Conversation
fglock
added a commit
that referenced
this pull request
Oct 2, 2026
Record PR #1628 and the current absence of reported branch checks in the high-impact issue plan. Generated with Codex (https://openai.com/codex) Co-Authored-By: Codex <158243242+openai-codex[bot]@users.noreply.github.com>
Checkpoint the in-progress five-issue batch before the final lifecycle algorithm review and acceptance runs. Generated with Codex (https://openai.com/codex) Co-Authored-By: Codex <158243242+openai-codex[bot]@users.noreply.github.com>
Reuse lifecycle root snapshots across cleanup targets and batch strong-cycle checks into one graph pass. Preserve acyclic weak targets that reach a strong cycle and add deterministic work-count regression coverage. Record final acceptance evidence and the remaining provider compatibility limits in the batch plan. Generated with Codex (https://openai.com/codex) Co-Authored-By: Codex <158243242+openai-codex[bot]@users.noreply.github.com>
Record merged-tree build and issue-specific results, including the complete Math::Decimal pure-Perl suite on both PerlOnJava backends and the pre-existing IO::Compress test failures reproduced on the parent snapshot. Generated with Codex (https://openai.com/codex) Co-Authored-By: Codex <158243242+openai-codex[bot]@users.noreply.github.com>
Record PR #1628 and the current absence of reported branch checks in the high-impact issue plan. Generated with Codex (https://openai.com/codex) Co-Authored-By: Codex <158243242+openai-codex[bot]@users.noreply.github.com>
Keep Config's POSIX d_dirfd capability consistent with PerlIO's fileno contract by assigning directory handles a simulated runtime descriptor and releasing it on closedir. Generated with Codex (https://openai.com/codex) Co-Authored-By: Codex <158243242+openai-codex[bot]@users.noreply.github.com>
Retire the distribution-wide PERLONJAVA_SKIP preference so jcpan exercises Catalyst::Runtime's regular test phase. Remove PerlOnJava's old signed preference from existing CPAN homes while preserving user-owned preferences, and cover that retirement in a unit test. Generated with Codex (https://openai.com/codex) Co-Authored-By: Codex <158243242+openai-codex[bot]@users.noreply.github.com>
Match Perl's behavior when a named helper tail-calls from a sort comparator, while preserving errors for tail calls written directly in the comparator. Generated with Codex (https://openai.com/codex) Co-Authored-By: Codex <158243242+openai-codex[bot]@users.noreply.github.com>
Keep accepted IO::Socket handles on deleted Symbol::gensym globs out of Perl's package stash. This lets the final lexical owner close the stream when its scope ends and allows peers to observe EOF. Generated with Codex (https://openai.com/codex) Co-Authored-By: Codex <158243242+openai-codex[bot]@users.noreply.github.com>
Record the Catalyst normal-test results, the accepted-socket fix, saved Mojolicious follow-up, and the remaining acceptance gaps. Exclude parser-area failures and issue #1269 as agreed. Generated with Codex (https://openai.com/codex) Co-Authored-By: Codex <158243242+openai-codex[bot]@users.noreply.github.com>
Record that Catalyst Runtime ran its normal upstream test phase, classify its HTTP exception failures under the parser exclusion, and narrow follow-up to the remaining in-scope failures. Generated with Codex (https://openai.com/codex) Co-Authored-By: Codex <158243242+openai-codex[bot]@users.noreply.github.com>
Track package-rooted targets during final teardown so transient references left by destructor method calls do not look like rescued objects. Preserve resurrection warnings for objects actually retained by DESTROY, and cover the weak-owner regression on both backends. Generated with Codex (https://openai.com/codex) Co-Authored-By: Codex <158243242+openai-codex[bot]@users.noreply.github.com>
Preserve rooted aggregate parents during incremental DOM parsing, honor lexical syntax warning suppression for prototype diagnostics, and retain enclosing conditional ancestry so nested gotos can reach sibling if/elsif arms. Add project-owned regressions and record the acceptance follow-up. Generated with Codex (https://openai.com/codex) Co-Authored-By: Codex <158243242+openai-codex[bot]@users.noreply.github.com>
Resolve close and fileno against existing symbolic IO slots so probing an unopened name does not materialize a typeglob. Add focused coverage for both operations and their effect on subsequent glob lookup. Generated with [Codex](https://openai.com/codex) Co-Authored-By: Codex <158243242+codex@users.noreply.github.com>
Return the typeglob for scalar assignments to its CODE slot, while keeping other slot assignments on their assigned value. Preserve selected glob references so select() can reflect later stash detachment and aliasing. Add coverage for code-slot assignment results and selected-glob identity. Generated with [Codex](https://openai.com/codex) Co-Authored-By: Codex <158243242+codex@users.noreply.github.com>
Return undef when scalar-backed getc reaches EOF. Treat system() without arguments as a child wait instead of raising a missing-command error. Add standard-Perl coverage for both behaviors. Generated with [Codex](https://openai.com/codex) Co-Authored-By: Codex <158243242+codex@users.noreply.github.com>
Trigger pull request checks for the exact source tree that passed the final full UAT and baseline comparison. Generated with Codex (https://openai.com/codex) Co-Authored-By: Codex <codex@openai.com>
Match Perl's behavior when a named helper tail-calls from a sort comparator, while preserving errors for tail calls written directly in the comparator. Generated with Codex (https://openai.com/codex) Co-Authored-By: Codex <158243242+openai-codex[bot]@users.noreply.github.com>
Ensure the release acceptance suite and its child test processes use the freshly built target JAR rather than a stale user-level JAR. Generated with OpenAI Codex (https://openai.com/codex) Co-Authored-By: OpenAI Codex <codex@openai.com>
Keep subprocess output capture finite while continuing to drain child streams. This prevents runaway test diagnostics from exhausting the JVM heap and adds coverage for output truncation and continued pipe draining. Allow the normal Catalyst::Runtime acceptance suite up to four hours to complete. Generated with OpenAI Codex (https://openai.com/codex) Co-Authored-By: OpenAI Codex <codex@openai.com>
Rebuild the installed launcher distribution after the tested JAR and verify that the retired Catalyst distropref is absent before acceptance begins. Generated with OpenAI Codex (https://openai.com/codex) Co-Authored-By: OpenAI Codex <openai-codex[bot]@users.noreply.github.com>
Nested calls can replace temporary aliases to unblessed arrays or hashes while strong container edges still retain them. Defer weak-reference cleanup to the statement-boundary reachability sweep and cover the DOM-style nested list case. Generated with Codex (https://openai.com/codex) Co-Authored-By: Codex <158243242+openai-codex[bot]@users.noreply.github.com>
Snapshot in-progress acceptance planning and numeric transliteration regression before resuming implementation. Generated with OpenAI Codex (https://openai.com/codex) Co-Authored-By: OpenAI Codex <codex@openai.com>
Keep transliteration results byte strings for numeric inputs, preserve upgraded string channels for in-place and non-destructive forms, and retain UTF8 flags across dualvar and version-string values. Generated with OpenAI Codex (https://openai.com/codex) Co-Authored-By: OpenAI Codex <codex@openai.com>
Record the clean rebase, post-rebase build, focused backend checks, and the remaining final acceptance work in the design plan. Generated with [OpenAI Codex](https://openai.com/codex) Co-Authored-By: OpenAI Codex <chatgpt-codex@users.noreply.github.com>
This was referenced Oct 5, 2026
Move all new reachability and lifecycle fixes and tests to independent WIP #1642. Record incomplete acceptance and confirmed reproductions #1643 and #1644. Generated with [OpenAI Codex](https://openai.com/codex) Co-Authored-By: OpenAI Codex <158243242+openai-codex[bot]@users.noreply.github.com>
Generated with [OpenAI Codex](https://openai.com/codex) Co-Authored-By: OpenAI Codex <158243242+openai-codex[bot]@users.noreply.github.com>
fglock
changed the title
fglock
force-pushed
the
feature/top-five-high-impact-issues-20261002
branch
from
October 5, 2026 11:46
64e06d5 to
3bfd2ad
Compare
Generated with [OpenAI Codex](https://openai.com/codex) Co-Authored-By: OpenAI Codex <158243242+openai-codex[bot]@users.noreply.github.com>
Owner
Author
|
Committed UAT handoff records the scope, exact acceptance limits, follow-up issues and next steps. UAT split source Full release acceptance remains incomplete. Follow #1643, #1644 and #1646; historical run 11 results do not qualify either split for release. |
10 tasks
fglock
marked this pull request as draft
October 5, 2026 13:37
Record the rejected 2026-10-05 Perl core run, distinguish the socket sandbox failures from unresolved test regressions, and give the next owner reproducible fix priorities and acceptance criteria. Generated with [OpenAI Codex](https://openai.com/codex) Co-Authored-By: OpenAI Codex <158243242+openai-codex[bot]@users.noreply.github.com>
fglock
changed the title
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.
Scope and handoff
UAT batch for Object::Pad accessors (#1177), directory/handle behavior (#1187), bundled provider fixtures (#1278), UTF-8 flag compatibility, sort helper tail calls, bounded diagnostics, and coherent CPAN launcher/preferences.
Durable handoff: CPAN UAT handoff. It records scope, exact test outcomes, unresolved failures and the next-owner fix list. Acceptance plan.
All reachability/lifecycle changes, fixes and their new tests moved together to independent draft #1647, tracked by #1642. This includes #1597's pipe ownership fix, #1336-related lifecycle work, DOM parent lifetime and accepted sockets. Those issues must not be auto-closed by this PR. No master baseline tests were deleted.
Latest Perl core UAT: not accepted
The 2026-10-05 run completed all 575 files but is not green: 3 files failed and 1 was incomplete. It reported 680,763 passing assertions, which does not override the failures. Full log:
../PerlOnJava/logs/test_20261005_140000_1628.log; baseline comparison:build/reports/uat-20261005/compare_20261005_140000_1628.txtin the working checkout.comp/proto.t: emitted tests 1–215, but not test 216 or the TAP plan, then exited 0.comp/redef.t: test 18 failed; prototype mismatch warning was not suppressed bylocal $^W = 0under-w.op/stat.t: test 94 failed for chained-d -r _on a directory handle. This is not a TTY flake.io/socket.t: bind andsetsockoptfailed in the restricted full-run environment. Focused rerun with extended access passed 25/25; rerun the complete suite with extended access.The durable handoff lists reduction steps and acceptance criteria for each issue. Experimental edits that did not resolve these failures were removed. Do not mark this UAT accepted until fixes are covered by project regressions, both backends pass, and a new full run is clean.
CPAN release qualification remains open
Historical release acceptance run 11 ended at the outer eight-hour guard. Template/PPR/XLSX passed; Mojolicious dom.t passed in the combined candidate but other failures remain. Catalyst exceeded prerequisite setup guard before its normal suite; DateTime setup and ExifTool completion also remain open.
Confirmed, self-contained follow-up tickets: #1643 (mixed arithmetic UTF-8 flags), #1644 (interpreter Test::Mojo app undefined). The handoff distinguishes these from unresolved proxy/restful failures and artificial short-timeout false positives. Authorized exclusions remain
area:parser, #1269 and unsupportedfork.Split code commit
1acfa1f80: fullnice -n 19 makepassed (exit 0), including all unit shards and Joni packaging;nice -n 19 make check-linkspassed. This PR is for UAT; it does not establish release acceptance.Fixes #1177
Fixes #1187
Fixes #1278
AI-assisted with OpenAI Codex.
Remaining qualification tracker: #1646. The independent lifecycle draft passed its own make and focused regression gates; it remains WIP for the unresolved Catalyst cost.