docs(i18n): close historical dynamic census and preserve native regression evidence

This commit is contained in:
cesnimda
2026-11-02 23:03:19 +01:00
parent 80d3849125
commit c7e618a548
4 changed files with 11 additions and 0 deletions
+11
View File
@@ -0,0 +1,11 @@
# Current localisation continuation
2 November 2026. Historical 188 dynamic-call census finite key-domain review COMPLETE; full I18N-00 / European locale release acceptance remains INCOMPLETE. Original source-bound snapshots are retained under 00/. Machine register schema 2 binds inventory indices, source hashes and selected owner dependencies; 108 original plus 80 supplemental calls cover all 188 indices without overlap. Static review does not establish all translator sensitivity/context, every visible/backend string, grammar/plural quality, runtime coverage or native review.
Independent fixes now include descriptive attachment preview labels; session-change feedback outside private providers; safe inherited-name handling for completeness/timeline/pipeline/processing/interview/purpose labels; and original historical checklist-text fallback for unrecognised system keys. No user data, canonical Career Profile ownership, CV/letter revision or final design contract was rewritten. No dependencies or migrations added.
Latest implementation qualification: processing monitor 17 tests; checklist/workspace 37 tests; interview/document index 41 tests. Each selected execution passed with zero failures/pending, scoped lint and full typecheck. They are disjoint selected suites, not complete application acceptance. Earlier diagnostic failures and overlapping executions remain separately recorded in their reports. A fresh combined Next 16.4.1 static build after the final fixes passed, exit 0, including TypeScript and all three generated pages. No deployment, live API/provider/model or final integrated security test implied.
Sanitized native reports: evidence/processing-label-fallbacks.json.gz, evidence/checklist-unknown-system-keys.json.gz, evidence/interview-document-fallbacks.json.gz. Earlier scope reports and prior evidence remain intact. Generated AST graph stays untracked; optional SQL/PDF extractor gaps do not change application acceptance.
Next independent work: complete direct-access and literal review, every-string context/sensitivity triage, European catalogue/territory/source qualification and approved glossary/translation QA preparation. Full registry/persistence/plural engine architecture still needs prerequisite acceptance and the documented decisions; no proposed locale is published. Native translators/market priorities, Montenegrin tag identity, final design, provider rights, staging/paid scans, hosted CI and production/recovery decisions remain distinct gates. Final integrated security testing follows the relevant application changes.