ci: give the jobs that run the unit suite a reaping PID 1 #91

Merged

claude merged 1 commit from ci/reaping-init into main 2026-10-16 13:41:18 +00:00

Conversation 0 Commits 1 Files changed 2 +16 -3

claude commented 2026-10-16 13:26:59 +00:00

Member

Copy link

Summary

SessionCloseTest times out on main and on every PR since #86. The job container's PID 1 is tail -f /dev/null, which never reaps. A killed grandchild of the CLI stays a zombie, the JDK reports it as alive (JDK-8284282), and zio-process killTreeForcibly waits on its onExit() for ever.

• container: init: true on ci.yml › test and publish.yml › publish. Docker then runs docker-init as PID 1.
• The value is a literal on purpose: the runner reads init before it evaluates expressions, and drops ${{ }} silently.

This fixes CI only, to unblock 0.7.0. The SDK still hangs a close in a container without an init; that fix is filed as its own issue.

Stacked on #89 (its commit is the first one here). Merge #89 first.

Verification

Probe on this runner (v13.1.0, branch ci/init-probe, runs 1417 and the literal re-run):

Job	PID 1	Orphan zombie	SessionCloseTest
no container:	tail -f /dev/null	stays	1 of 2 fail (timeout)
init: ${{ matrix.init }}	tail -f /dev/null	stays	1 of 2 fail
init: true	/sbin/docker-init -- tail -f /dev/null	reaped	2 of 2 pass

🤖 Generated with Claude Code

https://claude.ai/code/session_01XE9uNVmKpVQiZvc3kLY2zh

## Summary `SessionCloseTest` times out on main and on every PR since #86. The job container's PID 1 is `tail -f /dev/null`, which never reaps. A killed grandchild of the CLI stays a zombie, the JDK reports it as alive (JDK-8284282), and zio-process `killTreeForcibly` waits on its `onExit()` for ever. - `container: init: true` on `ci.yml` › `test` and `publish.yml` › `publish`. Docker then runs `docker-init` as PID 1. - The value is a literal on purpose: the runner reads `init` before it evaluates expressions, and drops `${{ }}` silently. This fixes CI only, to unblock 0.7.0. The SDK still hangs a close in a container without an init; that fix is filed as its own issue. Stacked on #89 (its commit is the first one here). Merge #89 first. ## Verification Probe on this runner (v13.1.0, branch `ci/init-probe`, runs 1417 and the literal re-run): | Job | PID 1 | Orphan zombie | SessionCloseTest | |---|---|---|---| | no `container:` | `tail -f /dev/null` | stays | 1 of 2 fail (timeout) | | `init: ${{ matrix.init }}` | `tail -f /dev/null` | stays | 1 of 2 fail | | `init: true` | `/sbin/docker-init -- tail -f /dev/null` | reaped | 2 of 2 pass | 🤖 Generated with [Claude Code](https://claude.com/claude-code) https://claude.ai/code/session_01XE9uNVmKpVQiZvc3kLY2zh

claude added 2 commits 2026-10-16 13:26:59 +00:00

Keep the kyo experiment out of every publish ... 8484ae442c

KX-D7 holds kyo as an unpublished experiment, and build.mill said so,
but kyo extended SharedModule, which mixed in PublishModule. The publish
selector __:PublishModule.publishArtifacts therefore took it, and
claude-code-query-kyo_3:0.7.0-SNAPSHOT reached the registry. Tagging
v0.7.0 would have added a release of it that the registry cannot
overwrite.

SharedModule now carries the compile settings only. PublishedModule adds
PublishModule, the version and the POM, and every module except kyo
extends it. ./mill resolve __:PublishModule.publishArtifacts no longer
lists kyo.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01XE9uNVmKpVQiZvc3kLY2zh

ci: give the jobs that run the unit suite a reaping PID 1 ... 4e6cf970d0

The job container's PID 1 is tail -f /dev/null, which never reaps. A
killed grandchild of the CLI then stays a zombie, the JDK reports it
alive (JDK-8284282), and zio-process killTreeForcibly waits on its
onExit for ever: SessionCloseTest hit its 60 s timeout on main and on
every PR since #86.

container.init: true makes Docker run docker-init as PID 1. Probed on
this runner (v13.1.0): with it PID 1 is docker-init, an orphan is
reaped, and SessionCloseTest passes; without it, the test times out.
The value is a literal because the runner drops an expression there.

This fixes CI only. The SDK still hangs a close in a container without
an init; that fix is filed separately.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01XE9uNVmKpVQiZvc3kLY2zh

claude referenced this pull request 2026-10-16 13:28:03 +00:00

ZIO: closing a query or session hangs for ever in a container whose PID 1 does not reap #92

claude merged commit c729fa8a24 into main 2026-10-16 13:41:18 +00:00

claude deleted branch ci/reaping-init 2026-10-16 13:41:18 +00:00

claude referenced this pull request from a commit 2026-10-16 13:41:20 +00:00

Merge pull request 'ci: give the jobs that run the unit suite a reaping PID 1' (#91) from ci/reaping-init into main

claude referenced this pull request 2026-10-16 13:42:42 +00:00

Release 0.7.0 #90

Sign in to join this conversation.

Reviewers

No reviewers

Labels

Clear labels   

bug

Something isn't working

  

documentation

Improvements or additions to documentation

  

duplicate

This issue or pull request already exists

  

enhancement

New feature or request

  

good first issue

Good for newcomers

  

help wanted

Extra attention is needed

  

invalid

This doesn't seem right

  

question

Further information is requested

  

wontfix

This will not be worked on

No labels

bug

documentation

duplicate

enhancement

good first issue

help wanted

invalid

question

wontfix

Milestone

Clear milestone

No items

No milestone

Projects

Clear projects

No items

No project

Assignees

Clear assignees

No assignees

1 participant

Notifications

Subscribe

Due date

The due date is invalid or out of range. Please use the format "yyyy-mm-dd".

No due date set.

Dependencies

No dependencies set

Reference

iterative-works/claude-code-query!91

No description provided.