Skip to content

chore: bump umm-actually to v0.4.7 - #28

Merged
aliasunder merged 2 commits into
mainfrom
bump-umm-0.4.7
Sep 19, 2026
Merged

aliasunder merged 2 commits into
mainfrom
bump-umm-0.4.7

Conversation

@aliasunder

Copy link
Copy Markdown
Owner

Summary

  • Bump umm-actually from v0.4.6 to v0.4.7 (SHA 9c9353c)
  • Wire new conventions_budget_tokens input to vars.UMM_CONVENTIONS_BUDGET_TOKENS — controls the token budget for the conventions file section (default 8000; empty = 8000)

What changed in v0.4.7

  • conventions_budget_tokens input: configurable token budget for the conventions file section of the review prompt (replaces the hardcoded 8000-token cap)
  • Empty-string default safety for the new input (mirrors the existing request_timeout_seconds pattern)
  • Inlined single-use schema helper per decomposition convention
  • buildUserPrompt custom-budget test coverage
  • README typo fix (permission-checks → checks)

- Pin SHA: 9c9353c (v0.4.7)
- Wire new `conventions_budget_tokens` input to `vars.UMM_CONVENTIONS_BUDGET_TOKENS`

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
@umm-actually

umm-actually Bot commented Sep 19, 2026

Copy link
Copy Markdown

Add issue.pull_request to the concurrency noop condition
Low severity · ci · high confidence

.github/workflows/umm_review.yml:20 — beyond the diff's line ranges, in code the changes touch or depend on.

Pre-existing: the workflow comment claims the concurrency noop condition “mirrors both legs of the job guard,” but its issue_comment branch omits the guard's github.event.issue.pull_request check. An owner comment starting with exactly @umm review on a plain issue shares its number space with PRs, so the run gets the same group as the same-numbered PR's review run and cancel-in-progress: true cancels that in-flight review — even though the job then skips (no issue.pull_request). This is precisely the guard-failing-run-cancels-live-review failure the noop suffix exists to prevent.

Failure scenario: While PR #42's review run is in flight (concurrency group umm-review-42), the owner posts @umm review on plain issue #42 (not the PR). The new issue_comment run's group evaluates to umm-review-42 with no -noop-<run_id> suffix, so the new run cancels the in-flight review; the job guard then skips the new run because github.event.issue.pull_request is absent, leaving PR #42 without a review.

Suggested fix
Add the guard's missing check inside the negation so the issue_comment branch reads `!(github.event.issue.pull_request && github.event.comment.user.login == github.repository_owner && startsWith(github.event.comment.body, '@umm review'))`, matching the job `if` exactly.

umm-actually · deepseek/deepseek-v4-flash-0731

@umm-actually

umm-actually Bot commented Sep 19, 2026 •

Copy link
Copy Markdown

umm-actually re-reviewed at bb694dd

No new findings (1 tracked finding(s) across all runs).


umm-actually · deepseek/deepseek-v4-flash-0731

The issue_comment leg of the noop condition was missing
github.event.issue.pull_request, so an owner '@umm review' on a
plain issue with the same number as a PR would cancel the PR's
in-flight review — then skip, leaving no review posted.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
@aliasunder
aliasunder merged commit d860d81 into main Sep 19, 2026
7 checks passed
@aliasunder
aliasunder deleted the bump-umm-0.4.7 branch September 19, 2026 03:41
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