Skip to content

release 3.26.0#311

Open
aelttil wants to merge 17 commits into
mainfrom
dev
Open

release 3.26.0#311
aelttil wants to merge 17 commits into
mainfrom
dev

Conversation

@aelttil

@aelttil aelttil commented Jun 10, 2026

Copy link
Copy Markdown
Contributor

No description provided.

chrlesur and others added 17 commits May 14, 2026 13:36
…AJ changelog et memory bank

- Fix ContentSplitter : ne coupe plus au milieu des blocs de code
- Nouvelles methodes : _find_code_block_ranges, _is_inside_code_block, _find_safe_split_point
- Migration modele traduction : qwen3-2507:30b-a3b vers qwen3.6:35b (contexte 250k)
- Changelog : ajout 3 entrees manquantes (stockage IOPS, sauvegarde IaaS, DB sizing)
- Memory Bank : mise a jour complete (nouveaux services, etat couverture 97%)
…ranslation hashes

Migration complète des chemins d'images et nettoyage du système de traduction :

Images (208 fichiers migrés) :
- ./images/ → @site/docs/.../images/ (165 fichiers FR + i18n)
- ../images/ → @site/docs/.../images/ (43 fichiers FR + i18n)
- Fix cross-dir ../console/images/ dans iaas_vmware/quickstart.md
- Suppression de 11 dossiers images orphelins dans i18n/

Traductions & hash :
- Resync translation-meta.json (157/157 fichiers synchronisés)
- extract_changelog.py : ajout mise à jour automatique des hash
- Suppression de oldies/translate.js (script legacy)
- scripts/README.md : suppression section legacy, ajout extract_changelog

Fixes MDX :
- batch.md : échappement {batch_id} et {file_id} (FR + 4 langues)
- contact_partner.md EN : import manquant marketplaceContactEmail
- usinggpu.md IT : typo nvidasmi → nvidiasmi

Documentation :
- README.md : section traduction enrichie, règle images @site/docs
- changelog_produits.md : tronqué avant v4.0 (~2600 → ~420 lignes)
- memory-bank/documentationStatus.md : état mis à jour

Build : ✅ SUCCESS (toutes les langues)
Managed Kubernetes / Managed Core Kubernetes SLA is 99.95%
Changed storage pre-requisites for Managed Kubernetes / Managed Core Kubernetes
The file was committed before the .gitignore rule existed (or before
a previous git rm --cached). The rule on line 19 of .gitignore had no
effect on the already-tracked file. Removing it from the index keeps
the local copy intact but stops future modifications from being
committed.

NOTE: any API key that was previously committed to history must still
be considered compromised and rotated separately on the LLMaaS side.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
- README.md: collapse the duplicated Translation System section to a
  short pitch + one critical warning (never edit i18n/ manually) + a
  link to scripts/README.md (single source of truth).
- scripts/README.md: add a Python venv step in the Installation section
  to avoid polluting the system Python, and a new "Règles importantes"
  block grouping the warnings previously duplicated in the main README
  (i18n manual edits, absolute @site/docs image paths, .notranslation).

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
… SLA

- Intervalle minimum de réplication passé de 1 minute à 15 minutes
- Ajout d'un avertissement expliquant que le RPO est un objectif best effort
- Le RPO effectif dépend du taux de changement (change rate) des données
- Aucun SLA contractuel n'est associé au RPO
…es/it

- changelog.md: resync with FR source (debt from recent dev commits)
- iaas_opensource/concepts.md: translate RPO 15 min change from
  fix(openiaas) commit on this branch
- translation-meta.json: update SHA-256 hashes accordingly

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
…config

The README and .env.example previously duplicated the env vars list,
which had silently drifted from config.py defaults:
- TRANSLATION_MODEL was still qwen3:30b-a3b (obsolete, causes HTTP 404)
- CONCURRENT_TRANSLATIONS doc said 8 but real default is 4
- MAX_TOKENS_PER_BLOCK doc said 5000 but real default is 4000
- MODEL_TYPE / MAX_MODEL_CONTEXT_LENGTH / BUFFER_TOKENS were undocumented

scripts/translate_py/.env.example:
- Rewritten as self-documenting with inline comments (role, default,
  valid range) for every variable supported by config.py
- TRANSLATION_MODEL updated to qwen3.6:27b (current default)
- Added MODEL_TYPE, MAX_MODEL_CONTEXT_LENGTH, BUFFER_TOKENS
- Documented TRANSLATION_LOCAL_MODE and DOC_BASE_PATH

scripts/README.md:
- Removed the duplicated env vars block
- Points to .env.example as the single source of truth
- Adds a note explaining why duplication was removed

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
…o-15min

fix(openiaas): RPO réplication minimum 15 min + avertissement absence…
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants