Skip to content

Repair the base: lint debt, pin step, CI on ctc-master - #2

Merged
conrade-ctc merged 3 commits into
ctc-masterfrom
fork-hygiene
Aug 31, 2026
Merged

Repair the base: lint debt, pin step, CI on ctc-master#2
conrade-ctc merged 3 commits into
ctc-masterfrom
fork-hygiene

Conversation

@conrade-ctc

Copy link
Copy Markdown
Collaborator

Summary

  • Format the 4 files the pinned linters flag on a clean ctc-master. Whitespace only, +21/-10.
  • Bump the Bazel CppInterOp pin to match CMake's CPPINTEROP_GIT_TAG (9802d619).
  • Add ctc-master to the push triggers in both workflows. Mark the llvm22 and cling matrix cells non-gating.

The first inbound PR inherited 4 base-owned red cells. CI has never run on ctc-master before this PR.

Evidence:

  • ruff format --check . and clang-format --dry-run -Werror (pinned versions) pass at HEAD.
  • bazelisk build //... succeeds after each of the last two commits.
  • Both edited workflow files parse as YAML.

🤖 Done with the help of Claude Code (Fable 5, human in the loop)

Emery Conrad added 3 commits August 31, 2026 13:08
The Lint workflow checks the whole tree, but ctc-master never ran it
on push. Drift built up unseen in 4 files across 3 unrelated commits.
Format them so the gate starts clean for every future PR.

Co-developed-with-the-help-of: Claude Code (Fable 5, human in the loop)
CMakeLists.txt moved CPPINTEROP_GIT_TAG to 9802d619 but MODULE.bazel
still fetched 8d624c62. Update the Bazel archive_override to match,
per the file's own comment to keep both pins in step.

Co-developed-with-the-help-of: Claude Code (Fable 5, human in the loop)
Neither workflow ran on ctc-master pushes, so lint and test drift on
that branch stayed invisible until the next PR inherited it. Add
ctc-master to both push triggers.

Split the llvm22 and cling cells into a separate build-test-tracking
job. They track upstream LLVM ahead of our pin and can go red on
their own; a red cell there now reads as tracking noise, not a
blocker. A job that calls a reusable workflow cannot take
continue-on-error, so a job split is the only valid way to keep them
out of the gating job. The llvm21 cells and Lint stay gating.

Co-developed-with-the-help-of: Claude Code (Fable 5, human in the loop)
@github-actions

Copy link
Copy Markdown

Test Results

Configuration Result
macos-26-intel-llvm21-py3.14-cxx20 = 441 passed, 46 skipped, 85 xfailed, 32 xpassed, 100 warnings in 150.20s (0:02:30) =
macos-26-llvm21-py3.14-cxx20 ==== 442 passed, 39 skipped, 93 xfailed, 30 xpassed, 100 warnings in 55.11s ====
ubuntu-24.04-llvm21-py3.14-cxx20-vg ====== 558 passed, 26 skipped, 18 xfailed, 2 xpassed in 97.91s (0:01:37) =======
ubuntu-24.04-llvm22-py3.14-cxx20 test_regression.py s................................x................
ubuntu-24.04-llvm22-py3.14-cxx20-cling = 1 failed, 547 passed, 26 skipped, 27 xfailed, 3 xpassed in 81.17s (0:01:21) ==
ubuntu-24.04-llvm22-py3.14-cxx23 test_regression.py s................................x................

@conrade-ctc
conrade-ctc merged commit b29e239 into ctc-master Aug 31, 2026
6 of 9 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