Freegle / Iznik / 5487

70%

Build Type

push

circleci

Committed by web-flow

Commit Message

Retire per-group includearea/includepc subject-line settings (#230)

* refactor(subject): retire per-group includearea/includepc settings

Always construct post subjects as "KEYWORD: item (Area VaguePostcode)".
Removes the now-dead branching in the Go API constructLocationString,
the PHP v1 subject builder, and the showarea/showpc response plumbing.
Drops the two toggles from the ModTools group-settings page.

Existing groups.settings rows keep the JSON keys as harmless cruft; no
code reads them any more.

* fix(test): match VectorSearch 3-return signature in branches test

embedding_vectorsearch_branches_test.go was merged (PR #227) with the old
two-return signature, but VectorSearch had already been changed to return
(results, stats, err) in PR #221. This broke the Go test build on master
and any branch off it.

Discards the stats value with `_` in all seven call sites — none of the
branch tests inspect it.

* test(sentry): add missing Sentry error reporting test for OurUploadedImage

Adds the missing test for Sentry.captureMessage being called when
brokenImage is triggered with a connected target. Ensures error
reporting is properly tested.

Co-Authored-By: Claude Haiku 4.5 <noreply@anthropic.com>

* fix(auth): guard disableGoogleAutoselect window access for SSR/test teardown

Bare `window` identifier threw ReferenceError when the setTimeout-scheduled
retry fired after Vitest tore down the jsdom environment. Switch to
`typeof window === 'undefined'` early-return so the guard works with no
globals defined, and add a test that reproduces the teardown scenario.

Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>

* fix(spec): stable hoisted mock identity for Sentry.captureMessage spy

Three Sentry error-reporting assertions in OurUploadedImage's spec were
reading the captureMessage spy through either a dynamic
`await import('@sentry/browser')` or the top-level
`import * as Sentry` namespace. Under the mock resolver those
reads can surface a diffe... (continued)

Coverage Stats

8892 of 10271 branches covered (86.57%)

Branch coverage included in aggregate %.

1 of 1 new or added line in 1 file covered. (100.0%)

44 existing lines in 1 file now uncovered.

95399 of 143571 relevant lines covered (66.45%)

18.57 hits per line