ci: pin shared workflow to v1 #4

Merged

james.maes merged 4 commits from ci/pin-v1-develop into develop 2026-11-06 23:11:06 +00:00

Conversation 0 Commits 4 Files Changed 1

+1 -1

4 Commits

Author	SHA1	Message	Date
James MaesandClaude Opus 5.5	34a3666bbf	Merge develop: adopt the released Document Generation 1.0 changes ... develop (PR #5) already fixed the renderer BarcodeDeclaredCapabilityTest gaps, moved the family to QQQ 4.1.0-RC.22 and pinned the released Record Files 0.1.0-RC.1 and Scripting 1.0.0, which is what resolves the scripting bridge's "reviewed core isolation marker metadata" errors from run 6360. Take develop's implementation and docs wholesale so this branch only carries its own change: pinning the shared Maven workflow to v1. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>	2026-11-06 16:54:15 -06:00
James MaesandClaude Opus 5.5	6a3153456e	build: align the family on QQQ 4.1.0-RC.22 ... The published qbit-scripting 0.1.0-SNAPSHOT (and its tests jar) is built against QQQ Core 4.1.0-RC.22, whose Script entity carries the reviewed executionMode isolation marker. The scripting bridge pinned RC.21 through dependencyManagement, so ScriptingQBitProducer rejected the core script table ("Scripting requires the reviewed core isolation marker metadata") in every PublishedScriptContextAdapterTest. qbit-record-files and qbit-printing have also moved to RC.22, so move foundation and all bridges together rather than pinning older sibling snapshots. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>	2026-11-06 16:53:07 -06:00
James MaesandClaude Opus 5.5	207e3ce034	fix(renderer): make the renderer test suite pass on the shared CI workflow ... - BackgroundImageQualification expected the patched OpenHTML provider at the old fixed version 1.1.105-qqq-p1, but PR #1 publishes the vendored module at the family revision. The exact-GAV check now takes the expected version from the build's own ${revision} (surefire system property) and fails closed if absent; group, artifact, single-class-definition and companion checks are unchanged. - barHeight was advertised and accepted for every Okapi family but silently ignored by fixed-geometry encoders. It is now advertised/accepted only where it changes the symbol; the 16 fixed-geometry families and gs1-databar in STACKED mode refuse it with UNSUPPORTED_BARCODE_OPTION (catalog metadata carries notWithMode=STACKED). - UNSUPPORTED_BARCODE_OPTION now crosses the worker boundary as its closed code (foundation already models it), and PDF VALIDATE checks literal (template-free) barcode options of in-process encoders. Resolves the 3 known BarcodeDeclaredCapabilityTest failures and the PDF_PROVIDER_GAV_MISMATCH error; no tests skipped or weakened. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>	2026-11-06 11:37:12 -06:00
james.maesandClaude Opus 5.5	063b1e8136	ci: pin shared workflow to v1 ... Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>	2026-11-06 15:55:32 +00:00