Skip to content

Nightly: maven_reactor_implicit_subprojects fails on Maven 4.0.0-rc-7 ("the root is pinned") since 4fe0859 #2158

Open

Description

[agent] PR shepherd: the nightly e2e-extended leg e2e_vendor_jvm_build, maven, 4.0.0-rc-7, --ignored maven_reactor fails on main.

Failing test: maven_reactor_implicit_subprojects_warm_and_offline_build_and_revert (crates/socket-patch-cli/tests/e2e_vendor_jvm_build.rs:687)

thread 'maven_reactor_implicit_subprojects_warm_and_offline_build_and_revert' panicked at crates/socket-patch-cli/tests/e2e_vendor_jvm_build.rs:687:5:
the root is pinned
test result: FAILED. 2 passed; 1 failed

When: the scheduled CI run on 4fe0859 (2026-11-19 05:42 UTC) fails it. The same leg passed in the scheduled run on 89c482d (2026-11-18). Push CI doesn't run this leg, so the PR checks stayed green.

Likely cause (not confirmed): 6eaa807 (#2100, "Pin a Maven patch only where Maven selects its base version") is the only Maven change in 89c482d..4fe0859 that touches pinning. The test's project relies on Maven 4 implicit subprojects (a root pom with no <modules> list). If the new selection reader doesn't find those modules, it can't be sure the root selects the base version, so it would leave the root unpinned. The other two reactor tests in the same leg pass.

Next step: run cargo test -p socket-patch-cli --test e2e_vendor_jvm_build -- --ignored maven_reactor_implicit against Maven 4.0.0-rc-7. Then either teach the selection reader about implicit subprojects or change the test's expectation if leaving the root unpinned is now intended.


Generated by Claude Code

Activity

  1. mikolalysenko commented on Nov 19, 2026

    @mikolalysenko
    CollaboratorAuthor

    [agent] PR shepherd: the same scheduled run on 4fe0859 has a second Maven 4.0.0-rc-7 leg failing, e2e-extended (ubuntu-latest, e2e_redirect_maven_build e2e_vendor_maven_build, maven, 4.0.0-rc-7) (job 135367641894). It may have the same cause, so I'm noting it here rather than opening a separate issue:

    • mediation::maven_vendored_reactor_pins_each_module_at_the_version_it_selects panics at crates/socket-patch-cli/tests/e2e_vendor_maven_build.rs:857, assert!(!proj.join(".mvn").exists()) after vendor --revert. The revert reports removed for victim@1.10.0 and @1.11.0, but .mvn is still there.
    • mediation::maven_vendored_parent_pin_goes_when_a_module_moves_on_and_reverts_clean panics at :1008 on the same assertion.

    So on Maven 4 rc-7 a vendored revert leaves .mvn behind (or Maven 4 creates it during the build). Whoever fixes this issue should check both legs.


    Generated by Claude Code

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Assignees

No one assigned

    Labels

    Type

    No type

    Projects

    No projects

      Milestone

      No milestone

      Relationships

      None yet

      Development

      No branches or pull requests

      Issue actions