Skip to content

unsent_saves(): scan every review lane, not just HEAD - #94

Merged
vijay-tylertech merged 1 commit into
mainfrom
fix/unsent-saves-all-lanes
Aug 30, 2026
Merged

vijay-tylertech merged 1 commit into
mainfrom
fix/unsent-saves-all-lanes

Conversation

@vijay-tylertech

Copy link
Copy Markdown
Contributor

git log HEAD --not --remotes is bounded by HEAD's ancestry, so with the checkout on
any branch other than the reviewer's lane their saves become invisible. The docstring
claimed the opposite -- "deliberately not filtered to this branch either" -- which is
what made this hard to spot.

Measured today with the repo left on main: 0 unsent saves reported while 6 sat on
review/vijay-tylertech/08282026-121644, and the Save page read "Nothing waiting.
Everything reviewed has been sent in."

That is the worst possible failure for this indicator specifically. It is the one
thing a reviewer checks to confirm their work still exists, and it told them it did
not -- while the six commits holding four knowledge-file edits were sitting in the
repo the whole time.

--branches=review/* rather than --branches: machinery branches (feature/, fix/)
are not saves, and listing them would report the tool's own development as the
reviewer's pending work.

Verified: 6 reported from main, matching git log --branches='review/*' --not --remotes exactly, where the old form reported 0.

Claude-Session: https://claude.ai/code/session_01VVNYqGza2dFdeWFX7qXnwb

`git log HEAD --not --remotes` is bounded by HEAD's ancestry, so with the checkout on
any branch other than the reviewer's lane their saves become invisible. The docstring
claimed the opposite -- "deliberately not filtered to this branch either" -- which is
what made this hard to spot.

Measured today with the repo left on `main`: 0 unsent saves reported while 6 sat on
`review/vijay-tylertech/08282026-121644`, and the Save page read "Nothing waiting.
Everything reviewed has been sent in."

That is the worst possible failure for this indicator specifically. It is the one
thing a reviewer checks to confirm their work still exists, and it told them it did
not -- while the six commits holding four knowledge-file edits were sitting in the
repo the whole time.

`--branches=review/*` rather than `--branches`: machinery branches (feature/, fix/)
are not saves, and listing them would report the tool's own development as the
reviewer's pending work.

Verified: 6 reported from `main`, matching `git log --branches='review/*' --not
--remotes` exactly, where the old form reported 0.

Claude-Session: https://claude.ai/code/session_01VVNYqGza2dFdeWFX7qXnwb
@vijay-tylertech
vijay-tylertech requested a review from a team as a code owner August 30, 2026 21:27
@vijay-tylertech
vijay-tylertech merged commit 967bc92 into main Aug 30, 2026
1 check passed
@vijay-tylertech
vijay-tylertech deleted the fix/unsent-saves-all-lanes branch August 30, 2026 21:27
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