Repository navigation
feat(t27b): tri t27b watch merges a stacked PR chain as steward.t27 decides (Closes #6285) - #6287
Merged
Merged
Conversation
…ecides (Closes #6285) watch_action(verdict, parent) in specs/tri/t27b/steward.t27 says what to do with one PR now: merge on READY, wait, retarget once the stacked parent merged, stop on red, skip a closed PR. t27b.py reads GitHub and runs the action; an unreadable read is no action (Q30). gen/c steward.c is the t27c lab's gen-c output (sha256 75f05574...), test-report 91/91. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
… first) (Closes #6285) Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Contributor
Contributor
PR DashboardGenerated at: 2026-10-04 21:44:27 UTC
Summary
Seal Status
|
6 tasks
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Closes #6285
What
tri t27b watch N...merges a t27b PR stack in order. Each look at the first open PR:tri t27b readyverdictgh pr merge --merge, next PR--interval(150 s)gh pr edit --base master+gh pr update-branch(merge commit, no force-push)The table is
watch_action(verdict, parent)inspecs/tri/t27b/steward.t27, with 10 new tests (91 in all). Python reads GitHub and runs the action. It has no verdict branch of its own, and a CI check enforces that.Why
On 2026-10-04 the steward merged #6247 -> #6249 -> #6254 -> #6258 with throwaway watchers under /tmp. One of them guessed: a single unreadable
readyread skipped the retarget of #6254, and the watcher then stopped at RETARGET (skill slip Q33). Here the parent's state is read fromgh pr list --head <base> --state allon every look, and an unreadable read means no action and a second look.Evidence
gen/c/tri/t27b/steward.cist27c gen-coutput from the t27c Railway lab, sha25675f0557421494de87305da65b0cfa673a40ba1e4a71569fcc35468e1760ef34e. On the lab,t27c test-report specs/tri/t27b/steward.t27gives pass 91, FAIL 0, rate 100.0%. Locally,t27b test --checkgives 91 passed and 1 invariant held.scripts/ci/test_the_t27b_steward_decides_in_t27.py: PASS.scripts/ci/test_a_t27b_tick_reads_before_it_acts.py: PASS, with 4 new watch checks: each planted PR's action, a fixture never merges for real, a stack walked in order (done -> merged -> red stops it), and an unreadable parent is no action.parent == 1->parent == 9in the generated C fails 3 checks across the two tests. The file is restored to the lab's sha.tri t27b watch 6286 --dry-run --onceprinted#6286 WAIT parent=NONE -> WAIT validate=QUEUED, ....Refs #6063
🤖 Generated with Claude Code