You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
Repository navigation
Make carry.py's check-sections Report the Insertion Refusals apply-sections Raises #3690
Follow-ups from the strict-review receipt on #3683, all in the section modes of scripts/carry.py and the docs describing them.
scripts/README.md (line 405) still says every refusal is raised while planning, before the first file is written. Insertion and overlap refusals now raise in apply_plans phase 1, so check-sections can report a file as cleanly plannable that apply-sections then refuses, for example a target ending inside an unclosed code fence. Either make check-sections run the same in-memory insertion, or correct the sentence.
With a target whose last section holds an unclosed code fence, outside_sections raises "section 'B' is absent" before the post-insert check runs. The operator is told an inserted section is absent and the real cause (the unclosed fence) is never named. The test for this asserts only CarryError, so it would not notice.
RESYNC.md step 1 says a missing section whose unchanged text the repository already carries is refused, but the scan is per file. A section moved unchanged from AGENTS.md to GOVERNANCE.md is inserted as a second copy. Either say "in the same file" in the docs, or scan every carried file.
Class: all introduced by #3683, deferred by its edit budget.
One more instance for this issue's fix, raised by the review of promotion #3693: scripts/README.md:405 says "Every refusal is raised while planning". An insertion refusal is raised in apply_plans' first pass instead, still before any file is written, so check-sections cannot surface it. Constructed case: a hub GUIDE.md holding ## Alpha and ## Beta, a target holding ## Alpha followed by an unclosed code fence. plan_unit reports {"missing": ["Beta"], "stale": ["Alpha", "Beta"]}, and apply_plans then refuses with "section 'Beta' is absent, so the untouched text cannot be compared", leaving the target unchanged. The README sentence should change with whatever makes check-sections run the insertion dry-run.
Activity
Carry a Newly Declared Verbatim Section in carry.py Instead of Refusi…
ptr727 commentedon Oct 31, 2026
One more instance for this issue's fix, raised by the review of promotion #3693:
scripts/README.md:405says "Every refusal is raised while planning". An insertion refusal is raised inapply_plans' first pass instead, still before any file is written, socheck-sectionscannot surface it. Constructed case: a hubGUIDE.mdholding## Alphaand## Beta, a target holding## Alphafollowed by an unclosed code fence.plan_unitreports{"missing": ["Beta"], "stale": ["Alpha", "Beta"]}, andapply_plansthen refuses with "section 'Beta' is absent, so the untouched text cannot be compared", leaving the target unchanged. The README sentence should change with whatever makescheck-sectionsrun the insertion dry-run.Promote Develop: Carry a New Verbatim Section, Read a Null GraphQL No…
ptr727 commentedon Nov 2, 2026
Released (2026-11-02): PR #3835 merged to develop and its branch is deleted, so this claim holds no files. The issue closes on promotion.
Backlog burndown claim (away run 2026-11-02, round 4, group R4G3). Branch:
burndown/3690-check-sections. Issues: #3690.Predicted file set:
scripts/carry.pytests/test_carry.pyscripts/README.mdRESYNC.mdMake carry.py's check-sections Refuse What apply-sections Would and N…