Repository navigation
fix: tell a release build about a newer release - #585
Merged
Merged
Conversation
ConfigReference.ToMarkdown uses raw strings, which take the line endings of the source file. On a Windows checkout with core.autocrlf those are CRLF, so ConfigReferenceTests failed there even right after regenerating the page. CI runs its tests on Linux and macOS and never saw it. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The update check could not announce anything. It asked WowLegacyCore/HermesProxy, whose last release is from 2023, and it returned for any build with commits since the last tag. Release binaries are built before their tag exists, so that was all of them: v4.7.0 reports 4.7.0+10(95ed7b0). It now asks this repository, and skips builds that aren't from master, which is what shipped binaries are. The tag is compared as a version instead of by date. The notice is a log warning now. It used to be console text followed by a 10 s sleep, which nobody met while the check was dead, but releases land several times a day, so a working check would have stalled nearly every start. The check moved after Log.Configure so the warning reaches the log file. Fixes #583 Co-Authored-By: Claude Opus 5.5 <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
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Fixes #583
The update check could not announce anything. It asked
WowLegacyCore/HermesProxy, whose last release is from 2023, and it returned for any build with commits since the last tag. Release binaries are built before their tag exists, so that was all of them: v4.7.0 reports4.7.0+10(95ed7b0).Changes
Branch.master, checked in v4.7.0 and v4.6.113.Thread.Sleep. Nobody met that while the check was dead, but releases land several times a day, so a working check would have stalled nearly every start.Log.Configure, so the warning can reach the log file.Behaviour change: a release build now calls the GitHub API at each start. On a machine that can't reach it, that waits up to 5 s, the client's timeout.
--no-version-checkturns it off, and the Docker image already has it off.Also here
ConfigReference.ToMarkdownnow normalises line endings. Its raw strings take the source file's, so on a Windows checkout withcore.autocrlfConfigReferenceTests.Page_MatchesTheSettingsfailed on every run, even right after regenerating the page. That came with #582. CI runs its tests on Linux and macOS and didn't see it.Tested
Real Release binaries against the live GitHub API, built from a clone without the v4.7.0 tag so they are older than the latest release:
mastermasterHermesProxy v4.7.0 is out and this is 4.6.114. Get it from https://github.com/Xian55/HermesProxy/releases/latest, started in 2 sUpdateCheckTestscovers the version comparison.dotnet test: 4722 passed, 34 skipped, on a CRLF checkout. 54 of 55 full runs passed. The other failedDynamicSetUpdateField_ComputedIndex_AllocatesNothing, in code this PR doesn't touch: Allocation test failed once in a full run #584.🤖 Generated with Claude Code