Repository navigation
π tracker: one-line PR sub-row header - #15444
Conversation
|
[APPROVALNOTIFIER] This PR is NOT APPROVED This pull-request has been approved by: The full list of commands accepted by this bot can be found here. DetailsNeeds approval from an approver in each of these files:Approvers can indicate their approval by writing |
Open PR overlap checkBase branch: Conflicts are checked with every PR merged onto the current tip of
|
There was a problem hiding this comment.
Read tracker-row.js and tracker.css at 90bcf72 against the base. Looks correct to me.
Verified
- The new
>-child CSS rules actually match:stageTrackerHeadColsHtml(tracker-row.js:829) emits<li class="ham-seg ham-head-pr">${prRefs.join('')}</li>, andrefis glyph +<a class="ham-pr-ref">/<span class="ham-pr-line">+ lineage chip + base pill concatenated as direct children β no wrapper. The repo-chip case wraps only the link in.ham-pr-line, and that class has its ownflex: 0 1 autorule, so it shrinks the same way. .ham-badge(flex: 0 0 auto) vs.ham-lineage(flex: 0 2 auto; min-width: 7em) have equal specificity on aham-badge ham-lineagechip; the lineage rule is declared later, so it wins and the chip gives way first as the body says..ham-titlegoingnowrap; ellipsisis safe for hover: both emit sites (tracker-row.js:1231,:1243) already settitle="${esc(title)}".- Lineage chip removed from the cap cell at
:992and added once to the PR cell at:976; the new test asserts it's drawn exactly once.static/index.htmlmatchesstatic-srcfor both the CSS and JS hunks. - Single-PR issue row (
:1201) reordered lineageβbase pill consistently with the sub-row.
Not run: go test is disabled in this pod.
β hive: agent=reviewer backend=copilot model=claude-fable-5.1 copilot=1.0.88
There was a problem hiding this comment.
Correctness review (read at head 79ef5c7, after the stated 90bcf72 moved on). Contract #15441 is met for the normal nested PR, the single-PR issue row and the top-level PR header; cascade checks out (tracker.css:116 display:block outranks .ham-lineage inline-flex at :395, so text-overflow applies). Two small things:
- Superseded sub-row can still wrap (low). The dimmed "adopted original" line goes through the same
ham-head-prcell (tracker-row.js:985) but its<span class="ham-superseded-note">gets no sizing rule: it inheritswhite-space: normalfrom.ham-head-cols .ham-seg(tracker.css:103) and defaultmin-width:auto, so when the cell is tight the note wraps onto extra lines while the title ellipsizes. Fix intracker.cssafter line 118:.ham-head-cols .ham-seg.ham-head-pr > .ham-pr-ref, .ham-head-cols .ham-seg.ham-head-pr > .ham-pr-line { flex: 0 1 auto; } .ham-head-cols .ham-seg.ham-head-pr > .ham-superseded-note { flex: 0 1 auto; min-width: 0; white-space: nowrap; overflow: hidden; text-overflow: ellipsis; } - Comment contradicts behaviour (info).
tracker.css:111says "the title gives way first (then the lineage chip)", but the rules give the chipflex-shrink: 2(:116) vs the link's1(:118), i.e. the chip gives way first β which is what #15441 and the PR body ask for. Just fix the comment.
β hive: agent=ci-maintainer backend=copilot model=claude-fable-5.1 copilot=1.0.88
|
stage 080-awaiting-merge-admission β 050-review-pending: required-check-pending:build-and-test |
|
Hive filed review findings that do not block this PR as backlog items:
Cap: 10 backlog item(s) per PR per day. |
There was a problem hiding this comment.
Delta re-review (correctness) β head ecf8abb9 vs previously reviewed 79ef5c71.
The delta is only the v5 merge (#15442, auto-merge current-base default); compare/3feba139...ecf8abb9 touches just the PR's five files, so no new code was introduced by the merge. Prior correctness notes are unchanged at this head (low, non-blocking):
src/pkg/dashboard/static-src/css/tracker.css:111β comment says the title gives way first, then the lineage chip; the rules do the opposite (.ham-lineageflex: 0 2 autoat :115 vs.ham-pr-titleflex: 0 1 autoat :130).src/pkg/dashboard/static-src/css/tracker.css:369β.ham-superseded-noteis rendered inside the now-nowrap PR cell (tracker-row.js:985) but has nowhite-space: nowrap/ellipsis rule, so a long superseded note can still wrap.
β hive: agent=adjudicator-2 backend=copilot model=claude-fable-5.1 copilot=1.0.88
There was a problem hiding this comment.
Delta re-review (79ef5c7 β ecf8abb) β the only new commit is Merge branch 'v5' pulling in #15442 (auto-merge current-base); none of the 9 delta files are PR-authored, so intent alignment against #15441 is unchanged. Still approve from this perspective; two prior low-severity notes remain as-is:
src/pkg/dashboard/static-src/css/tracker.css:110-113β the comment says "the title gives way first (then the lineage chip)", but the rules below (lineageflex: 0 2 autoat :115,.ham-pr-ref/.ham-pr-lineflex: 0 1 autoat :117) and #15441 ("the lineage chip β¦ giving way before the title") say the opposite. The code matches the issue; the comment is reversed./* The PR cell is one line: connector, author icon + #N + title, lineage chip and base pill sit side by side; the lineage chip gives way first (then the title), the pills and the connector never shrink (hivecommons/hive "consolidate from 3 lines of text to 1"). */src/pkg/dashboard/static-src/css/tracker.css:369β.ham-superseded-noteis rendered as a direct child of the now-nowrap PR cell (tracker-row.js:985) but has noflex/white-space/ellipsis rule, so a long superseded note can still wrap inside the one-line sub-row. Minor (dimmed, superseded rows only)..ham-superseded-note { color: var(--muted); font-size: var(--fs-xs); flex: 0 1 auto; min-width: 0; white-space: nowrap; overflow: hidden; text-overflow: ellipsis; }
β hive: agent=reviewer backend=copilot model=claude-fable-5.1 copilot=1.0.88
There was a problem hiding this comment.
Looks correct to me. Read the diff against the tree at ecf8abb, not just the hunks:
stageTrackerHeadColsHtml(tracker-row.js:829) emits<li class="ham-seg ham-head-pr">${prRefs.join('')}</li>, so the connector span,<a class="ham-pr-ref">, lineage chip and base pill really are direct children of the cell β the new.ham-head-pr > β¦selectors intracker.css:112-116match the markup, and theham-pr-linewrapper case is covered by the> .ham-pr-line { flex: 0 1 auto }rule.> .ham-badge { flex: 0 0 auto }and> .ham-lineage { flex: 0 2 auto }have equal specificity; the lineage rule is declared later and wins, so the chip (which has both classes,tracker-row.js:941) does shrink as intended.- The superseded branch (
tracker-row.js:980-989) never rendered a lineage chip before this change (its chips arg wasstageTrackerActedBadgeHtmlonly), so gatinglineageon!supersededpreserves behaviour rather than dropping anything. - Every
.ham-titleuse (tracker-row.js:1231,1243,1247) sits inside adisplay:flex.ham-head-main, sonowrap; overflow:hidden; text-overflow:ellipsisactually takes effect. go run ./pkg/dashboard/webstatic/gen β¦ -checkexits 0 on this head:static/index.htmlis in sync withstatic-src.
Only remaining nit is the one already filed as #15447 β the CSS comment at tracker.css:110-111 says the title gives way first, while the flex-shrink: 2 on the lineage chip (and the PR body) says the chip does. Not blocking.
β hive: agent=reviewer-2 backend=copilot model=claude-fable-5.1 copilot=1.0.88
There was a problem hiding this comment.
Style (advisory, no blocker) β read at ecf8abb.
Two comments in src/pkg/dashboard/static-src/css/tracker.css now describe the layout this PR removes:
tracker.css:107still says the PR cell "gives way to the chip, which lives in the cap cell" β the lineage / adopted-from chip now rides the PR cell (tracker-row.js:975-978); only the acted badge andstageTrackerChipsHtmlremain in the cap (tracker-row.js:992).tracker.css:119-122still explains the cap'salign-items: flex-endby "on a two-line title it sits on the second line" β titles no longer wrap (.ham-title/.ham-pr-titlearenowrap+ ellipsis,tracker.css:130,145), so that case cannot occur.
/* Lead = exactly three of the six equal chevrons (half of the strip
minus the fixed 4em cap and the six gaps) plus the two gaps between
them, so the PR cell starts at the PR in progress chevron; the PR cell takes what
is left; the acted badge and stage chips live in the cap cell. */
/* The cap's badges sit on the baseline of the one-line title
(align-items: flex-end inside the stretched cap cell), directly over
the strip they describe. */
Also, the new comments at tracker.css:113 and tracker-row.js:970 cite the issue by quote (hivecommons/hive "consolidate from 3 linesβ¦") β the file's convention is hivecommons/hive#NNNNN (e.g. tracker.css:16,238,317); hivecommons/hive#15441 would make it greppable.
β hive: agent=adjudicator backend=copilot model=claude-fable-5.1 copilot=1.0.88
|
@clubanderson is auto-promoted for contributor-lane merging on this PR: the last 5 merged PRs by this author in this hive's repositories have no revert or sentinel-alert evidence. CI, reviewer verdict, diff fences, soak, holds and branch protection still apply. |
|
stage 080-awaiting-merge-admission β 050-review-pending: required-check-pending:build-and-test |
There was a problem hiding this comment.
Read the diff against v5 at 2900a33 plus the surrounding tracker.css / tracker-row.js; this looks correct to me.
What I verified:
src/pkg/dashboard/static-src/css/tracker.css:114β the new row rule (.ham-head-cols .ham-seg.ham-head-pr, 3 classes) outranks thealign-items: flex-startbase at line 103 (2 classes), soalign-items: centerandflex-direction: rowdo take effect..ham-seg(line 282) already carriesmin-width: 0; overflow: hidden, and.ham-head-main(line 125) hasmin-width: 0, so thenowrap+text-overflow: ellipsison.ham-title(line 145) and.ham-pr-titleactually ellipsizes instead of overflowing the cell.tracker.css:115-116β> .ham-badge { flex: 0 0 auto }and> .ham-lineage { flex: 0 2 auto }have equal specificity; the lineage rule is later, so theham-badge ham-lineagechip shrinks first as the body describes.tracker.css:605-614β the β€1100px grid keeps.ham-head-prspanning the full row; the new flex-row rule composes with it rather than fighting it.tracker-row.js:977-982β in the repo-chip caserefis wrapped in.ham-pr-linebefore the lineage chip and base pill are appended, so they stay direct siblings of the cell, which is exactly what the new node assertion atstage_tracker_ui_test.go:725-733checks. The superseded early-return (tracker-row.js:984) is byte-identical tov5, so the dimmed adopted-original line is unchanged.make'sgo run ./pkg/dashboard/webstatic/gen β¦ -checkpasses on this head: the committedstatic/index.htmlis in sync withstatic-src/.- The new prelude symbols (
jsConstBlock,STAGE_TRACKER_LINEAGE_REASON_TIP,stageTrackerAgoTextintracker-motion.js:38) all exist, so the harness lifts real code.
One low-severity note, not a blocker: .ham-superseded-note (tracker.css:369) has no white-space/flex rule of its own and inherits white-space: normal from .ham-head-cols .ham-seg (line 103), so on a tight row the "Β· superseded by #N" text can still wrap inside its flex item and give that sub-row a second line. If the one-line goal is meant to cover adopted originals too, a white-space: nowrap; flex: 0 0 auto on it would close that gap. I did not measure it in a browser.
β hive: agent=reviewer-2 backend=copilot model=claude-fable-5.1 copilot=1.0.88
|
stage 060-reviewed β 050-review-pending: required-check-pending:build-and-test |
There was a problem hiding this comment.
Read the diff against v5 plus the callers it touches (stageTrackerSubRowHtml, the single-covering-PR branch in stageTrackerRowHtml, and stageTrackerHeadColsHtml at tracker-row.js:840). Looks correct to me, and the body matches what the code does.
What I verified:
- The lineage / adopted-from chip now lives in the PR cell instead of the chip cap (
tracker-row.js:985vs the old cap-arg at:1005), and the superseded branch never drew it before, sosuperseded ? ''keeps that row unchanged. - The one-PR issue row (
tracker-row.js:1214) only reorders link β lineage β base pill; nothing else in that expression changed. .ham-head-cols .ham-seg.ham-head-prisflex-direction: rowwithmin-width: 0; overflow: hidden,.ham-head-maingainsoverflow: hidden, and.ham-title/.ham-pr-titleareflex: 0 1 auto; min-width: 0β the ellipsis has a shrinkable ancestor on both the lead and PR cells. The issue-row.ham-titlechange is in scope per #15441 ("The issue row and the top-level PR header keep to one line the same way").static/index.htmlmirrorsstatic-srcline for line (33/33 changed lines).- The new harness block in
stage_tracker_ui_test.goasserts sibling order and "drawn once", which fails on the oldflex-direction: column+ cap-chip layout.
Two notes, neither blocking:
tracker.css:120-123β the comment above.ham-head-cols .ham-seg.capstill says the chip "hugs the bottom of the title block β¦ on a two-line title it sits on the second line". Titles no longer wrap, so that rationale is stale; worth a one-line edit while you're here.conflicts-with-open-pris real: #15457 rewrites the samestageTrackerHeadColsHtml([ref], β¦)line attracker-row.js:999(appending the lead button to the cap).git merge-treereports a conflict ontracker-row.jsandstatic/index.html. Resolution is mechanical β drop the lineage-chip term, keep thestageTrackerLeadButtonHtmlterm β whichever lands second.
β hive: agent=reviewer backend=copilot model=claude-fable-5.1 copilot=1.0.88
|
stage 080-awaiting-merge-admission β 050-review-pending: required-check-pending:build-and-test |
|
@clubanderson this PR is held at Why:
Next: the hive does not push to your branch. Push a new head (or reply here); CI and the review swarm re-run automatically and the PR returns to the normal path. |
There was a problem hiding this comment.
Looks correct to me. Read tracker-row.js and tracker.css at 8b68da1 against the v5 tree.
- The sub-row change is a move, not a duplication: the lineage/adopted-from chip leaves
stageTrackerHeadColsHtml's cap argument and joinsrefbefore the base pill (static-src/js/tracker-row.js, sub-row branch), and the single-PR issue row is reordered to match; the new harness assertion thatham-lineageappears exactly once in the head guards against it reappearing in the cap. .ham-titleis the top-level issue row's title (tracker-row.js:1244,:1256), so thenowrap/ellipsis rule reaches main rows too β the PR body states this is intended ("the issue row and a top-level PR header never wrap either"), and thetitle=attribute already carries the full text, so nothing is lost on hover.static/index.htmlcarries the same CSS block asstatic-src/css/tracker.css, so the generated artifact and source agree.
Open question, not verified: whether the ellipsized .ham-title reads well at the Dashboard Mobile Shots viewport β the body's measurements are at 1400 px only.
β hive: agent=reviewer-2 backend=copilot model=claude-fable-5.1 copilot=1.0.88
A PR nested under its issue drew its header on three lines: the β³ connector, then the author icon + #N + title, then the base-branch pill, because the PR cell of the ghost strip was a flex column and the title wrapped. The cell is now a single nowrap flex row β connector, repo chip, PR link, lineage / adopted-from chip, base pill β the title ellipsizes with its full text on hover, the lineage chip gives way before the title, and pills never shrink. Issue titles and the top-level PR header ellipsize the same way. Sub-row height 67 β 40 px (124 β 40 px with a long title and a lineage chip) at 1400 px wide. Signed-off-by: Andrew Anderson <andy@clubanderson.com>
8b68da1 to
9d9ee59
Compare
|
hive/awaiting-ci: rebased onto v5 (kept the v5 lead-button cap slot alongside this PR's lineage-chip move; regenerated index.html). CI pending β sweep will merge when green. π Hive Agent: β hive: agent=scanner backend=copilot model=claude-fable-5.1 copilot=1.0.88 |
|
@clubanderson this PR is back in your court (
Once the new head lands, CI and the review swarm re-run automatically and the PR returns to the normal path. |
|
merged by Hive trusted-author auto-merge: author @clubanderson holds role owner and GitHub permission admin on repo; CI green at 9d9ee59. |
|
Thank you for your contribution! Your PR has been merged. We'd love to hear how your experience was: share feedback |
There was a problem hiding this comment.
Read the diff against the tree at 9d9ee59 (tracker-row.js, tracker.css, the test harness). Looks correct to me.
What I checked:
- The lineage / adopted-from chip is a direct child of
.ham-seg.ham-head-pron all three render paths, so the new>child selectors attracker.css:115-118actually apply: sub-row (tracker-row.js:999-1002, appended after the.ham-pr-linewrapper closes when repo chips are on), single-PR issue row (:1225), and top-level PR row (:1263, no base pill).stageTrackerHeadColsHtml(:842) joinsprRefsstraight into the<li>with no wrapper. .ham-lineagecarriesham-badgetoo (:954,:977).tracker.css:115(.ham-badge { flex: 0 0 auto }) and:116(.ham-lineage { flex: 0 2 auto }) have equal specificity, and:116is later, so the chip shrinks and the plain base pill does not β the ordering is load-bearing but correct as written.- Every
.ham-titlelives inside.ham-head-main(tracker-row.js:1255,:1267,:1271), which now hasoverflow: hidden; min-width: 0(tracker.css:125), sonowrap + text-overflowellipsizes instead of overflowing the fixed-width lead cell; each hastitle=for hover. - Superseded sub-row (
:1004-1011) never drew the cap lineage chip before and doesn't now β no behaviour change there.
Info only, not blocking: two CSS comments now describe the old layout β tracker.css:104-107 ("gives way to the chip, which lives in the cap cell") and :119-122 ("on a two-line title it sits on the second line"). Titles can no longer be two lines and the lineage chip no longer lives in the cap.
β hive: agent=reviewer-2 backend=copilot model=claude-fable-5.1 copilot=1.0.88
Fixes #15441
Problem
A PR nested under its issue in the Activity Monitor drew its header on three lines β the
β³connector, thenicon #N title, then the base-branch pill β before its chevron strip, and a long title wrapped further. Owner: "consolidate from 3 lines of text to 1 to compact the row".Root cause: the ghost-strip PR cell (
.ham-head-cols .ham-seg.ham-head-pr) wasflex-direction: column, so each child was its own row, and.ham-pr-title/.ham-titlewerewhite-space: normal.Change
tracker.css: the PR cell is a single nowrap flex row (align-items: center,min-width: 0,overflow: hidden). Connector and pills areflex: 0 0 auto; the PR link / title areflex: 0 1 autowithmin-width: 0and ellipsize (full text already ontitle=); the lineage / adopted-from chip gives way first (flex-shrink: 2, floor7em) and ellipsizes..ham-title(issue slot) and.ham-pr-titlearenowrap; overflow: hidden; text-overflow: ellipsis, so the issue row and a top-level PR header never wrap either.tracker-row.js: the sub-row puts the lineage / adopted-from chip in the PR cell, after the PR link and before the base pill (it used to sit in the chip cap); the single-PR issue row is reordered the same way (link β lineage β base pill). Ids/classes asserted by tests (ham-linked-ico,ham-pr-ref,ham-pr-num,ham-pr-title,ham-pr-line,ham-badge,ham-lineage,ham-adopted-from,ham-head-pr) are unchanged.stage_tracker_ui_test.go: the node harness asserts the sub-row PR cell holds connector, PR link, lineage chip and base pill as siblings in that order with no block element between them, the PR link hover starts with the full title, the lineage chip is drawn once (not also in the cap), and the single-PR issue row keeps the same order. The prelude now lifts the real lineage-chip helpers.changelog.d/changed-tracker-subrow-one-line.md.Measured (headless Chromium, 1400 px wide, generated
index.html+tokens.css, fixture issue with two covering PRs)Before:
After:
Tests
go test -race ./pkg/dashboard/ -run 'StageTracker|Tracker|IndexHTML'and the full./pkg/dashboard/package pass locally.