Skip to content

Sync release/v5 with main - #263

Merged
Jake-Moore merged 42 commits into
release/v5from
main
Sep 4, 2026
Merged

Sync release/v5 with main#263
Jake-Moore merged 42 commits into
release/v5from
main

Conversation

@Jake-Moore

Copy link
Copy Markdown
Owner

Syncs release/v5 with main.

The whole content difference is one line in renovate.json:

+  "minimumReleaseAge": "7 days",

The 42 commits behind this are 39 merges plus three renovate.json commits, two of which
(219e0a0c, 1de71938) already have their content on release/v5. One of them is
Bring the Renovate config on main up to date with release/v5, so that reconciliation already ran
in the other direction. Nothing else has drifted.

Why the line matters

Renovate opened a pull request for slf4j-nop 2.0.19 twenty three minutes after the artifact
appeared on Maven Central, against a configuration that auto-merges minor and patch updates.
Consumers shade this library, so a bad version would reach all of them before anyone read the diff.

Worth knowing before merging

Renovate reads its configuration from the repository's default branch, which is main, and this
repository does not set baseBranches. The rule is already in force. This branch does not need
the line for Renovate to honour it; merging only keeps release/v5 from reading as though it
disagrees.

A cherry-pick of 479a31a7 is the alternative, and it leaves release/v5 with one commit instead
of 39 merges.

Jake-Moore and others added 30 commits August 29, 2026 02:58
The dependency dashboard reports "Error executing gradle wrapper update
command" under Repository Problems, and the "all non-major dependencies" rule
has no matchManagers, so it swept gradle-wrapper into the same branch as
everything else. Updating the wrapper makes Renovate run ./gradlew inside its
container, that fails here, and one failing member errors the whole batch.

The effect is visible in the PR history: the last successful all-minor-patch PR
was #242 on 2025-11-03, nine months ago, while individual major updates on their
own branches kept working fine. So lombok, gson, lettuce, HikariCP, snakeyaml,
slf4j, commons-text, httpclient5, amqp-client and the rest have all been frozen
without anything looking obviously broken.

gradle-wrapper now gets its own branch, so its failure is contained and the rest
of the batch can land. automerge is off for it since it is currently failing and
should be looked at rather than merged blind.

Note this only takes effect once it reaches main: Renovate reads its config from
the default branch, and this repo works on release/v5 with main updated
periodically.

Co-Authored-By: Claude Code <[email protected]>
(cherry picked from commit 6e60a6e)
Renovate reads its config from the DEFAULT branch. This repo works on release/v5
while main is the default, so every renovate.json change made today landed
somewhere Renovate never looks.

That means none of it was live here: the rules pinning the version ladder, the
switch to matchDepTypes for Gradle plugins, and grouping the wrapper with the
plugins. It explains why this repo kept producing the pre-fix PR shape while the
sibling picked the changes up immediately.

Same trap as the gradle-wrapper fix nine months ago, which also had to be
cherry-picked to main for the same reason.
main is the continuous timeline and release/v5 is the working line for this
release, so main should absorb it rather than receive individual files.

The two commits main had that release/v5 did not were both single-file
cherry-picks of renovate.json, made because Renovate reads its config from the
default branch. Each one bought a working Renovate config at the cost of drift,
which is the thing that made this merge necessary. A real merge fixes both: the
config is current AND the branches reconcile.
github-actions Bot and others added 12 commits August 31, 2026 07:11
Renovate opened a pull request for slf4j-nop 2.0.19 twenty three minutes
after the artifact appeared on Maven Central, against a configuration
that auto-merges minor and patch updates. Only a repository setting stood
between a one hour old dependency and a published release, and consumers
shade this library, so a bad version would reach all of them before
anyone read the diff.

Seven days is long enough for a compromised release to be reported and
yanked, and short enough that ordinary updates still land promptly.

The github-actions rule sets its own zero day age and is unaffected. That
exemption is deliberate and is left for a separate decision.

Co-Authored-By: Claude Code <[email protected]>
@coderabbitai

coderabbitai Bot commented Sep 4, 2026

Copy link
Copy Markdown

Important

Review skipped

Auto reviews are disabled on base/target branches other than the default branch.

Please check the settings in the CodeRabbit UI or the .coderabbit.yaml file in this repository. To trigger a single review, invoke the @coderabbitai review command.

⚙️ Run configuration

Configuration used: Organization UI

Review profile: ASSERTIVE

Plan: Team

Run ID: b2840dd8-557f-41d6-8ec0-49b8a938ba11

You can disable this status message by setting the reviews.review_status to false in the CodeRabbit configuration file.

Use the checkbox below for a quick retry:

  • 🔍 Trigger review

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

Comment @coderabbitai help to get the list of available commands.

@Jake-Moore
Jake-Moore merged commit da05589 into release/v5 Sep 4, 2026
2 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant