
Result 2026-10-16: all Okta sync tests and `security_check.py` pass in the sandbox.
This task changes no frontend file.
## T13.8: run the stack in Docker on a PC
**Item:** D22 (`docs/waves/decisions-2026-10-16.md`), as a way to check T13.7 and later builds before a deploy.
**Done when:**
- [x] one command starts the api, the database and the web server from the shipped compose file plus `docker-compose.local.yml`
- [x] `http://localhost:8080/api/health` answers 200 and `/api/projects` answers 401 without a session
- [x] every migration runs on Postgres (head `8e2cb3003f8a`)
- [x] the api log shows the `Okta sync` startup line
- [x] the steps are in `DEPLOYMENT.md`, "Run the stack on your own PC"
- [x] the shipped `docker-compose.yml` and nginx config are not changed
Result 2026-10-16 (Docker Desktop 29.7.2, Compose 5.4.0, Windows): passed. Three
things in the shipped files stop a local run. The override handles the first two
and the third is logged:
1. The web server publishes no port (the host's reverse proxy reaches it). The override publishes 8080 on loopback.
2. An existing `proxy` network on a PC can have no gateway, and Docker then does not publish ports. The override uses its own bridge. The stock nginx `default.conf` also answers "localhost" first and the app server answers only wp.controls.dev, so the override removes the stock file and makes the app server the default.
3. The backup container cannot start on Windows: `scripts/entrypoint.sh` has CRLF line endings (BL-070). The override turns the backup service off.
## T13.9: D23 sign in as an OAuth service app (private_key_jwt)
**Item:** D23 (`docs/waves/decisions-2026-10-16.md`). Extends T13.2.
**Done when:**
- [x] with `OKTA_SYNC_CLIENT_ID` and `OKTA_SYNC_PRIVATE_KEY`, the sync gets an access token and lists users with `Bearer` (`tests/okta_sync_oauth_check.py`)
- [x] the assertion verifies against the public key: `iss`, `sub`, `aud`, `kid`, life, `jti`
- [x] RSA and EC keys; JSON and base64 forms
- [x] a failed token request, a missing token, a DPoP token, a public key, or an unreadable key raises, and the users endpoint is never called
- [x] `OKTA_API_TOKEN` still works and wins when both are set
- [x] the startup line names the credential and holds no secret; no error message holds key text
- [x] mutation checks: Bearer to SSWS, wrong `aud`, DPoP accepted, public-key check removed: each makes the test fail
- [x] `.env.example`, `docker-compose.yml`, `DEPLOYMENT.md` describe the variables and the Okta app setup
Result 2026-10-16: sandbox, fake Okta. Not yet run against the real tenant.
No new dependency: PyJWT signs with `cryptography`, which Authlib already installs.
T13.9 follow-up (2026-10-16): the first live deploy failed with "OKTA_SYNC_PRIVATE_KEY is not a
JWK as JSON or as base64 of JSON". The first loader used strict base64, which rejects any
space or line break, and it gave one message for every failure. The loader now forgives quote
marks, spaces, line breaks, missing padding, the URL-safe alphabet, a byte-order mark and a
one-key JWK Set, and each failure names its step (counts only, no key text). The new checks
in `tests/okta_sync_oauth_check.py` fail against the first loader and pass now.
T13.9 follow-up 2 (2026-10-16): the second deploy decoded the value but the text was not JSON.
The loader now also names a PEM key, adds missing outer braces, unwraps a value that was encoded
twice, and its "not JSON" message gives the length and the kind of first character (counts only).
T13.9 follow-up 3 (2026-10-16): the third deploy decoded the base64 to 1,760 characters that start
with a brace, but Python's strict JSON reader refused them (PowerShell accepted the same text).
The loader now accepts control characters such as line breaks inside a string, curly quote marks and
a trailing comma, removes spaces inside the key's number fields, and a real syntax error reports the
line and column (position and rule only, never key text).
## T13.10: D24 DPoP for the Okta Management API
**Item:** D24 (`docs/waves/decisions-2026-10-16.md`). Extends T13.9.
**Done when:**
- [x] auto mode: a plain Bearer token is tried first; `invalid_dpop_proof` makes the sync retry with DPoP (`tests/okta_sync_dpop_check.py`)
- [x] each proof verifies: header (`dpop+jwt`, ES256, public jwk only), signature, `htm`, `htu` without query, current `iat`, one-time `jti`, `ath` on API calls, same key for token and API calls
- [x] nonce challenges at the token endpoint and at the API are answered once, and the retries are bounded
- [x] paging: a new proof for every page
- [x] mode `on` and mode `off`; a DPoP-bound token with mode off is refused; each run makes a new key
- [x] no secret in any message or the startup line; the SSWS path is unchanged
- [x] mutation checks: query kept in `htu`, `ath` omitted, `jti` reused, nonce not retried (token and API), private key in the proof header, and auto detection removed each make a test fail
- [x] `.env.example`, `docker-compose.yml`, `DEPLOYMENT.md` updated
Result 2026-10-16: sandbox, strict fake Okta. Not yet run against the real tenant.
## Live run and final verification (2026-10-17)
**Live deploy.** The api container runs the build from `main` (`1a3468e`) in Portainer. It signs in with the
service-app key (`private_key_jwt`) and DPoP. The first dry run (2026-10-17 18:31) fetched 2,564 Okta users and
checked 42 active local accounts. 41 matched (so `profile.login` is the right attribute) and 1 would be disabled:
`admin`, Okta status MISSING. Matt confirmed that `admin` is a leftover account, and switched the dry run off.
Live run log lines (api log):
- 18:46:21 `Okta sync disabled 'admin' (Okta status: MISSING)`, then `done: 2564 Okta user(s), 42 active local account(s) checked, 1 disabled`
- 19:01:50 `done: ... 41 active local account(s) checked, 0 disabled`
- 19:16:30 `done: ... 41 active local account(s) checked, 0 disabled`
This shows: the job runs on its schedule (every 900 seconds), it disables an account that Okta no longer lists, a
disabled account is not touched again and gets no new row, and the account count fell from 42 to 41.
The first session logs of fail-closed behavior are in the earlier runs: with a rejected token, a bad key, and an
unreadable key, the sync logged `SKIPPED, no account was changed` every time and no account changed.
**Admin recovery.** `python -m server.manage_users list | promote | enable | disable` works from the Portainer
console of `wp_api` (no `docker compose exec` prefix). `promote` changes an existing account. It does not create one.
**Full verification on the final tree (`1a3468e`).**
- Test suite on Matt's PC (Edge): 50 scripts, 48 pass. The two failures are not regressions:
`token_check.py` exits 2 (it needs `--out` or `--compare`), and `generalinfo_check.py` finds an old colour literal
(`rgba(`) in `wp-creation-styles.css`.
- `server/smoketest.py`: 27 of 27 checks pass. `server/seed_demo.py`: runs clean. Both ran in the sandbox against
a throwaway SQLite database. Note: use `http://127.0.0.1:PORT`, not `http://localhost:PORT`, because Python's cookie
jar does not send a cookie set for `localhost` (logged as BL-071).
- The Okta tests (`okta_sync_check`, `_job_`, `_schedule_`, `_dry_run_`, `_oauth_`, `_dpop_`, `_ui_`) and `security_check.py` pass.
**Still open for wave 13.**
- `T13.1` (idle timeout) is held back from `main` until T14.1 lands. The exit criterion "`AUTH_SESSION_HOURS` defaults to 2"
is not met, and it conflicts with T13.1's own text (8 hours). Matt has not chosen.
- ~~"Every auto-disable is individually audited with a clearly non-human actor" was not checked after the live run.~~
Checked 2026-10-17: the Admin Console activity list shows one row, actor `system:okta_sync`, action "user deprovisioned
by sync", item `admin`, detail `okta_status: MISSING, via: okta_sync`, at 11:46 AM local (18:46 UTC).
- `D18` cannot be called fully accounted for while T13.1 is outstanding.
- `BL-069` (a cap on one run and a last-admin guard) is open.
## T13.11: D25 the session limit is 8 hours
**Item:** D25 (`docs/waves/decisions-2026-10-16.md`).
**Done when:**
- [x] `AUTH_SESSION_HOURS` defaults to 8 in `server/auth.py`, `docker-compose.yml`, `server/.env.example`, `DEPLOY-login-portal.md` (`tests/session_default_check.py`)
- [x] the token expiry and the cookie max-age follow it; a set value overrides it
- [x] mutation check: putting the default back to 12 makes the test fail
- [x] the wave 13 exit criterion that said 2 is updated
Effect on the live system: after the next deploy, sessions last 8 hours instead of 12. To get the same result
before a deploy, set `AUTH_SESSION_HOURS` to `8` in the Portainer stack variables.
## Gate for T13.1 (2026-10-17)
T13.1 now has a written gate: `docs/waves/wave-14.md`. It lands in T14.4, after T14.1 (a `401` holds the outbox and does not
destroy it) and T14.2 (the offline window follows the server) are done. The reason is a fault found in the outbox on
2026-10-17: it marks a `401` as a permanent rejection, so an idle timeout would reject a crew's queued changes for good.
That fault already exists with the 8 hour limit, to a lesser degree.

188
docs/waves/wave-14.md Normal file
View File

@@ -0,0 +1,188 @@
# Wave 14: The outbox keeps every change, and the sign-in limit cannot destroy it
**Items:** `D26`, `D27`, `BL-061` (`B5` is context only: its wording rules stay, and this wave does not widen it)
**Depends on:** wave 13 merged (it is, except T13.1). Independent of wave 15.
**Decision record:** `docs/waves/decisions-2026-10-17.md` (D26 and D27 are **accepted** as of 2026-10-17, see "Open questions")
**Unblocks:** `T13.1` (idle timeout). T13.1 must not reach `main` before T14.1 and T14.2 are done.
**Branch rule:** one task for each pull request. Cut each branch from `main`.
## Why this wave exists
The outbox in `html/project-data.js` queues each save in `localStorage` and sends it later. A crew in the field
can work with no signal, and the changes go out when the network returns. The code was checked on 2026-10-17
against the 8 hour session limit (D25) and the idle timeout that T13.1 adds. It has four faults that lose or hide
work. They are small now. The idle timeout would make them common.
| # | Fault | Where | Result |
|---|---|---|---|
| 1 | A `401` (session ended) or `403` is treated like any other `4xx`. The change is marked **dead** and is never sent again. | `project-data.js`, `flush()` | A crew that works past the session limit, or past an idle timeout, comes back online and every queued change is rejected for good. The badge says "rejected by the project". The change is not lost from the device, but nothing can send it. |
| 2 | `qWrite()` swallows a storage error. | `project-data.js`, `qWrite()` | If the device storage is full, a change is **not queued and nobody is told**. |
| 3 | The queue belongs to the browser, not to a person. A queued change has no owner. | `project-data.js`, `enqueue()` | On a shared device, person B signs in and the queue of person A is sent under the name of person B. |
| 4 | The offline sign-in window is a fixed 12 hours in the browser. | `auth-guard.js`, around line 183 | It does not follow `AUTH_SESSION_HOURS` (8 since D25). It also slides: every online page load resets it. The browser and the server disagree about when the session ends. |
Also: the sync badge sits over page controls (`BL-061`) and cannot be used with a keyboard or a finger. The
badge is a `div` with `pointer-events:none`. A badge that must ask a person to sign in again needs a real link or button.
## Rules for this wave
- Do not use a `div` with `onclick` (`CLAUDE.md`, accessibility). The sign-in action is an `<a>` or a `<button>`.
- Status text goes in an `aria-live` region. The badge already has `role="status"`. Keep it.
- Server work is needed for T14.2 (one new field). Do not store session data in `localStorage` as the source of truth.
- Colors, spacing and type come from the shared tokens. Do not add a raw hex value to a page stylesheet.
- Frontend checks: run the app on a throwaway SQLite database, check at 390 px and 1440 px, capture before and after
screenshots, run the smoke test (use `http://127.0.0.1`, see `BL-071`).
- Do not fix other problems you see. Log them in `docs/waves/backlog.md`.
---
### T14.1: D26 a sign-in problem never destroys a queued change
- **Items:** `D26`
- **Depends on:** nothing (first task)
- **Blocks:** T14.2, T14.3, T13.1
- **Surface:** `html/`
- **Files:** `html/project-data.js`, `html/auth-guard.js`, a new `tests/outbox_check.py`
**Do:**
1. In `flush()`, treat `401` as **held**, not dead. Stop the run. Do not add to `tries`. Do not mark any change
dead. Keep the whole queue in its order. Set a "held for sign-in" state that the badge can read through
`ProjectData.syncStatus()` (add a field, for example `held: true`).
2. Keep `403` as a permanent refusal, but say why in the reason text ("you no longer have access to this project").
Keep `409` (archived project) and the other `4xx` as they are.
3. When the person signs in again (the page loads with a good session, or `wp-auth-ready` fires), clear the held
state and flush at once. The `online` event, the 20 second timer and the tab-focus trigger must **not** clear it
early: a held queue sends only after a good session check.
4. Make `qWrite()` return `true` or `false`. If it fails, `enqueue()` must tell the caller and raise the
`device full` state. The change stays in the page memory and the form stays open. The badge says the change is
**not saved on this device**. Do not drop it silently.
5. Give each queued change an `owner` (the signed-in username) when it is queued. At flush time, if `WP_USER` is not the
owner, hold that change and name the owner in the message. Do not send it. Do not discard it. (See open question 1.)
6. Make sure `auth-guard.js` does not delete the queue when it sends a person to `login.html` after a `401`. It must
keep `wp_sync_outbox_v1`.
**Done when:**
- [x] a queue of N changes meets a `401`: all N are still in the queue, none is dead, `tries` did not change, and the badge state is "held"
- [x] after a good sign-in, all N changes reach the server in their original order, and the queue is empty
- [x] a `403` marks only that change dead, with a reason that names the cause
- [x] a full device store (a forced `QuotaExceededError`) is reported to the caller and the badge. No change disappears silently
- [x] a change queued by person A is not sent when person B is signed in; it is sent after A signs in
- [x] `tests/outbox_check.py` covers each case above and fails when `flush()` is put back to the old rule (a mutation check)
- [x] the full suite passes (2026-10-17: 52 of 54 checks pass; the 2 others are `generalinfo_check` and `token_check`, known and not regressions). `mobile_check` covers 390 px. No before and after screenshots were taken: the badge looks the same until T14.3. The badge words for the new states are plain text. T14.3 gives them their final wording and the Sign in control.
**Built (2026-10-17):** commit `73f0b5e`. Open question 1 as accepted: another person's changes are held and the badge names the owner. A change queued before the sign-in check finished has no owner and can be sent by anyone, as before.
---
### T14.2: D27 the offline sign-in window follows the server
- **Items:** `D27`
- **Depends on:** T14.1
- **Blocks:** T13.1
- **Surface:** `server/` and `html/`
- **Files:** `server/app.py` (`whoami()`), `html/auth-guard.js`, `html/wp-format.js` (line 223 writes the cache), tests
**Do:**
1. Add one field to `/api/auth/me`: `session_expires_at` (the `exp` claim of the session, as a UTC time). Read the claim from the
token that `get_current_user` already validated. Do not add a database column. When T13.1 lands, this field is the
current `exp` and moves with each refresh.
2. In `auth-guard.js`, store `session_expires_at` in `wp_auth_cache`. Replace the fixed `12 * 3600 * 1000` with a rule
that uses it. The rule for this wave (see open question 2):
- offline, the app **opens** while the current time is before `session_expires_at`;
- after that time, offline, the app still opens in a **read and queue** mode: the person can see cached data and add
changes, which go to the outbox and are held (T14.1) until a good sign-in. The app does not go to `login.html`
while it has no network, because that page cannot load an Okta sign-in without a network;
- online, a `401` goes to `login.html` as now, and the outbox is kept (T14.1).
3. Stop sliding the window. A page load must not extend `wp_auth_cache`. Only the server can extend a session.
4. Show the person which mode they are in. One line: "Working offline. Sign in again when you have a signal to send your changes."
Use the same `aria-live` region as the badge.
**Done when:**
- [x] `/api/auth/me` returns `session_expires_at`, and it matches the token (`tests/session_expiry_check.py`, no database change)
- [x] the fixed 12 hours is gone from `auth-guard.js`
- [x] opening the app offline before the time works; opening it offline after the time gives read-and-queue mode and does not redirect
- [x] online with a dead session: redirect to `login.html?next=...`, queue kept, changes sent after sign-in (the send after sign-in is the T14.1 test, section 3)
- [x] a page load does not extend the cached window
- [ ] the mode line is announced by a screen reader (`aria-live`) and fits at 390 px. The line is in the badge, which has `role="status"` (checked). It is **not** checked at 390 px or with a real screen reader. Do this in T14.3, which rebuilds the badge.
**Built (2026-10-17):** `tests/session_expiry_check.py` (22 checks) fails when the old 12 hour rule is put back (7 failures). The mode line shows only in read-and-queue mode, not when the app opens offline before the session ends, because "sign in again" would be wrong while the session is still good. In read-and-queue mode the outbox does not send at all until a page load gets a good answer from the server. Saving display preferences refreshes the cached user and keeps the end time (`wpCacheAuth`).
---
### T14.3: D26 and BL-061 the sync badge shows the true state and gets out of the way
- **Items:** `D26`, `BL-061`
- **Depends on:** T14.1
- **Blocks:** nothing
- **Surface:** `html/`
- **Files:** `html/project-data.js` (`renderSyncBadge`), `html/wp-chrome.css` or the shared token file, `tests/a11y_check.py`, `tests/autosave_check.py` (it reads the badge strings)
**Do:**
1. Show these states, each with its own words that name the **project**, never a draft save (the `B5` rule that wave 4 set): `sending`, `retrying`,
`held for sign-in` (with a link or button "Sign in"), `not saved on this device` (store full), `rejected` (with the reason),
and `all sent` (hides after a short time).
2. Make the badge usable: the "Sign in" and "Dismiss" controls are `<a>` or `<button>` elements with a visible focus ring
and a touch target of at least 44 px. Remove `pointer-events:none` for the states that have controls.
3. Keep the badge clear of page controls (`BL-061`): do not show it on pages that have no outbox work (the console
pages), and place it so that it does not cover the bottom-right action buttons at 390 px or 1440 px.
4. Keep text contrast at 4.5:1 or better in every state.
5. Update the tests that read the badge text (`autosave_check.py` reads its symbols; `a11y_check.py` reads the element).
**Done when:**
- [ ] each state shows in a browser test, with its text, at 390 px and 1440 px
- [ ] "Sign in" works with the keyboard and by touch, and goes to `login.html?next=...`
- [ ] the badge covers no control on the User Directory at 1440 px or the bulk bar at 390 px (the two places `BL-061` named)
- [ ] no badge on the console pages
- [ ] the contrast check passes for every state
- [ ] `autosave_check.py` and `a11y_check.py` pass
---
### T14.4: verify the wave and release T13.1
- **Items:** `D26`, `D27`, `BL-061`, `T13.1`
- **Depends on:** T14.1, T14.2, T14.3
- **Blocks:** nothing
**Do:** Run the full suite, the smoke test and `seed_demo.py`. Then test the real case by hand on a phone or tablet:
sign in, set the device to airplane mode, make three changes in Field View, wait until the session limit has passed
(use `AUTH_SESSION_HOURS=1` on a test stack with a clock change, or a forced expiry), turn the network on, and check that the badge shows
"held for sign-in", that signing in sends all three, and that the server has them. Then rebase `feat/waves-11-13`
T13.1 (`6cde6e3`) onto `main`, run `tests/session_idle_check.py` and the same hand test with the idle timeout on, and merge T13.1 as its own pull request.
**Done when:**
- [ ] the hand test passes on one iPhone and one Android device (or the failure and the reason are written here)
- [ ] the full suite and the smoke test pass on the final tree
- [ ] T13.1 is on `main`, and `tests/session_idle_check.py` runs (it no longer skips)
- [ ] the wave 13 exit line about D18 is ticked
---
## Open questions
**Answered 2026-10-17 (Matt accepted the proposal for questions 1, 2 and 3).** Question 4 is still open and belongs to `T13.1`.
1. **Shared devices.** If person B signs in on a device that holds unsent changes from person A, what must happen?
Proposed: hold A's changes, name A in the badge, send them when A signs in. Do not send them as B and do not discard them.
Other choices: allow B to discard A's changes after a confirmation; or send them as B (not proposed, wrong author in the log).
2. **Offline window.** Proposed for this wave: the app opens offline until the session ends, and after that works in read-and-queue mode.
The 7 day offline window (`BL-062`, Q2) is a **later** wave. Do you accept this order, or must 7 days come first?
3. **Device full.** Proposed: do not drop the change, keep it in the page memory, and tell the person. Is "tell the person and keep the
form open" enough, or must the app also stop new edits?
4. **Idle time.** T13.1 proposes 30 minutes. It is a proposed value, not a decision. It matters here because every minute of idle
time offline ends the session on the server side.
## Wave 14 exit criteria
- [ ] a session that ends while changes are queued loses none of them
- [ ] a full device store is never silent
- [ ] a change is sent only under the signed-in person who made it
- [ ] the offline window follows the server (no fixed 12 hours in the browser)
- [ ] the badge shows the true state, is usable, and covers no control
- [ ] T13.1 is merged to `main` and `D18` is fully accounted for

View File

@@ -34,12 +34,12 @@
/* Every container admin.js paints a table into is a scrollport of its own, so a
sticky header always has something to stick to rather than sliding up behind
the app bar. Same rule as console.css's .tscroll. */
#comments-admin, #audit-admin, #notif-box, #usage-admin, #projects-table, #defmem-table{
#comments-admin, #audit-admin, #notif-box, #activity-admin, #projects-table, #defmem-table{
overflow:auto; max-height:min(70vh,640px); overscroll-behavior:contain; }
/* If admin.js wraps its table in its own .tscroll, the outer box steps aside so
one table never ends up with two scrollbars. */
#comments-admin:has(.tscroll), #audit-admin:has(.tscroll), #notif-box:has(.tscroll),
#usage-admin:has(.tscroll), #projects-table:has(.tscroll), #defmem-table:has(.tscroll){
#activity-admin:has(.tscroll), #projects-table:has(.tscroll), #defmem-table:has(.tscroll){
overflow:visible; max-height:none; }
/* Comment text and audit detail are the two columns you are actually here to
@@ -53,7 +53,7 @@
border:1px solid var(--border-strong); border-radius:0; margin-bottom:var(--s3); }
@media (max-width:900px){
#comments-admin, #audit-admin, #notif-box, #usage-admin, #projects-table, #defmem-table{
#comments-admin, #audit-admin, #notif-box, #activity-admin, #projects-table, #defmem-table{
max-height:none; }
}
</style>
@@ -175,16 +175,38 @@
<div id="audit-admin" class="note">Click refresh to load.</div>
</div>
<!-- USAGE LOGS -->
<!-- ACTIVITY &amp; USAGE (CR-019) -->
<div class="card">
<h2>Usage logs</h2>
<div class="sub">Engagement recorded by both tools — the work package creator and the SOP wizard —
sessions, actions and counts, with a download per tool (D5). Note: stored locally per browser,
so this reflects activity on <strong>this</strong> machine.</div>
<h2>Activity &amp; usage</h2>
<div class="sub">Who is using the suite, which tools, and how often — recorded server-side on every
sign-in and page open, kept indefinitely. Filter below, or export a CSV: raw (real usernames) for
internal use, or sanitized (each user replaced with a stable, non-reversible id) for feeding into
Power BI or another external reporting tool without carrying real identities.</div>
<div class="toolbar">
<button onclick="loadUsage()">Refresh</button>
<label for="act-from">From</label>
<input type="date" id="act-from" onchange="loadActivity()">
<label for="act-to">To</label>
<input type="date" id="act-to" onchange="loadActivity()">
<select id="act-project" onchange="loadActivity()"><option value="">All projects</option></select>
<input id="act-user" placeholder="Username…" oninput="loadActivity()">
<select id="act-tool" onchange="loadActivity()">
<option value="">All tools</option>
<option value="launcher">Launcher</option>
<option value="wizard">SOP wizard</option>
<option value="creator">Work package creator</option>
<option value="field_view">Field view</option>
<option value="admin">Admin console</option>
<option value="directory">User directory</option>
</select>
<button onclick="loadActivity()">Refresh</button>
</div>
<div id="usage-admin" class="note">Click refresh to load.</div>
<div id="activity-banner" class="banner" style="display:none"></div>
<div id="activity-admin" class="note">Loading…</div>
<div class="toolbar" style="margin-top:var(--s3)">
<button onclick="exportActivity(false)">Download CSV (raw)</button>
<button onclick="exportActivity(true)">Download CSV (sanitized)</button>
</div>
<div id="activity-export-banner" class="banner" style="display:none"></div>
</div>
<!-- DB SNAPSHOT -->
@@ -215,7 +237,6 @@
</div>
</div>
<script src="wp-usage.js"></script>
<script src="console-util.js"></script>
<script src="wp-dialog.js"></script>
<script src="admin.js"></script>

View File

@@ -24,7 +24,7 @@ function reveal(){
loadNotifications();
loadComments();
loadAudit();
loadUsage();
loadActivity();
}
function showDenied(){
document.getElementById('admin-denied').style.display='';
@@ -189,6 +189,19 @@ async function loadProjects(){
banner.style.display='none';
_adminProjects = json;
renderProjects();
populateActivityProjectFilter();
}
// The activity project filter reuses the same project list the Projects card
// already fetched — no second /api/projects call just to fill a <select>.
function populateActivityProjectFilter(){
const sel = document.getElementById('act-project');
if(!sel) return;
const cur = sel.value;
sel.innerHTML = '<option value="">All projects</option>' +
_adminProjects.slice().sort((a,b)=>String(a.name||'').localeCompare(String(b.name||'')))
.map(p => '<option value="'+uesc(p.id)+'">'+uesc(p.name||p.number||p.id)+'</option>').join('');
sel.value = cur;
}
function renderProjects(){
@@ -465,6 +478,90 @@ function renderAudit(){
'</tr>').join('')+'</tbody></table>';
}
// ── activity & usage (CR-019) ────────────────────────────────────────────────────
// Server-side, per-user activity — distinct from the "Activity log" card above
// (that's AuditLog: business mutations) and from the "Usage logs" card below
// (that's per-browser localStorage, D5/T7.10, on the way out per T11.6). This is
// /api/usage/summary and /api/usage/export: real rows, aggregated server-side,
// filterable by date/project/user/tool, and exportable raw or sanitized.
function _activityFilters(){
const q = new URLSearchParams();
const from = document.getElementById('act-from').value; if(from) q.set('from', from);
const to = document.getElementById('act-to').value; if(to) q.set('to', to);
const proj = document.getElementById('act-project').value; if(proj) q.set('project_id', proj);
const user = (document.getElementById('act-user').value||'').trim(); if(user) q.set('username', user);
const tool = document.getElementById('act-tool').value; if(tool) q.set('tool', tool);
return q;
}
async function loadActivity(){
const banner = document.getElementById('activity-banner');
const box = document.getElementById('activity-admin');
if(!box) return;
banner.style.display='none';
box.textContent = 'Loading…';
const { status, json } = await api('GET', '/api/usage/summary?'+_activityFilters().toString());
if(status===403){
banner.className='banner bad'; banner.style.display='';
banner.textContent = '✕ Your account can’t see suite-wide activity here — this needs admin, or Project Super User on at least one project.';
box.innerHTML=''; return;
}
if(status!==200 || !json){
banner.className='banner bad'; banner.style.display='';
banner.textContent = '✕ Could not load activity ('+apiError(status, json, 'load failed')+').';
box.innerHTML=''; return;
}
renderActivity(json);
}
function renderActivity(sum){
const box = document.getElementById('activity-admin');
if(!sum.event_count){ box.innerHTML = '<div class="note">No activity matches this filter.</div>'; return; }
const fmt = s => s ? wpFormatDateTime(s) : '—';
const bucket = (label, obj, take) => {
const entries = Object.entries(obj).sort((a,b)=> b[0].localeCompare(a[0])).slice(0, take);
if(!entries.length) return '';
return '<table class="kv" style="margin-top:8px"><caption style="text-align:left;font-weight:600;margin-bottom:4px">'+
uesc(label)+'</caption>'+entries.map(([k,v]) => '<tr><th>'+uesc(k)+'</th><td>'+v+' active user'+(v===1?'':'s')+'</td></tr>').join('')+
'</table>';
};
const perTool = Object.entries(sum.by_tool||{});
const toolTable = perTool.length ? '<table class="users"><thead><tr><th>Tool</th><th>Events</th></tr></thead><tbody>'+
perTool.map(([t,n]) => '<tr><td>'+uesc(t||'(login)')+'</td><td>'+n+'</td></tr>').join('')+'</tbody></table>' :
'<div class="note">No per-tool data.</div>';
const perUser = Object.entries(sum.per_user_last_active||{}).sort((a,b)=> String(b[1]).localeCompare(String(a[1])));
const userTable = perUser.length ? '<table class="users"><thead><tr><th>User</th><th>Last active</th></tr></thead><tbody>'+
perUser.map(([u,t]) => '<tr><td><strong>'+uesc(u)+'</strong></td><td style="color:var(--muted)">'+fmt(t)+'</td></tr>').join('')+
'</tbody></table>' : '<div class="note">No per-user data.</div>';
box.innerHTML =
'<div class="note">'+sum.event_count+' event'+(sum.event_count===1?'':'s')+' matched.</div>'+
'<div class="row" style="gap:var(--s4);flex-wrap:wrap;align-items:flex-start">'+
'<div>'+bucket('Active users by day', sum.active_users.by_day, 30)+'</div>'+
'<div>'+bucket('Active users by week', sum.active_users.by_week, 12)+'</div>'+
'<div>'+bucket('Active users by month', sum.active_users.by_month, 12)+'</div>'+
'</div>'+
'<h2 style="margin-top:16px">By tool</h2>'+toolTable+
'<h2 style="margin-top:16px">Per-user last active</h2>'+userTable;
}
async function exportActivity(sanitize){
const banner = document.getElementById('activity-export-banner');
banner.className='banner'; banner.style.display=''; banner.textContent='Preparing export…';
const q = _activityFilters();
if(sanitize) q.set('sanitize', 'true');
const { status, json } = await api('GET', '/api/usage/export?'+q.toString());
if(status!==200 || typeof json !== 'string'){
banner.className='banner bad';
banner.textContent = '✕ Export failed ('+apiError(status, json, 'export failed')+').';
return;
}
const blob = new Blob([json], { type:'text/csv' });
const a = document.createElement('a');
a.href = URL.createObjectURL(blob);
a.download = 'usage_export_'+(sanitize?'sanitized':'raw')+'_'+new Date().toISOString().slice(0,10)+'.csv';
document.body.appendChild(a); a.click(); a.remove();
setTimeout(() => URL.revokeObjectURL(a.href), 1000);
banner.className='banner ok';
banner.textContent = '✓ Downloaded the '+(sanitize?'sanitized':'raw')+' export.';
}
// ── notifications / email settings ──────────────────────────────────────────────
let _settings = {};
async function loadSettings(){
@@ -656,48 +753,6 @@ async function loadNotifications(){
'</tr>').join('')+'</tbody></table>';
}
// ── usage logs (read from this browser's localStorage) ──────────────────────────
// D5 / T7.10: the report for BOTH tools' recorded usage, in the one place an
// operator-facing readout belongs - behind the same admin gate as this whole
// page (gateByRole() below shows nothing else either). Data comes from
// wp-usage.js, the single implementation; the keys predate the move, so
// everything recorded before it is still here.
function loadUsage(){
const box = document.getElementById('usage-admin');
if(!box) return;
const tools = [
['Work package creator', WPUsage.KEYS.creator, 'wp-iwp-usage'],
['SOP wizard', WPUsage.KEYS.wizard, 'wp-suite-usage'],
];
let html = '';
tools.forEach(([label, key, prefix]) => {
const evs = (WPUsage.load(key).events) || [];
html += '<h2 style="margin-top:16px">' + uesc(label) + '</h2>';
if(!evs.length){
html += '<div class="note">No usage recorded in this browser yet.</div>';
return;
}
const byEvent = {}, sessions = new Set();
let first = evs[0].ts, last = evs[0].ts;
evs.forEach(e => {
byEvent[e.event] = (byEvent[e.event]||0)+1;
if(e.session) sessions.add(e.session);
if(e.ts < first) first = e.ts; if(e.ts > last) last = e.ts;
});
const fmt = v => v ? wpFormatDateTime(v) : '—';
html += '<table class="kv">'+
'<tr><th>Sessions</th><td>'+sessions.size+'</td></tr>'+
'<tr><th>Events</th><td>'+evs.length+'</td></tr>'+
'<tr><th>Range</th><td style="font-weight:600">'+fmt(first)+' → '+fmt(last)+'</td></tr></table>';
html += '<table class="users"><thead><tr><th>Event</th><th>Count</th></tr></thead><tbody>';
Object.keys(byEvent).sort().forEach(k => html += '<tr><td>'+uesc(k)+'</td><td>'+byEvent[k]+'</td></tr>');
html += '</tbody></table>';
html += '<div class="toolbar" style="margin-top:8px"><button onclick="WPUsage.download(WPUsage.KEYS.'+
(key === WPUsage.KEYS.creator ? 'creator' : 'wizard')+', ' + jsq(prefix) + ')">Download the full event log</button></div>';
});
box.innerHTML = html;
}
// ── access control: admins only ─────────────────────────────────────────────────
// auth-guard.js requires a login and sets window.WP_USER (firing 'wp-auth-ready').
// Show the console for admins; otherwise show the "Admins only" notice.

View File

@@ -125,16 +125,69 @@
// Nothing replaces it. Every signed-in page mounts the drawer, so there is no page
// left that would need a floating fallback pill.
// ── CR-019: usage ping ───────────────────────────────────────────────────
// One page_open event per authenticated load, sent from exactly ONE place
// (here) rather than from each page's own script - the shared-chrome lesson
// S4 and the token-drift lesson S5 both taught this codebase the hard way.
// Fire-and-forget: never blocks reveal(), never retries, never surfaces an
// error to the person using the app - a missed usage ping is not something
// anyone here should notice happening.
var TOOL_BY_PAGE = {
'index.html': 'launcher',
'work-package-suite.html': 'wizard',
'wp-creation-index.html': 'creator',
'field.html': 'field_view',
'admin.html': 'admin',
'users.html': 'directory'
};
function pingUsage() {
var page = (location.pathname.split('/').pop() || 'index.html');
var tool = TOOL_BY_PAGE[page] || page.replace(/\.html$/, '');
try {
fetch('/api/usage/ping', {
method: 'POST',
headers: { 'Content-Type': 'application/json' },
body: JSON.stringify({ tool: tool })
}).catch(function () {});
} catch (e) {}
}
function proceed(user) {
clearTimeout(safety);
window.WP_USER = user;
reveal();
if (window.WP_USER) {
window.wpFlags(); // start the feature-flag fetch; pages await it as needed
pingUsage();
try { document.dispatchEvent(new CustomEvent('wp-auth-ready', { detail: window.WP_USER })); } catch (e) {}
}
}
// D27. The offline sign-in window is the SERVER session, not a timer of the
// browser's own. /api/auth/me returns `session_expires_at` (the session token's
// exp), and that is what the cache keeps. Only a good answer from the server
// changes it: loading a page offline, or saving preferences, never extends it.
// (This replaced a fixed 12 hour window that slid forward on every page load and
// disagreed with AUTH_SESSION_HOURS.)
function readAuthCache() {
try { return JSON.parse(localStorage.getItem('wp_auth_cache') || 'null'); } catch (e) { return null; }
}
function writeAuthCache(user, expMs) {
try { if (user) localStorage.setItem('wp_auth_cache', JSON.stringify({ user: user, exp: expMs || 0, at: Date.now() })); } catch (e) {}
}
// For callers that refresh the cached user (the preferences dialog): the user
// changes, the end of the session does not.
window.wpCacheAuth = function (user) {
var c = readAuthCache();
writeAuthCache(user, c && c.exp);
};
// 'online' (a good answer from the server), 'offline' (opened from the cache
// before the session ended) or 'readqueue' (opened from the cache after it ended:
// cached data can be read and changes are queued, but nothing is sent until a
// good sign-in). The sync badge shows the one line for 'readqueue'.
window.WP_SESSION_MODE = 'online';
fetch('/api/auth/me', { headers: { 'Accept': 'application/json' } })
.then(function (r) {
if (r.status === 401 || r.status === 403) { try { localStorage.removeItem('wp_auth_cache'); } catch (e) {} goToLogin(); return; }
@@ -143,17 +196,26 @@
var user = data && data.user;
// Remember the last good auth so the PWA can open offline. The server is
// still the real gate; offline writes queue in the outbox until reconnect.
try { if (user) localStorage.setItem('wp_auth_cache', JSON.stringify({ user: user, at: Date.now() })); } catch (e) {}
var expMs = data && data.session_expires_at ? Date.parse(data.session_expires_at) : 0;
writeAuthCache(user, isNaN(expMs) ? 0 : expMs);
proceed(user);
});
})
.catch(function () {
// Offline / API unreachable: fall back to a recent cached auth if present,
// so the app (and the field view) still open without a network.
try {
var c = JSON.parse(localStorage.getItem('wp_auth_cache') || 'null');
if (c && c.user && (Date.now() - (c.at || 0)) < 12 * 3600 * 1000) { proceed(c.user); return; }
} catch (e) {}
// Offline / API unreachable: open from the cached sign-in so the app (and the
// field view) still work without a network. With no cache this device has
// never signed in, so there is nothing to open: go to the sign-in page.
var c = readAuthCache();
if (c && c.user) {
// Before the session ends: open as normal. After it (or when the cache has
// no end time, an older cache): read and queue. The page does NOT redirect,
// because login.html cannot reach Okta without a network.
var live = !!c.exp && Date.now() < c.exp;
window.WP_SESSION_MODE = live ? 'offline' : 'readqueue';
proceed(c.user);
try { document.dispatchEvent(new CustomEvent('wp-session-mode', { detail: window.WP_SESSION_MODE })); } catch (e) {}
return;
}
goToLogin();
});
})();

View File

@@ -245,25 +245,64 @@
// local cache still updates immediately, so rendering never waits on the network.
var OUTBOX_KEY = 'wp_sync_outbox_v1';
var _flushTimer = null, _backoff = 0, _flushing = false;
// D26. `_held` is set when the server answers 401: the session ended, so the
// queue waits for a good sign-in. It is cleared ONLY by `wp-auth-ready` (a page
// load with a good session check), never by the online event, the timer or tab
// focus, because none of those prove the session is back.
var _held = false;
// D26. Changes the device could not store (localStorage full) stay here, in page
// memory, and are still sent. They are lost if the page closes, so the badge says
// they are not saved on this device. They are never dropped without a word.
var _mem = [];
var _deviceFull = false;
function qRead() { try { return JSON.parse(localStorage.getItem(OUTBOX_KEY) || '[]') || []; } catch (e) { return []; } }
function qWrite(list) { try { localStorage.setItem(OUTBOX_KEY, JSON.stringify(list)); } catch (e) {} }
// Returns true when the list is stored. A storage error used to be swallowed here,
// so a full device lost the change and nobody knew (D26, fault 2).
function qWrite(list) {
try { localStorage.setItem(OUTBOX_KEY, JSON.stringify(list)); return true; }
catch (e) { return false; }
}
// Stored ops plus any held only in memory, stored ones first (send order).
function qAll() {
var q = qRead();
if (!_mem.length) return q;
var ids = {};
q.forEach(function (o) { ids[o.opId] = true; });
return q.concat(_mem.filter(function (o) { return !ids[o.opId]; }));
}
// Remove one op everywhere it can be held.
function qDrop(opId) {
_mem = _mem.filter(function (o) { return o.opId !== opId; });
qWrite(qRead().filter(function (o) { return o.opId !== opId; }));
}
function readQueueMode() { try { return window.WP_SESSION_MODE === 'readqueue'; } catch (e) { return false; } }
function sameUser(a, b) { return String(a || '').toLowerCase() === String(b || '').toLowerCase(); }
// Append an op, coalescing by (kind,key) so only the latest write per entity is
// queued. A delete supersedes any pending upsert for the same id.
// queued. A delete supersedes any pending upsert for the same id. Returns true
// when the change is stored on this device, false when it is only in memory.
function enqueue(op) {
var q = qRead();
if (op.kind === 'wp-del') {
q = q.filter(function (o) { return !(o.key === op.key && (o.kind === 'wp' || o.kind === 'wp-del')); });
} else {
q = q.filter(function (o) { return !(o.kind === op.kind && o.key === op.key); });
}
var supersedes = (op.kind === 'wp-del')
? function (o) { return o.key === op.key && (o.kind === 'wp' || o.kind === 'wp-del'); }
: function (o) { return o.kind === op.kind && o.key === op.key; };
var q = qRead().filter(function (o) { return !supersedes(o); });
_mem = _mem.filter(function (o) { return !supersedes(o); });
op.opId = op.kind + ':' + op.key + ':' + Date.now().toString(36) + Math.random().toString(36).slice(2, 6);
op.tries = 0;
op.owner = currentUser(); // D26: sent only when this person is signed in
q.push(op);
qWrite(q);
var stored = qWrite(q);
if (stored) {
// Room again: bring anything held in memory onto the device.
if (_mem.length && qWrite(q.concat(_mem))) _mem = [];
_deviceFull = _mem.length > 0;
} else {
_mem.push(op); _deviceFull = true;
}
notifySync();
scheduleFlush(0);
return stored;
}
function opRequest(op) {
@@ -275,41 +314,65 @@
});
}
// Apply `fn` to the op with this id, wherever it is held (device or memory).
function patchOp(opId, fn) {
var q = qRead(), hit = false;
for (var i = 0; i < q.length; i++) { if (q[i].opId === opId) { fn(q[i]); hit = true; break; } }
if (hit) qWrite(q);
for (var j = 0; j < _mem.length; j++) { if (_mem[j].opId === opId) { fn(_mem[j]); break; } }
}
function bumpTries(opId, err) {
var q = qRead();
for (var i = 0; i < q.length; i++) { if (q[i].opId === opId) { q[i].tries = (q[i].tries || 0) + 1; q[i].lastErr = err; break; } }
qWrite(q);
patchOp(opId, function (o) { o.tries = (o.tries || 0) + 1; o.lastErr = err; });
}
// Permanently-failed op (a 4xx client error) — keep it for visibility but stop
// retrying, so a rejected write can't loop forever. `err` is the server's own
// explanation when it sent one; it is what the sync badge shows the user.
function markDead(opId, err) {
var q = qRead();
for (var i = 0; i < q.length; i++) { if (q[i].opId === opId) { q[i].dead = true; q[i].lastErr = err; break; } }
qWrite(q);
patchOp(opId, function (o) { o.dead = true; o.lastErr = err; });
}
// Attempt every live op; successes are removed, 4xx client errors are marked
// dead (won't succeed on retry), transient failures (429/5xx/network) stay queued.
// True when this op is waiting for another person to sign in (D26). An op with no
// owner (queued before the sign-in check finished) can be sent by anyone, which is
// what every op did before owners existed.
function foreignOp(o) {
var me = currentUser();
return !!o.owner && !sameUser(o.owner, me);
}
function liveOps() { return qAll().filter(function (o) { return !o.dead; }); }
function sendableOps() { return liveOps().filter(function (o) { return !foreignOp(o); }); }
// Attempt every live op that belongs to the signed-in person; successes are
// removed. A 401 HOLDS the queue (D26): the session ended, which says nothing
// about the changes, so nothing is marked dead and no try is counted. A 403 or
// other 4xx is marked dead (it will not succeed on retry); transient failures
// (429/5xx/network) stay queued.
function flush() {
if (_flushing) return Promise.resolve();
var q = qRead().filter(function (o) { return !o.dead; });
// D27: opened offline after the session ended (read and queue). Changes are kept
// and wait; nothing is sent until a page load gets a good answer from the server.
if (_held || readQueueMode()) { notifySync(); return Promise.resolve(); }
var q = sendableOps();
if (!q.length) { notifySync(); return Promise.resolve(); }
_flushing = true; notifySync();
var chain = Promise.resolve(), anyFail = false;
var chain = Promise.resolve(), anyFail = false, halted = false;
q.forEach(function (op) {
chain = chain.then(function () {
if (halted) return; // a 401 stops the run; later ops keep their place
return opRequest(op).then(function (r) {
var status = r ? r.status : 0;
var done = r && (r.ok || (op.kind === 'wp-del' && status === 404)); // 404 on delete = already gone
if (done) { qWrite(qRead().filter(function (o) { return o.opId !== op.opId; })); }
if (done) { qDrop(op.opId); }
else if (status === 401) { halted = true; _held = true; }
else if (status >= 400 && status < 500 && status !== 429) {
// Refused once, refused forever — so the only useful thing left is the
// reason. A 409 here is the archived-project gate, whose detail tells
// the user the project is read-only and how to get it unarchived; a
// bare "HTTP 409" would leave them staring at a change that vanished.
return r.json().catch(function () { return null; }).then(function (j) {
var why = (j && typeof j.detail === 'string' && j.detail) || ('HTTP ' + status);
var detail = (j && typeof j.detail === 'string' && j.detail) || '';
var why = status === 403
? 'You no longer have access to this project.' + (detail ? ' (' + detail + ')' : '')
: (detail || ('HTTP ' + status));
markDead(op.opId, why);
});
}
@@ -320,7 +383,8 @@
return chain.then(function () {
_flushing = false;
notifySync();
if (qRead().filter(function (o) { return !o.dead; }).length) {
if (_held) return; // wait for wp-auth-ready; no timer retries a dead session
if (sendableOps().length) {
_backoff = anyFail ? Math.min((_backoff || 5000) * 2, 60000) : 0;
scheduleFlush(_backoff || 15000);
} else { _backoff = 0; }
@@ -336,14 +400,21 @@
// `failed` stays the total not-getting-through count (what listeners already
// read); `dead` splits out the ops the server has permanently refused, with the
// first reason it gave, because those two states need different words.
// D26 adds: `held` (the session ended; the queue waits for a sign-in), `others`
// and `otherOwner` (changes that belong to someone else, waiting for them),
// `deviceFull` and `unsaved` (changes that exist only in page memory).
function syncCounts() {
var q = qRead(), pending = 0, failed = 0, dead = 0, reason = '';
var q = qAll(), pending = 0, failed = 0, dead = 0, reason = '', others = 0, otherOwner = '';
for (var i = 0; i < q.length; i++) {
if (q[i].dead) { dead++; if (!reason && q[i].lastErr) reason = String(q[i].lastErr); }
else if (foreignOp(q[i])) { others++; if (!otherOwner) otherOwner = String(q[i].owner); }
else if ((q[i].tries || 0) >= 3) failed++;
else pending++;
}
return { pending: pending, failed: failed + dead, dead: dead, reason: reason, syncing: _flushing };
return {
pending: pending, failed: failed + dead, dead: dead, reason: reason, syncing: _flushing,
held: _held, others: others, otherOwner: otherOwner, deviceFull: _deviceFull, unsaved: _mem.length
};
}
ProjectData.syncStatus = syncCounts;
function notifySync() {
@@ -395,7 +466,20 @@
// Every string here now names the project explicitly. The draft indicator
// (WPAutosave.mountIndicator) is the one that speaks for the form, and the two
// can no longer be read as each other.
if (c.dead) {
// D26. These three are the states that used to be invisible or wrong. T14.3
// gives them their final wording and controls; until then they are plain text.
if (c.deviceFull) {
el.textContent = '✕ ' + c.unsaved + ' change' + (c.unsaved === 1 ? ' is' : 's are') + ' not saved on this device. Keep this page open until it is sent to the project.';
el.style.color = 'var(--wp-status-error-text)'; el.style.borderColor = 'var(--wp-status-error-border-a)'; el.style.background = 'var(--wp-status-error-bg)'; el.style.display = 'inline-flex';
} else if (readQueueMode()) {
// D27. The one line for read-and-queue mode. T14.3 owns the final wording
// and the sign-in control.
el.textContent = 'Working offline. Sign in again when you have a signal to send your changes.';
el.style.color = 'var(--wp-status-warning-text)'; el.style.borderColor = 'var(--cds-support-warning)'; el.style.background = 'var(--wp-status-warning-bg)'; el.style.display = 'inline-flex';
} else if (c.held) {
el.textContent = '⚠ Sign in again to send your changes to the project. They are kept on this device.';
el.style.color = 'var(--wp-status-warning-text)'; el.style.borderColor = 'var(--cds-support-warning)'; el.style.background = 'var(--wp-status-warning-bg)'; el.style.display = 'inline-flex';
} else if (c.dead) {
el.innerHTML = '<span>✕ ' + c.dead + ' change' + (c.dead === 1 ? '' : 's') + ' rejected by the project — not saved</span>' +
(c.reason ? '<span style="font-weight:400">' + esc(c.reason) + '</span>' : '');
el.style.color = 'var(--wp-status-error-text)'; el.style.borderColor = 'var(--wp-status-error-border-a)'; el.style.background = 'var(--wp-status-error-bg)'; el.style.display = 'inline-flex';
@@ -405,6 +489,9 @@
} else if (c.pending) {
el.textContent = '↻ Sending ' + c.pending + ' change' + (c.pending === 1 ? '' : 's') + ' to the project…';
el.style.color = 'var(--cds-text-secondary)'; el.style.borderColor = 'var(--cds-border-subtle)'; el.style.background = 'var(--cds-layer)'; el.style.display = 'inline-flex';
} else if (c.others) {
el.textContent = '⚠ ' + c.others + ' change' + (c.others === 1 ? ' belongs' : 's belong') + ' to ' + c.otherOwner + '. ' + (c.others === 1 ? 'It goes' : 'They go') + ' to the project when ' + c.otherOwner + ' signs in.';
el.style.color = 'var(--wp-status-warning-text)'; el.style.borderColor = 'var(--cds-support-warning)'; el.style.background = 'var(--wp-status-warning-bg)'; el.style.display = 'inline-flex';
} else {
el.textContent = '✓ Everything sent to the project';
el.style.color = 'var(--wp-status-success-text)'; el.style.borderColor = 'var(--wp-status-success-border-a)'; el.style.background = 'var(--wp-status-success-bg)'; el.style.display = 'inline-flex';
@@ -418,7 +505,12 @@
window.addEventListener('online', function () { _backoff = 0; scheduleFlush(0); });
window.addEventListener('storage', function (e) { if (e.key === OUTBOX_KEY) { notifySync(); scheduleFlush(0); } });
document.addEventListener('visibilitychange', function () { if (!document.hidden) scheduleFlush(0); });
setInterval(function () { if (qRead().filter(function (o) { return !o.dead; }).length) scheduleFlush(0); }, 20000);
setInterval(function () { if (sendableOps().length) scheduleFlush(0); }, 20000);
// D26: the one thing that releases a held queue. auth-guard fires this after a
// good /api/auth/me (or the offline cache), and the next flush finds out for
// itself whether the session is really good: a 401 holds the queue again.
document.addEventListener('wp-session-mode', function () { notifySync(); });
document.addEventListener('wp-auth-ready', function () { _held = false; _backoff = 0; notifySync(); scheduleFlush(0); });
} catch (e) {}
if (document.readyState === 'loading') { document.addEventListener('DOMContentLoaded', function () { notifySync(); scheduleFlush(0); }); }
else { setTimeout(function () { notifySync(); scheduleFlush(0); }, 0); }
@@ -428,7 +520,7 @@
// project so re-completing updates the same row.
ProjectData.pushSOP = function (projectId, sop, state) {
if (!projectId) return Promise.resolve(null);
enqueue({
return Promise.resolve(enqueue({
kind: 'sop', key: 'sop__' + projectId,
body: {
id: 'sop__' + projectId, project_id: projectId,
@@ -436,22 +528,21 @@
number: (sop && sop.project && sop.project.number) || '',
complete: true, created_by: currentUser(), data: { sop: sop, state: state }
}
});
return Promise.resolve(true);
}));
};
// Upsert a single Work Package. The local cache stays the source of truth for
// immediate rendering; the outbox guarantees the write reaches the server.
ProjectData.pushWP = function (p, projectId) {
if (!p || !p.id) return Promise.resolve(null);
enqueue({ kind: 'wp', key: p.id, body: pkgToServer(p, projectId) });
return Promise.resolve(true);
// D26: resolves false when the device could not store the change (it is still
// sent from memory, but it is lost if the page closes first).
return Promise.resolve(enqueue({ kind: 'wp', key: p.id, body: pkgToServer(p, projectId) }));
};
ProjectData.removeWP = function (id) {
if (!id) return Promise.resolve();
enqueue({ kind: 'wp-del', key: id });
return Promise.resolve(true);
return Promise.resolve(enqueue({ kind: 'wp-del', key: id }));
};
// Force a flush now and resolve when the queue drains (or a round-trip is done).

View File

@@ -24,7 +24,10 @@
The directory is one wide table, so the column exceptions live here: email is
the one cell long enough to stretch a row, and the two role dropdowns need
room for "Assistant Project Manager" without pushing Actions off screen. */
#users-table table td:nth-child(3){ max-width:230px; overflow:hidden; text-overflow:ellipsis; }
/* The email cell, by its class: CR-020's selection column moved it from third to
fourth on the managers' table, and a position-based rule would have followed
the wrong column. */
#users-table table td.ell{ max-width:230px; overflow:hidden; text-overflow:ellipsis; }
#users-banner:not(:empty), #scope-banner:not(:empty){ margin-bottom:var(--s3); }
#nu-projects{ margin-top:var(--s2); }
#nu-projects .pickrow{ padding:var(--s1) var(--s1); }
@@ -33,6 +36,45 @@
#nu-project-list{ max-height:200px; overflow:auto; border:1px solid var(--border); }
.whoami-chip{ font-size:12px; color:var(--muted); }
.whoami-chip strong{ color:var(--text); }
/* CR-020: row selection and the bulk bar. Tokens only (the token rule). */
#sel-banner:not(:empty), #bulk-banner:not(:empty), #csv-banner:not(:empty){ margin-bottom:var(--s3); }
#bulk-banner ul, #csv-banner ul{ margin:var(--s1) 0 0; padding-left:var(--s5); font-weight:400;
max-height:220px; overflow:auto; }
#csv-btn[hidden]{ display:none; }
table.grid th.selcell, table.grid td.selcell{ width:1%; padding:0 var(--s1); text-align:center; }
/* The label is the hit area: the 16px box alone is under the 24px floor. */
.selhit{ display:inline-flex; align-items:center; justify-content:center;
min-width:28px; min-height:28px; cursor:pointer; }
.selhit input{ width:16px; height:16px; margin:0; accent-color:var(--accent); cursor:pointer; }
.selhit input:focus-visible{ outline:2px solid var(--accent); outline-offset:2px; }
.selhit input:disabled{ cursor:default; }
/* Not --accent-soft: that is the admin pill's own fill (console.css, hover note). */
table.grid tbody tr.is-selected td{ background:var(--cds-layer-selected); }
table.grid tbody tr.is-selected td.selcell{ box-shadow:inset 3px 0 0 var(--accent); }
.bulkbar{ border:1px solid var(--accent); border-left:3px solid var(--accent);
background:var(--surface); padding:var(--s2) var(--s3); margin:0 0 var(--s3); }
.bulkbar[hidden]{ display:none; }
.bulkhead{ display:flex; align-items:center; justify-content:space-between; gap:var(--s2);
flex-wrap:wrap; margin-bottom:var(--s2); }
.bulkgroup{ display:flex; flex-wrap:wrap; align-items:center; gap:var(--s2);
padding:var(--s2) 0; border-top:1px solid var(--border); }
.bulkgroup > .bulklabel{ font-size:12px; font-weight:600; color:var(--muted); min-width:110px; }
.bulkgroup select{ height:var(--ctl); padding:0 var(--s1) 0 var(--s2); font:inherit; font-size:13px;
border:1px solid var(--border-strong); border-radius:0; background:var(--surface); color:var(--text);
max-width:100%; }
.bulkgroup select:focus-visible{ outline:2px solid var(--accent); outline-offset:-2px; }
/* The bulk delete dialog lists every affected username; a long list scrolls
inside the dialog instead of pushing its buttons off the screen. */
#wp-dlg-msg{ max-height:50vh; overflow:auto; }
@media (max-width:620px){
.bulkgroup > .bulklabel{ min-width:0; flex:1 1 100%; }
.bulkgroup select, .bulkgroup button{ flex:1 1 100%; }
}
@media (pointer:coarse){
.selhit{ min-width:44px; min-height:44px; }
.bulkbar button, .bulkgroup select{ min-height:44px; }
}
</style>
</head>
<body>
@@ -63,15 +105,33 @@
<div class="sub" id="people-sub"></div>
<div class="toolbar">
<button onclick="loadUsers()">Refresh</button>
<input id="user-search" placeholder="Search name / username / email / job function…" oninput="renderUsers()">
<select id="user-filter" onchange="renderUsers()">
<input id="user-search" placeholder="Search name / username / email / job function…" oninput="filterChanged()"
aria-label="Search the directory">
<select id="user-filter" onchange="filterChanged()" aria-label="Show">
<option value="">Everyone</option>
<option value="active">Active only</option>
<option value="disabled">Disabled only</option>
<option value="mine">Accounts I manage</option>
</select>
<!-- CR-020 (T12.4): a list of usernames from a file becomes the selection.
Managers only; users.js shows it. -->
<button type="button" id="csv-btn" onclick="pickCsv()" hidden
title="A CSV with one username (or email) per line, first column. The action is chosen in the bulk bar.">Select from CSV…</button>
<input type="file" id="csv-file" accept=".csv,.txt,text/csv,text/plain" hidden>
</div>
<div id="users-banner"></div>
<div id="csv-banner"></div>
<!-- CR-020: bulk actions on the selected rows. Present only while at least one
row is selected (the hidden attribute also takes it out of the tab order). -->
<div id="bulk-bar" class="bulkbar" role="region" aria-label="Bulk actions on the selected users" hidden>
<div class="bulkhead">
<strong id="bulk-count">0 selected</strong>
<button type="button" onclick="clearSelection(true)">Clear selection</button>
</div>
<div id="bulk-actions"></div>
</div>
<div id="sel-banner"></div>
<div id="bulk-banner"></div>
<div id="users-table"><div class="note">Loading…</div></div>
</div>

View File

@@ -23,6 +23,10 @@
let _users = []; // the directory as the server scoped it
let _scope = null; // GET /api/auth/user-scope
let _meId = null;
// CR-020: the rows checked for a bulk action. Client-side UI state only - every
// bulk action is decided and audited by the server, row by row (T12.1).
let _selected = new Set();
let _shownSelectable = []; // ids the current filter shows AND this caller may select
// ── boot ──────────────────────────────────────────────────────────────────────
async function boot(){
@@ -40,6 +44,7 @@ async function boot(){
renderScope();
}
renderCreateForm();
renderBulkActions();
loadUsers();
}
@@ -90,11 +95,309 @@ async function loadUsers(){
}
banner('users-banner','', '');
_users = json;
// Fresh data: any earlier selection may name accounts that changed or are gone.
clearSelection(false);
renderUsers();
}
function manages(){ return !!(_scope && _scope.can_manage_users); }
// ── selection (CR-020, T12.2) ─────────────────────────────────────────────────
// A row is selectable when this caller may change the account. Your own row is
// not: every bulk action that could touch it (disable, demote, delete) is refused
// for your own account, so offering it would only produce a guaranteed failure.
function selectable(u){ return manages() && !!u.manageable && u.id !== _meId; }
// The search box and the filter both decide which rows are shown. Changing either
// clears the selection (decided 2026-10-06), so no row hidden by the filter can
// stay selected and be acted on without being seen.
function filterChanged(){
const had = _selected.size;
clearSelection(false);
renderUsers();
banner('sel-banner', '', had ? 'Selection cleared: the filter changed. Select the rows again.' : '');
}
function clearSelection(announce){
const had = _selected.size;
_selected = new Set();
updateSelectionUI();
banner('sel-banner', '', (announce && had) ? 'Selection cleared.' : '');
}
function setRowSelected(id, on){
if(on) _selected.add(id); else _selected.delete(id);
}
// Select-all acts on the rows the filter shows, never on the whole directory.
function toggleSelectAll(on){
_shownSelectable.forEach(id => setRowSelected(id, on));
banner('sel-banner', '', '');
updateSelectionUI();
}
// Reflect _selected into the page without re-rendering the table, so keyboard
// focus stays on the checkbox that was just used.
function updateSelectionUI(){
const wrap = document.getElementById('users-table');
if(wrap){
wrap.querySelectorAll('input.selbox').forEach(cb => {
const on = _selected.has(cb.dataset.id);
cb.checked = on;
const tr = cb.closest('tr');
if(tr) tr.classList.toggle('is-selected', on);
});
const all = document.getElementById('sel-all');
if(all){
const n = _shownSelectable.filter(id => _selected.has(id)).length;
all.checked = n > 0 && n === _shownSelectable.length;
all.indeterminate = n > 0 && n < _shownSelectable.length;
all.disabled = !_shownSelectable.length;
}
}
const bar = document.getElementById('bulk-bar');
if(bar){
const n = _selected.size;
bar.hidden = n === 0;
const c = document.getElementById('bulk-count');
if(c) c.textContent = n + ' selected';
}
}
// One listener for every checkbox in the table (event delegation): no ids are
// bound into inline handlers, so there is nothing to escape.
document.addEventListener('change', e => {
const t = e.target;
if(!(t instanceof HTMLInputElement)) return;
if(t.id === 'sel-all'){ toggleSelectAll(t.checked); return; }
if(t.classList.contains('selbox')){
setRowSelected(t.dataset.id, t.checked);
banner('sel-banner', '', '');
updateSelectionUI();
}
});
// ── bulk actions (CR-020, T12.3) ──────────────────────────────────────────────
// The toolbar offers only what this caller may do: the roles they may grant, the
// projects they administer, the per-project roles they may hand out. The server
// re-checks every row anyway (T12.1); this is so nobody is offered a choice that
// can only fail.
function renderBulkActions(){
const box = document.getElementById('bulk-actions');
const csvBtn = document.getElementById('csv-btn');
if(csvBtn) csvBtn.hidden = !manages();
if(!box || !manages()) return;
const grantable = PERM_ROLES.filter(r => (_scope.grantable_roles || []).indexOf(r) >= 0);
const projects = (_scope.managed_projects || []).slice()
.sort((a,b) => (a.archived?1:0) - (b.archived?1:0));
const projRoles = PROJECT_SCOPED_ROLES.filter(r => (_scope.grantable_project_roles || []).indexOf(r) >= 0);
const group = (label, inner) => '<div class="bulkgroup" role="group" aria-label="'+uesc(label)+'">'+
'<span class="bulklabel">'+uesc(label)+'</span>'+inner+'</div>';
box.innerHTML =
group('Permissions',
'<select id="bulk-role" aria-label="New permissions role">'+
grantable.map(r => '<option value="'+r+'"'+(r==='project_user'?' selected':'')+'>'+uesc(PERM_LABELS[r])+'</option>').join('')+
'</select>'+
'<button type="button" onclick="bulkRole()">Set permissions</button>') +
group('Sign-in',
'<button type="button" onclick="bulkActive(true)">Enable</button>'+
'<button type="button" onclick="bulkActive(false)">Disable</button>') +
(projects.length ? group('Project access',
'<select id="bulk-project" aria-label="Project">'+
projects.map(p => '<option value="'+uesc(p.id)+'">'+uesc((p.name||'(unnamed)')+(p.number?' '+p.number:'')+
(p.archived?' (archived)':''))+'</option>').join('')+
'</select>'+
'<select id="bulk-project-role" aria-label="Role on that project, for Add">'+
'<option value="">Role: same as account</option>'+
projRoles.map(r => '<option value="'+r+'">Role: '+uesc(PERM_LABELS[r])+' here</option>').join('')+
'</select>'+
'<button type="button" onclick="bulkProject(\'add\')">Add to project</button>'+
'<button type="button" onclick="bulkProject(\'remove\')">Remove from project</button>') : '') +
group('Delete',
'<button type="button" class="danger" onclick="bulkDelete()">Delete selected…</button>');
}
function selectedIds(){ return _users.filter(u => _selected.has(u.id)).map(u => u.id); }
function people(n){ return n + ' user' + (n === 1 ? '' : 's'); }
function selVal(id){ return (document.getElementById(id) || {}).value || ''; }
function selText(id){
const s = document.getElementById(id);
return (s && s.selectedIndex >= 0) ? s.options[s.selectedIndex].text : '';
}
// One path for every bulk action: confirm through the dialog kit with the exact
// count, send, report row by row, reload. A reload clears the selection, which is
// right: the rows may have changed underneath it.
async function runBulk(path, body, ask){
const ids = selectedIds();
if(!ids.length) return;
if(!(await wpConfirmDialog({title: ask.title, okLabel: ask.ok, message: ask.message(ids.length)}))) return;
return sendBulk(path, body, ask.title, ids);
}
async function sendBulk(path, body, title, ids){
bannerHtml('bulk-banner', '', uesc(title)+': working on '+uesc(people(ids.length))+'…');
const { status, json } = await api('POST', path, Object.assign({ user_ids: ids }, body));
if(status !== 200 || !json || !Array.isArray(json.results)){
bannerHtml('bulk-banner', 'bad', '✕ '+uesc(apiError(status, json, title+' failed'))+' Nothing was changed.');
return;
}
showBulkResult(json, title);
loadUsers();
}
// T12.5: delete is permanent (a hard delete, as for one account) and one bad
// multi-select now removes many, so it asks more than the other actions do:
// every affected username is listed, and nothing is sent until the word DELETE
// is typed, whatever the count (confirmed 2026-10-06). Your own account cannot
// be selected; if it ever reached the request, the server refuses that row and
// the report names it.
async function bulkDelete(){
const chosen = _users.filter(u => _selected.has(u.id));
if(!chosen.length) return;
const n = chosen.length;
const typed = await wpPromptDialog({
title: 'Delete '+people(n),
message: 'Delete '+(n === 1 ? 'this account' : 'these '+n+' accounts')+'? Each one, and every project '+
'assignment it has, is removed. This cannot be undone. Disable them instead if you only want to '+
'block sign-in.\n\n'+chosen.map(u => '• '+u.username+(u.full_name ? ' ('+u.full_name+')' : '')).join('\n'),
label: 'Type DELETE to confirm',
value: '',
okLabel: 'Delete '+people(n),
validate: v => v === 'DELETE' ? '' : 'Type DELETE, in capital letters, to confirm.',
});
if(typed !== 'DELETE') return;
return sendBulk('/api/auth/bulk-users/delete', {}, 'Delete accounts', chosen.map(u => u.id));
}
// Every row is accounted for: changed, already so, or failed with the server's
// reason. A batch that partly failed never reads as a success (CR-020).
function showBulkResult(json, title){
const failed = json.results.filter(r => r.status === 'failed');
const parts = [json.ok+' changed'];
if(json.unchanged) parts.push(json.unchanged+' already set');
if(failed.length) parts.push(failed.length+' failed');
const kind = !failed.length ? 'ok' : (json.ok + json.unchanged ? 'warn' : 'bad');
const lead = (failed.length ? (kind === 'bad' ? '✕ ' : '⚠ ') : '✓ ')+uesc(title)+': '+parts.join(', ')+'.';
const list = failed.length
? '<ul>'+failed.map(r => '<li><strong>'+uesc(r.username || r.user_id)+'</strong>: '+uesc(r.reason)+'</li>').join('')+'</ul>'
: '';
bannerHtml('bulk-banner', kind, lead + list);
}
function bannerHtml(id, kind, html){
const el = document.getElementById(id);
if(!el) return;
el.className = 'banner' + (kind ? ' '+kind : '');
el.innerHTML = html;
}
function bulkRole(){
const role = selVal('bulk-role');
if(!role) return;
return runBulk('/api/auth/bulk-users/role', { role }, {
title: 'Set permissions', ok: 'Change permissions',
message: n => 'Set permissions to '+(PERM_LABELS[role] || role)+' for '+people(n)+'?\n\n'+
'Each account is checked on its own. Any that cannot be changed are listed afterwards.'});
}
function bulkActive(on){
return runBulk('/api/auth/bulk-users/active', { is_active: !!on }, {
title: on ? 'Enable accounts' : 'Disable accounts', ok: on ? 'Enable' : 'Disable',
message: n => (on ? 'Enable sign-in for ' : 'Disable sign-in for ')+people(n)+'?'+
(on ? '' : '\n\nA disabled account cannot sign in. Nothing else about it changes.')});
}
function bulkProject(op){
const pid = selVal('bulk-project');
if(!pid) return;
const role = op === 'add' ? selVal('bulk-project-role') : '';
const pname = selText('bulk-project');
return runBulk('/api/auth/bulk-users/projects', { project_id: pid, op, role }, {
title: op === 'add' ? 'Add to project' : 'Remove from project',
ok: op === 'add' ? 'Add' : 'Remove',
message: n => op === 'add'
? 'Add '+people(n)+' to '+pname+' with '+(role ? PERM_LABELS[role]+' on that project' : 'their account role')+'?\n\n'+
'Their other projects, and their roles there, stay as they are.'
: 'Remove '+people(n)+' from '+pname+'?\n\nTheir other projects stay as they are.'});
}
// ── CSV upload (CR-020, T12.4) ────────────────────────────────────────────────
// The file holds usernames only (decided 2026-10-06). The server reads it and
// answers line by line; the accepted accounts become the ordinary selection, so
// the action, its confirmation and its row-by-row report are the same ones the
// checkboxes use. A CSV cannot reach an account a checkbox could not: the bulk
// route re-checks every row.
const CSV_MAX_BYTES = 1024 * 1024;
// Blob.text() where the browser has it, FileReader where it does not (older
// tablet browsers).
function readFileText(file){
if(typeof file.text === 'function') return file.text();
return new Promise((resolve, reject) => {
const r = new FileReader();
r.onload = () => resolve(String(r.result || ''));
r.onerror = () => reject(r.error);
r.readAsText(file);
});
}
function pickCsv(){
const f = document.getElementById('csv-file');
if(f){ f.value = ''; f.click(); }
}
document.addEventListener('change', e => {
if(e.target && e.target.id === 'csv-file' && e.target.files && e.target.files[0]) loadCsv(e.target.files[0]);
});
async function loadCsv(file){
if(file.size > CSV_MAX_BYTES){
bannerHtml('csv-banner', 'bad', '✕ '+uesc(file.name)+' is larger than 1 MB. Split it and load the parts one at a time.');
return;
}
let text = '';
try { text = await readFileText(file); }
catch(_){ bannerHtml('csv-banner', 'bad', '✕ Could not read '+uesc(file.name)+'.'); return; }
bannerHtml('csv-banner', '', 'Reading '+uesc(file.name)+'…');
const { status, json } = await api('POST', '/api/auth/bulk-users/resolve', { text });
if(status !== 200 || !json || !Array.isArray(json.rows)){
bannerHtml('csv-banner', 'bad', '✕ '+uesc(apiError(status, json, 'Could not read the list')));
return;
}
// Show everyone, so every accepted account is a visible, checked row: a
// selection must never include a row the filter hides.
const s = document.getElementById('user-search'), f = document.getElementById('user-filter');
if(s) s.value = '';
if(f) f.value = '';
await loadUsers();
const want = new Set(json.user_ids || []);
_shownSelectable.forEach(id => { if(want.has(id)) _selected.add(id); });
updateSelectionUI();
const missed = [...want].filter(id => !_selected.has(id)).length;
const rejected = json.rows.filter(r => r.status !== 'ok');
const lead = (rejected.length || missed ? '⚠ ' : '✓ ')+uesc(file.name)+': '+json.read+' line'+(json.read===1?'':'s')+' read, '+
_selected.size+' selected'+(rejected.length ? ', '+rejected.length+' rejected' : '')+'.'+
(missed ? ' '+missed+' accepted account'+(missed===1?' is':'s are')+' no longer in the directory; reload and try again.' : '')+
(_selected.size ? ' Pick an action in the bulk bar to apply it to the selected accounts.' : '');
const list = rejected.length
? '<ul>'+rejected.map(r => '<li>Line '+r.line+' ('+uesc(r.text)+'): '+uesc(r.reason)+'</li>').join('')+'</ul>'
: '';
bannerHtml('csv-banner', (rejected.length || missed) ? 'warn' : 'ok', lead + list);
}
function selCell(u){
if(!selectable(u)){
const why = u.id === _meId ? 'Your own account: change it one setting at a time'
: (u.manage_blocked_reason || 'You cannot change this account');
return '<td class="selcell"><span class="selhit" title="'+uesc(why)+'">'+
'<input type="checkbox" disabled aria-label="'+uesc('Cannot select '+u.username+': '+why)+'"></span></td>';
}
return '<td class="selcell"><label class="selhit">'+
'<input type="checkbox" class="selbox" data-id="'+uesc(u.id)+'"'+(_selected.has(u.id)?' checked':'')+
' aria-label="'+uesc('Select '+u.username)+'"></label></td>';
}
function renderUsers(){
const wrap = document.getElementById('users-table');
const q = ((document.getElementById('user-search')||{}).value||'').trim().toLowerCase();
@@ -116,7 +419,8 @@ function renderUsers(){
(u.project_role||'')+' '+roleLabel(u.role)).toLowerCase().indexOf(q) >= 0;
});
const count = '<div class="note">'+list.length+' of '+total+' '+(total===1?'person':'people')+'</div>';
if(!list.length){ wrap.innerHTML = count+'<div class="note">Nothing matches.</div>'; return; }
_shownSelectable = list.filter(selectable).map(u => u.id);
if(!list.length){ wrap.innerHTML = count+'<div class="note">Nothing matches.</div>'; updateSelectionUI(); return; }
const head = manages()
? ['Username','Name','Email',
@@ -130,9 +434,16 @@ function renderUsers(){
'Status'];
const ths = head.map(h => Array.isArray(h)
? '<th title="'+uesc(h[1])+'">'+uesc(h[0])+'</th>' : '<th>'+uesc(h)+'</th>').join('');
// CR-020: a selection column for anyone who manages accounts. Select-all names
// how many rows it covers, because it covers only what the filter shows.
const selTh = manages()
? '<th class="selcell"><label class="selhit"><input type="checkbox" id="sel-all" aria-label="'+
uesc('Select all '+_shownSelectable.length+' shown that you can change')+'"></label></th>'
: '';
const rows = list.map(manages() ? managerRow : readonlyRow).join('');
wrap.innerHTML = count+'<div class="tscroll"><table class="grid"><thead><tr>'+ths+
wrap.innerHTML = count+'<div class="tscroll"><table class="grid"><thead><tr>'+selTh+ths+
'</tr></thead><tbody>'+rows+'</tbody></table></div>'+legend();
updateSelectionUI();
}
function legend(){
@@ -189,6 +500,7 @@ function managerRow(u){
if(!can && !me) actions.push('<span class="note" style="margin:0" title="'+uesc(why)+'">read-only</span>');
return '<tr'+(can||me ? '' : ' class="is-locked"')+'>'+
selCell(u)+
'<td><strong>'+uesc(u.username)+'</strong>'+(me?'<span class="me-tag">you</span>':'')+'</td>'+
'<td>'+uesc(u.full_name||'')+'</td>'+
// The address is truncated with the full value on the title: a long one used to

View File

@@ -2207,20 +2207,18 @@ function loadStepComments(){
}
}
// ── USAGE ANALYTICS ─────────────────────────────────────────────────────────
// Lightweight usage analytics stored in localStorage so the tool owner can review
// engagement over time. No field VALUES are stored (field-edit events record only
// the field id), keeping captured data non-sensitive.
// D5 / T7.10: the analytics implementation lives in wp-usage.js and the report
// on the admin console. The wizard's own copy of showAnalytics() never had a
// caller here - the button lived on the creator - and once B7 dissolved the
// frame the duplicate sat in the same document as five colliding globals.
// This page only records; dwell tracking keeps its page-local state below.
// ── USAGE ANALYTICS (retired, T11.6) ────────────────────────────────────────
// This used to write to wp-usage.js's per-browser localStorage log (D5/T7.10),
// read back by the admin console's old "Usage logs" panel. CR-019 replaced
// both with real, server-side, per-user activity (UsageEvent + the Activity &
// usage card) - see decisions-2026-09-17.md for why this per-browser data was
// never a source the new report could adopt. track() is now a no-op; kept
// (rather than deleting its handful of call sites, including the dwell-timer
// plumbing below) so this stays a one-line change instead of touching every
// caller for the same outcome.
let _stepEnter = Date.now();
function track(event, detail){
WPUsage.track(WPUsage.KEYS.wizard, event, detail);
}
function track(event, detail){ /* retired, T11.6 — see comment above */ }
function trackStepDwell(){
const ms = Date.now() - _stepEnter;
if(ms > 400 && ms < 1000*60*60) track('step_dwell', {step: currentStep, ms});

View File

@@ -11,7 +11,6 @@
<!-- Addressable state (S3). Parses before the app scripts, which read the URL
during their own boot. -->
<script src="wp-url.js"></script>
<script src="wp-usage.js"></script>
<script src="wp-list-import.js"></script>
<!-- Autosave, unsaved-work guard, draft recovery (S2). -->
<script src="wp-autosave.js"></script>

View File

@@ -3993,11 +3993,17 @@ function openSopModal(){
document.getElementById('sop-modal').classList.add('open'); track('view_sop');
}
function closeSopModal(){ document.getElementById('sop-modal').classList.remove('open'); }
// D5 / T7.10: the analytics implementation lives in wp-usage.js - ONE copy for
// the whole suite - and its report lives on the admin console, where an
// operator-facing readout belongs. This page only records. Same key, same
// event shape: everything recorded before the move is still readable after it.
function track(event,detail){ if(devMode) return; WPUsage.track(WPUsage.KEYS.creator, event, detail); }
// CR-019 / T11.6 (2026-09-23): this used to write to wp-usage.js's per-browser
// localStorage log (D5/T7.10). Retired along with the admin console's old
// "Usage logs" panel, its only reader - real, server-side, per-user activity
// now exists (UsageEvent, the Activity & usage card, T11.1-T11.5) and this
// data was never reliably tied to a real identity anyway, so it was not a
// source the new report could adopt (decisions-2026-09-17.md). track() stays
// as a no-op rather than deleting its ~35 call sites throughout this file:
// removing every call individually is a much larger, riskier diff for the
// same outcome, and a call site here still documents the moment worth
// recording if usage analytics are ever rebuilt server-side.
function track(event,detail){ /* retired, T11.6 — see comment above */ }
// ── COMMENTS ─────────────────────────────────────────────────────────────────
const COMMENTS_KEY='wp_iwp_comments_v1';

View File

@@ -11,7 +11,6 @@
<!-- Addressable state (S3). Parses before the app scripts, which read the URL
during their own boot. -->
<script src="wp-url.js"></script>
<script src="wp-usage.js"></script>
<!-- Autosave, unsaved-work guard, draft recovery (S2). -->
<script src="wp-autosave.js"></script>
<!-- Which sections this project uses (CR-006). The same file the SOP wizard

View File

@@ -220,7 +220,8 @@
.then(function (res) {
if (!res.ok) { msg((res.j && res.j.detail) || ('Could not save (HTTP ' + res.status + ').'), false); return; }
if (res.j && res.j.user) window.WP_USER = res.j.user;
try { localStorage.setItem('wp_auth_cache', JSON.stringify({ user: window.WP_USER, at: Date.now() })); } catch (e) {}
// D27: refresh the cached user, but not the end of the session.
try { if (window.wpCacheAuth) window.wpCacheAuth(window.WP_USER); } catch (e) {}
msg('Saved. Reloading so every date on the page agrees…', true);
// Dates are formatted at render time all over the app; a reload is the
// honest way to apply the change everywhere at once.

View File

@@ -1,58 +0,0 @@
/* Usage analytics core — the ONE implementation (D5 / T7.10).
This existed three times: the creator's copy, the wizard's copy (which had no
caller — the button lived on the creator), and the admin console's own reader.
Once the creator stopped being an iframe (B7/T7.1) the first two sat in one
document as five colliding globals; an unreferenced duplicate is exactly what
produced D5. One core now; the pages keep only a thin track() wrapper because
page state (the creator's dev-mode pause) belongs to the page.
The storage KEYS are unchanged on purpose: everything recorded before this
file existed is still readable through it. No field VALUES are ever stored —
a field-edit event records the field id, nothing else.
Classic script, no modules: exposes window.WPUsage. */
'use strict';
(function () {
var SESSION = 's_' + Date.now().toString(36) + Math.random().toString(36).slice(2, 6);
function load(key) {
try { return JSON.parse(localStorage.getItem(key)) || { events: [] }; }
catch (e) { return { events: [] }; }
}
function save(key, data) {
try { localStorage.setItem(key, JSON.stringify(data)); }
catch (e) { /* storage unavailable — degrade silently */ }
}
function track(key, event, detail) {
try {
var d = load(key);
d.events.push({ ts: new Date().toISOString(), session: SESSION, event: event, detail: detail || null });
if (d.events.length > 5000) d.events = d.events.slice(-5000);
save(key, d);
} catch (e) { /* never let telemetry break the tool it watches */ }
}
function download(key, prefix) {
var blob = new Blob([JSON.stringify(load(key), null, 2)], { type: 'application/json' });
var a = document.createElement('a');
a.href = URL.createObjectURL(blob);
a.download = (prefix || 'wp-usage') + '-' + new Date().toISOString().slice(0, 10) + '.json';
document.body.appendChild(a);
a.click();
a.remove();
setTimeout(function () { URL.revokeObjectURL(a.href); }, 1000);
}
window.WPUsage = {
load: load,
save: save,
track: track,
download: download,
// The pre-D5 keys, verbatim — continuity of the recorded data is a done-when.
KEYS: { creator: 'wp_iwp_analytics_v1', wizard: 'wp_suite_analytics_v1' },
};
})();

View File

@@ -27,8 +27,8 @@ DATABASE_URL=postgresql+psycopg://wpsuite:CHANGE_ME@localhost:5432/wpsuite
# python -c "import secrets; print(secrets.token_urlsafe(48))"
AUTH_SECRET_KEY=CHANGE_ME_run_the_command_above
# How long a login lasts before re-authentication (hours). Default 12.
# AUTH_SESSION_HOURS=12
# How long a login lasts before re-authentication (hours). Default 8 (D25).
# AUTH_SESSION_HOURS=8
# ── Okta OIDC (required — this is the only sign-in path) ───────────────────────
# The Okta *authorization server* issuer, e.g. https://yourorg.okta.com/oauth2/default
@@ -55,6 +55,59 @@ OKTA_REDIRECT_URI=CHANGE_ME
# a custom claim, …) — no code change needed, just this value.
# OKTA_IDENTITY_CLAIM=preferred_username
# ── Okta deprovisioning sync (D18, wave 13) ────────────────────────────────────
# A scheduled job asks Okta who is still active and disables local accounts
# Okta no longer shows as active. It only ever DISABLES. It never re-enables.
# If Okta cannot be reached or answers with anything unexpected, the job changes
# NOTHING that cycle and logs the failure.
#
# A read-only Okta API token, created by the person who administers the Okta
# tenant (Security > API > Tokens). This is NOT the OIDC client secret above.
# It is a different credential with a different scope. Treat it like
# AUTH_SECRET_KEY: never commit it, never log it. Leave it unset and the sync
# stays off (the startup log says so).
# OKTA_API_TOKEN=CHANGE_ME
#
# Second way to sign in to the Management API (D23): an Okta "API Services" app that
# uses client authentication private_key_jwt. Set these two INSTEAD of OKTA_API_TOKEN.
# If OKTA_API_TOKEN is also set, the token is used.
# OKTA_SYNC_CLIENT_ID the client ID of that app (starts with 0oa)
# OKTA_SYNC_PRIVATE_KEY its PRIVATE key as a JWK, as JSON or as base64 of that JSON.
# In Portainer use base64: it has no quote marks to be mangled.
# The Okta app needs the scope okta.users.read granted (Okta API Scopes tab), an admin
# role such as Read-only Administrator assigned to it. If the app has "Require Demonstrating
# Proof of Possession (DPoP)" switched on, the sync detects that and uses DPoP by itself. The key is a secret: never commit it or paste it
# into chat or a ticket.
# OKTA_SYNC_DPOP: auto (default), on, or off. auto asks for a plain Bearer token first, and
# uses DPoP only if Okta answers invalid_dpop_proof. Set "on" to always use DPoP.
# Set "off" to never use it. You normally set nothing.
# OKTA_SYNC_DPOP=auto
# OKTA_SYNC_CLIENT_ID=0oaXXXXXXXXXXXXXXXXX
# OKTA_SYNC_PRIVATE_KEY=
# The Okta org URL that serves the Management API, e.g. https://yourorg.okta.com
# Optional. If unset, it is the scheme and host of OKTA_ISSUER. Must be https.
# OKTA_ORG_URL=https://your-org.okta.com
#
# Which Okta PROFILE attribute holds the value that matches users.username.
# Optional. By default it follows OKTA_IDENTITY_CLAIM: preferred_username reads
# profile.login, email reads profile.email, any other claim name reads the
# profile attribute of the same name. Set this only if your tenant builds the
# claim from a different attribute.
# OKTA_SYNC_PROFILE_ATTRIBUTE=login
#
# How often the sync runs, in seconds. Default 900 (15 minutes), proposed and not
# confirmed with IT. A value under 300 is raised to 300. Every api worker checks
# once a minute and one of them wins each cycle, so raising the worker count does
# not raise the call rate to Okta.
# OKTA_SYNC_INTERVAL_SECONDS=900
#
# Dry run (D22). Set to true for the first deploy. The sync then does everything
# except the change: it logs who it WOULD disable and writes nothing, and it writes
# no audit rows. Read the result with: docker compose logs api | grep "Okta sync"
# Remove it (or set it to false) to go live. Off by default.
# OKTA_SYNC_DRY_RUN=true
# ── Email notifications (optional) ─────────────────────────────────────────────
# WP-assignment emails are OFF by default and are turned on from the Admin
# console (Notifications & email card), where the SMTP host/port/from-address

View File

@@ -0,0 +1,46 @@
"""usage events (CR-019, wave 11)
Append-only navigation/session activity, separate from audit_log on purpose —
see the UsageEvent docstring in server/models.py. Retention is indefinite by
decision (docs/waves/decisions-2026-09-17.md); nothing here schedules a purge.
Revision ID: 8e2cb3003f8a
Revises: 1d60a608bb51
Create Date: 2026-09-23 00:00:00.000000
"""
from alembic import op
import sqlalchemy as sa
# revision identifiers, used by Alembic.
revision = '8e2cb3003f8a'
down_revision = '1d60a608bb51'
branch_labels = None
depends_on = None
def upgrade() -> None:
op.create_table('usage_events',
sa.Column('id', sa.String(length=40), nullable=False),
sa.Column('at', sa.DateTime(timezone=True), nullable=False),
sa.Column('username', sa.String(length=200), nullable=False),
sa.Column('project_id', sa.String(length=40), nullable=True),
sa.Column('tool', sa.String(length=40), nullable=False),
sa.Column('event', sa.String(length=40), nullable=False),
sa.Column('detail', sa.JSON(), nullable=False),
sa.PrimaryKeyConstraint('id')
)
op.create_index(op.f('ix_usage_events_at'), 'usage_events', ['at'], unique=False)
op.create_index(op.f('ix_usage_events_username'), 'usage_events', ['username'], unique=False)
op.create_index(op.f('ix_usage_events_project_id'), 'usage_events', ['project_id'], unique=False)
op.create_index(op.f('ix_usage_events_tool'), 'usage_events', ['tool'], unique=False)
op.create_index(op.f('ix_usage_events_event'), 'usage_events', ['event'], unique=False)
def downgrade() -> None:
op.drop_index(op.f('ix_usage_events_event'), table_name='usage_events')
op.drop_index(op.f('ix_usage_events_tool'), table_name='usage_events')
op.drop_index(op.f('ix_usage_events_project_id'), table_name='usage_events')
op.drop_index(op.f('ix_usage_events_username'), table_name='usage_events')
op.drop_index(op.f('ix_usage_events_at'), table_name='usage_events')
op.drop_table('usage_events')

View File

@@ -9,10 +9,16 @@ Run (prod): gunicorn -k uvicorn.workers.UvicornWorker -b 127.0.0.1:8000 serve
Interactive docs: http://<host>/api/docs
"""
import base64
import csv
import contextlib
import hashlib
import hmac
import io
import logging
import os
import re
import uuid
from datetime import datetime, timedelta, timezone
from typing import Any, Optional
from urllib.parse import urlparse
@@ -22,12 +28,12 @@ from fastapi.middleware.cors import CORSMiddleware
from fastapi.responses import JSONResponse, RedirectResponse
from fastapi.staticfiles import StaticFiles
from pydantic import BaseModel, ConfigDict, Field
from sqlalchemy import select, delete, func
from sqlalchemy import select, delete, func, or_
from sqlalchemy.orm import Session
from starlette.middleware.sessions import SessionMiddleware
from .db import Base, engine, get_db
from . import models, auth, notify, assets_db, okta_auth, okta_fake, files_http
from .db import Base, SessionLocal, engine, get_db
from . import models, auth, notify, assets_db, okta_auth, okta_fake, files_http, okta_sync
log = logging.getLogger("wpsuite.app")
@@ -42,7 +48,21 @@ if engine.dialect.name == "sqlite":
# Interactive docs are handy in dev but hand an attacker the full API map in prod,
# so enable them only on the SQLite dev fallback (production runs on Postgres).
_docs_enabled = engine.dialect.name == "sqlite"
@contextlib.asynccontextmanager
async def _lifespan(_app):
"""Start the Okta deprovisioning sync (D18 / T13.4) when the process starts and
stop it when the process ends. With no OKTA_API_TOKEN it starts nothing and
says so in the startup log."""
stop = okta_sync.start_scheduler(SessionLocal)
try:
yield
finally:
if stop is not None:
stop.set()
app = FastAPI(
lifespan=_lifespan,
title="Work Package Suite API",
docs_url="/api/docs" if _docs_enabled else None,
redoc_url=None,
@@ -756,6 +776,11 @@ async def okta_callback(request: Request, db: Session = Depends(get_db)):
user.failed_attempts = 0
user.locked_until = None
user.last_login_at = models.utcnow()
# CR-019: one login event per Okta sign-in, written here rather than
# inferred from session creation elsewhere, so there is exactly one
# source of truth for "did this person sign in" - not one per page load
# afterward (T11.2 covers that separately, as page_open events).
db.add(models.UsageEvent(id=gen_id("uev"), username=user.username, tool="", event="login"))
db.commit()
db.refresh(user)
@@ -809,12 +834,22 @@ if okta_fake.is_active():
@app.get("/api/auth/me")
def whoami(user: models.User = Depends(auth.get_current_user)):
def whoami(request: Request, user: models.User = Depends(auth.get_current_user)):
"""Who is logged in. The frontend guard calls this on every page load.
`role` is normalized here so no page has to know that a pre-roles account stores
'user' where it now means 'project_user'."""
return {"user": {**user.to_dict(), "role": auth.normalize_role(user.role)}}
'user' where it now means 'project_user'.
D27: `session_expires_at` is the `exp` claim of the session token, as a UTC time.
The browser keeps it so the offline sign-in window ends when the server session
ends. No column holds it: get_current_user has already validated this token, so
the claim is read back from the same cookie."""
claims = auth.is_request_authenticated(request) or {}
exp = claims.get("exp")
expires_at = (datetime.fromtimestamp(exp, tz=timezone.utc).strftime("%Y-%m-%dT%H:%M:%SZ")
if isinstance(exp, (int, float)) else None)
return {"user": {**user.to_dict(), "role": auth.normalize_role(user.role)},
"session_expires_at": expires_at}
# ── Display preferences (self-service) ─────────────────────────────────────────
@@ -1013,16 +1048,100 @@ def create_user(body: NewUserIn, actor: models.User = Depends(require_user_manag
return directory_entry(db, u, actor)
# ── One account-level change, used by the single-user routes AND the bulk routes ──
# CR-020 (T12.1): a bulk action is the single-user action, once per row. So each
# rule lives in exactly one function below, and both kinds of route call it: a
# bulk request cannot reach further than a one-at-a-time request, because there
# is no second copy of the rules for it to reach through. Every function checks
# all of its guards BEFORE it changes anything, so a row that is refused leaves
# nothing half-written, and writes its AuditLog row the way the single route
# always has. None of them commits; the route owns the transaction.
def guard_target(db: Session, actor: "models.User", u: "models.User") -> None:
"""The two checks every account change starts with: may the caller see this
account (404 if not, so existence does not leak), and may they change it."""
require_see_user(db, actor, u)
require_manage_user(db, actor, u)
def apply_user_active(db: Session, actor: "models.User", u: "models.User", is_active: bool) -> None:
if u.id == actor.id and not is_active:
raise HTTPException(status_code=400, detail="You cannot disable your own account")
u.is_active = is_active
log_event(db, actor, "user_enabled" if is_active else "user_disabled", "user", u.id,
summary=u.username, detail={"is_active": bool(is_active)})
def apply_user_role(db: Session, actor: "models.User", u: "models.User", role: str) -> None:
"""The role guards: never your own role, never the last active admin. The
grantable-roles check is the caller's, made once per request before this."""
if u.id == actor.id:
raise HTTPException(status_code=400, detail="You cannot change your own role")
if auth.is_admin(u) and role != auth.ROLE_ADMIN:
other_admins = db.scalars(
select(models.User.id).where(
(models.User.role == auth.ROLE_ADMIN)
& (models.User.id != u.id)
& (models.User.is_active.is_(True))
)
).all()
if not other_admins:
raise HTTPException(status_code=400, detail="Can't remove the last admin account")
old_role = u.role
u.role = role
detail = {"from": old_role, "to": role}
# Promoting someone to admin retires their default-member flag: an admin already
# reaches every project, so the flag would do nothing except sit there invisibly
# (the console shows admins no controls) and come back to life the day they are
# demoted. Same reasoning as clearing the role in set_user_auto_add.
if role == auth.ROLE_ADMIN and (u.auto_add_projects or u.auto_add_role):
u.auto_add_projects = False
u.auto_add_role = ""
detail["auto_add_cleared"] = True
log_event(db, actor, "role_changed", "user", u.id, summary=u.username, detail=detail)
def apply_user_delete(db: Session, actor: "models.User", u: "models.User") -> None:
if u.id == actor.id:
raise HTTPException(status_code=400, detail="You cannot delete your own account")
log_event(db, actor, "user_deleted", "user", u.id, summary=u.username, detail={"role": u.role})
db.delete(u)
def apply_user_projects(db: Session, actor: "models.User", u: "models.User",
managed: Optional[set], valid: set, roles: dict) -> None:
"""Replace the caller's slice of `u`'s memberships with `valid`, each carrying
its role from `roles` ('' = inherit). `valid` must already be checked to exist
and to sit inside `managed`."""
# Rebuild only the rows this caller owns. Scoping the DELETE is the whole of the
# "leave other jobs alone" guarantee — get it wrong and a super user's save
# silently revokes access everywhere else.
doomed = delete(models.ProjectMember).where(models.ProjectMember.user_id == u.id)
if managed is not None:
# in_() on an empty set is a valid always-false predicate, so a caller who
# administers nothing deletes nothing (require_manage_user already refused them).
doomed = doomed.where(models.ProjectMember.project_id.in_(managed))
db.execute(doomed)
for pid in valid:
db.add(models.ProjectMember(id=gen_id("pm"), user_id=u.id, project_id=pid,
role=roles.get(pid, "")))
log_event(db, actor, "project_access_changed", "user", u.id, summary=u.username,
detail={"projects": len(valid),
"scope": "all" if managed is None else "managed",
"overrides": {p: r for p, r in roles.items() if p in valid}})
def project_scoped_grantable(actor: "models.User") -> tuple:
"""Per-project roles `actor` may hand out. Only an app admin can make someone a
super user on a project. user_scope and get_user_projects still spell the same
rule inline for the console; left as they are to keep this diff to CR-020."""
return auth.PROJECT_SCOPED_ROLES if auth.is_admin(actor) else (auth.ROLE_PROJECT_ADMIN, auth.ROLE_PROJECT_USER)
@app.post("/api/auth/users/{user_id}/active")
def set_user_active(user_id: str, body: ActiveIn, actor: models.User = Depends(require_user_manager), db: Session = Depends(get_db)):
u = load_target_user(db, user_id)
require_see_user(db, actor, u)
require_manage_user(db, actor, u)
if u.id == actor.id and not body.is_active:
raise HTTPException(status_code=400, detail="You cannot disable your own account")
u.is_active = body.is_active
log_event(db, actor, "user_enabled" if body.is_active else "user_disabled", "user", u.id,
summary=u.username, detail={"is_active": bool(body.is_active)})
guard_target(db, actor, u)
apply_user_active(db, actor, u, body.is_active)
db.commit()
return directory_entry(db, u, actor)
@@ -1039,32 +1158,8 @@ def set_user_role(user_id: str, body: RoleIn, actor: models.User = Depends(requi
if body.role not in allowed:
raise HTTPException(status_code=400, detail=f"role must be one of {', '.join(allowed)}")
u = load_target_user(db, user_id)
require_see_user(db, actor, u)
require_manage_user(db, actor, u)
if u.id == actor.id:
raise HTTPException(status_code=400, detail="You cannot change your own role")
if auth.is_admin(u) and body.role != auth.ROLE_ADMIN:
other_admins = db.scalars(
select(models.User.id).where(
(models.User.role == auth.ROLE_ADMIN)
& (models.User.id != u.id)
& (models.User.is_active.is_(True))
)
).all()
if not other_admins:
raise HTTPException(status_code=400, detail="Can't remove the last admin account")
old_role = u.role
u.role = body.role
detail = {"from": old_role, "to": body.role}
# Promoting someone to admin retires their default-member flag: an admin already
# reaches every project, so the flag would do nothing except sit there invisibly
# (the console shows admins no controls) and come back to life the day they are
# demoted. Same reasoning as clearing the role in set_user_auto_add.
if body.role == auth.ROLE_ADMIN and (u.auto_add_projects or u.auto_add_role):
u.auto_add_projects = False
u.auto_add_role = ""
detail["auto_add_cleared"] = True
log_event(db, actor, "role_changed", "user", u.id, summary=u.username, detail=detail)
guard_target(db, actor, u)
apply_user_role(db, actor, u, body.role)
db.commit()
db.refresh(u)
return directory_entry(db, u, actor)
@@ -1121,12 +1216,8 @@ def set_user_auto_add(user_id: str, body: AutoAddIn, admin: models.User = Depend
@app.delete("/api/auth/users/{user_id}")
def delete_user(user_id: str, actor: models.User = Depends(require_user_manager), db: Session = Depends(get_db)):
u = load_target_user(db, user_id)
require_see_user(db, actor, u)
require_manage_user(db, actor, u)
if u.id == actor.id:
raise HTTPException(status_code=400, detail="You cannot delete your own account")
log_event(db, actor, "user_deleted", "user", u.id, summary=u.username, detail={"role": u.role})
db.delete(u)
guard_target(db, actor, u)
apply_user_delete(db, actor, u)
db.commit()
return {"deleted": user_id}
@@ -1179,8 +1270,7 @@ def set_user_projects(user_id: str, body: ProjectAssignIn, actor: models.User =
exactly as they were, because a payload that simply omits them would otherwise cut
someone off from a job the caller can't even see."""
u = load_target_user(db, user_id)
require_see_user(db, actor, u)
require_manage_user(db, actor, u)
guard_target(db, actor, u)
requested = [p for p in dict.fromkeys(body.project_ids) if p]
for pid in requested:
check_id(pid)
@@ -1193,28 +1283,227 @@ def set_user_projects(user_id: str, body: ProjectAssignIn, actor: models.User =
# A super user may hand out the project-scoped roles below their own; only an app
# admin can make someone a super user on a project. Anything unrecognised falls
# back to inheriting the account's own role.
allowed = auth.PROJECT_SCOPED_ROLES if auth.is_admin(actor) else (auth.ROLE_PROJECT_ADMIN, auth.ROLE_PROJECT_USER)
allowed = project_scoped_grantable(actor)
roles = {pid: r for pid, r in (body.roles or {}).items() if r in allowed}
# Rebuild only the rows this caller owns. Scoping the DELETE is the whole of the
# "leave other jobs alone" guarantee — get it wrong and a super user's save
# silently revokes access everywhere else.
doomed = delete(models.ProjectMember).where(models.ProjectMember.user_id == user_id)
if managed is not None:
# in_() on an empty set is a valid always-false predicate, so a caller who
# administers nothing deletes nothing (require_manage_user already refused them).
doomed = doomed.where(models.ProjectMember.project_id.in_(managed))
db.execute(doomed)
for pid in valid:
db.add(models.ProjectMember(id=gen_id("pm"), user_id=user_id, project_id=pid,
role=roles.get(pid, "")))
log_event(db, actor, "project_access_changed", "user", u.id, summary=u.username,
detail={"projects": len(valid),
"scope": "all" if managed is None else "managed",
"overrides": {p: r for p, r in roles.items() if p in valid}})
apply_user_projects(db, actor, u, managed, valid, roles)
db.commit()
return {"assigned": sorted(valid), "roles": {p: roles.get(p, "") for p in sorted(valid)}}
# ── Bulk user administration — CR-020 (T12.1) ─────────────────────────────────
# Four actions over a list of accounts: activate/disable, permissions role,
# project add/remove (with the per-project role), delete. Each row runs exactly
# the guards and the audit of the single-user route, through the same apply_*
# function above. Rules, decided 2026-09-17 and confirmed 2026-10-06:
# * a row that is refused is reported with its reason and the rest still run;
# a batch is never reported as a success when part of it failed;
# * one AuditLog row per changed account, the same as the single route writes -
# never one opaque "bulk" entry;
# * the caller's own account is refused only on its own row (disable, demote,
# delete), never by failing the whole batch;
# * project assignment adds to or removes from ONE project; it never replaces a
# user's whole project list.
# A row already in the requested state is reported "unchanged" and writes no
# audit row: nothing changed, so there is nothing to audit.
#
# The routes live under /api/auth/bulk-users/ rather than /api/auth/users/bulk/,
# because POST /api/auth/users/{user_id}/active would otherwise match "bulk" as
# a user id.
BULK_MAX_USERS = 500
class BulkUsersIn(BaseModel):
user_ids: list[str] = Field(default_factory=list, max_length=BULK_MAX_USERS)
class BulkActiveIn(BulkUsersIn):
is_active: bool
class BulkRoleIn(BulkUsersIn):
role: str
class BulkProjectIn(BulkUsersIn):
project_id: str
op: str # "add" | "remove"
role: str = "" # per-project role on add; '' = inherit the account's own
def run_bulk(db: Session, actor: "models.User", user_ids: list[str], action: str, apply) -> dict:
"""Run `apply(u)` once per distinct id, in the order given. `apply` returns
"ok" or "unchanged", or raises HTTPException with the reason the single-user
route would have given. The flush after every row matters: the session does
not autoflush, and a later row's guard (the last-admin count, for one) has to
see what an earlier row already changed."""
results = []
for uid in dict.fromkeys(i for i in user_ids if i):
row = {"user_id": uid, "username": "", "status": "", "reason": ""}
try:
check_id(uid)
u = load_target_user(db, uid)
row["username"] = u.username
guard_target(db, actor, u)
row["status"] = apply(u)
db.flush()
except HTTPException as exc:
row["status"] = "failed"
row["reason"] = exc.detail if isinstance(exc.detail, str) else str(exc.detail)
results.append(row)
db.commit()
counts = {s: sum(1 for r in results if r["status"] == s) for s in ("ok", "unchanged", "failed")}
return {"action": action, "requested": len(results), **counts, "results": results}
@app.post("/api/auth/bulk-users/active")
def bulk_set_active(body: BulkActiveIn, actor: models.User = Depends(require_user_manager), db: Session = Depends(get_db)):
def apply(u):
if bool(u.is_active) == bool(body.is_active):
return "unchanged"
apply_user_active(db, actor, u, body.is_active)
return "ok"
return run_bulk(db, actor, body.user_ids, "enable" if body.is_active else "disable", apply)
@app.post("/api/auth/bulk-users/role")
def bulk_set_role(body: BulkRoleIn, actor: models.User = Depends(require_user_manager), db: Session = Depends(get_db)):
allowed = grantable_roles(actor)
if body.role not in allowed:
raise HTTPException(status_code=400, detail=f"role must be one of {', '.join(allowed)}")
def apply(u):
# Compared as stored: a legacy 'user' row moved to 'project_user' is a real
# change of the stored value, so it is applied and audited, as one at a time.
if u.role == body.role:
return "unchanged"
apply_user_role(db, actor, u, body.role)
return "ok"
return run_bulk(db, actor, body.user_ids, "role", apply)
@app.post("/api/auth/bulk-users/projects")
def bulk_set_project(body: BulkProjectIn, actor: models.User = Depends(require_user_manager), db: Session = Depends(get_db)):
"""Add every listed account to one project (with a per-project role), or remove
them from it. Implemented as the single route's replace of the caller's slice,
computed per user, so the audit row is the same shape the dialog writes."""
if body.op not in ("add", "remove"):
raise HTTPException(status_code=400, detail="op must be 'add' or 'remove'")
check_id(body.project_id)
if not db.get(models.Project, body.project_id):
raise HTTPException(status_code=404, detail="Project not found")
managed = managed_project_ids(db, actor)
if managed is not None and body.project_id not in managed:
raise HTTPException(status_code=403, detail="You don't administer the users of that project")
role = (body.role or "").strip()
if body.op == "add" and role and role not in project_scoped_grantable(actor):
raise HTTPException(
status_code=400,
detail=f"role must be '' (inherit) or one of {', '.join(project_scoped_grantable(actor))}")
pid = body.project_id
def apply(u):
mine = [m for m in db.scalars(select(models.ProjectMember)
.where(models.ProjectMember.user_id == u.id)).all()
if managed is None or m.project_id in managed]
current = {m.project_id: (m.role or "") for m in mine}
if body.op == "add":
if current.get(pid) == role:
return "unchanged"
current[pid] = role
else:
if pid not in current:
return "unchanged"
current.pop(pid)
# Roles on the caller's OTHER projects are carried over exactly as stored -
# adding someone to one job must never reset their role on another.
apply_user_projects(db, actor, u, managed, set(current),
{p: r for p, r in current.items() if r})
return "ok"
return run_bulk(db, actor, body.user_ids, "project_" + body.op, apply)
class BulkResolveIn(BaseModel):
text: str = Field(default="", max_length=200_000)
_CSV_HEADER_WORDS = ("username", "user", "login", "email", "user name")
@app.post("/api/auth/bulk-users/resolve")
def bulk_resolve(body: BulkResolveIn, actor: models.User = Depends(require_user_manager), db: Session = Depends(get_db)):
"""CR-020 (T12.4): turn an uploaded list of usernames into the account ids a
bulk action will run on - a second way IN to the same four bulk routes, not a
third copy of their rules. The CSV holds usernames only (decided 2026-10-06);
the action is chosen in the console and applies to every accepted row.
Validate and report, the CR-005 pattern: every line is either accepted or
rejected WITH ITS LINE NUMBER and a reason - never silently skipped. A line is
rejected when it names nobody this caller can see (the same 'not found' a
hidden account gets everywhere else), an account they may not change (with
the same reason the single route would give), their own account, or a
username already listed on an earlier line. Nothing is changed here; the
bulk route the console calls next re-checks every row anyway."""
rows = []
seen: dict[str, int] = {}
visible = visible_user_ids(db, actor)
cache: dict = {}
for line_no, raw in enumerate((body.text or "").splitlines(), start=1):
line = raw.strip().lstrip("")
if not line:
continue
cell = re.split(r"[,;\t]", line, maxsplit=1)[0].strip().strip('"').strip()
if not cell:
rows.append({"line": line_no, "text": line[:120], "status": "rejected",
"reason": "no username in the first column"})
continue
if line_no == 1 and cell.lower() in _CSV_HEADER_WORDS:
continue # header row
row = {"line": line_no, "text": cell[:120], "status": "rejected", "reason": "",
"user_id": None, "username": ""}
if any(ord(ch) < 32 for ch in cell) or len(cell) > 120:
row["reason"] = "not a valid username"
rows.append(row)
continue
key = cell.lower()
if key in seen:
row["reason"] = f"duplicate of line {seen[key]}"
rows.append(row)
continue
seen[key] = line_no
u = db.scalars(select(models.User).where(func.lower(models.User.username) == key)).first()
if u is None:
# An offboarding list usually carries email addresses. Accepted only
# when exactly one account has that address, so a shared or reused
# address can never pick the wrong person.
by_mail = db.scalars(select(models.User).where(func.lower(models.User.email) == key)).all()
u = by_mail[0] if len(by_mail) == 1 else None
if u is None or (visible is not None and u.id not in visible):
row["reason"] = "User not found"
elif u.id == actor.id:
row["reason"] = "your own account - change it one setting at a time"
else:
problem = manage_user_problem(db, actor, u, cache)
if problem:
row["reason"] = problem
else:
row.update(status="ok", user_id=u.id, username=u.username)
rows.append(row)
accepted = [r for r in rows if r["status"] == "ok"]
return {"read": len(rows), "accepted": len(accepted), "rejected": len(rows) - len(accepted),
"user_ids": [r["user_id"] for r in accepted], "rows": rows}
@app.post("/api/auth/bulk-users/delete")
def bulk_delete(body: BulkUsersIn, actor: models.User = Depends(require_user_manager), db: Session = Depends(get_db)):
"""Hard delete, the same as DELETE /api/auth/users/{id} (decided 2026-09-17:
no soft delete invented for the bulk case). The console asks for the typed
word DELETE first; the server's guard is the per-row rules, as for one."""
def apply(u):
apply_user_delete(db, actor, u)
return "ok"
return run_bulk(db, actor, body.user_ids, "delete", apply)
# ── Projects ─────────────────────────────────────────────────────────────────
@app.post("/api/projects")
def upsert_project(body: ProjectIn, user: models.User = Depends(auth.get_current_user), db: Session = Depends(get_db)):
@@ -3421,6 +3710,177 @@ def create_feedback(body: CommentIn, user: models.User = Depends(auth.get_curren
return _save_comment(body, db, user)
# ── CR-019: usage/activity metrics ──────────────────────────────────────────
class UsageIn(BaseModel):
tool: str = ""
project_id: Optional[str] = None
@app.post("/api/usage/ping")
def usage_ping(body: UsageIn, user: models.User = Depends(auth.get_current_user), db: Session = Depends(get_db)):
"""One usage_events row per authenticated page load. Called exactly once,
from auth-guard.js after wp-auth-ready fires (T11.2) - never duplicated
per page's own script, the same lesson S4's per-page nav already taught
this codebase. `user` comes from the session via get_current_user, never
from anything the client claims - identity here is a server-enforced
fact, matching every other write in this file, not a client-reported one."""
tool = (body.tool or "").strip()[:40]
db.add(models.UsageEvent(
id=gen_id("uev"), username=user.username, project_id=body.project_id,
tool=tool, event="page_open",
))
db.commit()
return {"ok": True}
def _parse_date_q(v: str, end: bool = False) -> Optional[datetime]:
"""Accepts a plain YYYY-MM-DD (what a <input type=date> sends) or a full
ISO datetime. A date-only `to` means "through the end of that day", not
midnight at its start - otherwise a range of "today" would match nothing
from today at all."""
if not v:
return None
try:
d = datetime.fromisoformat(v)
except ValueError:
return None
if d.tzinfo is None:
d = d.replace(tzinfo=timezone.utc)
if end and len(v) <= 10: # date-only
d = d + timedelta(days=1) - timedelta(microseconds=1)
return d
def _usage_query(db: Session, caller: "models.User", date_from, date_to, project_id, username, tool):
"""Shared by the summary and export endpoints so the two can never
disagree about which rows a filter set matches - the export is a raw
dump of exactly what the summary counted, not a separately-derived view."""
stmt = select(models.UsageEvent)
managed = managed_project_ids(db, caller)
if managed is not None:
# A project_super_user (never an app admin, who gets managed=None) is
# scoped to events tied to a project they administer, plus their OWN
# suite-wide activity (admin console opens etc. carry no project_id) -
# never another user's activity outside what they manage.
if managed:
stmt = stmt.where(or_(
models.UsageEvent.project_id.in_(managed),
models.UsageEvent.username == caller.username,
))
else:
stmt = stmt.where(models.UsageEvent.username == caller.username)
df = _parse_date_q(date_from)
dt = _parse_date_q(date_to, end=True)
if df:
stmt = stmt.where(models.UsageEvent.at >= df)
if dt:
stmt = stmt.where(models.UsageEvent.at <= dt)
if project_id:
stmt = stmt.where(models.UsageEvent.project_id == project_id)
if username:
stmt = stmt.where(models.UsageEvent.username == username)
if tool:
stmt = stmt.where(models.UsageEvent.tool == tool)
return stmt.order_by(models.UsageEvent.at)
@app.get("/api/usage/summary")
def usage_summary(
date_from: Optional[str] = Query(None, alias="from"),
date_to: Optional[str] = Query(None, alias="to"),
project_id: Optional[str] = Query(None),
username: Optional[str] = Query(None),
tool: Optional[str] = Query(None),
caller: models.User = Depends(require_user_manager),
db: Session = Depends(get_db),
):
"""CR-019. Same gate as the User Directory (require_user_manager): an app
admin or a project_super_user with at least one managed project. Filters
combine. Aggregated in Python over the filtered row set rather than a SQL
GROUP BY - correct and simple at today's scale; if usage_events grows
into the millions (plausible, given retention is indefinite by decision),
the day/week/month bucketing here is the first thing to move server-side
into SQL. Not done now because nothing currently requires it."""
rows = db.scalars(_usage_query(db, caller, date_from, date_to, project_id, username, tool)).all()
by_day: dict[str, set] = {}
by_week: dict[str, set] = {}
by_month: dict[str, set] = {}
per_user_last: dict[str, datetime] = {}
per_tool: dict[str, int] = {}
for e in rows:
by_day.setdefault(e.at.date().isoformat(), set()).add(e.username)
by_week.setdefault(e.at.strftime("%G-W%V"), set()).add(e.username)
by_month.setdefault(e.at.strftime("%Y-%m"), set()).add(e.username)
if e.username not in per_user_last or e.at > per_user_last[e.username]:
per_user_last[e.username] = e.at
per_tool[e.tool] = per_tool.get(e.tool, 0) + 1
return {
"active_users": {
"by_day": {k: len(v) for k, v in sorted(by_day.items())},
"by_week": {k: len(v) for k, v in sorted(by_week.items())},
"by_month": {k: len(v) for k, v in sorted(by_month.items())},
},
"per_user_last_active": {u: models._iso(t) for u, t in sorted(per_user_last.items())},
"by_tool": dict(sorted(per_tool.items(), key=lambda kv: -kv[1])),
"event_count": len(rows),
}
def _pseudonym(username: str) -> str:
"""A stable per-user id for the sanitized export — the SAME input always
produces the SAME output, within one export and across separate export
runs, so an external system (Power BI or similar) can still group and
trend "by user" without ever receiving a real name. HMAC rather than a
plain hash: a plain sha256(username) is trivially reversed against a
wordlist of the handful of usernames this app actually has; keying it
with AUTH_SECRET_KEY (already a real secret, already required in
production — see auth.py) means recovering a username from its
pseudonym requires the signing key, not just guessing."""
digest = hmac.new(auth.SECRET_KEY.encode(), username.encode(), hashlib.sha256).hexdigest()
return "u_" + digest[:16]
@app.get("/api/usage/export")
def usage_export(
date_from: Optional[str] = Query(None, alias="from"),
date_to: Optional[str] = Query(None, alias="to"),
project_id: Optional[str] = Query(None),
username: Optional[str] = Query(None),
tool: Optional[str] = Query(None),
sanitize: bool = Query(False),
caller: models.User = Depends(require_user_manager),
db: Session = Depends(get_db),
):
"""CR-019. Same gate, same filters, same underlying row set as
usage_summary() (_usage_query) — the export can never show a different
slice of data than what the console counted for the same filters.
sanitize=true replaces `username` with a stable pseudonym (_pseudonym)
and — deliberately — the `detail` column is not exported in EITHER mode.
Every event this app writes today (login, page_open) leaves `detail`
empty, so this costs nothing now, but it also means a future event type
that DOES populate `detail` can't accidentally leak a real name into a
sanitized file through a column nobody thought to scrub. If `detail`
is ever needed in the export, it has to be sanitized explicitly, not
assumed safe because the rest of the row was."""
rows = db.scalars(_usage_query(db, caller, date_from, date_to, project_id, username, tool)).all()
buf = io.StringIO()
w = csv.writer(buf)
w.writerow(["at", "username", "project_id", "tool", "event"])
for e in rows:
who = _pseudonym(e.username) if sanitize else e.username
w.writerow([models._iso(e.at), who, e.project_id or "", e.tool, e.event])
filename = "usage_export_%s_%s.csv" % (
"sanitized" if sanitize else "raw", datetime.now(timezone.utc).strftime("%Y%m%d"),
)
return Response(
content=buf.getvalue(), media_type="text/csv",
headers={"Content-Disposition": f'attachment; filename="{filename}"'},
)
@app.get("/api/comments")
def list_comments(
source: Optional[str] = Query(None),

View File

@@ -59,8 +59,10 @@ log = logging.getLogger("wpsuite.auth")
COOKIE_NAME = "wp_session"
JWT_ALG = "HS256"
# How long a session lasts before the person must sign in again.
SESSION_HOURS = int(os.getenv("AUTH_SESSION_HOURS", "12"))
# How long a session lasts before the person must sign in again. Default 8 hours (D25).
# One flat limit from the moment of sign-in. When T13.1 lands, the same variable becomes
# the absolute ceiling and an idle timer is added under it.
SESSION_HOURS = int(os.getenv("AUTH_SESSION_HOURS", "8"))
# ── permissions roles ─────────────────────────────────────────────────────────
ROLE_ADMIN = "admin"
@@ -191,8 +193,7 @@ def decode_token(token: str) -> Optional[dict]:
if not token or len(token) > _MAX_TOKEN_CHARS:
return None
try:
# Every token this app issues carries exp and sub (create_token,
# maybe_refresh_token); one without them is not ours.
# Every token this app issues carries exp and sub (create_token); one without them is not ours.
claims = jwt.decode(token, SECRET_KEY, algorithms=[JWT_ALG],
options={"require": ["exp", "sub"]})
except Exception: # noqa: BLE001 — see docstring: every failure is "not signed in"

View File

@@ -337,6 +337,45 @@ class AuditLog(Base):
}
class UsageEvent(Base):
"""CR-019: append-only record of who used the suite, when, and which tool —
navigation/session activity, not business mutations. Deliberately a SEPARATE
table from AuditLog rather than a new `action` value there: AuditLog answers
"who changed what" and is read by people auditing a specific record's
history; mixing in a `page_open` row for every authenticated page load
would make that trail noisy for its existing purpose. This table answers a
different question — "who is active, and on what" — and CR-019's admin
console reads from here, not from AuditLog.
Not a ForeignKey to `users`, matching AuditLog's own reasoning: a user who
is later removed should still show up in historical activity rather than
silently vanishing from it, and `D18`'s deprovisioning sync only ever sets
`is_active=False` — it never deletes a row — so this is defensive symmetry
rather than a live concern today.
Retention is indefinite (decided 2026-09-17, `decisions-2026-09-17.md`) —
nothing purges rows written here; that is a deliberate product decision,
not an oversight to fix later."""
__tablename__ = "usage_events"
id: Mapped[str] = mapped_column(String(40), primary_key=True)
at: Mapped[datetime] = mapped_column(DateTime(timezone=True), default=utcnow, index=True)
username: Mapped[str] = mapped_column(String(200), default="", index=True)
project_id: Mapped[Optional[str]] = mapped_column(String(40), nullable=True, index=True)
# creator | wizard | field_view | dashboard | admin | directory | ...
tool: Mapped[str] = mapped_column(String(40), default="", index=True)
# page_open | login
event: Mapped[str] = mapped_column(String(40), default="", index=True)
detail: Mapped[dict] = mapped_column(JSON, default=dict)
def to_dict(self) -> dict:
return {
"id": self.id, "at": _iso(self.at), "username": self.username,
"project_id": self.project_id, "tool": self.tool, "event": self.event,
"detail": self.detail or {},
}
class AppSetting(Base):
"""Admin-editable application settings (feature flags, SMTP config, …) stored
as key -> JSON value. Read/written via /api/settings (admin only). Secrets like

815
server/okta_sync.py Normal file
View File

@@ -0,0 +1,815 @@
"""Okta Management API client - T13.2, wave 13 (D18).
This module asks Okta one question: who is in the org, and what is each person's
status? It is the first half of the deprovisioning sync. It does not touch the
database and it does not disable anyone. T13.3 does that, using the list this
module returns.
Trust boundary. Sign-in uses the OIDC client (OKTA_CLIENT_ID and OKTA_CLIENT_SECRET,
see okta_auth.py). This module uses a different credential, OKTA_API_TOKEN, a
read-only Okta API token. The two never mix. A rotation or a leak of one does not
touch the other.
Failure rule (D18, decision 4). This is the most important rule in this file:
**an error is never an empty list.** If anything goes wrong, this module raises
OktaSyncError. It never returns [] to mean "the call failed". T13.3 reads an empty
list as "nobody is active", so a function that returned [] on a network error
would disable every account on an Okta outage. The cases that raise are:
- the token or the org URL is not set
- a network error or a timeout
- any HTTP status that is not 200 (401, 403 and 429 included)
- a body that is not a JSON list, or an item that is not shaped as expected
- a user with no usable identity value
- a list with no users at all (a real org is never empty; an empty list means
the token cannot see the users)
- a page limit reached (a cut-off list would make real users look missing)
- a "next page" link that points to a different host
T13.3 (the sync job, `run_sync` below) uses that list to disable local accounts.
Its rules are in its own docstring. In short: it only ever disables, it never
re-enables, it writes one audit row per change, and any failure means NO change.
The token is read from the environment each time it is used. It is never logged,
never put in an error message, and never returned to a browser. Errors carry the
exception class name and the HTTP status only. httpx exceptions can include the
request URL, so the URL is not copied into the message either.
"""
import base64
import hashlib
import json
import logging
import os
import re
import threading
import time
import uuid
from datetime import datetime, timedelta, timezone
from dataclasses import dataclass, field
from typing import Callable, Optional
from urllib.parse import urlparse
import httpx
import jwt
from cryptography.hazmat.primitives.asymmetric import ec
from sqlalchemy import select, update
from sqlalchemy.exc import IntegrityError
from sqlalchemy.orm import Session
from . import models, okta_auth
log = logging.getLogger("wpsuite.okta_sync")
def _make_logs_visible() -> None:
"""Make this module's INFO lines reach `docker compose logs api`.
Nothing in this app configures Python logging, and gunicorn only configures its
own loggers. So the root logger sits at WARNING with no handler, and every
INFO line from a `wpsuite.*` logger is dropped. Found on 2026-10-16: with the
exact production launch command, `grep "Okta sync"` found nothing on a healthy
start, because "configured", "scheduled" and "done" are INFO lines. Warnings
and errors (a disable, a skipped run) showed, but the lines that say the sync
is alive did not. An operator cannot tell "off" from "on and quiet".
This adds one stderr handler to THIS module's logger only, and only when the
root logger has none, so it never doubles up if logging is configured later and
it does not change what any other module prints. (BL-028 and the general
question of app-wide logging are separate items.)"""
if log.handlers or logging.getLogger().handlers:
return
handler = logging.StreamHandler()
handler.setFormatter(logging.Formatter("%(asctime)s %(levelname)s %(name)s: %(message)s"))
log.addHandler(handler)
log.setLevel(logging.INFO)
_make_logs_visible()
# Okta's own page maximum for this endpoint. Asking for the maximum keeps a
# 20-odd account org to a single call.
PAGE_LIMIT = 200
# A cut-off list is more dangerous than an error, so the loop is bounded and
# raises when the bound is reached. 50 pages of 200 is 10,000 people.
MAX_PAGES = 50
TIMEOUT_SECONDS = 15.0
# The whole fetch, all pages together, must finish inside this. T13.4 relies on it:
# a run is always shorter than the shortest allowed interval, so one cycle can
# never still be running when the next is due.
TOTAL_DEADLINE_SECONDS = 120.0
# Which Okta *profile attribute* matches each OIDC claim that OKTA_IDENTITY_CLAIM
# may name. The sign-in flow reads a claim from the ID token. This API returns
# the profile instead. For the default claim, preferred_username, Okta's own
# default maps it to profile.login. A claim that is not in this table is read
# from profile[<claim>] (for example a custom attribute).
# OKTA_SYNC_PROFILE_ATTRIBUTE overrides the whole table. Use it if the tenant
# builds the claim from a different attribute.
_CLAIM_TO_PROFILE = {
"preferred_username": "login",
"email": "email",
}
class OktaSyncError(Exception):
"""The Okta user list could not be fetched with confidence. T13.3 must take no
action on any account when it sees this."""
@dataclass(frozen=True)
class OktaUser:
identity: str # the value that matches users.username (lower case is NOT applied here)
status: str # Okta's status, upper case: ACTIVE, SUSPENDED, DEPROVISIONED, ...
# ── configuration ─────────────────────────────────────────────────────────────
def _token() -> str:
return os.getenv("OKTA_API_TOKEN", "").strip()
def org_url() -> str:
"""The Okta org base URL, for example https://yourorg.okta.com. The Management
API lives here, not at the authorization server issuer. OKTA_ORG_URL wins if
set. Otherwise it is the scheme and host of OKTA_ISSUER."""
explicit = os.getenv("OKTA_ORG_URL", "").strip().rstrip("/")
if explicit:
return explicit
issuer = (okta_auth.ISSUER or "").strip()
parsed = urlparse(issuer)
if parsed.scheme and parsed.netloc:
return f"{parsed.scheme}://{parsed.netloc}"
return ""
# D23: a second way to prove who we are to Okta. Okta apps of type "API Services"
# use OAuth 2.0 with client authentication private_key_jwt: this app holds a private
# key (a JWK), signs a short JWT with it, trades the JWT for a one-hour access token
# that carries only the scope okta.users.read, and sends that as a Bearer token. The
# private key never leaves this process. An SSWS API token (OKTA_API_TOKEN) still
# works and wins if both are set, so a deploy that has only a token behaves as before.
OAUTH_SCOPE = "okta.users.read"
ASSERTION_LIFETIME_SECONDS = 300
def _client_id() -> str:
return os.getenv("OKTA_SYNC_CLIENT_ID", "").strip()
def _key_raw() -> str:
return os.getenv("OKTA_SYNC_PRIVATE_KEY", "").strip()
def auth_method() -> str:
""""token" (SSWS API token), "oauth" (service app with a private key), or "" (none)."""
if _token():
return "token"
if _client_id() and _key_raw():
return "oauth"
return ""
def _parse_json_forgiving(text: str):
"""Parse the key text as JSON, forgiving what copying does to it. Returns
(object, None) or (None, "problem at line L column C: what"). The problem names a
position and a rule, never any text of the key. Forgiven: line breaks and other control
characters inside a string (a long value wrapped over several lines), curly quote
marks, and a comma before a closing brace. Spaces inside the key's own number fields
are removed, because those fields are base64 text."""
attempts = [text]
cleaned = (text.replace("\u201c", '"').replace("\u201d", '"').replace("\u2018", "'").replace("\u2019", "'"))
cleaned = re.sub(r",\s*([}\]])", r"\1", cleaned)
if cleaned != text:
attempts.append(cleaned)
last = ""
for t in attempts:
try:
obj = json.loads(t, strict=False)
except json.JSONDecodeError as exc:
last = f"{exc.msg} at line {exc.lineno} column {exc.colno}"
continue
except ValueError:
last = "unreadable"
continue
if isinstance(obj, dict):
for k in ("n", "e", "d", "p", "q", "dp", "dq", "qi", "x", "y"):
if isinstance(obj.get(k), str):
obj[k] = re.sub(r"\s+", "", obj[k])
return obj, None
return None, last
def _load_jwk() -> dict:
"""OKTA_SYNC_PRIVATE_KEY is the private key as a JWK. Two forms are accepted:
the JSON itself (it starts with "{"), or that JSON encoded as base64. Use base64
in a Portainer environment field: it has no quote marks to be mangled.
The reader forgives the usual copy and paste damage: quote marks around the whole
value, spaces and line breaks inside it, missing "=" padding, the URL-safe base64
alphabet, a byte-order mark at the start of the file, and a JWK Set
({"keys": [one key]}) in place of a single key. Raises OktaSyncError that says
which step failed. A message never holds any text of the key, only counts."""
raw = _key_raw()
if not raw:
raise _fail("OKTA_SYNC_PRIVATE_KEY is not set")
if len(raw) >= 2 and raw[0] == raw[-1] and raw[0] in "\"'" and raw[0] not in raw[1:-1]:
raw = raw[1:-1].strip()
if raw.startswith(("{", '"', "-----BEGIN")):
text = raw
else:
compact = re.sub(r"\s+", "", raw)
compact += "=" * (-len(compact) % 4)
decoded = None
for decode in (base64.b64decode, base64.urlsafe_b64decode):
try:
decoded = decode(compact)
break
except Exception: # noqa: BLE001
continue
if decoded is None:
raise _fail(f"OKTA_SYNC_PRIVATE_KEY is not valid base64 ({len(compact)} characters after "
"removing spaces; it should be one long line of letters, digits, + / =)")
try:
text = decoded.decode("utf-8-sig")
except UnicodeDecodeError:
raise _fail("OKTA_SYNC_PRIVATE_KEY decodes from base64 but the result is not text") from None
text = text.lstrip("\ufeff").strip()
via = "" if raw.startswith(("{", '"', "-----BEGIN")) else " after decoding base64"
if text.startswith("-----BEGIN"):
raise _fail("OKTA_SYNC_PRIVATE_KEY holds a PEM key (-----BEGIN ...). A JWK is needed: "
"JSON with the fields kty, kid, n, e and d")
if text.startswith('"'):
text = "{" + text.rstrip(",") + "}" # the fields were copied without the outer braces
jwk, problem = _parse_json_forgiving(text)
if problem is not None and not text.startswith("{"):
# One more layer? A value that was encoded twice decodes to base64 text.
try:
squeezed = re.sub(r"\s+", "", text)
again = base64.b64decode(squeezed + "=" * (-len(squeezed) % 4)).decode("utf-8-sig").strip()
jwk, problem = _parse_json_forgiving(again)
except Exception: # noqa: BLE001
pass
if problem is not None:
shape = ("starts with a brace" if text.startswith("{") else
"starts with a letter or digit" if text[:1].isalnum() else "starts with a symbol")
raise _fail(f"OKTA_SYNC_PRIVATE_KEY is not JSON{via} ({len(text)} characters, {shape}; {problem})")
if isinstance(jwk, dict) and isinstance(jwk.get("keys"), list):
if len(jwk["keys"]) != 1:
raise _fail("OKTA_SYNC_PRIVATE_KEY is a key set with more than one key; paste one key")
jwk = jwk["keys"][0]
if not isinstance(jwk, dict) or jwk.get("kty") not in {"RSA", "EC"}:
raise _fail("OKTA_SYNC_PRIVATE_KEY is JSON but not an RSA or EC key")
if "d" not in jwk:
raise _fail("OKTA_SYNC_PRIVATE_KEY holds a PUBLIC key; the private key is needed to sign")
return jwk
def _client_assertion(jwk: dict, client_id: str, audience: str) -> str:
"""The signed JWT that proves we hold the key. Okta's rules: iss and sub are the
client ID, aud is the token endpoint, and the lifetime is at most 60 minutes."""
alg = jwk.get("alg") or ("RS256" if jwk["kty"] == "RSA" else "ES256")
algorithms = jwt.algorithms.get_default_algorithms()
if alg not in algorithms:
raise _fail(f"the key names an algorithm that is not supported ({alg})")
try:
key = algorithms[alg].from_jwk(json.dumps(jwk))
now = int(time.time())
return jwt.encode(
{"iss": client_id, "sub": client_id, "aud": audience, "iat": now,
"exp": now + ASSERTION_LIFETIME_SECONDS, "jti": uuid.uuid4().hex},
key, algorithm=alg, headers={"kid": jwk["kid"]} if jwk.get("kid") else None)
except OktaSyncError:
raise
except Exception as exc: # noqa: BLE001
raise _fail(f"the key could not sign ({type(exc).__name__})") from None
def _dpop_mode() -> str:
"""OKTA_SYNC_DPOP (D24): "auto" (the default), "on" or "off".
auto: ask for a plain Bearer token first. If Okta answers invalid_dpop_proof, the
Okta app requires DPoP, and the sync asks again with DPoP. Nothing to set.
on: always use DPoP.
off: never use DPoP. A DPoP-bound token is then refused with a clear message."""
v = os.getenv("OKTA_SYNC_DPOP", "").strip().lower()
if v in {"1", "true", "yes", "on"}:
return "on"
if v in {"0", "false", "no", "off"}:
return "off"
return "auto"
def _url_no_query(url: str) -> str:
p = urlparse(url)
return f"{p.scheme}://{p.netloc}{p.path}"
class _Dpop:
"""DPoP (RFC 9449): proof that the caller holds a key, sent with every request, so
a stolen access token alone is useless. One throwaway EC key (P-256) is made for each
run and is never stored or logged. Each request gets its own signed proof: the method
(htm), the URL without its query (htu), a one-time id (jti), the time (iat), the
server's latest nonce if it gave one, and for an API call a hash of the access token
(ath). The public half of the key travels in the proof header."""
def __init__(self):
self._key = ec.generate_private_key(ec.SECP256R1())
pub = json.loads(jwt.algorithms.ECAlgorithm.to_jwk(self._key.public_key()))
self._pub = {k: pub[k] for k in ("kty", "crv", "x", "y")}
self.nonce: Optional[str] = None
def proof(self, method: str, url: str, access_token: Optional[str] = None) -> str:
claims = {"jti": uuid.uuid4().hex, "htm": method.upper(), "htu": _url_no_query(url),
"iat": int(time.time())}
if self.nonce:
claims["nonce"] = self.nonce
if access_token:
claims["ath"] = base64.urlsafe_b64encode(
hashlib.sha256(access_token.encode()).digest()).rstrip(b"=").decode()
return jwt.encode(claims, self._key, algorithm="ES256",
headers={"typ": "dpop+jwt", "jwk": self._pub})
def take_nonce(self, resp: httpx.Response) -> Optional[str]:
n = resp.headers.get("DPoP-Nonce")
if n:
self.nonce = n
return n
class _Credential:
"""What is sent with each Management API call: an SSWS token, a Bearer token, or a
DPoP-bound token with a fresh proof for every request."""
def __init__(self, scheme: str, token: str, dpop: Optional[_Dpop] = None):
self.scheme, self.token, self.dpop = scheme, token, dpop
def headers(self, method: str, url: str) -> dict:
h = {"Accept": "application/json"}
if self.dpop is not None:
h["Authorization"] = f"DPoP {self.token}"
h["DPoP"] = self.dpop.proof(method, url, self.token)
else:
h["Authorization"] = f"{self.scheme} {self.token}"
return h
def _error_code(resp: httpx.Response) -> str:
"""Okta's OAuth error code (invalid_client, invalid_scope, invalid_dpop_proof, ...).
It tells an admin what to fix and holds no secret. Kept only if it is a short code."""
try:
c = resp.json().get("error", "")
if isinstance(c, str) and re.fullmatch(r"[a-z_]{1,40}", c):
return c
except Exception: # noqa: BLE001
pass
return ""
def _access_token(client: httpx.Client, base: str) -> _Credential:
"""Trade a signed assertion for an access token. Raises OktaSyncError for every
failure. Neither the key, the assertion nor the token is logged or put in a message.
Each attempt gets a new assertion, because Okta accepts an assertion id only once."""
client_id = _client_id()
if not client_id:
raise _fail("OKTA_SYNC_CLIENT_ID is not set")
endpoint = f"{base}/oauth2/v1/token"
jwk = _load_jwk()
mode = _dpop_mode()
dpop: Optional[_Dpop] = _Dpop() if mode == "on" else None
nonce_retries = 0
resp = None
for _ in range(4):
headers = {"Accept": "application/json"}
if dpop is not None:
headers["DPoP"] = dpop.proof("POST", endpoint)
try:
resp = client.post(
endpoint, headers=headers, follow_redirects=False,
data={"grant_type": "client_credentials", "scope": OAUTH_SCOPE,
"client_assertion_type": "urn:ietf:params:oauth:client-assertion-type:jwt-bearer",
"client_assertion": _client_assertion(jwk, client_id, endpoint)})
except httpx.HTTPError as exc:
raise _fail(f"network error on the token request ({type(exc).__name__})") from None
if dpop is not None:
dpop.take_nonce(resp)
if resp.status_code == 200:
break
code = _error_code(resp)
if mode == "auto" and dpop is None and code == "invalid_dpop_proof":
dpop = _Dpop() # the Okta app requires DPoP: ask again with it
continue
if dpop is not None and code == "use_dpop_nonce" and dpop.nonce and nonce_retries < 2:
nonce_retries += 1 # Okta asks for its nonce in the proof: ask again with it
continue
raise _fail(f"Okta answered HTTP {resp.status_code} to the token request"
+ (f", error {code}" if code else "")
+ (" (DPoP was used)" if dpop is not None else ""))
else:
raise _fail("the token request did not succeed after retries")
try:
body = resp.json()
except ValueError:
raise _fail("the token response is not JSON") from None
tok = body.get("access_token") if isinstance(body, dict) else None
if not isinstance(tok, str) or not tok:
raise _fail("the token response has no access token")
kind = str(body.get("token_type", "")).lower()
if kind == "dpop":
if dpop is None:
raise _fail("the token is DPoP-bound but OKTA_SYNC_DPOP is off")
return _Credential("DPoP", tok, dpop)
if kind == "bearer":
return _Credential("Bearer", tok)
raise _fail("the token response is neither a Bearer nor a DPoP token")
def is_configured() -> bool:
return bool(auth_method() and org_url().startswith("https://"))
def describe() -> str:
"""One line for the startup log. Never contains the token."""
method = auth_method()
if not method:
return ("Okta sync DISABLED - set OKTA_API_TOKEN, or OKTA_SYNC_CLIENT_ID and "
"OKTA_SYNC_PRIVATE_KEY. Deprovisioning is not detected.")
base = org_url()
if not base:
return "Okta sync DISABLED - no OKTA_ORG_URL and OKTA_ISSUER has no host. Deprovisioning is not detected."
if not base.startswith("https://"):
return f"Okta sync DISABLED - org URL {base!r} is not https. The token is never sent over http."
mode = "DRY RUN, no account is changed" if dry_run() else "live"
how = ("API token" if method == "token"
else f"OAuth service app (private_key_jwt), DPoP {_dpop_mode()}")
return (f"Okta sync configured ({mode}) - org {base}, credential: {how}, "
f"identity claim {okta_auth.IDENTITY_CLAIM!r}")
def dry_run() -> bool:
"""OKTA_SYNC_DRY_RUN (D22). When it is 1, true, yes or on, the sync does its whole
job except the change: it logs who it WOULD disable and writes nothing. Off by
default, so a deploy that does not set it behaves as before. Read at each run,
so the value in force is the one the worker started with."""
return os.getenv("OKTA_SYNC_DRY_RUN", "").strip().lower() in {"1", "true", "yes", "on"}
def _profile_attribute() -> str:
override = os.getenv("OKTA_SYNC_PROFILE_ATTRIBUTE", "").strip()
if override:
return override
claim = okta_auth.IDENTITY_CLAIM
return _CLAIM_TO_PROFILE.get(claim, claim)
# ── the client ────────────────────────────────────────────────────────────────
def _fail(reason: str) -> "OktaSyncError":
# One place to build the error, so no call site can add a URL or a header by
# accident.
return OktaSyncError(f"Okta user list could not be fetched: {reason}")
def _same_origin(url: str, base: str) -> bool:
a, b = urlparse(url), urlparse(base)
return (a.scheme, a.netloc) == (b.scheme, b.netloc)
def _parse_user(item, attribute: str) -> OktaUser:
if not isinstance(item, dict):
raise _fail("a list item is not an object")
status = item.get("status")
profile = item.get("profile")
if not isinstance(status, str) or not status.strip():
raise _fail("a user has no status")
if not isinstance(profile, dict):
raise _fail("a user has no profile")
value = profile.get(attribute)
if not isinstance(value, str) or not value.strip():
raise _fail(f"a user has no value for profile attribute {attribute!r}")
return OktaUser(identity=value.strip(), status=status.strip().upper())
def _get_page(client: httpx.Client, cred: _Credential, url: str, params: Optional[dict]) -> httpx.Response:
"""One GET, with the credential's headers. With DPoP, each call has its own proof,
and if the server answers use_dpop_nonce the call is repeated once with its nonce."""
for attempt in (0, 1):
resp = client.get(url, params=params, headers=cred.headers("GET", url), follow_redirects=False)
if cred.dpop is None:
return resp
nonce = cred.dpop.take_nonce(resp)
asks = resp.status_code in (400, 401) and nonce and (
"use_dpop_nonce" in resp.headers.get("WWW-Authenticate", "") or "use_dpop_nonce" in resp.text)
if attempt == 0 and asks:
continue
return resp
return resp
def fetch_users(http: Optional[httpx.Client] = None) -> "list[OktaUser]":
"""Return every user Okta lists, as OktaUser values. Raises OktaSyncError on
any problem; never returns an empty list. `http` is for tests (a client built
on httpx.MockTransport). In production it is None and a short-lived client is
made here."""
method = auth_method()
base = org_url()
if not method:
raise _fail("no Okta credential is set (OKTA_API_TOKEN, or OKTA_SYNC_CLIENT_ID and OKTA_SYNC_PRIVATE_KEY)")
if not base.startswith("https://"):
raise _fail("the org URL is missing or is not https")
attribute = _profile_attribute()
own_client = http is None
client = http if http is not None else httpx.Client(timeout=TIMEOUT_SECONDS)
users: "list[OktaUser]" = []
started = time.monotonic()
try:
cred = (_Credential("SSWS", _token()) if method == "token"
else _access_token(client, base))
url: Optional[str] = f"{base}/api/v1/users"
params: Optional[dict] = {"limit": PAGE_LIMIT}
for _ in range(MAX_PAGES):
if not url:
break
if time.monotonic() - started > TOTAL_DEADLINE_SECONDS:
raise _fail(f"took longer than {int(TOTAL_DEADLINE_SECONDS)} seconds in total")
try:
resp = _get_page(client, cred, url, params)
except httpx.HTTPError as exc:
# Class name only. The exception text can carry the URL.
raise _fail(f"network error ({type(exc).__name__})") from None
if resp.status_code != 200:
raise _fail(f"Okta answered HTTP {resp.status_code}")
try:
body = resp.json()
except ValueError:
raise _fail("the response is not JSON") from None
if not isinstance(body, list):
raise _fail("the response is not a list of users")
users.extend(_parse_user(item, attribute) for item in body)
nxt = resp.links.get("next", {}).get("url")
if not nxt:
url = None
break
if not _same_origin(nxt, base):
# Never send the token to a host that the Link header names.
raise _fail("a next-page link points to a different host")
url, params = nxt, None # the next link already carries its query
else:
# The loop used every page and the last one still named a next page.
raise _fail(f"more than {MAX_PAGES} pages; the list may be cut off")
finally:
if own_client:
client.close()
if not users:
raise _fail("Okta returned no users, which is not credible for a real org")
log.info("Okta user list fetched: %d user(s)%s", len(users),
" (with DPoP)" if cred.dpop is not None else "")
return users
# ── T13.3: the sync job ───────────────────────────────────────────────────────
# Which Okta statuses mean "this person is gone or was blocked on purpose".
# Decided by Matt, 2026-10-16: a literal "anything that is not ACTIVE" would also
# catch PASSWORD_EXPIRED, LOCKED_OUT and RECOVERY, which people leave on their
# own, and because this job never re-enables, one AD password expiry would lock a
# person out of this app until an admin stepped in. A user who is MISSING from the
# list disables too (Okta's default list leaves out DEPROVISIONED users).
DISABLING_STATUSES = frozenset({"SUSPENDED", "DEPROVISIONED"})
SYNC_ACTOR = "system:okta_sync"
SYNC_ACTION = "user_deprovisioned_by_sync"
@dataclass
class SyncResult:
ok: bool # False = the run took NO action at all
error: str = "" # why, when ok is False
okta_users: int = 0 # size of the Okta list
checked: int = 0 # active local accounts looked at
disabled: "list[str]" = field(default_factory=list) # usernames disabled this run
# (in a dry run: the ones it WOULD disable)
dry_run: bool = False # True = nothing was changed on purpose (D22)
def _gen_id(prefix: str) -> str:
# Same shape as server/app.py gen_id. Not imported: app.py will import this
# module (T13.4), and the reverse import would be circular.
return f"{prefix}_{uuid.uuid4().hex[:12]}"
def run_sync(db: Session, fetch: Optional[Callable[[], "list[OktaUser]"]] = None,
dry: Optional[bool] = None) -> SyncResult:
"""Compare the Okta user list to local accounts and disable the ones Okta no
longer vouches for.
The rules, none of which may be relaxed (D18):
1. FAIL CLOSED. If fetching raises ANY exception, or the answer cannot be
trusted, this function changes nothing and writes nothing. A failure is
never read as "nobody is active". The fetch happens before the database
is read, so there is no code path from a failed fetch to a change.
2. ONLY DISABLE. It sets is_active to False. No input makes it set True.
An account that is already disabled is not touched and gets no new
audit row, so a run that finds nothing new writes nothing.
3. One AuditLog row per disabled account, actor "system:okta_sync", so it
is never confused with an admin's action. Rows and changes are saved in
one transaction. If saving fails, none of it is kept.
4. Only changes are logged. A person who is still active in Okta and still
active here produces no row.
One guard beyond the rules above, inside decision 4's "anything the sync can't
confidently parse": if there are active local accounts and NONE of them
matches any Okta user, the identity attribute is almost certainly wrong (see
OKTA_SYNC_PROFILE_ATTRIBUTE). That is a setup mistake, not an offboarding of
the whole company, so the run is refused and nothing changes.
A local account matches an Okta user by username or by email, ignoring case,
the same two fields auth.find_user() uses. If the Okta list holds the same
identity twice with different statuses, the account is disabled only when
every one of them is a disabling status.
DRY RUN (D22, OKTA_SYNC_DRY_RUN): every step above runs, including the fetch and
both guards, but the account changes and the audit rows are skipped. The run logs
who it would disable. Use it for the first deploy, to see the match count and the
list before any account is touched.
`fetch` and `dry` are for tests. In production fetch is None and fetch_users() is
used, and dry is None so the environment decides.
"""
fetch = fetch or fetch_users
dry = dry_run() if dry is None else dry
# 1. The fetch. Nothing below runs unless it returned a list.
try:
okta_users = fetch()
except Exception as exc: # an error of ANY kind, not only OktaSyncError
msg = str(exc) if isinstance(exc, OktaSyncError) else f"unexpected {type(exc).__name__}"
log.error("Okta sync SKIPPED, no account was changed: %s", msg)
return SyncResult(ok=False, error=msg)
if not okta_users:
# fetch_users() already raises on an empty list. This covers a stand-in
# fetch (and any future change) that returns one: same rule.
msg = "the fetch returned no users"
log.error("Okta sync SKIPPED, no account was changed: %s", msg)
return SyncResult(ok=False, error=msg)
try:
by_identity: "dict[str, list[str]]" = {}
for u in okta_users:
by_identity.setdefault(u.identity.strip().lower(), []).append(u.status.strip().upper())
locals_ = list(db.scalars(select(models.User).where(models.User.is_active.is_(True))))
plan: "list[tuple[models.User, str]]" = []
matched = 0
for local in locals_:
statuses: "list[str]" = []
for key in {(local.username or "").strip().lower(), (local.email or "").strip().lower()}:
if key:
statuses.extend(by_identity.get(key, []))
if statuses:
matched += 1
if all(st in DISABLING_STATUSES for st in statuses):
plan.append((local, sorted(set(statuses))[0]))
else:
plan.append((local, "MISSING"))
if locals_ and matched == 0:
msg = (f"none of {len(locals_)} active local account(s) matches any of "
f"{len(okta_users)} Okta user(s); the identity attribute is probably wrong")
log.error("Okta sync SKIPPED, no account was changed: %s", msg)
return SyncResult(ok=False, error=msg, okta_users=len(okta_users), checked=len(locals_))
planned = [(local.username, status) for local, status in plan]
checked = len(locals_)
if not dry:
for local, status in plan:
local.is_active = False
db.add(models.AuditLog(
id=_gen_id("ev"), actor=SYNC_ACTOR, action=SYNC_ACTION,
entity_type="user", entity_id=local.id, project_id=None,
summary=local.username[:400],
detail={"okta_status": status, "via": "okta_sync"},
))
db.commit()
except Exception as exc:
db.rollback()
log.error("Okta sync FAILED, nothing was kept: unexpected %s", type(exc).__name__)
return SyncResult(ok=False, error=f"unexpected {type(exc).__name__}")
if dry:
for name, status in planned:
log.warning("Okta sync DRY RUN would disable %r (Okta status: %s)", name, status)
log.info("Okta sync DRY RUN done, no account was changed: %d Okta user(s), %d active "
"local account(s) checked, %d matched, %d would be disabled",
len(okta_users), checked, matched, len(planned))
return SyncResult(ok=True, okta_users=len(okta_users), checked=checked,
disabled=[name for name, _ in planned], dry_run=True)
for name, status in planned:
log.warning("Okta sync disabled %r (Okta status: %s)", name, status)
log.info("Okta sync done: %d Okta user(s), %d active local account(s) checked, %d disabled",
len(okta_users), checked, len(planned))
return SyncResult(ok=True, okta_users=len(okta_users), checked=checked,
disabled=[name for name, _ in planned])
# ── T13.4: running it on a schedule ───────────────────────────────────────────
# The api container runs gunicorn with 2 workers, and a later deploy may run more
# than one container. Each worker starts its own timer thread below. Without a
# gate, all of them would call Okta every cycle and write overlapping changes.
# The gate is one row in app_settings (key "okta_sync"). Its updated_at is "when
# the last run was claimed". A worker claims a run with ONE atomic conditional
# UPDATE: it succeeds only if the last claim is older than the interval. Exactly
# one worker wins each cycle, on SQLite and on Postgres, with no lock held.
#
# The row survives a restart, so a restart inside the interval does not start a
# second run. A slow run cannot overlap the next: a fetch is bounded to
# TOTAL_DEADLINE_SECONDS (120) and the interval can never be set below
# INTERVAL_MIN_SECONDS (300). Inside one process the timer thread runs one cycle at
# a time, so cycles cannot stack there either.
CLAIM_KEY = "okta_sync"
INTERVAL_DEFAULT_SECONDS = 900 # 15 minutes (D18, proposed, not confirmed)
INTERVAL_MIN_SECONDS = 300
TICK_SECONDS = 60 # how often each worker checks if a run is due
START_DELAY_SECONDS = 10 # let the app finish starting first
def interval_seconds() -> int:
"""OKTA_SYNC_INTERVAL_SECONDS, default 900. A value that is not a number uses
the default. A value under 300 is raised to 300 (see above for why)."""
raw = os.getenv("OKTA_SYNC_INTERVAL_SECONDS", "").strip()
if not raw:
return INTERVAL_DEFAULT_SECONDS
try:
value = int(raw)
except ValueError:
log.warning("OKTA_SYNC_INTERVAL_SECONDS=%r is not a whole number; using %d",
raw, INTERVAL_DEFAULT_SECONDS)
return INTERVAL_DEFAULT_SECONDS
if value < INTERVAL_MIN_SECONDS:
log.warning("OKTA_SYNC_INTERVAL_SECONDS=%d is under the minimum; using %d",
value, INTERVAL_MIN_SECONDS)
return INTERVAL_MIN_SECONDS
return value
def claim(db: Session, interval: int, now: Optional[datetime] = None) -> bool:
"""Try to claim this cycle. True for exactly one caller per interval."""
now = now or datetime.now(timezone.utc)
cutoff = now - timedelta(seconds=interval)
try:
row = db.get(models.AppSetting, CLAIM_KEY)
if row is None:
db.add(models.AppSetting(key=CLAIM_KEY, value={}, updated_at=now))
db.commit() # the insert is the claim
return True
except IntegrityError:
db.rollback() # another worker inserted first: it won
return False
db.rollback() # end the read; the UPDATE starts fresh
result = db.execute(
update(models.AppSetting)
.where(models.AppSetting.key == CLAIM_KEY, models.AppSetting.updated_at <= cutoff)
.values(updated_at=now)
)
db.commit()
return result.rowcount == 1
def run_if_due(session_factory: Callable[[], Session],
fetch: Optional[Callable[[], "list[OktaUser]"]] = None,
now: Optional[datetime] = None) -> Optional[SyncResult]:
"""One timer tick: claim the cycle, and if it is ours, run the sync. Returns
the SyncResult, or None when the cycle was not ours (or could not be claimed).
Nothing here can turn a failure into a change: run_sync() owns that rule, and
anything that goes wrong around it is logged and ends the tick."""
try:
with session_factory() as db:
if not claim(db, interval_seconds(), now=now):
return None
with session_factory() as db:
return run_sync(db, fetch=fetch)
except Exception as exc:
log.error("Okta sync tick failed, no account was changed: unexpected %s", type(exc).__name__)
return None
def start_scheduler(session_factory: Callable[[], Session]) -> "Optional[threading.Event]":
"""Start the timer thread. Returns a stop Event (set it to end the thread), or
None when the sync is off because OKTA_API_TOKEN or the org URL is missing.
Logs one line either way, so an operator can see it in `docker compose logs api`."""
log.info(describe())
if not is_configured():
return None
stop = threading.Event()
def loop():
if stop.wait(START_DELAY_SECONDS):
return
while not stop.is_set():
run_if_due(session_factory)
if stop.wait(TICK_SECONDS):
return
threading.Thread(target=loop, name="okta-sync", daemon=True).start()
log.info("Okta sync scheduled: every %d seconds (checked each %d s by every worker; "
"one wins each cycle)", interval_seconds(), TICK_SECONDS)
return stop

View File

@@ -254,27 +254,23 @@ def main():
if ("el.textContent = '" in ln or "el.innerHTML = '" in ln)
and "the project" not in ln and ("✓" in ln or "⚠" in ln or "↻" in ln or "✕" in ln)])
print("\n7. the analytics dwell listener still fires")
print("\n7. autosave's own registration survives the retired analytics tracker")
# This section used to also assert the per-browser usage-analytics dwell
# listener (wp-usage.js) still fired alongside autosave, to catch autosave
# accidentally replacing it. CR-019/T11.6 (2026-09-23) retired that tracker
# on purpose - track() is now a documented no-op and wp-usage.js is gone -
# so asserting it still records an event would be asserting the wrong
# thing. What's still worth checking here is autosave's own registration,
# which is what this section actually guards.
boot(page, base, tok, "/work-package-suite.html?project=projA&tab=sop")
time.sleep(0.8)
chk("the wizard still has its analytics dwell tracker",
page.eval("typeof trackStepDwell === 'function'"))
chk("...and the wizard registered autosave too",
chk("the wizard registered autosave",
page.eval("typeof WPAutosave === 'object' && typeof sopDraftId === 'function'"))
before = page.eval("""(() => {
try { return (JSON.parse(localStorage.getItem('wp_suite_analytics_v1')||'{}').events||[]).length; }
catch(e){ return -1; }
})()""")
# trackStepDwell only records a dwell longer than 400ms, so give it one.
time.sleep(0.9)
page.eval("typeof trackStepDwell==='function' && trackStepDwell()")
time.sleep(0.4)
after_n = page.eval("""(() => {
try { return (JSON.parse(localStorage.getItem('wp_suite_analytics_v1')||'{}').events||[]).length; }
catch(e){ return -1; }
})()""")
chk("calling it records an event (it was not replaced by the guard)",
after_n > before, "%s -> %s" % (before, after_n))
chk("...and the retired dwell tracker is a harmless no-op, not missing or throwing",
page.eval("""(() => {
if (typeof trackStepDwell !== 'function') return false;
try { trackStepDwell(); return true; } catch (e) { return false; }
})()"""))
finally:
page.close()
browser.close()

View File

@@ -207,9 +207,12 @@ def run(page, base, tok):
chk("auth resolved to the admin account", page.eval("(window.WP_USER||{}).role") == "admin")
chk("the directory is visible",
page.eval("getComputedStyle(document.getElementById('users-main')).display") != "none")
chk("manager table renders 9 columns",
page.eval("document.querySelectorAll('#users-table thead th').length") == 9,
# 10 since CR-020 (T12.2): the 9 data columns plus the row-selection column.
chk("manager table renders 10 columns (9 + row selection)",
page.eval("document.querySelectorAll('#users-table thead th').length") == 10,
page.eval("document.querySelectorAll('#users-table thead th').length"))
chk("...the first is the selection column",
page.eval("!!document.querySelector('#users-table thead th.selcell #sel-all')"))
chk("an admin sees every account in the fixture (7)",
page.eval("document.querySelectorAll('#users-table tbody tr').length") == 7,
page.eval("document.querySelectorAll('#users-table tbody tr').length"))
@@ -237,15 +240,22 @@ def run(page, base, tok):
# Your own row: permissions locked so you cannot demote yourself, job function
# still editable. Asserted on the two cells, not "no select in the row".
# The cells are found by their HEADER, not by position: CR-020's selection
# column moved every column one to the right and these asserted r.cells[3]
# and [4], which then read the username and the email (T12.6).
ROW = ("const r=[...document.querySelectorAll('#users-table tbody tr')]"
".find(r=>r.querySelector('.me-tag'));")
".find(r=>r.querySelector('.me-tag'));"
"const col=n=>[...document.querySelectorAll('#users-table thead th')]"
".findIndex(h=>h.textContent.trim().toLowerCase()===n);")
def own(q):
return "(()=>{" + ROW + "if(!r)return false;const c=r.cells[3];return " + q + "})()"
return "(()=>{" + ROW + "if(!r)return false;const c=r.cells[col('permissions')];return " + q + "})()"
chk("your own permissions cell is locked, not a dropdown",
page.eval(own("!c.querySelector('select') && !!c.querySelector('.tag')")))
chk("...and wears the Administrator pill", page.eval(own("!!c.querySelector('.tag.admin')")))
chk("...while your job function stays editable",
page.eval("(()=>{" + ROW + "return !!r && !!r.cells[4].querySelector('select')})()"))
page.eval("(()=>{" + ROW + "return !!r && !!r.cells[col('project role')].querySelector('select')})()"))
chk("...and your own row cannot be ticked for a bulk action",
page.eval("(()=>{" + ROW + "return !!r && !r.querySelector('input.selbox')})()"))
page.eval("[...document.querySelectorAll('#users-table tbody button')]"
".find(b=>/project/i.test(b.textContent)).click()")

326
tests/bulk_users_check.py Normal file
View File

@@ -0,0 +1,326 @@
#!/usr/bin/env python3
"""CR-020 (wave 12): bulk user administration, server side.
python tests/bulk_users_check.py
Self-contained and needs NO browser: a throwaway SQLite database in a temp
folder, the API driven in-process with FastAPI's TestClient, everything deleted
afterward. Your real wpsuite.db is never touched.
What it proves is the wave's own exit criteria, measured on the server, where
they are enforced:
- every bulk action works on a list, row by row, with the single-user rules;
- the caller's own account is refused on its own row, never the whole batch;
- a super user's batch cannot reach an account or project outside their scope;
- one AuditLog row per changed account, the same shape the single route writes;
- a partial failure is reported row by row, never as a clean success;
- project add/remove never touches the user's other projects or their roles;
- the single-user routes, now built on the same functions, behave as before.
Exit codes: 0 all passed, 1 a check failed, 2 the run could not start.
"""
import os
import shutil
import sys
import tempfile
import warnings
warnings.filterwarnings("ignore")
ROOT = os.path.dirname(os.path.dirname(os.path.abspath(__file__)))
TMP = tempfile.mkdtemp(prefix="wp-bulk-users-check-")
os.environ["DATABASE_URL"] = "sqlite:///" + os.path.join(TMP, "bulk.db").replace("\\", "/")
os.environ["AUTH_SECRET_KEY"] = "bulk-users-check-key-bulk-users-check-0000"
os.environ.pop("WP_OKTA_FAKE_DIRECTORY", None)
sys.path.insert(0, ROOT)
try:
from fastapi.testclient import TestClient
from sqlalchemy import select
from server import app as A, auth, models
from server.db import SessionLocal
except Exception as exc: # noqa: BLE001
print("could not start:", exc)
sys.exit(2)
RESULTS = []
def check(name, ok, detail=""):
RESULTS.append(bool(ok))
print(f" {'PASS' if ok else 'FAIL'} {name}" + (f" ({detail})" if detail and not ok else ""))
db = SessionLocal()
def account(uid, role="project_user"):
u = models.User(id=uid, username=uid[5:], role=role, email=uid[5:] + "@example.test")
db.add(u)
db.commit()
return u
def member(uid, pid, role=""):
db.add(models.ProjectMember(id=f"pm_{uid}_{pid}", user_id=uid, project_id=pid, role=role))
db.commit()
def client(user):
c = TestClient(A.app, base_url="http://127.0.0.1", raise_server_exceptions=False)
c.cookies.set(auth.COOKIE_NAME, auth.create_token(user))
return c
def fresh():
s = SessionLocal()
s.expire_all()
return s
def audit(action, entity_id=None):
s = fresh()
q = select(models.AuditLog).where(models.AuditLog.action == action)
if entity_id:
q = q.where(models.AuditLog.entity_id == entity_id)
return s.scalars(q.order_by(models.AuditLog.at)).all()
def roles_of(uid):
s = fresh()
return {m.project_id: (m.role or "") for m in
s.scalars(select(models.ProjectMember).where(models.ProjectMember.user_id == uid))}
def by_user(resp):
return {r["user_id"]: r for r in resp.json().get("results", [])}
# ── fixture ──────────────────────────────────────────────────────────────────
admin = account("user_admin", "admin")
admin2 = account("user_admin2", "admin")
superu = account("user_super") # super user on proj1 and proj4 by membership
pu = account("user_pu") # plain project user: manages nobody
for n in range(1, 10):
account(f"user_u{n}")
for pid in ("proj1", "proj2", "proj3", "proj4"):
db.add(models.Project(id=pid, name=pid))
db.commit()
member("user_super", "proj1", "project_super_user")
member("user_super", "proj4", "project_super_user")
member("user_pu", "proj1")
for n in (1, 2, 3, 4, 5):
member(f"user_u{n}", "proj1")
member("user_u6", "proj1"); member("user_u6", "proj2", "project_admin") # partly outside super's scope
member("user_u7", "proj2") # wholly outside it
member("user_u8", "proj1"); member("user_u8", "proj4", "project_super_user") # a role super may not grant
ca, cs, cp = client(admin), client(superu), client(pu)
def single_routes():
print("\n0. the single-user routes still behave as before (now on the shared functions)")
r = ca.post("/api/auth/users/user_admin/role", json={"role": "project_user"})
check("changing your own role: 400, same words", r.status_code == 400 and "own role" in r.json()["detail"])
r = ca.post("/api/auth/users/user_admin/active", json={"is_active": False})
check("disabling yourself: 400, same words", r.status_code == 400 and "disable your own" in r.json()["detail"])
r = ca.delete("/api/auth/users/user_admin")
check("deleting yourself: 400, same words", r.status_code == 400 and "delete your own" in r.json()["detail"])
r = cs.post("/api/auth/users/user_u6/role", json={"role": "project_admin"})
check("a super user changing an account outside their scope: 403", r.status_code == 403)
r = cs.put("/api/auth/users/user_u3/projects", json={"project_ids": ["proj1", "proj2"]})
check("a super user assigning a project they don't run: 403", r.status_code == 403)
r = ca.post("/api/auth/users/user_u9/role", json={"role": "project_admin"})
check("a single role change still works and returns the directory row",
r.status_code == 200 and r.json().get("role") == "project_admin")
one = audit("role_changed", "user_u9")
check("...and writes one role_changed row", len(one) == 1 and one[0].detail == {"from": "project_user", "to": "project_admin"},
str([e.detail for e in one]))
ca.post("/api/auth/users/user_u9/role", json={"role": "project_user"})
def bulk_role():
print("\n1. bulk permissions role")
r = ca.post("/api/auth/bulk-users/role", json={
"user_ids": ["user_u1", "user_u2", "user_admin", "user_nobody", "user_u1", "bad id!"],
"role": "project_admin"})
res = by_user(r)
check("200 even though some rows fail (a partial failure is a result, not an error)", r.status_code == 200)
check("u1 and u2 changed", res["user_u1"]["status"] == "ok" and res["user_u2"]["status"] == "ok")
check("your own row is refused with its reason; the rest still ran",
res["user_admin"]["status"] == "failed" and "own role" in res["user_admin"]["reason"])
check("an unknown account is refused as 'User not found'",
res["user_nobody"]["status"] == "failed" and res["user_nobody"]["reason"] == "User not found")
check("a malformed id is refused, not crashed on", res["bad id!"]["status"] == "failed")
check("a duplicated id is applied once", r.json()["requested"] == 5)
check("counts match the rows", (r.json()["ok"], r.json()["failed"]) == (2, 3), str(r.json()))
rows = audit("role_changed", "user_u1")
check("one role_changed audit row per changed account, same shape as the single route",
len(rows) == 1 and rows[0].detail == {"from": "project_user", "to": "project_admin"}
and rows[0].actor == "admin" and rows[0].entity_type == "user", str([e.detail for e in rows]))
r = ca.post("/api/auth/bulk-users/role", json={"user_ids": ["user_u1"], "role": "project_admin"})
check("a row already in that role is 'unchanged' and writes no audit row",
by_user(r)["user_u1"]["status"] == "unchanged" and len(audit("role_changed", "user_u1")) == 1)
r = cs.post("/api/auth/bulk-users/role", json={"user_ids": ["user_u3"], "role": "admin"})
check("a super user cannot grant admin in bulk either (grantable_roles, whole request 400)",
r.status_code == 400)
r = cs.post("/api/auth/bulk-users/role", json={"user_ids": ["user_u3", "user_u6", "user_u7"],
"role": "project_admin"})
res = by_user(r)
check("super user: an account wholly inside their projects changes", res["user_u3"]["status"] == "ok")
check("super user: an account also on a project they don't run is refused, with the reason",
res["user_u6"]["status"] == "failed" and "don't administer" in res["user_u6"]["reason"],
res["user_u6"]["reason"])
check("super user: an account they cannot see is 'User not found' (no existence leak)",
res["user_u7"]["status"] == "failed" and res["user_u7"]["reason"] == "User not found")
r = ca.post("/api/auth/bulk-users/role", json={"user_ids": ["user_admin2"], "role": "project_user"})
check("demoting another admin works while one active admin remains (the caller)",
by_user(r)["user_admin2"]["status"] == "ok")
def bulk_active():
print("\n2. bulk activate / disable")
r = ca.post("/api/auth/bulk-users/active", json={"user_ids": ["user_u4", "user_admin", "user_u5"],
"is_active": False})
res = by_user(r)
check("u4 and u5 disabled", res["user_u4"]["status"] == "ok" and res["user_u5"]["status"] == "ok"
and not fresh().get(models.User, "user_u4").is_active)
check("your own account is refused on its own row",
res["user_admin"]["status"] == "failed" and "disable your own" in res["user_admin"]["reason"])
check("one user_disabled audit row each, same shape as the single route",
[e.detail for e in audit("user_disabled", "user_u4")] == [{"is_active": False}])
r = ca.post("/api/auth/bulk-users/active", json={"user_ids": ["user_u4", "user_u2"], "is_active": True})
res = by_user(r)
check("re-enable: u4 ok, u2 (already active) unchanged",
res["user_u4"]["status"] == "ok" and res["user_u2"]["status"] == "unchanged")
def bulk_projects():
print("\n3. bulk project add / remove")
r = cs.post("/api/auth/bulk-users/projects", json={"user_ids": ["user_u1", "user_u2", "user_u8"],
"project_id": "proj1", "op": "add", "role": "project_admin"})
res = by_user(r)
check("add with a per-project role: every row ok", all(res[u]["status"] == "ok" for u in ("user_u1", "user_u2", "user_u8")),
str(res))
check("the role landed on that project", roles_of("user_u1").get("proj1") == "project_admin")
check("the user's OTHER project and its role are untouched - even a role the caller could not grant",
roles_of("user_u8").get("proj4") == "project_super_user", str(roles_of("user_u8")))
rows = audit("project_access_changed", "user_u1")
check("one project_access_changed row per account, in the single route's shape",
len(rows) == 1 and set(rows[0].detail) == {"projects", "scope", "overrides"}
and rows[0].detail["scope"] == "managed", str([e.detail for e in rows]))
r = cs.post("/api/auth/bulk-users/projects", json={"user_ids": ["user_u1"], "project_id": "proj2", "op": "add"})
check("a super user cannot add anyone to a project they don't run (403)", r.status_code == 403)
r = cs.post("/api/auth/bulk-users/projects", json={"user_ids": ["user_u1"], "project_id": "proj1",
"op": "add", "role": "project_super_user"})
check("a super user cannot hand out super user on a project (400)", r.status_code == 400)
r = ca.post("/api/auth/bulk-users/projects", json={"user_ids": ["user_u6"], "project_id": "proj3", "op": "add"})
check("admin adds u6 to proj3; proj1 and proj2 memberships and roles stay",
by_user(r)["user_u6"]["status"] == "ok"
and roles_of("user_u6") == {"proj1": "", "proj2": "project_admin", "proj3": ""}, str(roles_of("user_u6")))
r = cs.post("/api/auth/bulk-users/projects", json={"user_ids": ["user_u2", "user_u9"], "project_id": "proj1", "op": "remove"})
res = by_user(r)
check("remove: u2 taken off proj1", res["user_u2"]["status"] == "ok" and "proj1" not in roles_of("user_u2"))
check("remove: someone not on it is 'User not found' to a super user who cannot see them",
res["user_u9"]["status"] == "failed")
r = ca.post("/api/auth/bulk-users/projects", json={"user_ids": ["user_u9"], "project_id": "proj1", "op": "remove"})
check("remove: an admin removing someone not on the project gets 'unchanged'",
by_user(r)["user_u9"]["status"] == "unchanged")
r = ca.post("/api/auth/bulk-users/projects", json={"user_ids": ["user_u1"], "project_id": "proj1", "op": "replace"})
check("only add and remove exist: no replace-the-whole-list (400)", r.status_code == 400)
def bulk_delete():
print("\n4. bulk delete")
r = ca.post("/api/auth/bulk-users/delete", json={"user_ids": ["user_u5", "user_admin", "user_gone"]})
res = by_user(r)
check("u5 deleted", res["user_u5"]["status"] == "ok" and fresh().get(models.User, "user_u5") is None)
check("your own account is refused on its own row, the others still ran",
res["user_admin"]["status"] == "failed" and "delete your own" in res["user_admin"]["reason"])
check("an account that no longer exists is reported, not hidden",
res["user_gone"]["status"] == "failed" and res["user_gone"]["reason"] == "User not found")
check("one user_deleted audit row, same shape as the single route",
[e.detail for e in audit("user_deleted", "user_u5")] == [{"role": "project_user"}])
check("the deleted account's memberships went with it", roles_of("user_u5") == {})
def csv_resolve():
print("\n6. CSV upload: usernames in, a line-by-line report out (T12.4)")
text = ("username\n" # header, skipped
"u3\n"
"U3\n" # same person, different case: a duplicate
"nobody\n"
"\n" # blank lines are not rows
"u6\n" # visible to the super user, but also on a project they don't run
"u7\n" # not visible to the super user at all
"super\n" # the caller themself
"u8@example.test,extra column\n" # an email address, matched to exactly one account
",\n") # an empty first column
r = cs.post("/api/auth/bulk-users/resolve", json={"text": text})
j = r.json()
rows = {x["line"]: x for x in j["rows"]}
check("200 with a row for every non-blank line after the header",
r.status_code == 200 and sorted(rows) == [2, 3, 4, 6, 7, 8, 9, 10], sorted(rows))
check("a valid username is accepted", rows[2]["status"] == "ok" and rows[2]["user_id"] == "user_u3")
check("a repeated username is rejected as a duplicate of its first line, not applied twice",
rows[3]["status"] == "rejected" and rows[3]["reason"] == "duplicate of line 2")
check("an unknown username: 'User not found', with its line number", rows[4]["reason"] == "User not found")
check("an account the caller may not change: rejected with the single route's own reason",
rows[6]["status"] == "rejected" and "don't administer" in rows[6]["reason"], rows[6]["reason"])
check("an account the caller cannot see: 'User not found' (a CSV leaks no more than the table)",
rows[7]["reason"] == "User not found")
check("the caller's own account is rejected with a clear reason", "your own account" in rows[8]["reason"])
check("an email address resolves when exactly one account has it",
rows[9]["status"] == "ok" and rows[9]["user_id"] == "user_u8")
check("an empty first column is rejected, not skipped", rows[10]["reason"] == "no username in the first column")
check("the ids handed back are exactly the accepted rows",
j["user_ids"] == ["user_u3", "user_u8"] and (j["accepted"], j["rejected"]) == (2, 6), str(j["user_ids"]))
r = cs.post("/api/auth/bulk-users/active", json={"user_ids": j["user_ids"] + ["user_u7"], "is_active": False})
res = by_user(r)
check("the CSV's ids go through the same bulk route and the same guards: an id slipped in by hand "
"is still refused", res["user_u7"]["status"] == "failed" and res["user_u3"]["status"] == "ok")
ca.post("/api/auth/bulk-users/active", json={"user_ids": ["user_u3", "user_u8"], "is_active": True})
r = cp.post("/api/auth/bulk-users/resolve", json={"text": "u3"})
check("a plain project user cannot resolve a list (403)", r.status_code == 403)
r = cs.post("/api/auth/bulk-users/resolve", json={"text": "x" * 200_001})
check("an oversized upload is refused (422)", r.status_code == 422)
s = fresh()
s.add(models.User(id="user_twin1", username="twin1", email="shared@example.test"))
s.add(models.User(id="user_twin2", username="twin2", email="shared@example.test"))
s.commit()
r = ca.post("/api/auth/bulk-users/resolve", json={"text": "shared@example.test"})
check("an email shared by two accounts matches neither", r.json()["rows"][0]["reason"] == "User not found")
def gates():
print("\n7. who may call, and how much")
for path, body in (("role", {"role": "project_user"}), ("active", {"is_active": False}),
("projects", {"project_id": "proj1", "op": "add"}), ("delete", {})):
r = cp.post(f"/api/auth/bulk-users/{path}", json={"user_ids": ["user_u3"], **body})
check(f"a plain project user is refused /bulk-users/{path} (403)", r.status_code == 403, str(r.status_code))
anon = TestClient(A.app, base_url="http://127.0.0.1", raise_server_exceptions=False)
check("no session: 401", anon.post("/api/auth/bulk-users/delete", json={"user_ids": ["user_u3"]}).status_code == 401)
r = ca.post("/api/auth/bulk-users/active", json={"user_ids": [f"user_x{i}" for i in range(501)], "is_active": False})
check("more than 500 ids in one request is refused (422)", r.status_code == 422, str(r.status_code))
SECTIONS = [single_routes, bulk_role, bulk_active, bulk_projects, bulk_delete, csv_resolve, gates]
def main() -> int:
print("Wave 12 bulk user administration - server checks")
try:
for section in SECTIONS:
try:
section()
except Exception as exc: # noqa: BLE001 - a missing route must read as FAIL, not a traceback
check(f"section {section.__name__} ran to the end", False, f"{type(exc).__name__}: {exc}")
finally:
db.close()
A.engine.dispose()
shutil.rmtree(TMP, ignore_errors=True)
passed = sum(RESULTS)
print(f"\n{passed}/{len(RESULTS)} checks passed.")
return 0 if passed == len(RESULTS) else 1
if __name__ == "__main__":
sys.exit(main())

View File

@@ -0,0 +1,197 @@
#!/usr/bin/env python3
"""CR-020 (wave 12): bulk user administration in the browser, at 1440px and 390px.
python tests/bulk_users_ui_check.py
Self-contained like the other probes: a throwaway SQLite database seeded by
browser_check.seed(), its own uvicorn, headless Edge or Chrome over CDP. Your
real wpsuite.db is never touched. Run it alone, not back to back with another
probe (file-map.md: chained runs exhaust the browser's ports).
The server rules are proven in tests/bulk_users_check.py without a browser.
This file proves the page: selection, the toolbar, the CSV path and the typed
delete, end to end against the real server, operable by keyboard, and holding
together at 390px. It writes the after-screenshots CLAUDE.md asks for to
docs/reference/baseline/wave-12/; the before-shots are the existing
users-1440.png and after-wave9/users-390.png.
Exit 0 all passed, 1 a failure, 2 could not run.
"""
import json
import os
import sys
import tempfile
import time
import urllib.request
sys.path.insert(0, os.path.dirname(os.path.dirname(os.path.abspath(__file__))))
sys.path.insert(0, os.path.dirname(os.path.abspath(__file__)))
import cdp # noqa: E402
from browser_check import seed, start_server, chk, _PASS, _FAIL # noqa: E402
ROOT = os.path.dirname(os.path.dirname(os.path.abspath(__file__)))
SHOTS = os.path.join(ROOT, "docs", "reference", "baseline", "wave-12")
READY = "!!document.querySelector('#users-table table')"
def api(base, path, token):
req = urllib.request.Request(base + path, headers={"Cookie": "wp_session=" + token})
with urllib.request.urlopen(req, timeout=10) as r:
return json.loads(r.read().decode())
def user(base, token, username):
return next((u for u in api(base, "/api/auth/users", token) if u["username"] == username), None)
def press(page, key, code, vk):
"""A REAL key event (Input.dispatchKeyEvent), not a synthetic KeyboardEvent:
only a real one triggers a checkbox's native Space toggle - the lesson
launcher_check.py records."""
for t in ("keyDown", "keyUp"):
page.ws.call("Input.dispatchKeyEvent", {"type": t, "key": key, "code": code,
"windowsVirtualKeyCode": vk, "text": " " if t == "keyDown" else ""})
time.sleep(0.3)
def answer(page, ok=True, typed=None):
time.sleep(0.4)
if typed is not None:
page.eval("document.getElementById('wp-dlg-input').value=%s" % json.dumps(typed))
page.eval("document.getElementById('%s').click()" % ("wp-dlg-ok" if ok else "wp-dlg-cancel"))
time.sleep(1.2)
def tick(page, username):
page.eval("(() => { const b=[...document.querySelectorAll('input.selbox')]"
".find(x => x.dataset.id==='user_%s'); if(b) b.click(); })()" % username)
time.sleep(0.2)
BAR_390 = """(function(){
var bar=document.getElementById('bulk-bar'); var out={sw:document.documentElement.scrollWidth,under:[],clipped:[],n:0};
bar.querySelectorAll('button, select').forEach(function(el){
var r=el.getBoundingClientRect(); if(!r.width||!r.height) return; out.n++;
var id=el.tagName+' '+(el.textContent||el.id).trim().slice(0,20)+' '+Math.round(r.width)+'x'+Math.round(r.height);
if(Math.min(r.width,r.height)<24) out.under.push(id);
if(r.left<-2||r.right>392) out.clipped.push(id);
});
return JSON.stringify(out);
})()"""
def main():
exe = cdp.find_browser()
if not exe:
print("no headless-capable browser found; set WP_BROWSER.")
return 2
tmpdir = tempfile.mkdtemp(prefix="wpsuite-bulkui-")
db_path = os.path.join(tmpdir, "check.db")
server = browser = None
try:
tok = seed(db_path)
port = cdp.free_port()
base = "http://127.0.0.1:%d" % port
server = start_server(port, db_path)
if server is None:
print("server would not start")
return 2
browser = cdp.Browser(exe)
page = browser.page()
def visit(who, width, mobile=False):
page.clear_cookies()
page.set_cookie("wp_session", tok[who])
page.viewport(width, 900 if width > 500 else 844, mobile=mobile)
page.goto(base + "/users.html", wait_for=READY)
time.sleep(0.6)
print("\n1. the admin at 1440px")
visit("root", 1440)
page.screenshot(os.path.join(SHOTS, "users-1440.png"))
chk("a selection column with select-all", page.eval("!!document.getElementById('sel-all')"))
chk("no bulk bar until a row is selected", page.eval("document.getElementById('bulk-bar').hidden"))
page.eval("document.querySelector('input.selbox[data-id=\"user_pat\"]').focus()")
press(page, " ", "Space", 32)
chk("Space on a focused row checkbox selects it (keyboard operable)",
page.eval("document.querySelector('input.selbox[data-id=\"user_pat\"]').checked"))
chk("the bar appears, reachable by keyboard",
page.eval("!document.getElementById('bulk-bar').hidden && document.querySelector('#bulk-bar button').tabIndex >= 0"))
tick(page, "legacy")
chk("the bar counts the selection", page.eval("document.getElementById('bulk-count').textContent") == "2 selected")
page.screenshot(os.path.join(SHOTS, "users-1440-selected.png"))
page.eval("document.getElementById('bulk-role').value='project_admin';"
"document.querySelector('[onclick=\"bulkRole()\"]').click()")
time.sleep(0.4)
chk("the confirm names the count", "for 2 users" in (page.eval("document.getElementById('wp-dlg-msg').textContent") or ""))
answer(page)
chk("both accounts changed on the server",
user(base, tok["root"], "pat")["role"] == "project_admin" and user(base, tok["root"], "legacy")["role"] == "project_admin")
chk("the result is announced", "2 changed" in page.eval("document.getElementById('bulk-banner').textContent")
and page.eval("document.getElementById('bulk-banner').getAttribute('role')") in ("status", "alert"))
tick(page, "pat")
page.eval("const f=document.getElementById('user-filter'); f.value='active'; f.dispatchEvent(new Event('change'))")
time.sleep(0.3)
chk("changing the filter clears the selection, and says so",
page.eval("document.getElementById('bulk-bar').hidden") and
"filter changed" in page.eval("document.getElementById('sel-banner').textContent"))
print("\n2. the super user, a CSV with good and bad lines")
visit("sue", 1440)
page.eval("loadCsv(new File(['username\\npat\\nmix\\nbob\\nsue\\npat\\n'], 'list.csv', {type:'text/csv'}))")
time.sleep(1.5)
rep = page.eval("document.getElementById('csv-banner').textContent") or ""
chk("only the account she may change is selected", page.eval("document.getElementById('bulk-count').textContent") == "1 selected", rep)
chk("every bad line is reported by number and reason: another job, not found, herself, a duplicate",
all(s in rep for s in ("Line 3 (mix)", "Line 4 (bob): User not found", "Line 5 (sue)", "Line 6 (pat): duplicate of line 2")), rep)
page.eval("document.getElementById('bulk-project-role').value='project_admin';"
"document.querySelector('[onclick=\"bulkProject(\\'add\\')\"]').click()")
answer(page)
roles = api(base, "/api/auth/users/user_pat/projects", tok["root"])["roles"]
chk("add to project with a role landed on the server", roles.get("projA") == "project_admin", roles)
print("\n3. bulk delete needs the typed word")
visit("root", 1440)
tick(page, "bob")
page.eval("document.querySelector('[onclick=\"bulkDelete()\"]').click()")
time.sleep(0.4)
chk("the dialog lists the username", "bob" in page.eval("document.getElementById('wp-dlg-msg').textContent"))
page.eval("document.getElementById('wp-dlg-input').value='delete'; document.getElementById('wp-dlg-ok').click()")
time.sleep(0.4)
chk("a wrong word sends nothing", user(base, tok["root"], "bob") is not None
and page.eval("document.getElementById('wp-dlg-overlay').classList.contains('open')"))
answer(page, typed="DELETE")
chk("DELETE deletes, and the report says so", user(base, tok["root"], "bob") is None
and "1 changed" in page.eval("document.getElementById('bulk-banner').textContent"))
print("\n4. the gloved-hands width: 390px")
visit("root", 390, mobile=True)
page.screenshot(os.path.join(SHOTS, "users-390.png"))
tick(page, "pat")
m = json.loads(page.eval(BAR_390))
page.screenshot(os.path.join(SHOTS, "users-390-selected.png"))
chk("390px with the bar open: no sideways scrolling", m["sw"] <= 392, m["sw"])
chk("390px: no bar control clipped past the screen", not m["clipped"], m["clipped"])
chk("390px: every bar control meets the 24px floor", m["n"] > 0 and not m["under"], m["under"])
errs = [e for e in page.js_errors() if "beforeunload" not in e]
chk("no JavaScript errors in the whole run", not errs, errs[:3])
finally:
if browser is not None:
try:
browser.close()
except Exception:
pass
if server is not None:
server.terminate()
print("\n" + "-" * 54)
print("%d/%d checks passed." % (len(_PASS), len(_PASS) + len(_FAIL)))
for f in _FAIL:
print(" - " + f)
return 1 if _FAIL else 0
if __name__ == "__main__":
sys.exit(main())

View File

@@ -112,6 +112,33 @@ def main():
chk("cancel means no: bob is still an account",
st == 200 and any(u.get("username") == "bob" for u in (users or [])),
ascii_([u.get("username") for u in (users or [])]))
# CR-020 / T12.5: the bulk delete is a PROMPT-style dialog (typed DELETE),
# the kind this file's docstring says belongs back here.
print("\n4. bulk delete needs the typed word")
page.eval("(() => { const b=[...document.querySelectorAll('input.selbox')]"
".find(x => x.dataset.id==='user_bob'); b.click(); })()")
time.sleep(0.3)
chk("ticking a row shows the bulk bar",
page.eval("!document.getElementById('bulk-bar').hidden"))
page.eval("document.querySelector('[onclick=\"bulkDelete()\"]').click()")
time.sleep(0.4)
chk("the bulk delete asks through the kit's prompt, listing the username",
page.eval("(() => { const o=document.getElementById('wp-dlg-overlay');"
" return o.classList.contains('open')"
" && document.getElementById('wp-dlg-input-wrap').style.display !== 'none'"
" && /bob/.test(document.getElementById('wp-dlg-msg').textContent); })()"))
page.eval("document.getElementById('wp-dlg-input').value='delete';"
"document.getElementById('wp-dlg-ok').click()")
time.sleep(0.3)
chk("the wrong word keeps it open and names what to type",
page.eval("document.getElementById('wp-dlg-overlay').classList.contains('open')"
" && /DELETE/.test(document.getElementById('wp-dlg-err').textContent)"))
page.eval("document.getElementById('wp-dlg-cancel').click()")
time.sleep(0.6)
st, users = api(base, "/api/auth/users", tok["root"])
chk("cancel means no: bob is still an account after the bulk prompt",
st == 200 and any(u.get("username") == "bob" for u in (users or [])))
errs = [e for e in page.js_errors() if "beforeunload" not in e]
chk("no JavaScript errors, and no path reached a native dialog (they throw here)",
not errs, ascii_(errs[:3]))

252
tests/okta_sync_check.py Normal file
View File

@@ -0,0 +1,252 @@
#!/usr/bin/env python3
"""Does the Okta Management API client keep its promises? - D18 / T13.2.
The one promise that matters: a failure is never an empty list. T13.3 reads an
empty list as "nobody is active", so a client that returned [] when Okta was
down would disable the whole org. Every failure below must raise OktaSyncError.
Also checked: the client uses its own credential (never the OIDC client secret),
follows pagination, never sends the token to another host, never follows a
redirect, and never lets the token reach a log line, an error message, or
describe().
Self-contained: no server, no database, no network, no browser. httpx's
MockTransport stands in for Okta.
Exit 0 all passed, 1 a failure.
"""
import logging
import os
import sys
ROOT = os.path.dirname(os.path.dirname(os.path.abspath(__file__)))
sys.path.insert(0, ROOT)
import httpx # noqa: E402
from server import okta_auth, okta_sync # noqa: E402
TOKEN = "00tok-SECRET-api-token-xyz"
CLIENT_SECRET = "oidc-client-SECRET-abc"
ORG = "https://example.okta.com"
fails = 0
def chk(label, ok, extra=""):
global fails
print(("PASS " if ok else "FAIL ") + label + (f" [{extra}]" if (extra and not ok) else ""))
if not ok:
fails += 1
class Capture(logging.Handler):
def __init__(self):
super().__init__()
self.lines = []
def emit(self, record):
self.lines.append(record.getMessage())
cap = Capture()
logging.getLogger("wpsuite").addHandler(cap)
logging.getLogger("wpsuite").setLevel(logging.DEBUG)
def reset_env():
os.environ["OKTA_API_TOKEN"] = TOKEN
os.environ["OKTA_CLIENT_SECRET"] = CLIENT_SECRET
os.environ.pop("OKTA_ORG_URL", None)
os.environ.pop("OKTA_SYNC_PROFILE_ATTRIBUTE", None)
okta_auth.ISSUER = f"{ORG}/oauth2/default"
okta_auth.IDENTITY_CLAIM = "preferred_username"
def user(login, status="ACTIVE", **profile):
p = {"login": login, "email": login}
p.update(profile)
return {"id": "00u" + login, "status": status, "profile": p}
def client(handler):
return httpx.Client(transport=httpx.MockTransport(handler))
def raises(handler, label, **kw):
"""Run fetch_users against `handler`; it must raise OktaSyncError, and the
message must not hold the token."""
reset_env()
try:
out = okta_sync.fetch_users(http=client(handler))
except okta_sync.OktaSyncError as exc:
msg = str(exc)
clean = TOKEN not in msg and CLIENT_SECRET not in msg and "example.okta.com" not in msg
chk(f"{label}: raises, message has no token or host", clean, msg)
return msg
except Exception as exc: # the wrong exception type
chk(f"{label}: raises OktaSyncError", False, f"{type(exc).__name__}: {exc}")
return ""
chk(f"{label}: raises (did NOT raise, returned {out!r})", False)
return ""
def ok(body):
return lambda req: httpx.Response(200, json=body)
# ── 1. a good fetch ───────────────────────────────────────────────────────────
print("\n1. a good fetch")
reset_env()
seen = []
def good(req):
seen.append(req)
return httpx.Response(200, json=[user("alice", "ACTIVE"), user("bob", "suspended")])
users = okta_sync.fetch_users(http=client(good))
chk("returns every user", [u.identity for u in users] == ["alice", "bob"])
chk("status is upper case", [u.status for u in users] == ["ACTIVE", "SUSPENDED"])
chk("asks the users endpoint of the org", str(seen[0].url).startswith(f"{ORG}/api/v1/users"))
chk("asks for the page maximum", seen[0].url.params.get("limit") == "200")
chk("authenticates with the API token (SSWS)", seen[0].headers.get("authorization") == f"SSWS {TOKEN}")
chk("never sends the OIDC client secret", CLIENT_SECRET not in str(seen[0].headers))
# ── 2. pagination ─────────────────────────────────────────────────────────────
print("\n2. pagination")
reset_env()
calls = []
def paged(req):
calls.append(req)
if "after=p2" not in str(req.url):
return httpx.Response(200, json=[user("alice")],
headers={"Link": f'<{ORG}/api/v1/users?limit=200&after=p2>; rel="next"'})
return httpx.Response(200, json=[user("bob")])
users = okta_sync.fetch_users(http=client(paged))
chk("follows the next link and returns both pages", [u.identity for u in users] == ["alice", "bob"])
chk("two requests, no more", len(calls) == 2)
chk("page 2 also carries the token", calls[1].headers.get("authorization") == f"SSWS {TOKEN}")
# ── 3. the failure rule: an error is never an empty list ──────────────────────
print("\n3. every failure raises (never returns an empty list)")
for code in (401, 403, 404, 429, 500, 502, 503):
raises(lambda req, c=code: httpx.Response(c, json={"errorCode": "x"}), f"HTTP {code}")
raises(lambda req: httpx.Response(302, headers={"Location": "https://evil.example/"}), "a redirect (never followed)")
raises(lambda req: (_ for _ in ()).throw(httpx.ConnectError("boom to https://example.okta.com/api/v1/users")), "connection error")
raises(lambda req: (_ for _ in ()).throw(httpx.ReadTimeout("slow")), "timeout")
raises(lambda req: httpx.Response(200, text="<html>not json</html>"), "body is not JSON")
raises(ok({"errorCode": "E0000011"}), "body is an object, not a list")
raises(ok("a string"), "body is a string")
raises(ok([]), "an empty list (not credible for a real org)")
raises(ok(["alice"]), "an item that is not an object")
raises(ok([{"profile": {"login": "a"}}]), "an item with no status")
raises(ok([{"status": "ACTIVE"}]), "an item with no profile")
raises(ok([{"status": "ACTIVE", "profile": {}}]), "an item with no login")
raises(ok([{"status": "ACTIVE", "profile": {"login": " "}}]), "an item with a blank login")
raises(ok([user("alice"), {"status": "ACTIVE", "profile": {}}]), "one bad item among good ones (never skipped)")
# ── 4. misconfiguration raises ────────────────────────────────────────────────
print("\n4. misconfiguration")
reset_env()
os.environ["OKTA_API_TOKEN"] = ""
try:
okta_sync.fetch_users(http=client(good))
chk("no token raises", False)
except okta_sync.OktaSyncError:
chk("no token raises", True)
reset_env()
okta_auth.ISSUER = ""
try:
okta_sync.fetch_users(http=client(good))
chk("no org URL raises", False)
except okta_sync.OktaSyncError:
chk("no org URL raises", True)
reset_env()
os.environ["OKTA_ORG_URL"] = "http://example.okta.com"
sent = []
try:
okta_sync.fetch_users(http=client(lambda r: sent.append(r) or httpx.Response(200, json=[user("a")])))
chk("an http org URL raises", False)
except okta_sync.OktaSyncError:
chk("an http org URL raises", True)
chk("...and no request was made (token never sent over http)", sent == [])
# ── 5. never leaks the token to another host, never loops forever ─────────────
print("\n5. next-link safety")
reset_env()
hosts = []
def evil_link(req):
hosts.append(req.url.host)
return httpx.Response(200, json=[user("alice")],
headers={"Link": '<https://evil.example/steal>; rel="next"'})
raises(evil_link, "a next link to another host")
chk("...the other host was never contacted", hosts == ["example.okta.com"])
count = []
def endless(req):
count.append(1)
return httpx.Response(200, json=[user("alice")],
headers={"Link": f'<{ORG}/api/v1/users?after={len(count)}>; rel="next"'})
raises(endless, "an endless next chain (page cap)")
chk(f"...stops at the cap ({okta_sync.MAX_PAGES} requests)", len(count) == okta_sync.MAX_PAGES, str(len(count)))
# ── 6. org URL and identity attribute ─────────────────────────────────────────
print("\n6. org URL and identity attribute")
reset_env()
chk("org URL is the issuer's scheme and host", okta_sync.org_url() == ORG)
os.environ["OKTA_ORG_URL"] = "https://corp.okta.com/"
chk("OKTA_ORG_URL wins, trailing slash removed", okta_sync.org_url() == "https://corp.okta.com")
reset_env()
okta_auth.IDENTITY_CLAIM = "email"
body = [{"status": "ACTIVE", "profile": {"login": "alice", "email": "alice@x.com"}}]
chk("claim 'email' reads profile.email", okta_sync.fetch_users(http=client(ok(body)))[0].identity == "alice@x.com")
okta_auth.IDENTITY_CLAIM = "sAMAccountName"
body = [{"status": "ACTIVE", "profile": {"login": "alice@x.com", "sAMAccountName": "alice"}}]
chk("a custom claim reads profile[<claim>]", okta_sync.fetch_users(http=client(ok(body)))[0].identity == "alice")
os.environ["OKTA_SYNC_PROFILE_ATTRIBUTE"] = "login"
chk("OKTA_SYNC_PROFILE_ATTRIBUTE overrides", okta_sync.fetch_users(http=client(ok(body)))[0].identity == "alice@x.com")
# ── 7. the token never leaks ──────────────────────────────────────────────────
print("\n7. the token never leaks")
reset_env()
chk("describe() has no token", TOKEN not in okta_sync.describe() and "configured" in okta_sync.describe())
os.environ["OKTA_API_TOKEN"] = ""
chk("describe() says DISABLED without a token", "DISABLED" in okta_sync.describe())
reset_env()
every_log = " ".join(cap.lines)
chk("no log line holds the API token", TOKEN not in every_log)
chk("no log line holds the client secret", CLIENT_SECRET not in every_log)
# ── 8. the whole fetch is time-bounded (T13.4 relies on it) ──────────────────
print("\n8. the whole fetch is time-bounded")
reset_env()
saved = okta_sync.TOTAL_DEADLINE_SECONDS
okta_sync.TOTAL_DEADLINE_SECONDS = 0.0
def slow_pages(req):
return httpx.Response(200, json=[user("alice")],
headers={"Link": f'<{ORG}/api/v1/users?after=x>; rel="next"'})
raises(slow_pages, "a fetch that runs past the total deadline")
okta_sync.TOTAL_DEADLINE_SECONDS = saved
chk("the real deadline is under the shortest allowed interval",
okta_sync.TOTAL_DEADLINE_SECONDS < okta_sync.INTERVAL_MIN_SECONDS)
print()
print("ALL PASSED" if not fails else f"{fails} FAILED")
sys.exit(0 if not fails else 1)

View File

@@ -0,0 +1,277 @@
#!/usr/bin/env python3
"""Does the Okta sync use DPoP correctly? - D24 / T13.10.
A strict fake Okta (httpx.MockTransport) checks every DPoP proof the way RFC 9449 says a
server must: the header (typ dpop+jwt, ES256, a PUBLIC jwk only), the signature against
that jwk, htm, htu (no query), iat, a one-time jti, the nonce when the server asked for
one, and on an API call the ath hash of the access token and the same key as at the token
request. Also: auto mode finds out that DPoP is required, the nonce dance at the token
endpoint and at the API, paging, off mode, bounded retries, and no secret in any message.
Self-contained: no network, no server, no browser. Exit 0 all passed, 1 a failure.
"""
import base64
import hashlib
import json
import os
import sys
import time
from urllib.parse import parse_qs
ROOT = os.path.dirname(os.path.dirname(os.path.abspath(__file__)))
sys.path.insert(0, ROOT)
os.environ["DATABASE_URL"] = "sqlite://"
for v in ("OKTA_API_TOKEN", "OKTA_SYNC_PROFILE_ATTRIBUTE", "OKTA_SYNC_DPOP"):
os.environ.pop(v, None)
os.environ["OKTA_ORG_URL"] = "https://example.okta.com"
import httpx # noqa: E402
import jwt # noqa: E402
from cryptography.hazmat.primitives.asymmetric import rsa # noqa: E402
from server import okta_sync # noqa: E402
from server.okta_sync import OktaSyncError # noqa: E402
fails = 0
def chk(label, ok, extra=""):
global fails
print(("PASS " if ok else "FAIL ") + label + (f" [{extra}]" if (extra and not ok) else ""))
if not ok:
fails += 1
pk = rsa.generate_private_key(public_exponent=65537, key_size=2048)
CLIENT_JWK = json.loads(jwt.algorithms.RSAAlgorithm.to_jwk(pk))
CLIENT_JWK.update({"kid": "k1", "alg": "RS256", "use": "sig"})
CLIENT_PUB = jwt.algorithms.RSAAlgorithm.from_jwk(json.dumps({k: CLIENT_JWK[k] for k in ("kty", "n", "e")}))
os.environ["OKTA_SYNC_CLIENT_ID"] = "0oaCLIENT"
os.environ["OKTA_SYNC_PRIVATE_KEY"] = json.dumps(CLIENT_JWK)
TOKEN_URL = "https://example.okta.com/oauth2/v1/token"
USERS_URL = "https://example.okta.com/api/v1/users"
ACCESS = "ACCESS-TOKEN-VALUE-123"
def thumb(jwk):
return json.dumps({k: jwk[k] for k in ("crv", "kty", "x", "y")}, sort_keys=True, separators=(",", ":"))
class StrictOkta:
def __init__(self, require_dpop=False, token_nonce=False, api_nonce=False, always_bad_proof=False,
token_type_if_plain="Bearer", pages=1, endless_nonce=False):
self.endless_nonce = endless_nonce
self.require_dpop, self.token_nonce, self.api_nonce = require_dpop, token_nonce, api_nonce
self.always_bad_proof, self.token_type_if_plain, self.pages = always_bad_proof, token_type_if_plain, pages
self.token_calls = 0
self.api_calls = 0
self.problems = [] # anything the strict checks found wrong
self.jtis = set()
self.assertion_jtis = set()
self.token_key = None # thumbprint of the key used at the token request
self.api_keys = set()
self.dpop_on_token = 0
self.dpop_on_api = 0
self.issued_nonce = "nonce-" + os.urandom(3).hex()
self.api_seen = []
def _check_proof(self, req, url, method, ath=None, need_nonce=False):
proof = req.headers.get("dpop")
if not proof:
return None, "no proof"
try:
hdr = jwt.get_unverified_header(proof)
except Exception as e: # noqa: BLE001
return None, f"unreadable header {type(e).__name__}"
if hdr.get("typ") != "dpop+jwt" or hdr.get("alg") != "ES256":
self.problems.append(f"bad header typ/alg {hdr.get('typ')} {hdr.get('alg')}")
jwk = hdr.get("jwk")
if not isinstance(jwk, dict) or "d" in jwk or jwk.get("kty") != "EC":
self.problems.append("proof jwk missing or holds private part")
return None, "bad jwk"
try:
claims = jwt.decode(proof, jwt.algorithms.ECAlgorithm.from_jwk(json.dumps(jwk)),
algorithms=["ES256"], options={"verify_iat": False})
except Exception as e: # noqa: BLE001
self.problems.append(f"signature did not verify: {type(e).__name__}")
return None, "bad signature"
if claims.get("htm") != method:
self.problems.append(f"htm {claims.get('htm')} != {method}")
if claims.get("htu") != url:
self.problems.append(f"htu {claims.get('htu')} != {url}")
if abs(time.time() - int(claims.get("iat", 0))) > 60:
self.problems.append("iat not current")
if claims.get("jti") in self.jtis or not claims.get("jti"):
self.problems.append("jti missing or reused")
self.jtis.add(claims.get("jti"))
if ath is not None and claims.get("ath") != ath:
self.problems.append("ath wrong")
if ath is None and "ath" in claims:
self.problems.append("ath present on a token request")
if need_nonce and claims.get("nonce") != self.issued_nonce:
return jwk, "nonce"
return jwk, None
def handler(self, req: httpx.Request):
if req.url.path == "/oauth2/v1/token":
self.token_calls += 1
form = {k: v[0] for k, v in parse_qs(req.content.decode()).items()}
try:
a = jwt.decode(form["client_assertion"], CLIENT_PUB, algorithms=["RS256"], audience=TOKEN_URL)
if a["jti"] in self.assertion_jtis:
self.problems.append("client assertion jti reused")
self.assertion_jtis.add(a["jti"])
except Exception as e: # noqa: BLE001
self.problems.append(f"client assertion bad: {type(e).__name__}")
has = "dpop" in req.headers
if has:
self.dpop_on_token += 1
if self.always_bad_proof:
return httpx.Response(400, json={"error": "invalid_dpop_proof"})
if self.require_dpop and not has:
return httpx.Response(400, json={"error": "invalid_dpop_proof"})
if not has:
return httpx.Response(200, json={"token_type": self.token_type_if_plain, "access_token": ACCESS})
jwk, bad = self._check_proof(req, TOKEN_URL, "POST", need_nonce=self.token_nonce)
if bad == "nonce" or (self.endless_nonce and not bad):
return httpx.Response(400, json={"error": "use_dpop_nonce"}, headers={"DPoP-Nonce": self.issued_nonce})
if bad:
return httpx.Response(400, json={"error": "invalid_dpop_proof"})
self.token_key = thumb(jwk)
return httpx.Response(200, json={"token_type": "DPoP", "access_token": ACCESS})
if req.url.path == "/api/v1/users":
self.api_calls += 1
auth = req.headers.get("authorization", "")
self.api_seen.append(auth.split(" ")[0])
if auth.startswith("DPoP "):
self.dpop_on_api += 1
if auth != f"DPoP {ACCESS}":
self.problems.append("wrong DPoP access token")
ath = base64.urlsafe_b64encode(hashlib.sha256(ACCESS.encode()).digest()).rstrip(b"=").decode()
jwk, bad = self._check_proof(req, USERS_URL, "GET", ath=ath, need_nonce=self.api_nonce)
if bad == "nonce":
return httpx.Response(401, json={"errorCode": "invalid_token"},
headers={"DPoP-Nonce": self.issued_nonce,
"WWW-Authenticate": 'DPoP error="use_dpop_nonce"'})
if bad:
return httpx.Response(401, json={})
self.api_keys.add(thumb(jwk))
if thumb(jwk) != self.token_key:
self.problems.append("API proof key differs from the token request key")
elif auth == f"Bearer {ACCESS}":
if "dpop" in req.headers:
self.problems.append("DPoP header sent with a plain Bearer token")
else:
return httpx.Response(401, json={})
page = int(req.url.params.get("after", "0") or 0)
body = [{"status": "ACTIVE", "profile": {"login": f"user{page}"}}]
headers = {}
if page + 1 < self.pages:
headers["Link"] = f'<{USERS_URL}?after={page + 1}&limit=200>; rel="next"'
return httpx.Response(200, json=body, headers=headers)
return httpx.Response(404)
def client(self):
return httpx.Client(transport=httpx.MockTransport(self.handler))
def run(fake, mode=None):
if mode is None:
os.environ.pop("OKTA_SYNC_DPOP", None)
else:
os.environ["OKTA_SYNC_DPOP"] = mode
try:
return okta_sync.fetch_users(http=fake.client()), None
except OktaSyncError as e:
return None, str(e)
# 1. auto: a plain app needs no DPoP
f = StrictOkta()
users, err = run(f)
chk("auto, app without DPoP: works with one token call, Bearer", users and f.token_calls == 1 and f.dpop_on_token == 0 and f.api_seen == ["Bearer"], str(err))
chk("no problems found by the strict checks", not f.problems, str(f.problems))
# 2. auto: the app requires DPoP (what your tenant answered)
f = StrictOkta(require_dpop=True)
users, err = run(f)
chk("auto, app requires DPoP: plain try, then DPoP try, then users", users and f.token_calls == 2 and f.dpop_on_token == 1, f"{err} calls={f.token_calls}")
chk("API calls carry DPoP and the access token", f.api_calls == 1 and f.dpop_on_api == 1 and f.api_seen == ["DPoP"])
chk("API proof used the same key as the token request", len(f.api_keys) == 1 and f.api_keys == {f.token_key})
chk("every proof and assertion id is unique, no problems", not f.problems, str(f.problems))
# 3. nonce at the token endpoint
f = StrictOkta(require_dpop=True, token_nonce=True)
users, err = run(f)
chk("token nonce: plain, DPoP without nonce, DPoP with nonce", users and f.token_calls == 3, f"{err} calls={f.token_calls}")
chk("token nonce: no problems", not f.problems, str(f.problems))
# 4. nonce at the API
f = StrictOkta(require_dpop=True, api_nonce=True)
users, err = run(f)
chk("API nonce: the call is repeated once with the nonce", users and f.api_calls == 2 and f.dpop_on_api == 2, f"{err} calls={f.api_calls}")
chk("API nonce: no problems", not f.problems, str(f.problems))
# 5. paging: a new proof for every page, htu without the query
f = StrictOkta(require_dpop=True, pages=3)
users, err = run(f)
chk("3 pages with DPoP: 3 users, 3 API calls", users and len(users) == 3 and f.api_calls == 3, f"{err} {f.api_calls}")
chk("3 pages: htu had no query, jti unique, same key", not f.problems, str(f.problems))
# 6. mode on goes straight to DPoP
f = StrictOkta(require_dpop=True)
users, err = run(f, "on")
chk("mode on: DPoP from the first token call", users and f.token_calls == 1 and f.dpop_on_token == 1, f"{err} {f.token_calls}")
f = StrictOkta()
users, err = run(f, "true")
chk("mode on with an app that does not need it: still works", users is not None and not f.problems, f"{err} {f.problems}")
# 7. mode off
f = StrictOkta(require_dpop=True)
users, err = run(f, "off")
chk("mode off, app requires DPoP: stops after one call, names the error", users is None and f.token_calls == 1 and "invalid_dpop_proof" in err, str(err))
f = StrictOkta(token_type_if_plain="DPoP")
users, err = run(f, "off")
chk("mode off, DPoP-bound token: clear refusal, users never called", users is None and "DPoP-bound" in err and f.api_calls == 0, str(err))
f = StrictOkta(token_type_if_plain="mac")
users, err = run(f)
chk("unknown token type refused", users is None and "neither" in err, str(err))
# 8. bounded retries and a hopeless server
f = StrictOkta(always_bad_proof=True)
users, err = run(f)
chk("a server that always rejects the proof: bounded calls, clear error", users is None and f.token_calls <= 4 and "invalid_dpop_proof" in err and "DPoP was used" in err, f"{err} {f.token_calls}")
chk("no users call when the token failed", f.api_calls == 0)
f = StrictOkta(require_dpop=True, token_nonce=True, endless_nonce=True)
users, err = run(f)
chk("a server that asks for a nonce forever: bounded calls, clear error", users is None and f.token_calls <= 4 and "use_dpop_nonce" in err, f"{err} {f.token_calls}")
# 9. a new key on each run
f1, f2 = StrictOkta(require_dpop=True), StrictOkta(require_dpop=True)
run(f1); run(f2)
chk("each run makes its own DPoP key", f1.token_key and f2.token_key and f1.token_key != f2.token_key)
# 10. secrets and startup line
f = StrictOkta(always_bad_proof=True)
_, err = run(f)
chk("error holds no token, key or assertion", ACCESS not in err and CLIENT_JWK["d"][:12] not in err)
line = okta_sync.describe()
chk("startup line names the DPoP mode", "DPoP auto" in line, line)
os.environ["OKTA_SYNC_DPOP"] = "on"
chk("startup line follows the setting", "DPoP on" in okta_sync.describe())
os.environ.pop("OKTA_SYNC_DPOP")
# 11. the SSWS token path is untouched
os.environ["OKTA_API_TOKEN"] = "ssws-value"
f = StrictOkta()
def ssws_handler(req):
f.api_seen.append(req.headers.get("authorization", ""))
return httpx.Response(200, json=[{"status": "ACTIVE", "profile": {"login": "a"}}])
users = okta_sync.fetch_users(http=httpx.Client(transport=httpx.MockTransport(ssws_handler)))
chk("SSWS path unchanged: SSWS header, no DPoP", f.api_seen == ["SSWS ssws-value"] and users, str(f.api_seen))
os.environ.pop("OKTA_API_TOKEN")
print("\nDONE" if fails == 0 else f"\n{fails} FAILED")
sys.exit(1 if fails else 0)

View File

@@ -0,0 +1,134 @@
#!/usr/bin/env python3
"""Does the Okta sync dry run change nothing? - D22 / T13.7.
With OKTA_SYNC_DRY_RUN on, the sync must fetch, match and apply both guards exactly
as in a live run, report who it WOULD disable, and change no account and write no
audit row. With it off (the default) behavior is as before.
Self-contained: in-memory SQLite, no server, no network, no browser.
Exit 0 all passed, 1 a failure.
"""
import logging
import os
import sys
ROOT = os.path.dirname(os.path.dirname(os.path.abspath(__file__)))
sys.path.insert(0, ROOT)
from sqlalchemy import create_engine, select # noqa: E402
from sqlalchemy.orm import sessionmaker # noqa: E402
from sqlalchemy.pool import StaticPool # noqa: E402
from server import models, okta_sync # noqa: E402
from server.okta_sync import OktaSyncError, OktaUser, run_sync # noqa: E402
fails = 0
def chk(label, ok, extra=""):
global fails
print(("PASS " if ok else "FAIL ") + label + (f" [{extra}]" if (extra and not ok) else ""))
if not ok:
fails += 1
engine = create_engine("sqlite://", connect_args={"check_same_thread": False}, poolclass=StaticPool)
models.Base.metadata.create_all(engine)
Session = sessionmaker(bind=engine, autoflush=False, future=True)
def seed(names):
with Session() as db:
db.query(models.AuditLog).delete()
db.query(models.User).delete()
for n in names:
db.add(models.User(id="u_" + n, username=n, is_active=True, email=""))
db.commit()
def state():
with Session() as db:
return ({u.username: u.is_active for u in db.scalars(select(models.User))},
db.query(models.AuditLog).count())
def okta(*pairs):
return lambda: [OktaUser(identity=i, status=s) for i, s in pairs]
class Capture(logging.Handler):
def __init__(self):
super().__init__()
self.lines = []
def emit(self, r):
self.lines.append(r.getMessage())
cap = Capture()
logging.getLogger("wpsuite.okta_sync").addHandler(cap)
# 1. Flag parsing
for v, want in [("", False), ("false", False), ("0", False), ("true", True), ("TRUE", True),
("1", True), ("yes", True), ("on", True), (" on ", True), ("maybe", False)]:
os.environ["OKTA_SYNC_DRY_RUN"] = v
chk(f"OKTA_SYNC_DRY_RUN={v!r} -> {want}", okta_sync.dry_run() is want)
os.environ.pop("OKTA_SYNC_DRY_RUN")
chk("unset -> live", okta_sync.dry_run() is False)
# 2. Dry run reports the plan and changes nothing
seed(["alice", "bob", "carol", "dan"])
before = state()
with Session() as db:
res = run_sync(db, fetch=okta(("alice", "ACTIVE"), ("bob", "SUSPENDED"), ("carol", "DEPROVISIONED")), dry=True)
after = state()
chk("dry run ok", res.ok and res.dry_run)
chk("dry run lists bob, carol and dan (missing)", sorted(res.disabled) == ["bob", "carol", "dan"], str(res.disabled))
chk("dry run changed no account", before == after, f"{before} {after}")
chk("dry run wrote no audit row", after[1] == 0)
chk("log says DRY RUN would disable", any("DRY RUN would disable 'bob'" in l for l in cap.lines))
chk("summary line says no account was changed and counts 3 matched, 3 would be disabled",
any("DRY RUN done, no account was changed" in l and "3 matched" in l and "3 would be disabled" in l
for l in cap.lines))
# 3. The environment decides when no argument is given
os.environ["OKTA_SYNC_DRY_RUN"] = "true"
seed(["alice", "bob"])
with Session() as db:
res = run_sync(db, fetch=okta(("alice", "ACTIVE"), ("bob", "SUSPENDED")))
chk("env on -> dry run", res.dry_run and state()[0]["bob"] is True and state()[1] == 0)
# 4. Guards still apply in a dry run
seed(["alice", "bob"])
with Session() as db:
res = run_sync(db, fetch=okta(("zed", "ACTIVE")))
chk("zero-match guard still refuses in a dry run", (not res.ok) and "probably wrong" in res.error)
def boom():
raise OktaSyncError("down")
with Session() as db:
res = run_sync(db, fetch=boom)
chk("fetch failure still fails closed in a dry run", (not res.ok) and state() == ({"alice": True, "bob": True}, 0))
# 5. Live run (flag off) is unchanged
os.environ.pop("OKTA_SYNC_DRY_RUN")
seed(["alice", "bob"])
with Session() as db:
res = run_sync(db, fetch=okta(("alice", "ACTIVE"), ("bob", "SUSPENDED")))
st = state()
chk("live run disables and logs one audit row", (not res.dry_run) and st[0]["bob"] is False and st[1] == 1)
# 6. Startup line names the mode
os.environ["OKTA_API_TOKEN"] = "tok-secret-9f3a"
os.environ["OKTA_ORG_URL"] = "https://example.okta.com"
os.environ["OKTA_SYNC_DRY_RUN"] = "true"
chk("startup line says DRY RUN", "DRY RUN" in okta_sync.describe())
os.environ.pop("OKTA_SYNC_DRY_RUN")
chk("startup line says live", "(live)" in okta_sync.describe())
chk("startup line never holds the token", "tok-secret-9f3a" not in okta_sync.describe())
print("\nDONE" if fails == 0 else f"\n{fails} FAILED")
sys.exit(1 if fails else 0)

View File

@@ -0,0 +1,218 @@
#!/usr/bin/env python3
"""Does the Okta deprovisioning sync fail closed? - D18 / T13.3.
The rule that must never be skipped or weakened: when Okta cannot be asked, or
its answer cannot be trusted, the sync changes NO account and writes NO audit
row. Treating "could not reach Okta" as "nobody is active" would lock the whole
company out on one bad afternoon.
Also checked: the sync only ever disables (it never re-enables, under any
input), writes one audit row per change with a non-human actor, leaves an
already-disabled account alone, logs only changes, and keeps temporary Okta
states (PASSWORD_EXPIRED, LOCKED_OUT, RECOVERY) from disabling anyone.
Self-contained: its own in-memory SQLite, no server, no network, no browser.
Exit 0 all passed, 1 a failure.
"""
import os
import sys
ROOT = os.path.dirname(os.path.dirname(os.path.abspath(__file__)))
sys.path.insert(0, ROOT)
import httpx # noqa: E402
from sqlalchemy import create_engine, select # noqa: E402
from sqlalchemy.orm import sessionmaker # noqa: E402
from sqlalchemy.pool import StaticPool # noqa: E402
from server import models, okta_auth, okta_sync # noqa: E402
from server.okta_sync import OktaSyncError, OktaUser, run_sync # noqa: E402
fails = 0
def chk(label, ok, extra=""):
global fails
print(("PASS " if ok else "FAIL ") + label + (f" [{extra}]" if (extra and not ok) else ""))
if not ok:
fails += 1
engine = create_engine("sqlite://", connect_args={"check_same_thread": False}, poolclass=StaticPool)
models.Base.metadata.create_all(engine)
Session = sessionmaker(bind=engine, autoflush=False, future=True)
def seed(rows):
"""rows: list of (username, is_active[, email]). Wipes users and audit rows first."""
with Session() as db:
db.query(models.AuditLog).delete()
db.query(models.User).delete()
for r in rows:
name, active = r[0], r[1]
db.add(models.User(id="u_" + name, username=name, is_active=active,
email=(r[2] if len(r) > 2 else "")))
db.commit()
def state():
with Session() as db:
users = {u.username: u.is_active for u in db.scalars(select(models.User))}
audits = [a.to_dict() for a in db.scalars(select(models.AuditLog))]
return users, audits
def fixed(*pairs):
return lambda: [OktaUser(identity=i, status=s) for i, s in pairs]
# ── 1. the failure rule ───────────────────────────────────────────────────────
print("\n1. FAIL CLOSED: a failed or untrusted fetch changes nothing")
BASE = [("alice", True), ("bob", True), ("carol", True)]
def boom(exc):
def f():
raise exc
return f
failing = [
("OktaSyncError (auth failure)", boom(OktaSyncError("HTTP 401"))),
("RuntimeError", boom(RuntimeError("boom"))),
("ConnectionError", boom(ConnectionError("down"))),
("TimeoutError", boom(TimeoutError("slow"))),
("ValueError", boom(ValueError("bad"))),
("KeyError", boom(KeyError("x"))),
("fetch returns an empty list", lambda: []),
("fetch returns None", lambda: None),
("fetch returns garbage (strings)", lambda: ["alice", "bob"]),
("fetch returns garbage (dicts)", lambda: [{"identity": "alice"}]),
]
for label, fetch in failing:
seed(BASE)
before = state()
with Session() as db:
res = run_sync(db, fetch=fetch)
after = state()
chk(f"{label}: ok is False", res.ok is False)
chk(f"{label}: zero accounts changed", after[0] == before[0] == {"alice": True, "bob": True, "carol": True})
chk(f"{label}: zero audit rows", after[1] == [] and res.disabled == [])
print("\n ...and through the real client, not a stand-in")
okta_auth.ISSUER = "https://example.okta.com/oauth2/default"
os.environ["OKTA_API_TOKEN"] = "tok"
for code in (401, 403, 429, 500):
seed(BASE)
c = httpx.Client(transport=httpx.MockTransport(lambda r, k=code: httpx.Response(k, json={})))
with Session() as db:
res = run_sync(db, fetch=lambda c=c: okta_sync.fetch_users(http=c))
u, a = state()
chk(f"real client, HTTP {code}: nothing changed, no audit rows",
(not res.ok) and all(u.values()) and a == [])
seed(BASE)
os.environ["OKTA_API_TOKEN"] = ""
with Session() as db:
res = run_sync(db) # fetch=None: the default path, no token set
u, a = state()
chk("default fetch with no token: nothing changed, no audit rows", (not res.ok) and all(u.values()) and a == [])
print("\n ...and when saving fails")
seed(BASE)
with Session() as db:
real_commit = db.commit
def bad_commit():
raise RuntimeError("disk full")
db.commit = bad_commit
res = run_sync(db, fetch=fixed(("alice", "ACTIVE"))) # bob and carol would be disabled
u, a = state()
chk("a failed commit keeps no change and no audit row", (not res.ok) and all(u.values()) and a == [])
print("\n ...and when no local account matches any Okta user (wrong identity attribute)")
seed(BASE)
with Session() as db:
res = run_sync(db, fetch=fixed(("zed@x.com", "ACTIVE"), ("yan@x.com", "ACTIVE")))
u, a = state()
chk("zero matches is refused, not read as 'everyone left'", (not res.ok) and all(u.values()) and a == [])
chk("...and the error says why", "identity attribute" in res.error)
# ── 2. what it does disable ───────────────────────────────────────────────────
print("\n2. a good run disables the right accounts")
seed([("alice", True), ("bob", True), ("carol", True), ("dave", True),
("erin", True), ("frank", True), ("gina", True), ("hank", True)])
okta = fixed(("alice", "ACTIVE"), ("bob", "SUSPENDED"),
# carol is MISSING from the list
("dave", "DEPROVISIONED"), ("erin", "PASSWORD_EXPIRED"),
("frank", "LOCKED_OUT"), ("gina", "RECOVERY"), ("hank", "ACTIVE"))
with Session() as db:
res = run_sync(db, fetch=okta)
users, audits = state()
chk("run is ok", res.ok, res.error)
chk("missing, SUSPENDED and DEPROVISIONED are disabled",
sorted(res.disabled) == ["bob", "carol", "dave"] and
[users[n] for n in ("bob", "carol", "dave")] == [False, False, False])
chk("ACTIVE stays active", users["alice"] is True and users["hank"] is True)
chk("PASSWORD_EXPIRED, LOCKED_OUT and RECOVERY do NOT disable (temporary states)",
users["erin"] is True and users["frank"] is True and users["gina"] is True)
chk("one audit row per disabled account", len(audits) == 3, str(len(audits)))
row = next(a for a in audits if a["summary"] == "bob")
chk("audit actor is the sync job, not a person", {a["actor"] for a in audits} == {"system:okta_sync"})
chk("audit action is user_deprovisioned_by_sync", {a["action"] for a in audits} == {"user_deprovisioned_by_sync"})
chk("audit row names the user and the Okta status found",
row["entity_type"] == "user" and row["entity_id"] == "u_bob" and row["detail"].get("okta_status") == "SUSPENDED")
chk("a missing user is recorded as MISSING",
next(a for a in audits if a["summary"] == "carol")["detail"].get("okta_status") == "MISSING")
print("\n ...run it again: only changes are logged")
with Session() as db:
res2 = run_sync(db, fetch=okta)
users2, audits2 = state()
chk("second run is ok, disables nobody", res2.ok and res2.disabled == [])
chk("second run writes no new audit rows (no duplicates)", len(audits2) == 3)
chk("accounts unchanged by the second run", users2 == users)
# ── 3. never re-enables, never touches a disabled account ─────────────────────
print("\n3. never re-enables, never touches an already-disabled account")
seed([("alice", True), ("hank", False), ("gina", False)])
with Session() as db:
res = run_sync(db, fetch=fixed(("alice", "ACTIVE"), ("hank", "ACTIVE"))) # gina missing
users, audits = state()
chk("a disabled account that is ACTIVE in Okta stays disabled", users["hank"] is False)
chk("a disabled account that is missing from Okta stays as it is", users["gina"] is False)
chk("no audit row for either (nothing changed)", audits == [] and res.disabled == [])
src = open(os.path.join(ROOT, "server", "okta_sync.py")).read()
chk("no assignment of is_active to True anywhere in okta_sync.py",
"is_active = True" not in src and "is_active=True" not in src.replace("is_active.is_(True)", ""))
# ── 4. matching ───────────────────────────────────────────────────────────────
print("\n4. matching")
seed([("ALICE", True), ("bob", True, "Bob@X.com"), ("carol", True)])
with Session() as db:
res = run_sync(db, fetch=fixed(("alice", "ACTIVE"), ("bob@x.com", "ACTIVE"), ("carol", "SUSPENDED")))
users, _ = state()
chk("username match ignores case", users["ALICE"] is True)
chk("email match works, ignoring case", users["bob"] is True)
chk("a matching SUSPENDED user is disabled", users["carol"] is False)
seed([("alice", True), ("bob", True), ("carol", True)])
with Session() as db:
run_sync(db, fetch=fixed(("alice", "ACTIVE"), ("bob", "ACTIVE"), ("bob", "SUSPENDED"),
("carol", "SUSPENDED"), ("carol", "DEPROVISIONED")))
users, _ = state()
chk("an identity listed twice, one entry active: stays active", users["bob"] is True)
chk("an identity listed twice, every entry disabling: disabled", users["carol"] is False)
# ── 5. the secret never reaches the audit row or the result ───────────────────
print("\n5. no secret in what is stored or returned")
seed(BASE)
os.environ["OKTA_API_TOKEN"] = "SECRET-TOKEN-123"
c = httpx.Client(transport=httpx.MockTransport(lambda r: httpx.Response(401, json={})))
with Session() as db:
res = run_sync(db, fetch=lambda: okta_sync.fetch_users(http=c))
chk("the failure message has no token", "SECRET-TOKEN-123" not in res.error and "example.okta.com" not in res.error)
print()
print("ALL PASSED" if not fails else f"{fails} FAILED")
sys.exit(0 if not fails else 1)

View File

@@ -0,0 +1,288 @@
#!/usr/bin/env python3
"""Does the Okta sync sign in as an OAuth service app (private_key_jwt)? - D23 / T13.9.
Checks, with a fake Okta (httpx.MockTransport) and freshly made keys:
- the token request carries the right form fields and a JWT that VERIFIES against the
public key, with iss and sub = client ID, aud = the token endpoint, kid in the header,
a short lifetime, and a one-time jti
- the users request then uses the Bearer token, not SSWS
- every failure raises OktaSyncError and the users endpoint is never called
- the key is accepted as JSON and as base64, RSA and EC; a public key is refused
- an SSWS API token still works and wins when both are set
- the key, the assertion and the token never appear in an error message or the startup line
Self-contained: no network, no server, no browser. Exit 0 all passed, 1 a failure.
"""
import base64
import json
import os
import sys
from urllib.parse import parse_qs
ROOT = os.path.dirname(os.path.dirname(os.path.abspath(__file__)))
sys.path.insert(0, ROOT)
os.environ["DATABASE_URL"] = "sqlite://"
for v in ("OKTA_API_TOKEN", "OKTA_SYNC_CLIENT_ID", "OKTA_SYNC_PRIVATE_KEY", "OKTA_SYNC_PROFILE_ATTRIBUTE"):
os.environ.pop(v, None)
os.environ["OKTA_ORG_URL"] = "https://example.okta.com"
import httpx # noqa: E402
import jwt # noqa: E402
from cryptography.hazmat.primitives.asymmetric import ec, rsa # noqa: E402
from server import okta_sync # noqa: E402
from server.okta_sync import OktaSyncError # noqa: E402
fails = 0
def chk(label, ok, extra=""):
global fails
print(("PASS " if ok else "FAIL ") + label + (f" [{extra}]" if (extra and not ok) else ""))
if not ok:
fails += 1
def make_rsa():
pk = rsa.generate_private_key(public_exponent=65537, key_size=2048)
jwk = json.loads(jwt.algorithms.RSAAlgorithm.to_jwk(pk))
jwk.update({"kid": "rsa-kid-1", "use": "sig", "alg": "RS256"})
pub = {k: jwk[k] for k in ("kty", "n", "e")}
return jwk, jwt.algorithms.RSAAlgorithm.from_jwk(json.dumps(pub)), "RS256"
def make_ec():
pk = ec.generate_private_key(ec.SECP256R1())
jwk = json.loads(jwt.algorithms.ECAlgorithm.to_jwk(pk))
jwk.update({"kid": "ec-kid-1", "use": "sig", "alg": "ES256"})
pub = {k: jwk[k] for k in ("kty", "crv", "x", "y")}
return jwk, jwt.algorithms.ECAlgorithm.from_jwk(json.dumps(pub)), "ES256"
class FakeOkta:
def __init__(self, pub=None, alg="RS256", token_status=200, token_body=None, user_status=200):
self.pub, self.alg = pub, alg
self.token_status, self.token_body, self.user_status = token_status, token_body, user_status
self.token_calls, self.user_calls, self.seen_auth, self.form, self.verified = 0, 0, [], {}, None
def handler(self, req: httpx.Request):
if req.url.path == "/oauth2/v1/token":
self.token_calls += 1
self.form = {k: v[0] for k, v in parse_qs(req.content.decode()).items()}
if self.pub is not None and "client_assertion" in self.form:
try:
self.verified = jwt.decode(self.form["client_assertion"], self.pub, algorithms=[self.alg],
audience="https://example.okta.com/oauth2/v1/token")
except Exception as e: # noqa: BLE001
self.verified = e
body = self.token_body if self.token_body is not None else {
"token_type": "Bearer", "access_token": "ACCESS-TOKEN-VALUE", "expires_in": 3600}
return httpx.Response(self.token_status, json=body)
if req.url.path == "/api/v1/users":
self.user_calls += 1
self.seen_auth.append(req.headers.get("authorization", ""))
return httpx.Response(self.user_status, json=[
{"status": "ACTIVE", "profile": {"login": "alice"}}])
return httpx.Response(404)
def client(self):
return httpx.Client(transport=httpx.MockTransport(self.handler))
def setenv(jwk=None, as_b64=False, client_id="0oaCLIENT"):
os.environ["OKTA_SYNC_CLIENT_ID"] = client_id
if jwk is None:
os.environ.pop("OKTA_SYNC_PRIVATE_KEY", None)
else:
raw = json.dumps(jwk)
os.environ["OKTA_SYNC_PRIVATE_KEY"] = base64.b64encode(raw.encode()).decode() if as_b64 else raw
def fetch_err(fake):
try:
okta_sync.fetch_users(http=fake.client())
return None
except OktaSyncError as e:
return str(e)
# 1. Happy path, RSA, JSON form
rsa_jwk, rsa_pub, _ = make_rsa()
setenv(rsa_jwk)
chk("method is oauth", okta_sync.auth_method() == "oauth")
chk("is_configured with client id and key", okta_sync.is_configured())
f = FakeOkta(rsa_pub, "RS256")
users = okta_sync.fetch_users(http=f.client())
chk("returns the user", [u.identity for u in users] == ["alice"])
chk("token request fields", f.form.get("grant_type") == "client_credentials"
and f.form.get("scope") == "okta.users.read"
and f.form.get("client_assertion_type") == "urn:ietf:params:oauth:client-assertion-type:jwt-bearer")
v = f.verified
chk("assertion verifies against the public key", isinstance(v, dict), repr(v))
if isinstance(v, dict):
chk("iss and sub are the client id", v.get("iss") == "0oaCLIENT" and v.get("sub") == "0oaCLIENT")
chk("aud is the token endpoint", v.get("aud") == "https://example.okta.com/oauth2/v1/token")
chk("lifetime is at most 5 minutes", 0 < v["exp"] - v["iat"] <= 300)
chk("jti present", bool(v.get("jti")))
hdr = jwt.get_unverified_header(f.form["client_assertion"])
chk("kid and alg in the header", hdr.get("kid") == "rsa-kid-1" and hdr.get("alg") == "RS256")
chk("users call uses Bearer, not SSWS", f.seen_auth == ["Bearer ACCESS-TOKEN-VALUE"], str(f.seen_auth))
f2 = FakeOkta(rsa_pub)
okta_sync.fetch_users(http=f2.client())
chk("each assertion has a new jti", f.verified["jti"] != f2.verified["jti"])
# 2. base64 form and EC key
setenv(rsa_jwk, as_b64=True)
f = FakeOkta(rsa_pub)
chk("base64 form works", [u.identity for u in okta_sync.fetch_users(http=f.client())] == ["alice"])
ec_jwk, ec_pub, _ = make_ec()
setenv(ec_jwk)
f = FakeOkta(ec_pub, "ES256")
okta_sync.fetch_users(http=f.client())
chk("EC key works and verifies", isinstance(f.verified, dict), repr(f.verified))
# 3. Failures raise and never reach the users endpoint
setenv(rsa_jwk)
for label, fake in [
("token endpoint 401", FakeOkta(token_status=401, token_body={"error": "invalid_client"})),
("token endpoint 400 invalid_scope", FakeOkta(token_status=400, token_body={"error": "invalid_scope"})),
("token endpoint 500", FakeOkta(token_status=500, token_body={})),
("no access_token in body", FakeOkta(token_body={"token_type": "Bearer"})),
("DPoP token type", FakeOkta(token_body={"token_type": "DPoP", "access_token": "x"})),
("body is not an object", FakeOkta(token_body=[1, 2])),
]:
err = fetch_err(fake)
chk(f"{label}: raises, users never called", err is not None and fake.user_calls == 0, str(err))
e = fetch_err(FakeOkta(token_status=401, token_body={"error": "invalid_client"}))
chk("error carries the Okta error code and the status", "401" in e and "invalid_client" in e, e)
e = fetch_err(FakeOkta(token_status=400, token_body={"error": "<script>bad value with spaces</script>"}))
chk("an unsafe error value is dropped", "script" not in e and "400" in e, e)
# 4. Bad key material
pub_only = {k: rsa_jwk[k] for k in ("kty", "n", "e", "kid")}
for label, raw in [("public key only", json.dumps(pub_only)),
("not JSON, not base64", "{{{ not json"),
("base64 of something else", base64.b64encode(b"hello").decode()),
("wrong kty", json.dumps({"kty": "oct", "k": "AAAA"}))]:
os.environ["OKTA_SYNC_PRIVATE_KEY"] = raw
fake = FakeOkta()
err = fetch_err(fake)
chk(f"{label}: refused before any call", err is not None and fake.token_calls == 0 and fake.user_calls == 0, str(err))
if label == "public key only":
chk("public key: the message says it is a PUBLIC key", err is not None and "PUBLIC" in err, str(err))
chk(f"{label}: message holds no key text", err is not None and rsa_jwk["d"][:12] not in err and "hello" not in err)
setenv(rsa_jwk, client_id="")
os.environ["OKTA_SYNC_CLIENT_ID"] = ""
chk("no client id: not configured", okta_sync.auth_method() == "" and not okta_sync.is_configured())
os.environ["OKTA_SYNC_CLIENT_ID"] = "0oaCLIENT"
os.environ.pop("OKTA_SYNC_PRIVATE_KEY")
chk("no key: not configured", okta_sync.auth_method() == "")
# 4b. Copy and paste damage is forgiven, and the messages say which step failed
good = json.dumps(rsa_jwk)
b64 = base64.b64encode(good.encode()).decode()
variants = {
"double quotes around base64": '"' + b64 + '"',
"single quotes around base64": "'" + b64 + "'",
"line breaks every 76 characters": "\n".join(b64[i:i + 76] for i in range(0, len(b64), 76)),
"spaces and a trailing newline": " " + b64[:50] + " " + b64[50:] + " \r\n",
"padding removed": b64.rstrip("="),
"URL-safe alphabet": base64.urlsafe_b64encode(good.encode()).decode(),
"byte-order mark in the file": base64.b64encode(b"\xef\xbb\xbf" + good.encode()).decode(),
"JWK set with one key (JSON)": json.dumps({"keys": [rsa_jwk]}),
"JWK set with one key (base64)": base64.b64encode(json.dumps({"keys": [rsa_jwk]}).encode()).decode(),
}
for label, raw in variants.items():
os.environ["OKTA_SYNC_PRIVATE_KEY"] = raw
try:
f = FakeOkta(rsa_pub)
okta_sync.fetch_users(http=f.client())
chk(f"forgiven: {label}", isinstance(f.verified, dict) and f.user_calls == 1)
except OktaSyncError as e:
chk(f"forgiven: {label}", False, str(e))
# fields copied without the outer braces, and a value encoded twice, are forgiven
inner = json.dumps(rsa_jwk)[1:-1]
for label, raw in [("fields without braces (text)", inner),
("fields without braces (base64)", base64.b64encode(inner.encode()).decode()),
("encoded twice", base64.b64encode(b64.encode()).decode())]:
os.environ["OKTA_SYNC_PRIVATE_KEY"] = raw
try:
f = FakeOkta(rsa_pub)
okta_sync.fetch_users(http=f.client())
chk(f"forgiven: {label}", isinstance(f.verified, dict))
except OktaSyncError as e:
chk(f"forgiven: {label}", False, str(e))
pem = "-----BEGIN PRIVATE KEY-----\nMIIEvQIBADANBgkq\n-----END PRIVATE KEY-----"
for label, raw in [("PEM text", pem), ("PEM in base64", base64.b64encode(pem.encode()).decode())]:
os.environ["OKTA_SYNC_PRIVATE_KEY"] = raw
err = fetch_err(FakeOkta()) or ""
chk(f"PEM is named: {label}", "PEM" in err and "MIIEvQ" not in err, err)
# the pretty-printed, hand-copied shapes that Python's strict reader refuses
def b64s(t):
return base64.b64encode(t.encode()).decode()
pretty = json.dumps(rsa_jwk, indent=2)
wrapped = pretty.replace(rsa_jwk["n"], rsa_jwk["n"][:100] + "\n" + rsa_jwk["n"][100:]) # line break inside a value
wrapped_sp = pretty.replace(rsa_jwk["d"], rsa_jwk["d"][:60] + " \r\n " + rsa_jwk["d"][60:])
curly = pretty.replace('"', "\u201c", 1)
curly = curly.replace('": ', "\u201d: ") # curly quotes after the first
trailing = pretty[:-2] + ",\n}"
for label, text in [("pretty-printed, 36 lines", pretty),
("line break inside the n value", wrapped),
("space and line break inside the d value", wrapped_sp),
("trailing comma before the closing brace", trailing)]:
for how, raw in [("base64", b64s(text)), ("plain", text)]:
os.environ["OKTA_SYNC_PRIVATE_KEY"] = raw
try:
f = FakeOkta(rsa_pub)
okta_sync.fetch_users(http=f.client())
chk(f"forgiven ({how}): {label}", isinstance(f.verified, dict))
except OktaSyncError as e:
chk(f"forgiven ({how}): {label}", False, str(e))
curly_ok = json.dumps(rsa_jwk, indent=2).replace('"kty"', "\u201ckty\u201d")
os.environ["OKTA_SYNC_PRIVATE_KEY"] = b64s(curly_ok)
try:
f = FakeOkta(rsa_pub)
okta_sync.fetch_users(http=f.client())
chk("forgiven: curly quote marks", isinstance(f.verified, dict))
except OktaSyncError as e:
chk("forgiven: curly quote marks", False, str(e))
broken = pretty.replace('"kty": "RSA",', '"kty": "RSA" "x",', 1)
os.environ["OKTA_SYNC_PRIVATE_KEY"] = b64s(broken)
err = fetch_err(FakeOkta()) or ""
chk("a real syntax error names the line and column", "line " in err and "column " in err and "starts with a brace" in err, err)
chk("the position message holds no key text", rsa_jwk["d"][:10] not in err and rsa_jwk["n"][:10] not in err, err)
os.environ["OKTA_SYNC_PRIVATE_KEY"] = json.dumps({"keys": [rsa_jwk, rsa_jwk]})
chk("a key set with two keys is refused", "more than one key" in (fetch_err(FakeOkta()) or ""))
for label, raw, want in [
("garbage characters", "!!!not*base64!!!", "not valid base64"),
("base64 of non-JSON", base64.b64encode(b"hello world").decode(), "not JSON after decoding base64 (11 characters, starts with a letter or digit;"),
("base64 of bytes that are not text", base64.b64encode(b"\xff\xfe\xfd\xfc").decode(), "not text"),
("broken JSON", "{ not json", "not JSON"),
]:
os.environ["OKTA_SYNC_PRIVATE_KEY"] = raw
err = fetch_err(FakeOkta()) or ""
chk(f"clear message: {label}", want in err, err)
chk(f"no key text in message: {label}", raw[:8] not in err.replace("OKTA_SYNC_PRIVATE_KEY", ""), err)
# 5. An SSWS token still works and wins
setenv(rsa_jwk)
os.environ["OKTA_API_TOKEN"] = "ssws-token-value"
chk("method is token when both are set", okta_sync.auth_method() == "token")
f = FakeOkta()
okta_sync.fetch_users(http=f.client())
chk("SSWS used, no token request made", f.seen_auth == ["SSWS ssws-token-value"] and f.token_calls == 0, str(f.seen_auth))
os.environ.pop("OKTA_API_TOKEN")
# 6. Secrets stay out of the startup line
setenv(rsa_jwk)
line = okta_sync.describe()
chk("startup line names the method", "OAuth service app" in line, line)
chk("startup line holds no key part", rsa_jwk["d"][:12] not in line and rsa_jwk["n"][:12] not in line)
os.environ.pop("OKTA_SYNC_CLIENT_ID"); os.environ.pop("OKTA_SYNC_PRIVATE_KEY")
chk("startup line says DISABLED with no credential", "DISABLED" in okta_sync.describe())
print("\nDONE" if fails == 0 else f"\n{fails} FAILED")
sys.exit(1 if fails else 0)

View File

@@ -0,0 +1,242 @@
#!/usr/bin/env python3
"""Does the Okta sync run on a schedule without doubling up? - D18 / T13.4.
The api container runs 2 gunicorn workers, and each starts its own timer. Checked
here: exactly one of them wins each cycle (including a race on a real file
database), a restart inside the interval does not start a second run, the
interval is configurable and cannot go below the minimum, the timer thread starts
and stops, nothing starts when the sync is not configured, and a failure inside
the scheduled path still changes nothing.
Self-contained: a throwaway SQLite file, threads, no server, no network, no browser.
Exit 0 all passed, 1 a failure.
"""
import os
import sys
import tempfile
import threading
import time
from datetime import datetime, timedelta, timezone
ROOT = os.path.dirname(os.path.dirname(os.path.abspath(__file__)))
sys.path.insert(0, ROOT)
from sqlalchemy import create_engine, select # noqa: E402
from sqlalchemy.orm import sessionmaker # noqa: E402
from server import models, okta_auth, okta_sync # noqa: E402
from server.okta_sync import OktaSyncError, OktaUser # noqa: E402
fails = 0
def chk(label, ok, extra=""):
global fails
print(("PASS " if ok else "FAIL ") + label + (f" [{extra}]" if (extra and not ok) else ""))
if not ok:
fails += 1
tmp = tempfile.mkdtemp(prefix="oktasync_")
engine = create_engine(f"sqlite:///{os.path.join(tmp, 't.db')}",
connect_args={"check_same_thread": False, "timeout": 20}, future=True)
models.Base.metadata.create_all(engine)
Session = sessionmaker(bind=engine, autoflush=False, future=True)
INTERVAL = 900
os.environ["OKTA_SYNC_INTERVAL_SECONDS"] = str(INTERVAL)
T0 = datetime(2026, 10, 16, 12, 0, 0, tzinfo=timezone.utc)
def wipe(users=()):
with Session() as db:
db.query(models.AuditLog).delete()
db.query(models.User).delete()
db.query(models.AppSetting).filter(models.AppSetting.key == okta_sync.CLAIM_KEY).delete()
for name, active in users:
db.add(models.User(id="u_" + name, username=name, is_active=active))
db.commit()
def users_state():
with Session() as db:
return ({u.username: u.is_active for u in db.scalars(select(models.User))},
db.query(models.AuditLog).count())
# ── 1. interval ───────────────────────────────────────────────────────────────
print("\n1. interval")
def iv(v):
if v is None:
os.environ.pop("OKTA_SYNC_INTERVAL_SECONDS", None)
else:
os.environ["OKTA_SYNC_INTERVAL_SECONDS"] = v
return okta_sync.interval_seconds()
chk("default is 900 seconds (15 minutes)", iv(None) == 900 and iv("") == 900)
chk("an env value is used", iv("600") == 600 and iv("3600") == 3600)
chk("not a number: the default", iv("soon") == 900 and iv("1.5") == 900)
chk("under the minimum is raised to 300 (a slow cycle can never overlap the next)",
iv("60") == 300 and iv("0") == 300 and iv("-5") == 300)
chk("a fetch is bounded below the minimum interval",
okta_sync.TOTAL_DEADLINE_SECONDS < okta_sync.INTERVAL_MIN_SECONDS)
iv(str(INTERVAL))
# ── 2. the claim ──────────────────────────────────────────────────────────────
print("\n2. the claim gate")
wipe()
with Session() as db:
chk("first ever claim wins (the insert is the claim)", okta_sync.claim(db, INTERVAL, now=T0) is True)
with Session() as db:
chk("a second claim in the same instant loses", okta_sync.claim(db, INTERVAL, now=T0) is False)
with Session() as db:
chk("one second before the interval is up: loses",
okta_sync.claim(db, INTERVAL, now=T0 + timedelta(seconds=INTERVAL - 1)) is False)
with Session() as db:
chk("when the interval is up: wins",
okta_sync.claim(db, INTERVAL, now=T0 + timedelta(seconds=INTERVAL)) is True)
with Session() as db:
chk("and then the next caller loses again",
okta_sync.claim(db, INTERVAL, now=T0 + timedelta(seconds=INTERVAL + 1)) is False)
# ── 3. the race ───────────────────────────────────────────────────────────────
print("\n3. exactly one of many workers wins (a real race on a file database)")
def race(n=8):
barrier = threading.Barrier(n)
wins, errors = [], []
def worker():
try:
with Session() as db:
barrier.wait()
wins.append(okta_sync.claim(db, INTERVAL, now=datetime.now(timezone.utc)))
except Exception as exc: # a lost race may surface as an error
errors.append(type(exc).__name__)
wins.append(False)
ts = [threading.Thread(target=worker) for _ in range(n)]
[t.start() for t in ts]
[t.join() for t in ts]
return sum(1 for w in wins if w), errors
wipe()
won, errs = race()
chk("no row yet, 8 workers at once: exactly 1 wins (the insert race)", won == 1, f"{won} {errs}")
bad = []
for i in range(15):
with Session() as db: # age the claim so the next race is due again
db.query(models.AppSetting).filter(models.AppSetting.key == okta_sync.CLAIM_KEY).update(
{"updated_at": datetime.now(timezone.utc) - timedelta(seconds=INTERVAL + 5)})
db.commit()
won, errs = race()
if won != 1:
bad.append((i, won, errs))
chk("row due, 8 workers at once, 15 rounds: exactly 1 wins every round", not bad, str(bad))
# ── 4. one tick ───────────────────────────────────────────────────────────────
print("\n4. a tick runs the sync once per cycle")
wipe([("alice", True), ("bob", True)])
calls = []
def fetch_ok():
calls.append(1)
return [OktaUser("alice", "ACTIVE"), OktaUser("bob", "SUSPENDED")]
r1 = okta_sync.run_if_due(Session, fetch=fetch_ok, now=T0)
chk("first tick claims and runs", r1 is not None and r1.ok and r1.disabled == ["bob"])
r2 = okta_sync.run_if_due(Session, fetch=fetch_ok, now=T0 + timedelta(seconds=30))
chk("a tick 30 seconds later does nothing (not due)", r2 is None and len(calls) == 1)
r3 = okta_sync.run_if_due(Session, fetch=fetch_ok, now=T0 + timedelta(seconds=INTERVAL))
chk("a tick when due runs again", r3 is not None and r3.ok and len(calls) == 2)
print("\n ...a restart inside the interval does not start a second run")
wipe([("alice", True), ("bob", True)])
calls.clear()
okta_sync.run_if_due(Session, fetch=fetch_ok, now=T0)
# "restart": a brand-new caller with nothing in memory, 2 minutes later
r = okta_sync.run_if_due(Session, fetch=fetch_ok, now=T0 + timedelta(seconds=120))
chk("the claim row outlives the process: no duplicate run", r is None and len(calls) == 1)
print("\n ...two workers tick in the same instant")
wipe([("alice", True), ("bob", True)])
calls.clear()
results = []
barrier = threading.Barrier(2)
def tick():
barrier.wait()
results.append(okta_sync.run_if_due(Session, fetch=fetch_ok))
ts = [threading.Thread(target=tick) for _ in range(2)]
[t.start() for t in ts]
[t.join() for t in ts]
chk("Okta is asked once, not twice", len(calls) == 1, str(len(calls)))
chk("exactly one worker got a result", sum(1 for r in results if r is not None) == 1)
_, audits = users_state()
chk("one audit row, not two", audits == 1, str(audits))
# ── 5. fail closed through the scheduled path ─────────────────────────────────
print("\n5. FAIL CLOSED through the scheduled wrapper (not just run_sync)")
for label, exc in (("OktaSyncError", OktaSyncError("HTTP 500")), ("RuntimeError", RuntimeError("x")),
("TimeoutError", TimeoutError("t"))):
wipe([("alice", True), ("bob", True)])
def failing(e=exc):
raise e
r = okta_sync.run_if_due(Session, fetch=failing, now=T0)
u, a = users_state()
chk(f"{label} in the scheduled path: ok is False, zero changes, zero audit rows",
r is not None and r.ok is False and all(u.values()) and a == 0)
wipe([("alice", True), ("bob", True)])
def broken_factory():
raise RuntimeError("database is down")
chk("the database failing at claim time: the tick ends quietly, returns None",
okta_sync.run_if_due(broken_factory, fetch=fetch_ok, now=T0) is None)
u, a = users_state()
chk("...and nothing changed", all(u.values()) and a == 0)
# ── 6. the thread ─────────────────────────────────────────────────────────────
print("\n6. the timer thread")
os.environ["OKTA_API_TOKEN"] = ""
okta_auth.ISSUER = "https://example.okta.com/oauth2/default"
chk("no token: nothing starts", okta_sync.start_scheduler(Session) is None)
chk("...no thread named okta-sync", not any(t.name == "okta-sync" for t in threading.enumerate()))
os.environ["OKTA_API_TOKEN"] = "tok"
wipe([("alice", True), ("bob", True)])
okta_sync.START_DELAY_SECONDS = 0.05
okta_sync.TICK_SECONDS = 0.05
ran = []
real_fetch = okta_sync.fetch_users
okta_sync.fetch_users = lambda http=None: (ran.append(1) or
[OktaUser("alice", "ACTIVE"), OktaUser("bob", "DEPROVISIONED")])
stop = okta_sync.start_scheduler(Session)
chk("configured: a stop Event comes back", stop is not None)
time.sleep(1.2) # about 20 ticks
u, a = users_state()
chk("the thread ran the sync on its own (bob was disabled without a manual call)", u["bob"] is False)
chk("many ticks, one run (the claim gate holds)", len(ran) == 1 and a == 1, f"ran={len(ran)} audits={a}")
th = [t for t in threading.enumerate() if t.name == "okta-sync"]
chk("one timer thread is alive", len(th) == 1 and th[0].is_alive())
chk("the thread is a daemon (never blocks shutdown)", th[0].daemon)
stop.set()
th[0].join(timeout=3)
chk("setting the stop Event ends the thread", not th[0].is_alive())
okta_sync.fetch_users = real_fetch
os.environ["OKTA_API_TOKEN"] = ""
print()
print("ALL PASSED" if not fails else f"{fails} FAILED")
sys.exit(0 if not fails else 1)

153
tests/okta_sync_ui_check.py Normal file
View File

@@ -0,0 +1,153 @@
#!/usr/bin/env python3
"""Does an auto-disable read clearly in the admin console? - D18 / T13.5.
T13.5 builds no new UI. It confirms that an account the Okta sync disabled shows
up legibly in the surface admins already use (the Activity log card on
admin.html), without special-casing, and that the disabled account is refused on
its very next request. This runs the REAL sync against a throwaway database,
then reads the real page in a headless browser at 1440px and 390px.
It does not rely on CR-019's Activity & usage card: only the Activity log card
(AuditLog) is read.
Boots its own throwaway SQLite + uvicorn + headless browser; run it alone.
Screenshots go to a folder printed at the end (WP_SHOTS_DIR overrides it).
Exit 0 all passed, 1 a failure, 2 could not run.
"""
import json
import os
import re
import sys
import tempfile
import time
sys.path.insert(0, os.path.dirname(os.path.dirname(os.path.abspath(__file__))))
sys.path.insert(0, os.path.dirname(os.path.abspath(__file__)))
import cdp # noqa: E402
from browser_check import seed, start_server, chk, _PASS, _FAIL # noqa: E402
from qa_gate_check import api # noqa: E402
def ascii_(v, n=300):
return re.sub(r"\s+", " ", str(v)).encode("ascii", "replace").decode()[:n]
ROWS_JS = """JSON.stringify([...document.querySelectorAll('#audit-admin tbody tr')].map(
tr => [...tr.children].map(td => td.textContent.trim())))"""
def main():
exe = cdp.find_browser()
if not exe:
print("no headless-capable browser found; set WP_BROWSER.")
return 2
tmpdir = tempfile.mkdtemp(prefix="wpsuite-syncui-")
shots = os.environ.get("WP_SHOTS_DIR") or os.path.join(tmpdir, "shots")
db_path = os.path.join(tmpdir, "check.db")
server = browser = None
try:
tok = seed(db_path)
print("\n1. run the real sync against the throwaway database")
from server.db import SessionLocal
from server import okta_sync
from server.okta_sync import OktaUser
okta = lambda: [OktaUser("root", "ACTIVE"), OktaUser("sue", "ACTIVE"),
OktaUser("pat", "SUSPENDED"), OktaUser("mix", "ACTIVE"),
OktaUser("sam", "ACTIVE"), OktaUser("legacy", "ACTIVE")] # bob is missing
with SessionLocal() as db:
res = okta_sync.run_sync(db, fetch=okta)
chk("the sync disabled pat (SUSPENDED) and bob (missing)",
res.ok and sorted(res.disabled) == ["bob", "pat"], ascii_((res.ok, res.error, res.disabled)))
port = cdp.free_port()
base = "http://127.0.0.1:%d" % port
server = start_server(port, db_path)
if not server:
print("the server did not start")
return 2
print("\n2. a disabled account is refused on its very next request")
st, _ = api(base, "/api/auth/me", tok["pat"])
chk("pat (auto-disabled) is refused", st in (401, 403), st)
st, _ = api(base, "/api/auth/me", tok["bob"])
chk("bob (auto-disabled) is refused", st in (401, 403), st)
st, _ = api(base, "/api/auth/me", tok["mix"])
chk("mix (still active in Okta) still works", st == 200, st)
print("\n3. the audit rows through the API")
st, audit = api(base, "/api/audit?entity_type=user&limit=50", tok["root"])
mine = [e for e in (audit or []) if e.get("actor") == "system:okta_sync"]
chk("an admin gets both rows", st == 200 and len(mine) == 2, ascii_((st, len(mine))))
chk("each row names the action, the user, and the Okta status",
{(e["summary"], e["detail"].get("okta_status")) for e in mine} == {("pat", "SUSPENDED"), ("bob", "MISSING")},
ascii_([(e["summary"], e["detail"]) for e in mine]))
st, audit_sue = api(base, "/api/audit?entity_type=user&limit=50", tok["sue"])
chk("a project super user does not see them (they belong to no project)",
st == 200 and not [e for e in audit_sue if e.get("actor") == "system:okta_sync"])
print("\n4. the admin console, as an admin")
browser = cdp.Browser(exe)
page = browser.page()
for width, label in ((1440, "1440"), (390, "390")):
page.clear_cookies()
page.set_cookie("wp_session", tok["root"])
page.viewport(width, 900, mobile=(width < 500))
page.goto(base + "/admin.html")
time.sleep(2.0)
page.eval("void loadAudit()")
time.sleep(1.2)
rows = json.loads(page.eval(ROWS_JS) or "[]")
sync_rows = [r for r in rows if "system:okta_sync" in r]
chk("[%s] the Activity log lists both auto-disables" % label, len(sync_rows) == 2,
ascii_(rows[:4]))
if sync_rows:
r = next((x for x in sync_rows if "pat" in x), sync_rows[0])
# columns: When, Who, Action, Type, Item, Detail
chk("[%s] Who reads 'system:okta_sync', not a person" % label, r[1] == "system:okta_sync", r)
chk("[%s] Action reads 'user deprovisioned by sync'" % label,
r[2] == "user deprovisioned by sync", r)
chk("[%s] Type and Item name the account" % label, r[3] == "user" and r[4] == "pat", r)
chk("[%s] Detail gives the Okta status found" % label,
"SUSPENDED" in r[5] and "okta" in r[5].lower(), r)
miss = next((x for x in sync_rows if "bob" in x), None)
chk("[%s] a person missing from Okta reads 'MISSING'" % label,
bool(miss) and "MISSING" in miss[5], miss)
# filter + search work on it with no special-casing
page.eval("document.getElementById('audit-search').value='okta_sync'; renderAudit();")
rows = json.loads(page.eval(ROWS_JS) or "[]")
chk("[%s] searching 'okta_sync' finds exactly the sync rows" % label,
len(rows) == 2 and all("system:okta_sync" in r for r in rows), ascii_(rows))
page.eval("document.getElementById('audit-search').value='';"
"document.getElementById('audit-type').value='user'; renderAudit();")
rows = json.loads(page.eval(ROWS_JS) or "[]")
chk("[%s] the 'User accounts' filter keeps them" % label,
len([r for r in rows if "system:okta_sync" in r]) == 2)
page.eval("document.getElementById('audit-type').value=''; renderAudit();")
over = page.eval("document.documentElement.scrollWidth - document.documentElement.clientWidth")
chk("[%s] no sideways scroll on the page" % label, (over or 0) <= 0, over)
card = page.eval("(()=>{const e=document.getElementById('audit-admin');"
"e.scrollIntoView();return e.getBoundingClientRect().width>0})()")
chk("[%s] the card is on screen" % label, bool(card))
time.sleep(0.4)
page.screenshot(os.path.join(shots, "admin-activity-log-%s.png" % label), full_page=False)
errs = [e for e in page.js_errors() if "beforeunload" not in e]
chk("no JavaScript errors", not errs, ascii_(errs[:3]))
print("\n screenshots: " + shots)
finally:
if browser:
browser.close()
if server:
server.terminate()
print("\n" + "-" * 54)
print("%d/%d checks passed." % (len(_PASS), len(_PASS) + len(_FAIL)))
for f in _FAIL:
print(" - " + f)
return 1 if _FAIL else 0
if __name__ == "__main__":
sys.exit(main())

226
tests/outbox_check.py Normal file
View File

@@ -0,0 +1,226 @@
#!/usr/bin/env python3
"""Does a sign-in problem destroy queued work? - D26 / T14.1.
The outbox in html/project-data.js queues every save and sends it later. It used to
treat a 401 (the session ended) like any other 4xx: the change was marked dead and
never sent again. It also hid a full device store and sent a change under whoever
was signed in. This check drives the real page in a real browser. It replaces
window.fetch for the write endpoints only, so no real server state is needed.
1. a 401 holds the whole queue: nothing dead, no try counted, state is "held"
2. online, tab focus and the timer do not release a held queue
3. a good sign-in (wp-auth-ready) sends every change, in the original order
4. a 403 kills only that change and says why
5. a full device store is reported to the caller and the badge; nothing is dropped
6. a change is sent only under the person who made it
7. auth-guard.js never deletes the outbox
Exit 0 all passed, 1 a failure, 2 could not run.
"""
import json
import os
import re
import subprocess
import sys
import tempfile
import time
sys.path.insert(0, os.path.dirname(os.path.dirname(os.path.abspath(__file__))))
sys.path.insert(0, os.path.dirname(os.path.abspath(__file__)))
import cdp # noqa: E402
from browser_check import seed, start_server, chk, _PASS, _FAIL # noqa: E402
KEY = "wp_sync_outbox_v1"
HERE = os.path.dirname(os.path.abspath(__file__))
HTML = os.path.join(os.path.dirname(HERE), "html")
# Replaces fetch for POST /api/wps, POST /api/sops and DELETE /api/wps/<id> only.
# window.__plan maps a work package id to a status; the default is window.__status.
STUB = """(() => {
if (!window.__realFetch) window.__realFetch = window.fetch.bind(window);
window.__calls = []; window.__status = 200; window.__plan = {};
window.fetch = function (url, opts) {
const m = ((opts && opts.method) || 'GET').toUpperCase();
const u = String(url);
if ((m === 'POST' && /\\/api\\/(wps|sops)$/.test(u)) || (m === 'DELETE' && /\\/api\\/wps\\//.test(u))) {
let id = '';
try { id = JSON.parse(opts.body).id; } catch (e) { id = decodeURIComponent(u.split('/').pop()); }
window.__calls.push(id);
const st = (id in window.__plan) ? window.__plan[id] : window.__status;
return Promise.resolve(new Response(JSON.stringify(st >= 400 ? { detail: 'stub ' + st } : { ok: true }),
{ status: st, headers: { 'Content-Type': 'application/json' } }));
}
return window.__realFetch(url, opts);
};
localStorage.removeItem('wp_sync_outbox_v1');
return true;
})()"""
def boot(page, base, tok, user="root"):
page.clear_cookies()
page.set_cookie("wp_session", tok[user])
page.goto(base + "/field.html")
for _ in range(40):
if page.eval("!!(window.WP_USER && window.ProjectData)"):
break
time.sleep(0.3)
time.sleep(0.5)
page.eval(STUB)
def queue(page):
return json.loads(page.eval("localStorage.getItem(%r) || '[]'" % KEY))
def status(page):
return page.eval("ProjectData.syncStatus()")
def push(page, ids):
page.eval("(() => { %s; return true; })()" % ";".join(
"ProjectData.pushWP({id:%r, name:%r}, 'projA')" % (i, i) for i in ids))
def main():
exe = cdp.find_browser()
if not exe:
print("no headless-capable browser found; set WP_BROWSER.")
return 2
tmpdir = tempfile.mkdtemp(prefix="wpsuite-outbox-")
db_path = os.path.join(tmpdir, "check.db")
server = None
try:
tok = seed(db_path)
port = cdp.free_port()
base = "http://127.0.0.1:%d" % port
server = start_server(port, db_path)
if server is None:
print("the test server would not start.")
return 2
print("\nThe outbox keeps every change - D26 / T14.1\nTarget: %s" % base)
browser = cdp.Browser(exe)
page = browser.page()
try:
ids = ["wp_o1", "wp_o2", "wp_o3"]
print("\n1. a 401 holds the whole queue")
boot(page, base, tok)
page.eval("window.__status = 401")
push(page, ids)
page.eval("ProjectData.flushSync()")
q = queue(page)
st = status(page)
chk("all 3 changes are still queued", [o["key"] for o in q] == ids, str([o["key"] for o in q]))
chk("none is marked dead", not any(o.get("dead") for o in q) and st["dead"] == 0)
chk("no try was counted", all((o.get("tries") or 0) == 0 for o in q), str([o.get("tries") for o in q]))
chk("the state is held", st["held"] is True, json.dumps(st))
chk("the run stopped at the first 401", len(page.eval("window.__calls")) == 1,
str(page.eval("window.__calls")))
badge = page.eval("document.getElementById('wp-sync-badge').textContent")
chk("the badge asks for a new sign-in", "Sign in again" in badge, badge)
print("\n2. only a good sign-in releases it")
page.eval("window.__status = 200")
page.eval("window.dispatchEvent(new Event('online'))")
page.eval("document.dispatchEvent(new Event('visibilitychange'))")
page.eval("ProjectData.flushSync()")
time.sleep(1.0)
chk("online, focus and a manual flush send nothing while held",
len(page.eval("window.__calls")) == 1 and len(queue(page)) == 3,
str(page.eval("window.__calls")))
print("\n3. a good sign-in sends all of them, in order")
page.eval("document.dispatchEvent(new CustomEvent('wp-auth-ready', {detail: window.WP_USER}))")
for _ in range(30):
if not queue(page):
break
time.sleep(0.3)
calls = page.eval("window.__calls")
chk("the queue is empty", queue(page) == [], str(queue(page)))
chk("the changes arrived in the original order", calls[1:] == ids or calls[-3:] == ids, str(calls))
chk("the state is no longer held", status(page)["held"] is False)
print("\n4. a 403 kills only that change, with a reason")
boot(page, base, tok)
page.eval("window.__plan = {wp_o2: 403}")
push(page, ids)
page.eval("ProjectData.flushSync()")
q = queue(page)
dead = [o for o in q if o.get("dead")]
chk("only wp_o2 is dead", [o["key"] for o in dead] == ["wp_o2"], str([o["key"] for o in q]))
chk("the others were sent", [o["key"] for o in q if not o.get("dead")] == [])
chk("the reason names the cause", dead and "no longer have access" in dead[0].get("lastErr", ""),
dead[0].get("lastErr", "") if dead else "")
chk("a 403 does not hold the queue", status(page)["held"] is False)
print("\n5. a full device store is never silent")
boot(page, base, tok)
page.eval("window.__status = 500") # keep it queued so the badge state stays visible
page.eval("""(() => {
window.__setItem = Storage.prototype.setItem;
Storage.prototype.setItem = function (k, v) {
if (k === 'wp_sync_outbox_v1') { const e = new Error('full'); e.name = 'QuotaExceededError'; throw e; }
return window.__setItem.call(this, k, v);
};
return true; })()""")
stored = page.eval("ProjectData.pushWP({id:'wp_full', name:'full'}, 'projA')")
st = status(page)
chk("the caller is told it was not stored", stored is False, repr(stored))
chk("the state says device full, 1 unsaved", st["deviceFull"] is True and st["unsaved"] == 1, json.dumps(st))
badge = page.eval("document.getElementById('wp-sync-badge').textContent")
chk("the badge says it is not saved on this device", "not saved on this device" in badge, badge)
page.eval("window.__status = 200")
page.eval("ProjectData.flushSync()")
chk("the change is still sent from memory", "wp_full" in page.eval("window.__calls"),
str(page.eval("window.__calls")))
page.eval("Storage.prototype.setItem = window.__setItem")
chk("the state clears once it is sent", status(page)["unsaved"] == 0)
print("\n6. a change is sent only under the person who made it")
boot(page, base, tok)
page.eval("window.__status = 503") # keep the change queued while we look at it
push(page, ["wp_owned"])
time.sleep(0.5)
page.eval("window.__status = 200")
q0 = queue(page)
owner = (q0[0].get("owner") if q0 else None)
chk("the change records its owner", owner == "root", repr(owner))
page.eval("window.__calls = []; window.__me = window.WP_USER; window.WP_USER = {username: 'pat'}")
page.eval("ProjectData.flushSync()")
st = status(page)
chk("person B sends nothing", page.eval("window.__calls") == [], str(page.eval("window.__calls")))
chk("the change is kept", [o["key"] for o in queue(page)] == ["wp_owned"])
chk("the badge names the owner", st["others"] == 1 and st["otherOwner"] == "root", json.dumps(st))
page.eval("window.WP_USER = window.__me")
page.eval("ProjectData.flushSync()")
chk("it is sent after the owner signs in", page.eval("window.__calls") == ["wp_owned"]
and queue(page) == [], str(page.eval("window.__calls")))
finally:
page.close()
browser.close()
print("\n7. auth-guard.js never deletes the outbox")
src = open(os.path.join(HTML, "auth-guard.js"), encoding="utf-8").read()
code = re.sub(r"/\*.*?\*/|//[^\n]*", "", src, flags=re.S)
chk("no code path touches the outbox key", KEY not in code)
chk("nothing clears all of localStorage", "localStorage.clear" not in code)
finally:
if server is not None:
server.kill()
try:
server.wait(timeout=10)
except subprocess.TimeoutExpired:
pass
try:
from server.db import engine
engine.dispose()
except Exception:
pass
print("\n%d passed, %d failed" % (len(_PASS), len(_FAIL)))
return 1 if _FAIL else 0
if __name__ == "__main__":
sys.exit(main())

View File

@@ -0,0 +1,66 @@
#!/usr/bin/env python3
"""Is the session limit 8 hours by default? - D25 / T13.11.
The default of AUTH_SESSION_HOURS is 8. The token expiry and the cookie max-age both
follow it, a set value overrides it, and the shipped files name the same default.
Self-contained: no network, no server, no browser. Exit 0 all passed, 1 a failure.
"""
import importlib
import os
import re
import sys
from datetime import datetime, timezone
from types import SimpleNamespace
ROOT = os.path.dirname(os.path.dirname(os.path.abspath(__file__)))
sys.path.insert(0, ROOT)
os.environ["DATABASE_URL"] = "sqlite://"
os.environ["AUTH_SECRET_KEY"] = "session-default-check-secret-0123456789abcdef"
fails = 0
def chk(label, ok, extra=""):
global fails
print(("PASS " if ok else "FAIL ") + label + (f" [{extra}]" if (extra and not ok) else ""))
if not ok:
fails += 1
def load(value):
if value is None:
os.environ.pop("AUTH_SESSION_HOURS", None)
else:
os.environ["AUTH_SESSION_HOURS"] = value
from server import auth
return importlib.reload(auth)
import jwt # noqa: E402
user = SimpleNamespace(id="u1", username="a", role="project_user", token_version=0)
auth = load(None)
chk("default is 8 hours", auth.SESSION_HOURS == 8, str(auth.SESSION_HOURS))
claims = jwt.decode(auth.create_token(user), auth.SECRET_KEY, algorithms=["HS256"])
hours = (datetime.fromtimestamp(claims["exp"], timezone.utc) - datetime.fromtimestamp(claims["iat"], timezone.utc)).total_seconds() / 3600
chk("a new token lasts 8 hours", abs(hours - 8) < 0.01, str(hours))
auth = load("2")
chk("AUTH_SESSION_HOURS=2 overrides the default", auth.SESSION_HOURS == 2)
auth = load("12")
chk("AUTH_SESSION_HOURS=12 overrides the default", auth.SESSION_HOURS == 12)
load(None)
def read(p):
return open(os.path.join(ROOT, p), encoding="utf-8").read()
chk("docker-compose.yml defaults to 8", "AUTH_SESSION_HOURS: ${AUTH_SESSION_HOURS:-8}" in read("docker-compose.yml"))
chk(".env.example says 8", re.search(r"^# AUTH_SESSION_HOURS=8$", read("server/.env.example"), re.M) is not None)
chk("DEPLOY-login-portal.md says 8", "defaults to 8" in read("DEPLOY-login-portal.md"))
for p in ("docker-compose.yml", "server/.env.example", "DEPLOY-login-portal.md", "server/auth.py"):
chk(f"no leftover default of 12 in {p}", not re.search(r"AUTH_SESSION_HOURS[^\n]{0,30}\b12\b", read(p)))
print("\nDONE" if fails == 0 else f"\n{fails} FAILED")
sys.exit(1 if fails else 0)

View File

@@ -0,0 +1,220 @@
#!/usr/bin/env python3
"""Does the offline sign-in window follow the server? - D27 / T14.2.
auth-guard.js used to let the app open offline for a fixed 12 hours after the last
good check, and every page load reset that time. The server session lasts
AUTH_SESSION_HOURS (8). This check proves the browser now uses the server's own end
time.
1. /api/auth/me returns session_expires_at and it matches the token's exp
2. the browser keeps it, and the fixed 12 hours is gone from the code
3. offline before that time: the app opens as normal
4. offline after that time: read and queue. No redirect, the mode line shows,
a new change is queued and held, not sent
5. an older cache with no end time is treated as ended (no hidden 12 hours)
6. a page load offline does not extend the cached window
7. no cache at all, offline: sign-in page
8. online with a dead session: sign-in page with next=, and the outbox is kept
9. refreshing the cached user keeps the end of the session
Exit 0 all passed, 1 a failure, 2 could not run.
"""
import base64
import calendar
import json
import os
import re
import subprocess
import sys
import tempfile
import time
sys.path.insert(0, os.path.dirname(os.path.dirname(os.path.abspath(__file__))))
sys.path.insert(0, os.path.dirname(os.path.abspath(__file__)))
import cdp # noqa: E402
from browser_check import seed, start_server, chk, _PASS, _FAIL # noqa: E402
HERE = os.path.dirname(os.path.abspath(__file__))
HTML = os.path.join(os.path.dirname(HERE), "html")
KEY = "wp_sync_outbox_v1"
# Makes /api/auth/me fail like a missing network when localStorage.x_me_offline is set.
OFFLINE_SHIM = """(() => {
const real = window.fetch.bind(window);
window.fetch = function (u, o) {
let off = false; try { off = !!localStorage.getItem('x_me_offline'); } catch (e) {}
if (off && String(u).indexOf('/api/auth/me') !== -1) return Promise.reject(new TypeError('Failed to fetch'));
return real(u, o);
};
})();"""
STUB = """(() => {
if (!window.__realFetch) window.__realFetch = window.fetch.bind(window);
window.__calls = [];
window.fetch = function (url, opts) {
const m = ((opts && opts.method) || 'GET').toUpperCase();
if (m === 'POST' && /\\/api\\/(wps|sops)$/.test(String(url))) {
window.__calls.push(String(url));
return Promise.resolve(new Response('{}', { status: 200, headers: { 'Content-Type': 'application/json' } }));
}
return window.__realFetch(url, opts);
};
return true;
})()"""
def token_exp(token):
body = token.split(".")[1]
body += "=" * (-len(body) % 4)
return int(json.loads(base64.urlsafe_b64decode(body))["exp"])
def cache(page):
return json.loads(page.eval("localStorage.getItem('wp_auth_cache') || 'null'") or "null")
def load(page, base, path="/field.html"):
page.goto(base + path)
time.sleep(2.0)
def main():
exe = cdp.find_browser()
if not exe:
print("no headless-capable browser found; set WP_BROWSER.")
return 2
tmpdir = tempfile.mkdtemp(prefix="wpsuite-expiry-")
db_path = os.path.join(tmpdir, "check.db")
server = None
try:
tok = seed(db_path)
port = cdp.free_port()
base = "http://127.0.0.1:%d" % port
server = start_server(port, db_path)
if server is None:
print("the test server would not start.")
return 2
print("\nThe offline window follows the server - D27 / T14.2\nTarget: %s" % base)
browser = cdp.Browser(exe)
page = browser.page()
try:
page.ws.call("Page.addScriptToEvaluateOnNewDocument", {"source": OFFLINE_SHIM})
page.clear_cookies()
page.set_cookie("wp_session", tok["root"])
want = token_exp(tok["root"])
print("\n1-2. the server sends its own end time and the browser keeps it")
load(page, base)
me = page.eval("fetch('/api/auth/me').then(r => r.json())")
got = (me or {}).get("session_expires_at")
chk("/api/auth/me returns session_expires_at", bool(got), json.dumps(me)[:200])
chk("it is a UTC time that matches the token exp",
bool(got) and bool(re.match(r"^\d{4}-\d\d-\d\dT\d\d:\d\d:\d\dZ$", got)) and
abs(calendar.timegm(time.strptime(got, "%Y-%m-%dT%H:%M:%SZ")) - want) <= 1,
"%s vs %s" % (got, want))
c = cache(page)
chk("the cache holds the server's end time",
bool(c) and abs((c.get("exp") or 0) / 1000.0 - want) <= 1, json.dumps(c)[:200])
src = open(os.path.join(HTML, "auth-guard.js"), encoding="utf-8").read()
code = re.sub(r"/\*.*?\*/|//[^\n]*", "", src, flags=re.S)
chk("the fixed 12 hours is gone from auth-guard.js", "12 * 3600" not in code and "43200000" not in code)
user = json.dumps(c["user"]) if c else "{}"
now_ms = int(time.time() * 1000)
def set_cache(obj):
page.eval("localStorage.setItem('wp_auth_cache', %s); localStorage.setItem('x_me_offline','1'); true"
% json.dumps(json.dumps(obj)))
print("\n3. offline before the end time: opens as normal")
future = now_ms + 3600 * 1000
set_cache({"user": c["user"], "exp": future, "at": 111})
load(page, base)
chk("no redirect", page.eval("location.pathname") == "/field.html", page.eval("location.href"))
chk("the user is the cached user", page.eval("window.WP_USER && window.WP_USER.username") == "root")
chk("mode is offline, not read-and-queue", page.eval("window.WP_SESSION_MODE") == "offline",
page.eval("window.WP_SESSION_MODE"))
print("\n6. a page load offline does not extend the window")
c2 = cache(page)
chk("the end time did not move", c2 and c2.get("exp") == future, json.dumps(c2))
chk("the cache was not rewritten", c2 and c2.get("at") == 111, json.dumps(c2))
print("\n4. offline after the end time: read and queue")
set_cache({"user": c["user"], "exp": now_ms - 3600 * 1000, "at": now_ms})
load(page, base)
chk("no redirect to the sign-in page", page.eval("location.pathname") == "/field.html",
page.eval("location.href"))
chk("the user is the cached user", page.eval("window.WP_USER && window.WP_USER.username") == "root")
chk("mode is read and queue", page.eval("window.WP_SESSION_MODE") == "readqueue",
page.eval("window.WP_SESSION_MODE"))
badge = page.eval("(document.getElementById('wp-sync-badge')||{}).textContent || ''")
chk("the mode line shows", "Working offline. Sign in again when you have a signal to send your changes." in badge,
badge)
chk("the line sits in the live region",
page.eval("(document.getElementById('wp-sync-badge')||{getAttribute(){}}).getAttribute('role')") == "status")
page.eval(STUB)
page.eval("localStorage.removeItem(%r); ProjectData.pushWP({id:'wp_rq', name:'rq'}, 'projA'); true" % KEY)
page.eval("ProjectData.flushSync()")
time.sleep(0.5)
q = json.loads(page.eval("localStorage.getItem(%r) || '[]'" % KEY))
chk("a new change is queued", [o.get("key") for o in q] == ["wp_rq"], json.dumps(q)[:200])
chk("and held: nothing was sent", page.eval("window.__calls") == [], str(page.eval("window.__calls")))
print("\n5. an older cache with no end time counts as ended")
set_cache({"user": c["user"], "at": now_ms})
load(page, base)
chk("read and queue, no redirect, no hidden window",
page.eval("location.pathname") == "/field.html" and page.eval("window.WP_SESSION_MODE") == "readqueue",
page.eval("location.href") + " " + str(page.eval("window.WP_SESSION_MODE")))
print("\n9. refreshing the cached user keeps the end of the session")
set_cache({"user": c["user"], "exp": future, "at": 1})
page.eval("window.wpCacheAuth({username: 'renamed'}); true")
c3 = cache(page)
chk("the user changed", c3 and c3["user"].get("username") == "renamed", json.dumps(c3))
chk("the end time did not", c3 and c3.get("exp") == future, json.dumps(c3))
print("\n7. no cache, offline: the sign-in page")
page.eval("localStorage.removeItem('wp_auth_cache'); localStorage.setItem('x_me_offline','1'); true")
load(page, base)
chk("sent to login.html", page.eval("location.pathname") == "/login.html", page.eval("location.href"))
print("\n8. online with a dead session: sign-in page and the outbox is kept")
page.clear_cookies()
page.set_cookie("wp_session", tok["root"])
page.goto(base + "/field.html")
time.sleep(1.5)
page.eval("localStorage.removeItem('x_me_offline'); localStorage.setItem(%r, %s); true" % (
KEY, json.dumps(json.dumps([{"kind": "wp", "key": "wp_keep", "opId": "wp:wp_keep:1", "tries": 0,
"owner": "root", "body": {"id": "wp_keep"}}]))))
page.clear_cookies()
page.set_cookie("wp_session", "not-a-valid-token")
load(page, base)
loc = page.eval("location.pathname + location.search")
chk("online with a dead session: redirect to login.html?next=",
loc.startswith("/login.html") and "next=" in loc, loc)
q = json.loads(page.eval("localStorage.getItem(%r) || '[]'" % KEY))
chk("the queued change is still there", [o.get("key") for o in q] == ["wp_keep"], json.dumps(q)[:200])
finally:
page.close()
browser.close()
finally:
if server is not None:
server.kill()
try:
server.wait(timeout=10)
except subprocess.TimeoutExpired:
pass
try:
from server.db import engine
engine.dispose()
except Exception:
pass
print("\n%d passed, %d failed" % (len(_PASS), len(_FAIL)))
return 1 if _FAIL else 0
if __name__ == "__main__":
sys.exit(main())

236
tests/session_idle_check.py Normal file
View File

@@ -0,0 +1,236 @@
#!/usr/bin/env python3
"""Does the sliding session with a hard ceiling work? - D18 / T13.1 (checked in T13.6).
T13.1 shipped with its done-when list unchecked and no test of its own; the old
session tests had no idea about idle time. This is that test.
1. Continuous activity keeps a session alive past 30 minutes, but never past the
absolute ceiling (AUTH_SESSION_HOURS) measured from the ORIGINAL sign-in.
2. A session idle for more than 30 minutes is refused on its next request.
3. The cookie is not rewritten on every request (the refresh is throttled).
4. A token from before this shipped (no login_at) still gets a hard ceiling,
from its iat.
5. A refresh carries identity and login_at forward unchanged.
6. The two variables are documented, and both defaults are flagged as proposed.
Self-contained: throwaway SQLite, in-process test client, no network, no browser.
Exit 0 all passed, 1 a failure.
"""
import os
import re
import sys
import tempfile
from datetime import datetime, timedelta, timezone
ROOT = os.path.dirname(os.path.dirname(os.path.abspath(__file__)))
sys.path.insert(0, ROOT)
os.environ["DATABASE_URL"] = "sqlite:///" + os.path.join(tempfile.mkdtemp(prefix="wp-idle-"), "t.db").replace("\\", "/")
os.environ["AUTH_SECRET_KEY"] = "session-idle-check-secret-not-for-production"
os.environ.pop("AUTH_IDLE_MINUTES", None)
os.environ.pop("AUTH_SESSION_HOURS", None)
import jwt # noqa: E402
from fastapi.testclient import TestClient # noqa: E402
from server import app as appmod, auth, models # noqa: E402
from server.db import Base, SessionLocal, engine # noqa: E402
if not hasattr(auth, "maybe_refresh_token"):
print("SKIP: T13.1 (sliding session) is not on this branch. It waits for T14.1.")
sys.exit(0)
fails = 0
def chk(label, ok, extra=""):
global fails
print(("PASS " if ok else "FAIL ") + label + (f" [{extra}]" if (extra and not ok) else ""))
if not ok:
fails += 1
Base.metadata.create_all(bind=engine)
with SessionLocal() as db:
db.add(models.User(id="u1", username="alice", email="a@x.test", role=auth.ROLE_ADMIN))
db.commit()
USER = db.get(models.User, "u1")
db.expunge(USER)
# A route the auth gate covers. NOT /api/auth/me: everything under /api/auth/ is
# exempt from the gate (auth._EXEMPT_PREFIXES), so no session refresh happens there
# and a probe on it would pass without testing anything.
PROBE = "/api/projects"
IDLE = auth.IDLE_MINUTES
CEIL_H = auth.SESSION_HOURS
NOW = datetime.now(timezone.utc)
def mint(login_at=None, iat=None, exp=None, **extra):
"""A token with exactly the claims given (None = leave the claim out)."""
p = {"sub": "u1", "username": "alice", "role": auth.ROLE_ADMIN, "ver": 0}
if iat is not None:
p["iat"] = iat
if login_at is not None:
p["login_at"] = login_at.timestamp()
if exp is not None:
p["exp"] = exp
p.update(extra)
return jwt.encode(p, auth.SECRET_KEY, algorithm=auth.JWT_ALG)
def raw_claims(tok):
return jwt.decode(tok, auth.SECRET_KEY, algorithms=[auth.JWT_ALG], options={"verify_exp": False, "verify_iat": False, "verify_nbf": False})
class FakeClock(datetime):
_now = NOW
@classmethod
def now(cls, tz=None):
return cls._now
real_dt = auth.datetime
def at(t):
FakeClock._now = t
auth.datetime = FakeClock
def real():
auth.datetime = real_dt
print("defaults: idle %d min, ceiling %d h" % (IDLE, CEIL_H))
# ── 1. continuous activity, then the ceiling ──────────────────────────────────
print("\n1. continuous activity stays alive past 30 minutes, but is cut off at the ceiling")
at(NOW)
tok = auth.create_token(USER)
c = raw_claims(tok)
chk("a new token carries login_at", "login_at" in c)
chk("a new token expires one idle window out, not at the ceiling",
abs(c["exp"] - (NOW + timedelta(minutes=IDLE)).timestamp()) < 2)
login_at = NOW
cur = c
alive_past_30 = False
rewrites = 0
last_exp = cur["exp"]
t = NOW
STEP = timedelta(minutes=1) # an active user: a request every minute
end = NOW + timedelta(hours=CEIL_H + 2)
cut_off_at = None
while t < end:
t += STEP
at(t)
if cur["exp"] <= t.timestamp(): # the session already lapsed
cut_off_at = t
break
if t - NOW > timedelta(minutes=IDLE):
alive_past_30 = True
new = auth.maybe_refresh_token(cur)
if new:
rewrites += 1
cur = raw_claims(new)
last_exp = cur["exp"]
real()
chk("with a request every minute the session is alive well past 30 minutes", alive_past_30)
chk("it is never extended past the ceiling (last exp <= login + %d h)" % CEIL_H,
last_exp <= (login_at + timedelta(hours=CEIL_H)).timestamp() + 1,
str(datetime.fromtimestamp(last_exp, timezone.utc)))
chk("it IS cut off once the ceiling is reached, however active the session",
cut_off_at is not None and cut_off_at <= login_at + timedelta(hours=CEIL_H) + timedelta(minutes=IDLE) + STEP,
str(cut_off_at))
chk("the claims after the last refresh still carry the ORIGINAL login_at",
abs(cur["login_at"] - login_at.timestamp()) < 1)
steps = int((CEIL_H * 60) / 1)
chk("the cookie was rewritten far less often than once per request (%d rewrites for ~%d requests)" % (rewrites, steps),
0 < rewrites < steps / 5, f"{rewrites}/{steps}")
# ── 2. idle ───────────────────────────────────────────────────────────────────
print("\n2. a session idle for more than %d minutes is refused on its next request" % IDLE)
client = TestClient(appmod.app)
idle_tok = mint(login_at=NOW - timedelta(minutes=IDLE + 1), iat=NOW - timedelta(minutes=IDLE + 1),
exp=NOW - timedelta(minutes=1))
client.cookies.set(auth.COOKIE_NAME, idle_tok)
r = client.get(PROBE)
chk("an idle session gets 401", r.status_code == 401, r.status_code)
live_tok = mint(login_at=NOW - timedelta(minutes=5), iat=NOW - timedelta(minutes=5),
exp=NOW + timedelta(minutes=IDLE - 5))
client.cookies.clear()
client.cookies.set(auth.COOKIE_NAME, live_tok)
chk("a session inside the window gets 200", client.get(PROBE).status_code == 200)
# ── 3. throttle ───────────────────────────────────────────────────────────────
print("\n3. the cookie is not rewritten on every request")
fresh = auth.create_token(USER)
client.cookies.clear()
client.cookies.set(auth.COOKIE_NAME, fresh)
cookies_set = 0
for _ in range(20):
r = client.get(PROBE)
if "set-cookie" in r.headers and auth.COOKIE_NAME in r.headers["set-cookie"]:
cookies_set += 1
chk("20 rapid requests on a fresh session: 0 cookie rewrites", cookies_set == 0, cookies_set)
aged = mint(login_at=NOW - timedelta(minutes=20), iat=NOW - timedelta(minutes=20),
exp=NOW + timedelta(minutes=2)) # 2 minutes of idle window left
client.cookies.clear()
client.cookies.set(auth.COOKIE_NAME, aged)
r = client.get(PROBE)
chk("an aged session gets a refreshed cookie on its next request",
r.status_code == 200 and auth.COOKIE_NAME in r.headers.get("set-cookie", ""))
m = re.search(auth.COOKIE_NAME + r"=([^;]+)", r.headers.get("set-cookie", ""))
newc = raw_claims(m.group(1)) if m else {}
chk("the refreshed cookie keeps the original login_at",
abs(newc.get("login_at", 0) - (NOW - timedelta(minutes=20)).timestamp()) < 1)
client.cookies.clear()
client.cookies.set(auth.COOKIE_NAME, m.group(1) if m else "")
r2 = client.get(PROBE)
chk("...and the very next request does not rewrite it again", auth.COOKIE_NAME not in r2.headers.get("set-cookie", ""))
# ── 4. tokens from before D18 ─────────────────────────────────────────────────
print("\n4. a token with no login_at still gets a hard ceiling (via iat)")
old = {"sub": "u1", "username": "alice", "role": auth.ROLE_ADMIN, "ver": 0,
"iat": (NOW - timedelta(hours=CEIL_H + 1)).timestamp(),
"exp": (NOW + timedelta(hours=3)).timestamp()} # the old flat 12 h exp, still valid
chk("past its iat-based ceiling: no refresh, the old exp is not extended", auth.maybe_refresh_token(old) is None)
young = {"sub": "u1", "username": "alice", "role": auth.ROLE_ADMIN, "ver": 0,
"iat": (NOW - timedelta(hours=1)).timestamp(),
"exp": (NOW + timedelta(minutes=2)).timestamp()}
ref = auth.maybe_refresh_token(young)
rc = raw_claims(ref) if ref else {}
chk("inside the ceiling: refreshed, and login_at is taken from iat (not 'now')",
bool(ref) and abs(rc["login_at"] - young["iat"]) < 1)
chk("the refreshed token cannot outlive iat + ceiling",
bool(ref) and rc["exp"] <= young["iat"] + CEIL_H * 3600 + 1)
nothing = {"sub": "u1", "username": "alice", "role": auth.ROLE_ADMIN, "ver": 0,
"exp": (NOW + timedelta(minutes=2)).timestamp()}
chk("no login_at and no iat: no refresh, and no crash", auth.maybe_refresh_token(nothing) is None)
junk = dict(young, login_at="not-a-number")
chk("an unparseable login_at falls back to iat, no crash", auth.maybe_refresh_token(junk) is not None)
# ── 5. identity carried forward ───────────────────────────────────────────────
print("\n5. a refresh carries the identity forward unchanged")
src = {"sub": "u9", "username": "zed", "role": "project_admin", "ver": 7,
"iat": (NOW - timedelta(minutes=20)).timestamp(),
"login_at": (NOW - timedelta(minutes=20)).timestamp(),
"exp": (NOW + timedelta(minutes=2)).timestamp()}
rc = raw_claims(auth.maybe_refresh_token(src))
chk("sub, username, role, ver are unchanged",
(rc["sub"], rc["username"], rc["role"], rc["ver"]) == ("u9", "zed", "project_admin", 7))
# ── 6. documentation ──────────────────────────────────────────────────────────
print("\n6. the two variables are documented, defaults flagged as proposed")
env = open(os.path.join(ROOT, "server", ".env.example"), encoding="utf-8").read()
dep = open(os.path.join(ROOT, "DEPLOYMENT.md"), encoding="utf-8").read()
for name in ("AUTH_IDLE_MINUTES", "AUTH_SESSION_HOURS"):
chk(f"{name} is in .env.example", name in env)
chk(f"{name} is in DEPLOYMENT.md", name in dep)
chk(".env.example calls the defaults proposed, not confirmed", re.search(r"propos", env, re.I) is not None)
chk("DEPLOYMENT.md calls the defaults proposed, not confirmed", re.search(r"propos", dep, re.I) is not None)
print()
print("ALL PASSED" if not fails else f"{fails} FAILED")
sys.exit(0 if not fails else 1)

View File

@@ -1,192 +0,0 @@
#!/usr/bin/env python3
"""Is there exactly one analytics implementation, reported from admin? — D5, T7.10.
Usage analytics existed twice (creator + wizard), five of the nine colliding
globals `creator-frame.md` counted, and the wizard's copy had no caller. One
core survives in wp-usage.js; the pages keep thin track() wrappers; the report
and its downloads live on the admin console behind the same role gate as the
rest of that page. Keys are unchanged, so pre-move data still reads.
Self-contained: throwaway SQLite, its own uvicorn, headless Edge or Chrome.
Exit 0 all passed, 1 a failure, 2 could not run.
"""
import json
import os
import re
import sys
import tempfile
import time
sys.path.insert(0, os.path.dirname(os.path.dirname(os.path.abspath(__file__))))
sys.path.insert(0, os.path.dirname(os.path.abspath(__file__)))
import cdp # noqa: E402
from browser_check import seed, start_server, chk, _PASS, _FAIL # noqa: E402
from sections_check import set_sop # noqa: E402
from stepper_check import dismiss_dialogs # noqa: E402
ROOT = os.path.dirname(os.path.dirname(os.path.abspath(__file__)))
HTML = os.path.join(ROOT, "html")
def ascii_(v, n=280):
return re.sub(r"\s+", " ", str(v)).encode("ascii", "replace").decode()[:n]
def settle(seconds=0.6):
time.sleep(seconds)
def strip_js(src):
src = re.sub(r"/\*.*?\*/", "", src, flags=re.S)
return "\n".join(re.sub(r"(?<![:'\"])//.*$", "", ln) for ln in src.split("\n"))
def main():
exe = cdp.find_browser()
if not exe:
print("no headless-capable browser found; set WP_BROWSER.")
return 2
# ── 1. the grep half: one implementation, no leftover controls ───────────
print("\n1. grep: one implementation, nothing unreferenced")
files = {}
for name in os.listdir(HTML):
if name.endswith((".js", ".html")):
files[name] = strip_js(open(os.path.join(HTML, name), encoding="utf-8").read())
core_defs = [n for n, src in files.items() if "window.WPUsage" in src]
chk("the WPUsage core is defined in wp-usage.js and only there",
core_defs == ["wp-usage.js"], ascii_(core_defs))
chk("the pages record THROUGH it - no page touches the storage keys directly",
all("wp_iwp_analytics_v1" not in src and "wp_suite_analytics_v1" not in src
for n, src in files.items()
if n not in ("wp-usage.js",) and n.endswith(".js")))
leftovers = {n: re.findall(r"analyticsLoad|analyticsSave|downloadAnalytics|showAnalytics"
r"|ANALYTICS_KEY|USAGE_KEY|usageLoad|downloadUsage", src)
for n, src in files.items() if n != "wp-usage.js"}
leftovers = {n: v for n, v in leftovers.items() if v}
chk("none of the five colliding globals survives anywhere; grep confirms",
not leftovers, ascii_(leftovers))
chk("no analytics control remains on the creator or the wizard; grep confirms",
"Usage data" not in files["wp-creation-index.html"]
and "showAnalytics" not in files["work-package-suite.html"])
chk("both storage keys survive, verbatim, in the core (data continuity)",
"wp_iwp_analytics_v1" in files["wp-usage.js"]
and "wp_suite_analytics_v1" in files["wp-usage.js"])
tmpdir = tempfile.mkdtemp(prefix="wpsuite-usage-")
db_path = os.path.join(tmpdir, "check.db")
server = None
browser = None
try:
tok = seed(db_path)
set_sop(db_path, {})
port = cdp.free_port()
base = "http://127.0.0.1:%d" % port
server = start_server(port, db_path)
browser = cdp.Browser(exe)
page = browser.page()
page.clear_cookies()
page.set_cookie("wp_session", tok["root"])
page.viewport(1440, 900)
# ── 2. recording still works from both tools ─────────────────────────
print("\n2. the tools still record")
page.goto(base + "/wp-creation-index.html?project=projA")
dismiss_dialogs(page)
settle(2.0)
# Plant a LEGACY-format event under the pre-move key: the done-when is
# that data recorded before this task is still readable after it.
page.eval("""(() => {
const d = JSON.parse(localStorage.getItem('wp_iwp_analytics_v1')) || {events: []};
d.events.unshift({ts: '2026-07-01T10:00:00Z', session: 's_legacy',
event: 'legacy_probe_event', detail: null});
localStorage.setItem('wp_iwp_analytics_v1', JSON.stringify(d));
})()""")
n0 = page.eval("WPUsage.load(WPUsage.KEYS.creator).events.length")
page.eval("track('probe_event')")
chk("the creator's track() still lands events under the old key",
page.eval("WPUsage.load(WPUsage.KEYS.creator).events.length") == n0 + 1)
page.goto(base + "/work-package-suite.html?tab=sop")
dismiss_dialogs(page)
settle(2.0)
n0 = page.eval("WPUsage.load(WPUsage.KEYS.wizard).events.length")
page.eval("track('probe_event_wizard')")
chk("the wizard's track() still lands events under its old key",
page.eval("WPUsage.load(WPUsage.KEYS.wizard).events.length") == n0 + 1)
wiz_errors = [e for e in page.js_errors() if "beforeunload" not in e]
chk("...and the wizard page throws no errors without its old globals "
"(the blocked-beforeunload console line is BL-020, filtered not hidden)",
not wiz_errors, ascii_(wiz_errors[:2]))
# ── 3. the report, on admin, behind the admin gate ───────────────────
print("\n3. the admin report")
page.goto(base + "/admin.html")
dismiss_dialogs(page)
settle(2.0)
page.eval("loadUsage()")
settle(0.5)
report = page.eval("(document.getElementById('usage-admin')||{textContent:''}).textContent")
chk("usage data is reachable from admin.html, both tools reported",
"Work package creator" in report and "SOP wizard" in report, ascii_(report, 160))
chk("the pre-move legacy event is readable in the report",
"legacy_probe_event" in report)
chk("this session's fresh events are in it too",
"probe_event" in report and "probe_event_wizard" in report)
chk("each tool offers its download from the report",
page.eval("[...document.querySelectorAll('#usage-admin button')].length") >= 2)
page.viewport(390, 844, mobile=True)
settle(0.6)
fits = page.eval("""(() => {
const box = document.getElementById('usage-admin');
return box && box.scrollWidth <= box.clientWidth + 2
&& document.documentElement.scrollWidth <= 392;
})()""")
chk("the report is usable at 390px - no sideways scrolling", bool(fits))
page.viewport(1440, 900)
settle(0.4)
# The same role gate as the rest of the console: a non-admin sees the
# denied card and no cards, this one included.
page.clear_cookies()
page.set_cookie("wp_session", tok["pat"])
page.goto(base + "/admin.html")
dismiss_dialogs(page)
settle(2.0)
chk("a non-administrator gets the denied notice, not the usage report",
page.eval("""(() => {
const denied = document.getElementById('admin-denied');
const usage = document.getElementById('usage-admin');
const visible = el => !!el && el.offsetParent !== null;
return visible(denied) && !visible(usage);
})()"""))
js_errors = [e for e in page.js_errors() if "beforeunload" not in e]
chk("no JavaScript errors anywhere in this run", not js_errors,
ascii_(js_errors[:2]))
finally:
if browser is not None:
try:
browser.close()
except Exception:
pass
if server is not None:
try:
server.terminate()
except Exception:
pass
print("\n" + "-" * 54)
print("%d/%d checks passed." % (len(_PASS), len(_PASS) + len(_FAIL)))
for f in _FAIL:
print(" - " + f)
return 1 if _FAIL else 0
if __name__ == "__main__":
sys.exit(main())