Gitea Actions does not reliably honor `on.pull_request.paths:` filters —
the workflow silently fails to trigger even when changed files match the
filter. PR #13 touches .github/workflows/pr-iphone-deploy.yml (explicitly
listed in the paths filter) and AGENTS.md, but only ci.yml fired.
The repo history shows this was already known: 65d85b6 'remplacer paths
trigger par dorny/paths-filter' moved to job-level filtering, but later
commits reverted to the trigger-level paths: block. The proven-working
pattern already lives in ci.yml (dorny/paths-filter@v3 in a 'changes' job).
Fix (mirror ci.yml):
- Remove trigger-level `paths:` from on.pull_request.
- Add a 'changes' job (ubuntu-latest, dorny/paths-filter@v3, single 'ios'
filter covering tabatago-swift/** and the workflow itself).
- Gate build-deploy: needs: changes + if: needs.changes.outputs.ios == 'true'.
- Gate wait-approval: if: needs.build-deploy.result == 'success' so a
filtered-out deploy doesn't launch a merge poll against an undeployed PR.
Everything else unchanged: concurrency, squash-merge, 2h timeout, the
self-merge fix from the previous commit on this branch (sentinel marker
+ python3 parser). This PR will self-validate: it touches the workflow
file, so the ios filter matches, so the deploy runs.
The build-deploy job posts a 'Prêt à tester' comment whose body literally
contains 'Reply **LGTM** pour merger' and 'Reply **KO**' as instructions to
the reviewer. wait-approval re-fetches ALL comments every 30s and greps any
body for \bLGTM\b (checked before KO) — so on the first poll (~30s after
deploy) the bot matched its OWN comment and squash-merged automatically,
with no human review. Regressed in #6937a36 (PR #12) which tightened the
regex to catch 'Tested, LGTM!' but didn't notice the bot body matched.
Fix (sentinel marker + python parse, identity-agnostic so it's safe whether
PR_API_TOKEN is a bot or a personal account):
- build-deploy: embed <!-- tabatago:ready-to-test --> in the bot comment
(HTML comment, invisible in rendered markdown, present in raw body).
- wait-approval: replace the fragile grep-over-JSON with a python3 heredoc
using json.loads + the existing \bLGTM\b / \bKO\b regexes. Skip any
comment whose body contains the sentinel marker. Check KO BEFORE LGTM so
a PR with both signals blocks (matches the existing 'KO blocks' intent).
LGTM still does {"Do":"squash"}; KO still blocks; 2h timeout unchanged.
Verified locally with three scenarios:
bot-only -> PENDING (was LGTM before — the bug)
bot + reviewer -> LGTM (squash-merge still fires)
bot + KO -> KO (block still fires, even with bot LGTM in history)
AGENTS.md §5 step 9, §5 pitfalls, and §8 anti-patterns updated so the doc
describes the marker rule and warns against re-introducing the self-match.
