chore(deps): upgrade moonbitlang/async 0.19.1 → 0.19.4#32
Merged
Conversation
Contributor
Author
|
CI on this isolated bump reproduces the windows-latest regression — same failure as the earlier combined attempt: Marked DEP-1
This PR stays red intentionally as the tracking vehicle. |
![chatgpt-codex-connector[bot]](https://avatars.githubusercontent.com/in/1144995?s=60&v=4){kind=small}
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: fcfea5e9ff
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
Bump every workspace manifest (root, tests, tests/isatty, tests/resize, tests/open, examples) so the workspace resolves to a single async version. No source or public API change; .mbti files are unchanged. Stacked on the output write lock branch; the lock only needed Semaphore (already in 0.19.1), so the dependency bump is isolated here for review. A prior combined attempt timed out the windows pty Tty::open test on 0.19.4; this isolated bump lets CI judge whether that was a real regression or a flake. See docs/plans/2026-06-15-async-0.19.4-upgrade.md. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
CI on the isolated bump reproduces the windows-latest Tty::open timeout, confirming a real 0.19.4 regression rather than a flake. Mark DEP-1 blocked pending investigation of the async event-loop change. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
tonyfettes
force-pushed
the
chore/async-0.19.4
branch
from
fcfea5e to
d3fe516
Compare
async 0.19.2 changed the Windows file-open worker to detect file kind via
GetFileInformationByHandle (for the new FileIdentity / fs.watch feature)
instead of GetFileType. That disk-file API fails on console (character)
devices, so @async/fs.open("CONIN$"/"CONOUT$") now errors and the whole
Windows Tty::open path broke under 0.19.4 (the tests/open pty test failed
on windows-latest).
Open CONIN$/CONOUT$ directly with CreateFileW (synchronous, non-overlapped)
and wrap them in @async/raw_fd.RawFdStream, which classifies them by
GetFileType (CharDevice) and never registers overlapped IO — exactly the
working 0.19.1 console read/write path.
Unify both platforms on RawFdStream: drop the hand-rolled ControllingTerminal
struct and its six manual trait impls, and move the Fd/Reader/Writer impls
for RawFdStream into the shared io.mbt/isatty.mbt next to the File/Pipe/stdio
impls. The unix path is exercised by the pty-backed Tty::open test.
Co-Authored-By: Claude Opus 4.8 (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.
Follow-up to #31 (now merged). Base is
main.What
Bumps
moonbitlang/async0.19.1→0.19.4across every workspace manifest (root,tests,tests/isatty,tests/resize,tests/open,examples) so the workspace resolves to a single async version. Dependency version only — no source change, and.mbtifiles are unchanged.Why separate from #31
The output write lock in #31 only needs
Semaphore, which already exists in0.19.1with an identical API, so #31 stayed on0.19.1. Isolating the dependency bump here keeps the lock feature and the upgrade independently reviewable and revertable.0.19.4reproducibly fails windows-latest: thetests/openTty::openpty test times out (TimeoutError, 176/177 passed). ubuntu-latest and macos-latest pass (and the full suite passes locally on macOS native: root 165 + pty module 133). This was confirmed on the isolated bump, so it is a real0.19.4regression on the Windows CONOUT$ write path, not a flake.DEP-1is markedblocked. Suspect: the async0.19.1..0.19.4event-loop diff, in particular the changedIoHandle::detach_from_event_loopsignature/behavior, against the CONOUT$ write+close path.Do not merge until Windows is resolved. Options: bisect 0.19.2/0.19.3 to pin the regressing release for an upstream report, investigate/work around the event-loop change, or hold until fixed upstream.
Plan:
docs/plans/2026-06-15-async-0.19.4-upgrade.md(DEP-1).🤖 Generated with Claude Code