Repository navigation
feat(layout): upgrade Taffy and Stylo with sizing and overflow fixes - #1745
Merged
Merged
Conversation
Sequential Navigation Soak A/B✅ HEAD completed the 200-navigation resilience and memory observation. One browser process, one target, and one CDP session navigate CSDN → SegmentFault → Huaban → example.com repeatedly. This run issued 200 Common ancestor Session resilience
A public-page failure is reported separately from an unrecoverable session. The soak requires all navigation attempts, zero failed recovery, zero lifecycle/network ordering violations, and complete resource evidence. Process-tree memory
HEAD memory by 50-navigation quarter
Raw reports and per-navigation boundary samples: workflow run and full Public-site timing and memory are observational. Use the A/B deltas and trend shape as evidence, not as a deterministic performance threshold. |
Spider Bench A/B✅ All benchmark browser service runs completed; results are informational. Public HEAD: 184 / 240 rows; 38 / 48 sites produced rows; 10 unexpected empty sites. ✅ Deterministic fixture contract is clean. Common ancestor Public 48-site run · informational
Public-site content, timing, and memory are noisy. This single A/B run reports evidence only; site outcome counts explain missing rows without treating them as a deterministic regression. CPU and memory timelinesThese are bounded 40-point views of the same complete process-tree samples used by the HTML report. CPU uses 100% per occupied logical core; memory lines are RSS first and PSS second when complete PSS samples are available. Basexychart-beta
title "Base CPU"
x-axis "Elapsed seconds" [1.036, 2.537, 3.538, 4.539, 5.04, 6.042, 7.545, 8.546, 10.047, 11.049, 12.55, 13.551, 15.051, 16.052, 17.553, 18.553, 19.053, 20.055, 21.057, 22.559, 23.562, 24.563, 26.065, 27.066, 28.567, 29.568, 31.071, 32.073, 33.573, 34.573, 36.075, 37.075, 38.578, 39.578, 41.079, 42.08, 43.583, 44.083]
y-axis "CPU percent" 0 --> 200
line [3.99, 97.92, 123.8, 127.97, 147.75, 135.78, 97.76, 135.88, 119.76, 101.97, 45.96, 59.91, 49.98, 7.99, 6, 4, 0, 0, 2, 115.79, 25.93, 107.78, 131.97, 113.87, 127.91, 111.83, 107.78, 127.63, 86.01, 33.99, 79.89, 120, 0, 14, 121.92, 55.89, 107.91, 165.93]
xychart-beta
title "Base memory: RSS then PSS"
x-axis "Elapsed seconds" [0.035, 1.036, 2.537, 3.538, 4.539, 5.04, 6.042, 7.545, 8.546, 10.047, 11.049, 12.55, 13.551, 15.051, 16.052, 17.553, 18.553, 19.053, 20.055, 21.057, 22.559, 23.562, 24.563, 26.065, 27.066, 28.567, 29.568, 31.071, 32.073, 33.573, 34.573, 36.075, 37.075, 38.578, 39.578, 41.079, 42.08, 43.583, 44.083, 44.274]
y-axis "Memory MiB" 0 --> 500
line [13.65, 62.7, 102.29, 116.18, 137.25, 145.59, 148.63, 129.84, 133.83, 161.35, 142.03, 114.2, 120.99, 140.36, 117.44, 120.94, 119.89, 119.93, 119.96, 119.93, 154.21, 132.16, 140.49, 162.7, 256.19, 200.93, 129.96, 165.23, 152.65, 124.36, 127.3, 159.19, 212.46, 156.96, 146.95, 155.03, 127.63, 158.85, 156.46, 0]
line [10.24, 59.1, 98.66, 112.56, 133.54, 141.97, 145.16, 126.36, 130.37, 157.75, 138.5, 110.59, 117.37, 136.75, 113.82, 117.32, 116.27, 116.31, 116.34, 116.31, 150.72, 128.61, 136.88, 159.11, 252.34, 197.48, 126.35, 161.63, 149.21, 120.75, 123.69, 155.86, 208.87, 153.34, 143.33, 151.43, 124.02, 155.25, 152.98, 0]
HEADxychart-beta
title "HEAD CPU"
x-axis "Elapsed seconds" [0.536, 1.038, 2.039, 3.542, 4.544, 5.544, 6.546, 8.05, 9.052, 10.054, 11.054, 12.056, 13.559, 14.561, 15.561, 16.562, 18.064, 19.067, 20.069, 21.069, 21.571, 22.072, 23.573, 24.573, 25.575, 26.578, 28.08, 29.082, 30.082, 31.084, 32.085, 33.586, 34.587, 35.587, 36.59, 37.09, 38.093, 39.095]
y-axis "CPU percent" 0 --> 200
line [4, 5.98, 105.85, 117.73, 145.61, 127.95, 117.76, 127.69, 71.85, 85.81, 115.93, 85.81, 87.81, 115.74, 23.99, 23.99, 115.78, 59.89, 97.76, 123.92, 101.72, 107.79, 131.83, 149.9, 61.88, 115.69, 109.9, 103.81, 81.94, 15.97, 128, 113.91, 9.99, 65.98, 119.71, 161.83, 71.82, 121.77]
xychart-beta
title "HEAD memory: RSS then PSS"
x-axis "Elapsed seconds" [0.036, 0.536, 1.038, 2.039, 3.542, 4.544, 5.544, 6.546, 8.05, 9.052, 10.054, 11.054, 12.056, 13.559, 14.561, 15.561, 16.562, 18.064, 19.067, 20.069, 21.069, 21.571, 22.072, 23.573, 24.573, 25.575, 26.578, 28.08, 29.082, 30.082, 31.084, 32.085, 33.586, 34.587, 35.587, 36.59, 37.09, 38.093, 39.095, 39.883]
y-axis "Memory MiB" 0 --> 500
line [13.58, 55.41, 64.55, 112.43, 143.89, 132.04, 148.03, 131.49, 156.05, 141.05, 140.83, 142.02, 133.49, 126.04, 139.49, 118.27, 119.02, 158.16, 125.91, 142.45, 187.07, 258.05, 215.55, 184.93, 213.75, 134.44, 154.66, 154.09, 142.68, 138.37, 138.6, 183.96, 163.97, 146.83, 144.3, 165.11, 145.08, 134.75, 152.43, 0]
line [10.33, 52.01, 61.15, 109.1, 140.49, 128.87, 142.99, 128.01, 152.58, 137.54, 137.37, 138.54, 130.01, 122.58, 136.03, 114.81, 115.55, 154.7, 122.67, 138.99, 183.63, 254.66, 211.73, 181.51, 211.44, 130.91, 151.13, 150.56, 139.18, 134.92, 135.06, 180.24, 160.73, 143.28, 140.8, 161.57, 141.36, 131.21, 148.9, 0]
Deterministic fixture · required contractThe fixture exercises 15 routes: 8 are expected to emit rows and 7 intentionally exercise empty, loading, or timeout behavior.
Full HTML, JSON, CSV, logs, and page snapshots: workflow run and full The public-web A/B is informational. The exact HEAD fixture contract runs as a separate required CI check; timing and public-site content deltas remain non-blocking. |
CI Regression ReportSource CI run · source state at render:
Release regression — ✅ HEAD/base failures 0/0; raw binary −0.254382%Package and image size
Startup latency and PSS
Concurrency matrix
Frontend differential — ✅ 1,020/1,020 cases matched; 0 issues
Agent episodes · Moli vs Chromium — ✅ 8/8 Moli episodes passed; 0 failures
Runtime and CDP session contracts — ✅ 26 contract cases; 0 failures
CDP smoke — ✅ 52/52 groups passed; 622 scenariosWorkers: 4 · cumulative group time: 205.28 s · failed groups: 0
WebMainBench · 545 pages — ✅ 545/545 completed; 0 unexpected failures
Frozen HTML · Linux · Diagnostics: All artifact fields are parsed by the trusted default-branch renderer; missing or invalid inputs remain visible as unavailable. |
ldm0
force-pushed
the
upgrade/taffy-stylo-20261020
branch
from
October 19, 2026 22:37
d3a5a5c to
4d81a00
Compare
Pin Taffy5453d6cd, whose remaining-growth subtraction uses the same rounded base-plus-increase value as its growability predicate. This prevents a zero-growth iteration from repeating forever. MOL-464: the new bounded CLI test fails after10s before the pin and passes afterward. The nested grid, percentage margin and Webfont reproducer also completes in0.17s; this trigger hangs in both1.1.17 and1.1.18. cargo fmt --all; cargo clippy --workspace --all-targets --all-features -- -D warnings; cargo nextest run --no-fail-fast: 20276 passed,18 skipped. A disk-full compilation attempt is retained separately; the successful full retry passes. Original-commit: 1639419
Pin the separate Taffy71617899 and4a618c7b fixes: resolve authored intrinsic item minima for auto-track contributions and before fixing the width used to measure row height. A smaller preferred width must not hide the minimum in either path. MOL-465: browser regression covers min-content/max-content with auto,0px and50px preferred widths, track positions and row heights. The two pre-fix failures were next-column x40 instead of60 and row height60 instead of20. The Webfont grid nontermination control still completes. cargo fmt --all; cargo clippy --workspace --all-targets --all-features -- -D warnings; cargo nextest run --no-fail-fast:20277 passed,18 skipped. Initial partial-fix failure logs retained. Original-commit: 3b602b4
Constrain auto-width inline-blocks by max-width:min-content while preserving intrinsic minimums, minimum precedence and automatic growth. Pin the separate Taffy sizing fixes, including the block-root auto-fill control. MOL-466: Taffy a34a7a82e5144b311b2fcb34feeb8412db9d0117. Browser before/after probes retained. Validation: cargo fmt --all; cargo clippy --workspace --all-targets --all-features -- -D warnings; cargo nextest run --no-fail-fast: 20278 passed,18 skipped. Original-commit: d401616
Resolve percentage padding against the flex containing block while measuring a content flex basis. This keeps the text content width intact before line breaking. The browser regression covers three container widths, four padding percentages and both box-sizing modes. MOL-467: Taffy 02c4920e93348e20cb197667a366e0a15e91db39. Browser before/after probes retained. Validation: cargo fmt --all; cargo clippy --workspace --all-targets --all-features -- -D warnings; cargo nextest run --no-fail-fast: 20279 passed,18 skipped. Original-commit: 8ab5696
Replay the 43 Moli patches onto upstream 2e4dbc55 while preserving the 0.13 interfaces consumed by Moli and stylo_taffy. Include the upstream grid loop bounds, definiteness tracking and flex wrapping fixes. Keep the intrinsic sizing regression fixes and repair the existing RTL scrollbar inset and wrapped-column baseline failures. Pin the published fork at 3c19aa124ebcef6320a4e84e141be19f4af659c3. Validation: cargo fmt --all; strict workspace Clippy; full Nextest (20279 passed, 18 skipped, NEXTEST_TEST_THREADS=16); Release build; Taffy native tests (5898 passed, 4 ignored); full WPT Cross plus old/new three-round retests of all 26 worsened case statuses. Original-commit: d95b404
Move Taffy to 0.15 and Stylo to 0.22 using fixed fork revisions, with the matching stylo_taffy adapter. Upgrade cssparser and the production HTML/XML parsers while preserving the vendored XML behavior. Adapt sizing inputs, baselines, grid positions, overflow and out-of-flow ownership. Preserve font-stretch CSSOM compatibility and intrinsic sizing through nested atomic inline probes. Validation: root cargo fmt --all; strict workspace/all-target/all-feature Clippy; Nextest 20280 passed, 18 skipped; Release build; five browser regression rounds; RTL and nested intrinsic Chromium geometry controls. Full WPT Cross is recorded separately with its Cases refresh. Original-commit: 281a920
Check strict lexical XML errors before xml5ever tree construction. Cover six malformed attribute/name forms both alone and in a valid XHTML prologue, plus distinct xml:lang/lang namespaces. Validation: cargo fmt --all; cargo clippy --workspace --all-targets --all-features -- -D warnings; cargo nextest run --no-fail-fast (20283 passed, 18 skipped). Release v3 and five browser regression rounds. Full WPT recorded separately. Original-commit: 0232344
Map native controls to compressible replaced leaves without the image-only replaced flag. Cover checkbox, radio and range normal alignment at 50 by 50 pixels. Validation: cargo fmt --all; cargo clippy --workspace --all-targets --all-features -- -D warnings; cargo nextest run --no-fail-fast (20283 passed, 18 skipped). Release v3 and five browser regression rounds. Full WPT recorded separately. Original-commit: 143e72e
Pin published Taffy and Stylo fixes for baseline provenance, intrinsic child bounds, grid/OOF ratio constraints, writing modes, selector matching, stylesheet nesting and calc type dependencies. Treat IFC ContentProbe inputs as intrinsic contribution measurements. Validation: cargo fmt --all; cargo clippy --workspace --all-targets --all-features -- -D warnings; cargo nextest run --no-fail-fast (20283 passed, 18 skipped). Release v3 and five browser regression rounds. Full WPT recorded separately. Original-commit: b002c0e
…ng roles Subtract horizontal margins for explicit inline-block fit-content widths and carry float min/max-content constraints across the IFC boundary. Cover six font-backed role/keyword combinations, including an inline break that exercises the float path. Validation: cargo fmt --all; cargo clippy --workspace --all-targets --all-features -- -D warnings; cargo nextest run --no-fail-fast (20284 passed, 18 skipped). The combined working tree passed these checks before this logical split. Frozen Release and focused browser/WPT validation also passed; final full WPT is running. Original-commit: 00cfb54
Pin the tested Taffy fork with intrinsic flex constraints, supported baseline grouping, wrapped stretch measurement and scroll origin corrections. Reuse the existing shared static fallback and remove duplicate grid coordinate projection. Validation: cargo fmt --all; cargo clippy --workspace --all-targets --all-features -- -D warnings; cargo nextest run --no-fail-fast (20284 passed, 18 skipped). The combined working tree passed these checks before this logical split. Frozen Release and focused browser/WPT validation also passed; final full WPT is running. Original-commit: fe7aa2f
Adopt the separately validated Taffy fixes for vertical min-content float-grid parents and wrapped RTL flex scroll padding. Preserve the upstream intrinsic keyword result, share one physical flow projection, and cover both browser seams with CLI regression tests. Validation: cargo fmt --all; cargo clippy --workspace --all-targets --all-features -- -D warnings; cargo nextest run --no-fail-fast (20286 passed, 18 skipped). The combined working tree passed these checks before this logical split. Frozen Release and focused browser/WPT validation also passed; final full WPT is running. Original-commit: 6374fa6
Adopt the separate Taffy distribution fallback correction. Remove the scroll-only branch, preserve flex-start versus safe center behavior in every writing mode, and add a browser regression that checks child positions for visible and scroll overflow. Validation: cargo fmt --all; cargo clippy --workspace --all-targets --all-features -- -D warnings; cargo nextest run --no-fail-fast (20287 passed, 18 skipped). The combined working tree passed these checks before this logical split. Frozen Release and focused browser/WPT validation also passed; final full WPT is running. Original-commit: 4d5786f
Allow quoted font-family names that contain valid identifiers separated by single spaces to serialize unquoted. Keep the required quoting for reserved keywords, invalid identifier starts, and repeated or special whitespace. The rebased branch failed two WPT cases consistently: quoted Twisty Tie / Times New Roman and Lucida Grande. Preserve the dash and digit guards from the original fix and update specified/computed regression coverage. Fixes: ab4dcb2 (PR #1647) Validation of the complete fix set, with unchanged source content across the separate logical commits: - cargo fmt --all - cargo clippy --workspace --all-targets --all-features -- -D warnings - cargo nextest run --no-fail-fast: 20374 passed, 18 skipped Original-commit: c873872
Aggregate physical descendants once instead of treating Taffy's logical scroll origin as a physical top-left corner. Share reachable scroll bounds between scrollbar admission and final projection, including vertical writing modes and reversed flex flows. Preserve grid track areas, per-axis clipping, transformed borders, normal-flow margins and end padding. Anchor vertical RTL block flow at the physical bottom while keeping relative insets physical. Add five native browser guards and a live Chromium differential covering 140 fixtures, two scrollbar modes and three fresh-document rounds. Validation: root fmt, strict workspace/all-target/all-feature Clippy, and Nextest (20380 passed, 18 skipped); Release; 840 Chromium comparisons; 960 flex geometry controls; MOL-463-468 probes; full 13631-case WPT Cross. Three old/new control rounds found no new stable case, assertion or harness regression. Cases are refreshed in the following commit. Original-commit: d433653
Consume lexmount/html5ever at 4574311c63e79bba7d4ddb98cb0562d8dad499ee and remove the XML vendor tree and obsolete ignore exceptions. All 11 production XML source files match the former vendor byte-for-byte; HTML, XML and Stylo still share registry markup5ever 0.40.0 and web_atoms 0.3.0. Rebased the style stack onto origin/main at ebae4ca. Validation: Release; cargo fmt --all; strict workspace/all-target/all-feature Clippy; cargo nextest run --no-fail-fast (20413 passed, 18 skipped); 312 paired XML WPT executions and 840 Chromium overflow comparisons with no differences. Full WPT and nine status/subtest-change replays found no new stable regression.
Refresh the complete 13631-case snapshot at WPT db95fafd1fcef8428805e41eb5705d444e8c67ce after the XML fork migration and main rebase. 9866 pass, 3487 fail, 240 timeout, 28 harness-stalled, 9 error, 1 crash; zero missing. Frozen Release SHA-256: b6408a72a6ad26dcd613ced3b0998d6fcf3c9d510d7657011057e294b1637f60. All nine apparent case/subtest/status regressions were replayed against both frozen builds. No new stable regression; the Trusted Types navigation delta is preexisting flaky behavior. Keep the full-run statuses in these lists.
Refresh all 13631 WPT cases at db95fafd1fcef8428805e41eb5705d444e8c67ce after rebasing the style stack and XML fork migration onto main 3ca6888. 9863 pass, 3483 fail, 249 timeout, 26 harness-stalled, 9 error, 1 crash; zero missing. Lists exactly partition frozen Release 8eefbbace618edfbe851b6f2ba2a6231db214e5fbc202bf35d0e16f23d606004. All 16 status/subtest changes versus the prior validated build were replayed against both binaries for three rounds (96 case executions). No new stable or possible flaky regression; existing variations are recorded without rewriting full-run outcomes. Root fmt, strict workspace/all-target/all-feature Clippy and complete Nextest (20427 passed, 18 skipped) pass. XML 312-case and Chromium overflow 840-case comparisons have no differences.
ldm0
force-pushed
the
upgrade/taffy-stylo-20261020
branch
from
October 20, 2026 21:16
4d81a00 to
9d80858
Compare
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
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.
Upgrade the layout/style stack to Taffy 0.15 and Stylo 0.22, with compatible adapters, intrinsic sizing fixes, physical scroll overflow, and regression coverage. Runtime meta-refresh, detached Window, hash navigation and CSP fixes remain in #1726.
Changes
Validation
Rebased onto
origin/main@3ca6888b9d6c88987fef6035c678656ed2dce235. The final source commit ise39838564c7f3f4ec288e76708c0242e22a226f2; the separately refreshed Cases commit is4d81a00ce0f3595f3a79156d5f965fdf82f13726.cargo fmt --allpassed.cargo clippy --workspace --all-targets --all-features -- -D warningspassed.cargo nextest run --no-fail-fast: 20,427 passed, 18 skipped.db95fafd1fcef8428805e41eb5705d444e8c67ce: 9,863 pass, 3,483 fail, 249 timeout, 26 harness-stalled, 9 error, 1 crash; 0 missing, covering all 13,631 cases. The committed lists exactly partition this Release's full run.8eefbbace618edfbe851b6f2ba2a6231db214e5fbc202bf35d0e16f23d606004.CSS Color snapshot expectations
Two known Fail records reflect newer CSS Color 5 behavior against the fixed WPT snapshot. CSS Color 5 permits zero-sum color mixes that the older invalid-color test rejects; serialization preserves a trailing
0%that the older contrast-color test omits. WPT #61205 updates the contrast-color expectation; zero-sum behavior is tracked separately in Mozilla Bug 2049645.This PR remains independent of #1726. The main CSP regression introduced by #1714,
5399724fb07, is fixed in that runtime PR rather than copied into the style stack.