Repository navigation
Update .tool-versions: node 24.18.0, gradle 9.7.1, java 17 - #1068
Merged
Merged
Conversation
The asdf pin was 24.3.0 while CI runs `node-version: 24`, which resolves to a much newer 24.x. That gap is not harmless: it recently masked a real CI failure entirely, because jest 30.5.0's breakage of nock interception reproduces on 24.18.0 but not on 24.3.0. Pinning closer to CI means local runs are worth something again. Verified with a clean `./build install` and repeated `./build test` runs on both versions. Both show the same stable core of two failures, the network-dependent wrapper-validation checksum tests, plus an occasional third flake in the same area that varies between runs and between node versions. No behaviour attributable to the bump. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
The java pin was not merely stale, it was broken. The init-scripts wrapper is on Gradle 9.7.1, which refuses to run a build on JVM 11: Gradle requires JVM 17 or later to run. Your build is currently configured to use JVM 11. So `./build init-scripts` could not have worked with the pinned toolchain. Note `./gradlew --version` does start under JVM 11, which is probably how the pin survived this long - only an actual build fails. Java 17 matches what ci-init-script-check.yml provisions, so local runs now mirror CI. Verified with a clean `./gradlew check` in sources/test/init-scripts: BUILD SUCCESSFUL in 3m57s. Java 21 was also tried and passes (3m53s), so a later move is unblocked; 17 is chosen only to track CI. Gradle follows the wrappers, which the samples and init-scripts tests moved to 9.7.1. That matches how this file has been maintained before, e.g. "Use Gradle 8.14 everywhere". Nothing in `./build` invokes the asdf gradle - every path goes through ./gradlew - so this keeps ad-hoc CLI use consistent rather than changing what the build does. One flaky run was seen along the way: two TestDevelocityInjection cases against Gradle 6.0.1 failed once and passed on re-run and on repeat clean builds. Unrelated to the version change. Co-Authored-By: Claude Opus 5 (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
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.
The pins in
.tool-versionshad drifted well behind what CI and the Gradle wrappers use. One ofthem was not just stale but broken.
node 24.3.0 β 24.18.0
CI runs
node-version: 24, which resolves to a much newer 24.x. That gap is not harmless: itrecently masked a real CI failure outright, because jest 30.5.0's breakage of nock interception
reproduces on 24.18.0 but not on 24.3.0. A local run that passes is worth something again.
Verified with a clean
./build installand repeated./build testruns on both versions. Both showthe same stable core of two failures β the network-dependent wrapper-validation checksum tests β
plus an occasional third flake in the same area that varies between runs and between node versions.
Nothing attributable to the bump.
java liberica-11 β liberica-17
This pin was broken. The init-scripts wrapper is on Gradle 9.7.1, which refuses to run a build on
JVM 11:
So
./build init-scriptscould not have worked with the pinned toolchain../gradlew --versiondoes start under JVM 11 and reports successfully, which is likely how the pin survived unnoticed β
only an actual build fails.
17 matches what
ci-init-script-check.ymlprovisions, so local runs mirror CI. Verified with aclean
./gradlew checkinsources/test/init-scripts: BUILD SUCCESSFUL in 3m57s.Java 21 was also tried and passes (3m53s), so moving further later is unblocked. 17 is chosen only
to track CI.
gradle 8.14.2 β 9.7.1
Follows the wrappers, which the workflow samples and init-scripts tests are now on. This matches how
the file has been maintained before (e.g. "Use Gradle 8.14 everywhere"). Nothing in
./buildinvokes the asdf gradle β every path goes through
./gradlewβ so this only keeps ad-hoc CLI useconsistent.
Notes
ci-check-and-unit-test.ymlstill provisionsgradle-version: '8.14.2'for its unit-test job.That may be equally stale, but it is a CI behaviour change rather than local tooling, so it is left
alone here.
One flaky run was seen while verifying: two
TestDevelocityInjectioncases against Gradle 6.0.1failed once, then passed on re-run and on repeat clean builds. Unrelated to these version changes.
π€ Generated with Claude Code