fix(docx-mcp): seed revision ids from side parts#216
Open
stevenobiajulu wants to merge 2 commits into
Open
Conversation
|
The latest updates on your projects. Learn more about Vercel for GitHub.
|
Codecov Report✅ All modified and coverable lines are covered by tests. 📢 Thoughts on this report? Let us know! |
8 tasks
Refines the side-part revision-id seed (introduced in 8417d7b for #171) to address peer-review feedback before merge. Why: 1. **Over-broad scan.** The previous `inferStartingRevisionIdState` collected every `w:id`/`id` attribute in the DOM. Non-revision IDs (e.g. `<w:comment w:id>`, `<w:footnote w:id>`, `<w:bookmarkStart w:id>`) share the attribute name but live in a separate ID space, so they were spuriously inflating the starting counter. Limit the scan to known revision-bearing WordprocessingML elements (`w:ins`, `w:del`, `w:moveFrom/To`, `w:*PrChange`, `w:cellIns/Del/Merge`, `w:customXml*RangeStart/End`). 2. **Incomplete part coverage.** Revision IDs are package-wide. The prior fixed list missed headers, footers, and the glossary part — pre-existing redlines in any of those would still collide with AI-emitted IDs. Replace the fixed list with `DocxZip.load` + dynamic enumeration of `word/header*.xml` and `word/footer*.xml`, plus a static list of `comments.xml`, `footnotes.xml`, `endnotes.xml`, `glossary/document.xml`. Drop `commentsExtended.xml` and `people.xml` from the seed list — they carry `w15:paraId` / `w15:author` identifiers, not revision `w:id` values, so under the narrowed scan they would be no-ops. 3. **First-call concurrency race.** With the scan now async, two concurrent first callers could each run the seed scan and the slower one could overwrite a counter that the faster caller had already advanced via emitted revisions, causing real collisions. Add a module-level `WeakMap<DocxSession, Promise<RevisionIdState>>` single-flight guard that mirrors the existing `baselinePromises` pattern in `SessionManager`. 4. **Parse robustness.** `parseXml` throws on malformed input. A corrupt optional side part (e.g. truncated `comments.xml`) would crash every tracked edit on the session. Wrap each per-part parse in `try/catch` and skip on failure — collision risk on an unrelated corrupt part is better than bricking the session. New regression tests: - non-revision `<w:comment w:id="500">` alone does not seed counter - header with `<w:ins w:id="900">` seeds `nextId` to 901 - footer with `<w:ins w:id="1234">` seeds `nextId` to 1235 - malformed `comments.xml` does not block seeding from `document.xml` - 3-way concurrent first callers resolve to a single seeded state Verification: - `npm -w @usejunior/docx-mcp run test:run -- src/session/manager.test.ts` — 36 passed - `npm -w @usejunior/docx-mcp run lint` - `npm -w @usejunior/docx-mcp run build` - `npm run lint:workspaces` — clean (one pre-existing warning in edit.test.ts) - `npm run test:run` — all workspaces green (1239 + 742 + 116 + 3 passed) Ref: #171 Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
Closes #171.
getRevisionContextForSessionnow seeds the sessionRevisionIdStatefromdocument.xmlplus available side parts that can carry package-widew:idrevision attributes. This prevents AI-emitted revision IDs from colliding with pre-existing revisions incomments.xml,footnotes.xml,endnotes.xml,commentsExtended.xml, orpeople.xml.Implementation
getRevisionContextForSessionasync so it can inspect optional side-part XML from the DOCX package.inferStartingRevisionIdState(...docs)helper and removes the documented limitation.document.xmlhasw:id=42butcomments.xmlhasw:id=500, and verifies the next allocated ID starts at501.Verification
npm -w @usejunior/docx-mcp run test:run -- src/session/manager.test.ts— 31 passednpm -w @usejunior/docx-mcp run lintnpm -w @usejunior/docx-mcp run buildgit diff --check