Conversation
A build that fails while configuring writes nothing but the daemon notice to stdout and reports what went wrong on stderr. runGradle handed that result back to the caller, so the test failed on whatever it asserted about the coverage the build never produced, and the actual cause never reached the test report. That is what makes the current Windows failures of gradle-multi-module unreadable. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
gradle-daemon-jvm.properties made the fixture's own Gradle daemon require Java 21, which is stricter than
the Java 21 toolchain the projects compile against: the daemon's JVM is picked by the launcher before the
settings script runs, so the toolchain repository that foojay registers there does not exist yet and the
JDK cannot be provisioned either. As long as the runners ran Gradle on Java 21 the launcher's own JVM
satisfied it. On Java 25 it no longer does, and only Linux keeps its JDKs somewhere the launcher detects
(/usr/lib/jvm); on Windows they live in the tool cache, so every nested build there failed with
Unable to download toolchain matching the requirements ({languageVersion=21, ...}) from 'null', due to:
No defined toolchain download url for WINDOWS on x86_64 architecture.
The file is the only one of its kind in the repository and arrived as a stray in an unrelated commit; no
test says anything about the daemon's JVM. The projects keep their Java 21 toolchain, which resolves
normally once the build is running.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
DreierF
force-pushed
the
ts/windows-system-test-diagnostics
branch
from
September 17, 2026 06:45
5579622 to
de44117
Compare
DreierF
changed the title
DreierF
marked this pull request as ready for review
September 17, 2026 07:02
Contributor
Author
|
Superseded: the two fixes now live on renovate/java-jdk-25.x (#975), which already carries the Java 25 bump. |
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.
Supersedes #975: contains the Java 25 bump plus the two changes that make it pass on Windows.
What broke
system-tests/gradle-multi-module/gradle-project/gradle/gradle-daemon-jvm.propertiespinned the fixture's Gradle daemon to Java 21. That is stricter than the Java 21 toolchain the fixture's projects compile against, because the daemon's JVM is chosen by the launcher before the settings script is evaluated — so the foojay toolchain repository declared there does not exist yet and the JDK cannot be provisioned either.While the runners ran Gradle on Java 21 the launcher's own JVM satisfied the requirement. On Java 25 it no longer does, and only the Linux runners keep their JDKs where the launcher detects them (
/usr/lib/jvm); on Windows they live in the tool cache, so every nested build failed withThe file is the only one of its kind in the repository and arrived as a stray in an unrelated commit (5a6b8cf); no test asserts anything about the daemon's JVM. The projects keep their Java 21 toolchain, which resolves normally once the build is running.
Why it took a CI round to find
A nested Gradle build that fails while configuring writes nothing but the single-use daemon notice to stdout and reports the failure on stderr.
runGradlehanded that result back to the caller, so the tests failed onExpected exactly one session, but got 0and the cause never reached the test report. It now reports a failed build instead, which is what made the error above visible.🤖 Generated with Claude Code