chore(deps): bump astro from 7.2.7 to 7.2.9 in /web - #668
Merged
Conversation
Bumps [astro](https://github.com/withastro/astro/tree/HEAD/packages/astro) from 7.2.7 to 7.2.9. - [Release notes](https://github.com/withastro/astro/releases) - [Changelog](https://github.com/withastro/astro/blob/main/packages/astro/CHANGELOG.md) - [Commits](https://github.com/withastro/astro/commits/astro@7.2.9/packages/astro) --- updated-dependencies: - dependency-name: astro dependency-version: 7.2.9 dependency-type: direct:production update-type: version-update:semver-patch ... Signed-off-by: dependabot[bot] <support@github.com>
dependabot
Bot
added
dependencies
|
Review the following changes in direct dependencies. Learn more about Socket for GitHub.
|
rainmanjam
added a commit
that referenced
this pull request
Sep 4, 2026
…exec failure (#696) * fix(ffmpeg): an interrupted header read arrived as an unrecognisable exec failure Found as an intermittent macos-latest failure on #668, a pull request that changes nothing but a JavaScript dependency: rawes_test.go:331: error = signal: killed, want the context error in the chain so the caller can recognise an interruption ProbeFile runs ffprobe for the header and then, when there is no duration, ffmpeg to count one. The COUNT branch folds ctx.Err() into its error and explains at length why it has to: killing a child through CommandContext yields a plain *exec.ExitError saying "signal: killed" and carrying NO context error to match on -- measured there, in that comment. The HEADER READ did not, while its own comment claimed the property: // %w on both branches. The caller has to be able to tell an interrupted // probe from a file ffprobe disliked -- one of those must not delete the // operator's upload -- and flattening the error with %s took that // distinction away. %w on the run error cannot deliver that, for exactly the reason the count branch recorded fifty lines below. So the requirement was right, the mechanism was wrong, and the comment asserted the outcome -- which is worse than saying nothing, because it is what the next reader relies on. MILDER THAN IT LOOKS, and worth saying so rather than overselling the fix: a bare *exec.ExitError matches none of Refused's sentinels, so the upload is not deleted. What is lost is the caller's ability to tell why the probe ended, and a log reader's. WHY IT FLAKES. TestACountThatWasCutShortIsNotAVerdictAboutTheFile cancels at 200ms meaning to land inside the count, and on a loaded runner ffprobe has not finished by then -- so the cancellation lands in the header read instead and the assertion runs against the branch that had nothing to find. THE NEW TEST'S FIRST VERSION WAS VACUOUS and is recorded here rather than quietly replaced. It used a context cancelled up front, and exec.Cmd.Start returns ctx.Err() DIRECTLY when the context is already done -- so context.Canceled was in the chain no matter what this code did. A mutation that wrapped the run error instead left it green. It now uses a slow stand-in ffprobe cancelled mid-flight, which is the shape that actually produces a bare "signal: killed", and both mutations turn it red. Claude-Session: https://claude.ai/code/session_01A8N3W5ct9SZtHK9sCDD9cL * test: skip the POSIX-only stand-in on Windows Missed on the first push and caught by the Windows leg. The new test writes a `#!/bin/sh` script to stand in for ffprobe, which Windows cannot exec -- so instead of being killed mid-flight the stand-in failed to start, the count branch was reached with no context error, and the assertion failed against a situation the test was not written about. Its sibling already carries exactly this skip and says why: "the branch it pins is not platform-specific; the stand-in is." I added the test beside it and did not read that far. Claude-Session: https://claude.ai/code/session_01A8N3W5ct9SZtHK9sCDD9cL * test: raise the skip budget by hand for the Windows stand-in The skip census refused the previous commit: 101 sites against a committed 100, internal/ffmpeg 25 against 24. Correctly -- a new skip is a new test that can decline to run and still print ok. This one is environmental and is the SAME skip its sibling already carries for the same stand-in: the fixture is a #!/bin/sh script that sleeps until it is killed, chosen so the cancellation lands inside the header read rather than racing a real ffprobe's exit, and Windows will not execute it. The branch under test is not platform-specific; the stand-in is. Issue #190 already names this family and its real fix -- a small helper binary built by the test -- which would take this site back out along with the two it lists. RAISED BY HAND, WHICH IS THE POINT. `-update-skips` regenerates the counts and DROPS the raisedBy array -- 142 lines recording why each previous raise was allowed. Running it as the message suggests would have silently destroyed the audit trail that makes this budget reviewable at all, which is the opposite of what the guard is for. The counts here are edited directly and the entry is appended, so the record of every raise survives. That regeneration behaviour is worth its own issue rather than a workaround buried here. Claude-Session: https://claude.ai/code/session_01A8N3W5ct9SZtHK9sCDD9cL
|
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.
Bumps astro from 7.2.7 to 7.2.9.
Release notes
Sourced from astro's releases.
Changelog
Sourced from astro's changelog.
Commits
ad7a332[ci] release (#17845)eb87a23Escape set:text values in Astro JSX (#17847)b441180fix(astro): type .html imports outside of .astro files (#17846)32e8b44Guard against stripping /index.html when stripped pathname no longer matches ...c35448efix(build): strip prerender-only entry specifiers from SSR manifest (#17841)7cadf10[ci] release (#17826)ecb4082Update Sharp to 0.35.4 (#17837)8bf6f1achore: split v5/v6 changelogs (#17825)db7c53bchore(deps): replace "find-process" with a smaller, lighter alternative (#17786)Dependabot will resolve any conflicts with this PR as long as you don't alter it yourself. You can also trigger a rebase manually by commenting
@dependabot rebase.Dependabot commands and options
You can trigger Dependabot actions by commenting on this PR:
@dependabot rebasewill rebase this PR@dependabot recreatewill recreate this PR, overwriting any edits that have been made to it@dependabot show <dependency name> ignore conditionswill show all of the ignore conditions of the specified dependency@dependabot ignore this major versionwill close this PR and stop Dependabot creating any more for this major version (unless you reopen the PR or upgrade to it yourself)@dependabot ignore this minor versionwill close this PR and stop Dependabot creating any more for this minor version (unless you reopen the PR or upgrade to it yourself)@dependabot ignore this dependencywill close this PR and stop Dependabot creating any more for this dependency (unless you reopen the PR or upgrade to it yourself)