feat: show stack position in the pull request list - #1041
Merged
Conversation
roborev: Combined Review (
|
roborev: Combined Review (
|
Maintainers working through a stack had to open each pull request to learn whether it was stacked and where it sat. The sidebar showed workspace, review, and CI indicators but nothing about stacks, so a stacked pull request looked the same as a standalone one. Each list row now carries a compact stack indicator: the stack icon followed by the pull request's position and the stack size, for example "4/7". It sits just before the CI cluster in the row's indicator group. The list API returns this placement directly so the sidebar does not fetch the full stack context for every row. Only stacks with more than one visible member are reported, matching when the detail view shows its stack chip. Members hidden as removed upstream are excluded and the remaining members are renumbered, so the list and the detail view always agree on position and size. The detail view's stack chip drops the word "Stacked" from its visible label. The icon already says that, so the chip now shows just the position and size. The accessible name keeps the full wording for screen readers. Generated with Claude Code (claude-fable-5-1) Co-authored-by: Claude Fable 5.1 <noreply@anthropic.com>
The stack-status e2e test looked up the detail chip with a page-wide button role query for "Stacked: 2/2". The sidebar row is also a button, and its accessible name now includes the new stack indicator label, so the query matched both elements. The row's label reflects the list response and does not change when the detail refresh replaces the stack, which made the "no longer 2/7" assertion fail in CI. Assert the chip's accessible name through its test id instead, which is the element the test is about. The testing context notes the row-name behavior so future chip assertions are scoped from the start. Generated with Claude Code (claude-fable-5-1) Co-authored-by: Claude Fable 5.1 <noreply@anthropic.com>
…udget The workspace recency typography test failed its first attempt in every e2e mock job since it landed, on main and on pull requests, and only passed on retry. It made four full page loads in one test. The 30 second limit applies to the whole test, and each load goes through the Vite dev server, which slows several-fold while 14 workers share it. On retry the server is quieter, so the same test passed and the run reported it as flaky. Split it into a desktop test and a phone test with two loads each. The assertions are unchanged. The testing context records the budget rule so new tests keep their navigation count small. Generated with Claude Code (claude-fable-5-1) Co-authored-by: Claude Fable 5.1 <noreply@anthropic.com>
mariusvniekerk
force-pushed
the
t3code/show-stack-counts-in-pr-list
branch
from
September 4, 2026 16:08
f6d1202 to
b62bf78
Compare
roborev: Combined Review (
|
The pull list endpoint has no size cap. The stack placement lookup expanded the merge request IDs into placeholders and bound each one twice, so a list of about 16,000 pull requests would exceed SQLite's 32,766 bound-variable limit and turn the whole list response into a 500. The sibling worktree link lookup already batches for the same reason. Bind the IDs once as a JSON array through json_each instead. The database test now includes an ID set of 70,000 to prove the lookup survives beyond the limit. Also add a phone-viewport e2e case for the sidebar stack indicator. The indicator is enabled on phone rows, but coverage ran only at desktop width, so the row's fit beside the review, CI, and time indicators was unchecked. Generated with Claude Code (claude-fable-5-1) Co-authored-by: Claude Fable 5.1 <noreply@anthropic.com>
roborev: Combined Review (
|
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.
Stacked pull requests were indistinguishable from standalone ones in the sidebar. You had to open each one to learn whether it was part of a stack and where it sat. Each list row now carries the stack icon with the pull request's position and stack size, placed just before the CI cluster, so a stack reads at a glance from the list.
Sidebar PR rows show a stack indicator such as "4/7" when the pull request belongs to a stack with more than one visible member. The hover title and accessible name read "Stacked: 4/7".
The detail view's stack chip drops the word "Stacked" and shows only the icon and count. The accessible name keeps the full wording.
The pull list API returns a per-row
stackplacement so the sidebar does not fetch stack context for every row. Members hidden as removed upstream are excluded and the rest renumbered, matching the detail stack context.Also splits the workspace recency typography e2e test into desktop and phone tests. It made four dev-server page loads inside the 30 second per-test budget and failed its first attempt in every CI job since it landed, passing only on retry.
Sidebar row with the new indicator before the CI cluster, and the shortened detail chip (synthetic fixture data):
🤖 Generated with Claude Code (claude-fable-5-1)