Skip to content

Add .gitattributes so Windows checkouts get LF endings - #550

Merged
Andrews54757 merged 1 commit into
Andrews54757:mainfrom
Nawid3333:pr/gitattributes
Sep 9, 2026
Merged

Add .gitattributes so Windows checkouts get LF endings#550
Andrews54757 merged 1 commit into
Andrews54757:mainfrom
Nawid3333:pr/gitattributes

Conversation

@Nawid3333

Copy link
Copy Markdown
Contributor

With core.autocrlf=true (the Windows default), a fresh clone gets CRLF in the working tree, and eslint-config-google's linebreak-style rule then reports an error on essentially every line — 2352 of them on this repo. That makes linting unusable on Windows, and it's invisible on Linux so it never showed up in your CI.

The repository content is already LF throughout, so this file changes checkout behaviour only — it produces no content diff:

  • * text=auto eol=lf checks out all text files with LF.
  • Binary assets are pinned so they are never converted: png, wasm, ort, zip, xpi.

Verified on Windows 11: a fresh clone with this file lints at 0 errors, where the same clone without it reports 2352 linebreak-style errors.

Pairs with #548 — together they make a Windows checkout buildable and lintable.

With core.autocrlf=true (the Windows default), git hands the working tree
CRLF, and eslint-config-google linebreak-style then reports an error on
every single line - 2352 of them across the repo, i.e. lint is unusable on
Windows. The repository content is already LF throughout, so this file
changes checkout behaviour only and produces no content diff.

* text=auto eol=lf checks out all text files with LF.
Binary assets are pinned so they are never converted: png, wasm, ort,
zip, xpi.

Verified on Windows 11: a fresh clone with this file lints at 0 errors,
where the same clone without it reports 2352.
Nawid3333 added a commit to Nawid3333/FastStream that referenced this pull request Sep 7, 2026
…point

Phase 10 is no longer "not started": Andrews54757#548, Andrews54757#549, Andrews54757#550 and Andrews54757#551 are open
upstream, with follow-up comments on Andrews54757#547 and Andrews54757#546. The unlisted AMO
submission pipeline ran end to end on 2026-09-07 (local sign plus the
publish-amo workflow). The baseline re-capture is also recorded as done -
the workspace-level baseline now holds the modernised fork at cd728ab
(852 files, spot-verified), with the original V1.3.77 upstream build
archived separately.
@Andrews54757
Andrews54757 merged commit a833c9a into Andrews54757:main Sep 9, 2026
@Nawid3333
Nawid3333 deleted the pr/gitattributes branch September 9, 2026 11:38
Nawid3333 added a commit to Nawid3333/FastStream that referenced this pull request Sep 9, 2026
…ributes (Andrews54757#550)

Both commits are this fork's own upstream PRs coming back home.

Andrews54757#550 (.gitattributes) is a content no-op: the fork's file and Andrew's are
byte-identical, since he merged the PR as submitted.

Andrews54757#548 lands only the half the fork did not already have. The fork fixed the
WIN32 branch of cleanGlobPath in ab0719d (that was the actual Windows
build break); upstream's merge also renamed the parameter in the
non-WIN32 branch, which is what arrives here. No behaviour change on
Windows, and it removes the last divergence from upstream in miniglob.mjs.

The fork's removal of the file's blanket /* eslint-disable */ survives the
merge -- checked, not assumed.

Verified after merging: lint 0, tsc 0, 90 unit tests, all four builds.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Nawid3333 added a commit to Nawid3333/FastStream that referenced this pull request Sep 9, 2026
…mes in the upstream candidates queue

A1 (Windows build) and A2 (.gitattributes) merged 2026-09-09 -- mark them
`merged` instead of `queued`, since both landed and were pulled back into
dev/mv3-modernization by the same day's upstream sync.

Add A2b for Andrews54757#549 (drop the unused contextualIdentities permission), which
Andrew closed: it is provisioned ahead of an unshipped Firefox API
(Bugzilla 1917842), not dead weight. Recorded so nobody re-proposes it on
the strength of the same grep that looked convincing the first time.

Add `rejected` to the status vocabulary.

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
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.

2 participants