fix(markdown): preserve visible text and share export style observations

[lanyue-llk]

Native Markdown can omit visible white text over an image, positioned sibling or pseudo-element. The visibility adapter infers the painted background from DOM ancestors and a default white canvas, then removes text when the inferred background matches its color. DOM ancestry does not establish the actual painted backdrop.

Remove this color-only suppression and its unused helpers. Explicit display, visibility and opacity filtering remains in place. The regression scenarios cover sibling pictures, pseudo-element backgrounds and positioned backdrops, alongside controls for genuinely hidden text. Existing expectations for color-only suppression are updated to preserve the text.

Markdown export also sampled the same document through six independent style observations. Collect the required properties and visibility provenance together, remove unused color/background samples, and preserve the adapter's field layout. A large catalog scenario checks cascading styles, meaningful images, hidden controls, dynamic updates and repeated exports.

The exporter already projects finite, single-run opacity animations with a forwards fill to their bounded final state. Apply that same rule when the animation has a positive delay; a finite delay does not make the final state unbounded. The regression includes a delayed reveal and retains controls for paused, repeating, reverse, zero-duration and non-opacity animations.

This is stacked after #1380, whose branch already contains the #1337 style-aware Markdown export. The affected export does not exist on main. This PR does not update #919, add site-specific extraction, modify benchmark answers or postprocess native output. The performance change removes known repeated work; historical logs do not establish it as the sole cause of an observed export timeout.

Validation: whitespace checks passed. Format, Clippy and the test suite are delegated to the repository's normal CI; no local test suite or benchmark calls were run. Historical benchmark outputs and verdicts remain unchanged by this product patch.

PR stack

This PR is the sixth layer of the GitHub stack and targets the preceding branch directly:

main → #1334 → #1335 → #1336 → #1337 → #1380 → #1428 → #1431 → #1432

Merge from left to right. The GitHub base of #1428 is codex/readme-web-fetch-benchmarks, not #919 and not main; its Files changed view therefore contains only this layer's product fix. #919 remains a separate rebased historical combination and is not a dependency of this stack. #1429 targets main independently because it only adds reusable CI infrastructure.

[github-actions]

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 Page.navigate commands per binary.

Common ancestor 1f99b9fee9 → HEAD 8b63b90fdd; benchmark order: base-first; workflow: success.

Session resilience

Metric	Base	HEAD
Attempted / planned	200 / 200	200 / 200
Direct observable pass	199 / 200	196 / 200
Failures / recovered	1 / 1	4 / 4
Recovery failures	0	0
Lifecycle order violations	0	0
Network order violations	0	0
Wall time	468.9 s	470.1 s

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

Metric	Base	HEAD	Δ
Boundary samples	200	200	0
Boundary RSS samples	200	200	0
Boundary PSS samples	200	200	0
Periodic peak RSS	264.9 MiB	289.2 MiB	+24.3 MiB
Periodic peak PSS	262.5 MiB	285.5 MiB	+23.0 MiB
RSS · first 10 avg	180.2 MiB	159.0 MiB	-21.3 MiB
RSS · last 10 avg	194.1 MiB	198.8 MiB	+4.6 MiB
RSS · last − first window	+13.9 MiB	+39.8 MiB	+25.9 MiB
RSS · warm slope / 100 nav	+4.5 MiB	+8.1 MiB	+3.6 MiB
PSS · first 10 avg	176.3 MiB	155.4 MiB	-21.0 MiB
PSS · last 10 avg	190.7 MiB	195.3 MiB	+4.7 MiB
PSS · last − first window	+14.3 MiB	+40.0 MiB	+25.7 MiB
PSS · warm slope / 100 nav	+4.6 MiB	+8.2 MiB	+3.6 MiB
Peak file descriptors	58	59	1
Peak threads	29	29	0

HEAD memory by 50-navigation quarter

HEAD range	Avg RSS	Final RSS	Peak RSS	Avg PSS	Final PSS	Peak FD
1–50	178.8 MiB	161.4 MiB	248.9 MiB	175.2 MiB	157.7 MiB	55
51–100	187.0 MiB	177.8 MiB	228.0 MiB	183.4 MiB	173.9 MiB	53
101–150	191.2 MiB	167.2 MiB	227.6 MiB	187.4 MiB	163.6 MiB	54
151–200	194.7 MiB	197.8 MiB	246.3 MiB	191.3 MiB	194.3 MiB	53

Raw reports and per-navigation boundary samples: workflow run and full sequential-navigation-soak-results artifact.

Public-site timing and memory are observational. Use the A/B deltas and trend shape as evidence, not as a deterministic performance threshold.

[github-actions]

CI Regression Report

Source CI run · source state at render: in_progress · artifacts: 6/6

Check	Status	Signal
Release regression	✅	HEAD/base failures 0/0; raw binary −0.011257%
Frontend differential	✅	1,020/1,020 cases matched; 0 issues
Agent episodes	✅	8/8 Moli episodes passed; 0 failures
Runtime/CDP contracts	✅	26 contract cases; 0 failures
CDP smoke	✅	52/52 groups passed; 621 scenarios
WebMainBench	✅	545/545 completed; 0 unexpected failures

Release regression — ✅ HEAD/base failures 0/0; raw binary −0.011257%

Package and image size

Metric	Base	HEAD	Delta	Delta %
Raw binary	149.03 MiB	149.02 MiB	−17.18 KiB	−0.011257%
Stripped binary	102.66 MiB	102.65 MiB	−16.00 KiB	−0.015220%
gzip binary	51.15 MiB	51.15 MiB	−3.20 KiB	−0.006102%
Rootfs	155.94 MiB	155.92 MiB	−20.00 KiB	−0.012525%
gzip rootfs	60.05 MiB	60.05 MiB	−6.71 KiB	−0.010905%

Startup latency and PSS

Case	Base p50	HEAD p50	Delta	Base PSS	HEAD PSS	Delta
Serve ready	27.58 ms	27.29 ms	−0.29 ms	4.67 MiB	4.30 MiB	−382.50 KiB
CDP first page	53.33 ms	51.26 ms	−2.07 ms	5.26 MiB	3.78 MiB	−1.48 MiB
CDP warm page	12.90 ms	12.16 ms	−0.74 ms	49.77 MiB	49.28 MiB	−505.00 KiB
CLI about:blank	16.33 ms	15.90 ms	−0.42 ms	3.75 MiB	2.63 MiB	−1.12 MiB
CLI local JS	20.19 ms	19.36 ms	−0.83 ms	3.06 MiB	2.15 MiB	−928.50 KiB

Concurrency matrix

Concurrency	Base p50	HEAD p50	Delta	Base PSS	HEAD PSS	Delta	HEAD failures	Unstable B/H
1	18.50 ms	17.97 ms	−0.53 ms	562.00 KiB	566.00 KiB	+4.00 KiB	0	0/0
5	21.05 ms	21.75 ms	+0.71 ms	22.15 MiB	27.26 MiB	+5.12 MiB	0	0/0
10	21.17 ms	20.73 ms	−0.44 ms	26.82 MiB	27.39 MiB	+583.50 KiB	0	0/0
25	20.78 ms	20.52 ms	−0.26 ms	27.65 MiB	13.20 MiB	−14.46 MiB	0	0/0
100	20.85 ms	20.58 ms	−0.28 ms	12.49 MiB	27.42 MiB	+14.93 MiB	0	0/0

Frontend differential — ✅ 1,020/1,020 cases matched; 0 issues

Match	DOM mismatch	Diagnostic mismatch	Moli error	Reference error	Infrastructure error	Mismatched frames	Duration
1,020	0	0	0	0	0	0	7m 16.94s

Agent episodes · Moli vs Chromium — ✅ 8/8 Moli episodes passed; 0 failures

Fast contract	Episodes	Passed	Assertions	Ready p50	Episode p50
Moli	8	8	47/47	3.48 ms	26.35 ms
Chromium	8	8	47/47	274.58 ms	90.97 ms

Operation p50	Chromium	Moli	Delta	Delta %
navigate	45.41 ms	6.15 ms	−39.26 ms	−86.45%
observe	1.00 ms	0.90 ms	−0.10 ms	−10.34%
fill	4.50 ms	3.67 ms	−0.84 ms	−18.55%
click	3.41 ms	4.26 ms	+0.85 ms	+25.03%

Idle-resource episode	Peak RSS	Peak PSS	Average CPU	Peak CPU	Peak processes	Sampler
Moli	65.68 MiB	62.49 MiB	0.72%	10.00%	1	✅
Chromium	1288.66 MiB	315.85 MiB	2.46%	99.23%	13	⚠️ PSS partial

Runtime and CDP session contracts — ✅ 26 contract cases; 0 failures

Suite	Cases	Failures	Median case p50	Median case p95	Median peak PSS p50
Synthetic fetch	13	0	19.00 ms	21.07 ms	595.00 KiB
Long-lived CDP session	13	0	18.62 ms	19.08 ms	—

CDP smoke — ✅ 52/52 groups passed; 621 scenarios

Workers: 4 · cumulative group time: 155.38 s · failed groups: 0

Slowest group	Scenarios	Duration	Status
scrollbar-visibility	22	20.41 s	✅
target-lifecycle	24	12.45 s	✅
dom-input	27	10.34 s	✅
chromium-cdp	56	9.88 s	✅
multi-page	45	9.20 s	✅

WebMainBench · 545 pages — ✅ 545/545 completed; 0 unexpected failures

Completed	Successful	Expected DNS failure	Unexpected failures	Panics	Crashes	Timeouts	Empty output
545/545	544	1	0	0	0	0	0

Frozen HTML · Linux · --wait done · 45 seconds per page · no retries. Content quality is evaluated separately.

Diagnostics: webmainbench-results — every page's Markdown, stderr, and exit status. Source run and artifacts

All artifact fields are parsed by the trusted default-branch renderer; missing or invalid inputs remain visible as unavailable.

[github-actions]

Spider Bench A/B

✅ All benchmark browser service runs completed; results are informational.

Public HEAD: 189 / 240 rows; 39 / 48 sites produced rows; 9 unexpected empty sites.

✅ Deterministic fixture contract is clean.

Common ancestor 1f99b9fee9dd → HEAD 8b63b90fdd5b; benchmark order: base-first; workflow: success.

Public 48-site run · informational

Metric	Base	HEAD	Δ
Run status	success	success	—
Extracted rows / contract	184 / 240	189 / 240	+5 (+2.72%)
Contract fill	76.67%	78.75%	+2.08 pp (+2.71%)
Sites with rows	38 / 48	39 / 48	+1 (+2.63%)
Sites meeting row contract	36 / 48	37 / 48	+1 (+2.78%)
Unexpected empty sites	10	9	-1 (-10.00%)
Partial-row sites	2	2	0 (0.00%)
Site outcomes	extracted: 36; HTTP error + rows: 2; empty: 7; HTTP error + empty: 1; snapshot error + empty: 2	extracted: 38; HTTP error + rows: 1; empty: 6; HTTP error + empty: 1; snapshot error + empty: 2	—
Observed duration	48.42 s	44.48 s	-3.94 s (-8.14%)
Estimated CPU	48.76 s	47.52 s	-1.25 s (-2.56%)
Peak PSS	282.44 MiB	275.06 MiB	-7.38 MiB (-2.61%)
Peak RSS	296.08 MiB	277.43 MiB	-18.64 MiB (-6.30%)
Peak threads	26	26	0 (0.00%)
Mismatched sites	0	0	0
Timeout with items	0	0	0

Category rows	Base	HEAD	Δ
news	35 / 40	35 / 40	0
stocks	20 / 40	20 / 40	0
tech	35 / 40	35 / 40	0
sports	27 / 40	32 / 40	+5
games	37 / 40	37 / 40	0
life	30 / 40	30 / 40	0

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 timelines

These 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.

Base

xychart-beta
    title "Base CPU"
    x-axis "Elapsed seconds" [1.037, 1.538, 2.54, 3.542, 4.043, 5.045, 5.545, 6.546, 8.047, 9.548, 11.05, 12.052, 13.554, 15.055, 16.057, 17.559, 19.06, 20.062, 21.564, 23.065, 24.568, 25.568, 27.07, 28.572, 29.573, 31.074, 32.576, 33.576, 35.076, 36.578, 37.579, 39.081, 40.582, 42.084, 43.084, 44.585, 46.086, 47.087]
    y-axis "CPU percent" 0 --> 500
    line [2, 91.79, 107.77, 123.76, 147.75, 123.7, 181.94, 157.74, 165.82, 209.95, 179.76, 147.59, 103.91, 113.85, 87.86, 3.99, 109.86, 129.81, 11.97, 4, 123.83, 120, 143.84, 109.87, 103.81, 145.78, 147.81, 143.99, 126.03, 83.89, 37.96, 145.88, 84.04, 3.99, 74.02, 103.97, 27.98, 135.8]

Loading

xychart-beta
    title "Base memory: RSS then PSS"
    x-axis "Elapsed seconds" [0.036, 1.037, 1.538, 2.54, 3.542, 4.043, 5.045, 5.545, 6.546, 8.047, 9.548, 11.05, 12.052, 13.554, 15.055, 16.057, 17.559, 19.06, 20.062, 21.564, 23.065, 24.568, 25.568, 27.07, 28.572, 29.573, 31.074, 32.576, 33.576, 35.076, 36.578, 37.579, 39.081, 40.582, 42.084, 43.084, 44.585, 46.086, 47.087, 48.424]
    y-axis "Memory MiB" 0 --> 500
    line [12.52, 62.09, 93.77, 121.77, 126.28, 138.91, 157.05, 177.14, 186.21, 176.63, 176.19, 185.32, 169.94, 188, 167.86, 158.4, 147, 159.37, 159.58, 148.73, 143.4, 186.06, 187.5, 164.27, 296.08, 206.02, 157.02, 164.18, 202.63, 176.53, 159.32, 154.29, 205.57, 178.39, 166.27, 183.64, 204, 158.24, 176.09, 0]
    line [9.97, 59.31, 91.04, 119.11, 123.5, 136.13, 154.34, 174.38, 183.44, 173.87, 173.41, 182.54, 167.16, 184.78, 164.67, 155.17, 143.79, 156.1, 156.48, 145.46, 140.21, 182.87, 184.65, 161.09, 282.44, 202.87, 152.9, 161.01, 199.61, 173.35, 156.08, 151.06, 202.54, 175.15, 163.02, 180.05, 198.74, 155, 172.84, 0]

Loading

HEAD

xychart-beta
    title "HEAD CPU"
    x-axis "Elapsed seconds" [1.037, 2.539, 3.539, 5.043, 6.044, 7.546, 8.548, 10.049, 11.05, 12.551, 13.553, 15.055, 16.056, 17.557, 18.559, 19.56, 20.06, 21.061, 22.563, 23.564, 24.566, 25.066, 26.066, 27.067, 28.57, 29.57, 31.071, 32.071, 33.573, 34.574, 36.075, 37.077, 38.579, 39.58, 41.08, 42.081, 43.584, 44.085]
    y-axis "CPU percent" 0 --> 500
    line [2, 127.78, 125.92, 95.75, 171.97, 161.73, 99.83, 159.91, 69.88, 61.92, 131.84, 105.88, 89.93, 35.96, 51.85, 0, 4, 159.85, 91.91, 111.95, 105.8, 109.82, 110.18, 137.92, 45.92, 160, 135.91, 103.99, 125.85, 23.99, 135.96, 125.72, 17.97, 124, 169.87, 83.88, 139.69, 193.59]

Loading

xychart-beta
    title "HEAD memory: RSS then PSS"
    x-axis "Elapsed seconds" [0.037, 1.037, 2.539, 3.539, 5.043, 6.044, 7.546, 8.548, 10.049, 11.05, 12.551, 13.553, 15.055, 16.056, 17.557, 18.559, 19.56, 20.06, 21.061, 22.563, 23.564, 24.566, 25.066, 26.066, 27.067, 28.57, 29.57, 31.071, 32.071, 33.573, 34.574, 36.075, 37.077, 38.579, 39.58, 41.08, 42.081, 43.584, 44.085, 44.484]
    y-axis "Memory MiB" 0 --> 500
    line [12.67, 62.46, 125.87, 135.52, 151.44, 182.59, 199.26, 180.25, 190, 166.07, 188.86, 195.48, 152.53, 156.3, 166.86, 159.26, 153.01, 152.09, 173.6, 161.45, 180.15, 212.04, 277.43, 195.55, 207.33, 157.77, 185.86, 168.58, 158.05, 163.15, 160.5, 207.1, 202.84, 167.74, 194.67, 196.73, 171.48, 202.61, 212.86, 0]
    line [9.72, 59.24, 122.58, 132.31, 148.43, 179.48, 195.17, 176.99, 186.76, 162.82, 185.23, 192.17, 149.24, 153.24, 163.58, 156, 149.75, 148.83, 171.04, 158.32, 176.97, 208.78, 275.06, 192.33, 204.07, 154.49, 182.89, 165.3, 154.73, 159.86, 157.27, 203.81, 199.59, 164.45, 191.38, 193.45, 168.21, 198.84, 209.58, 0]

Loading

Deterministic fixture · required contract

The fixture exercises 15 routes: 8 are expected to emit rows and 7 intentionally exercise empty, loading, or timeout behavior.

Metric	Base	HEAD	Δ
Run status	success	success	—
Extracted rows / contract	40 / 40	40 / 40	0 (0.00%)
Contract fill	100.00%	100.00%	0.00 pp (0.00%)
Sites with rows	8 / 15	8 / 15	0 (0.00%)
Sites meeting row contract	15 / 15	15 / 15	0 (0.00%)
Unexpected empty sites	0	0	0
Partial-row sites	0	0	0
Site outcomes	extracted: 7; HTTP error + rows: 1; empty: 3; timeout + empty: 4	extracted: 7; HTTP error + rows: 1; empty: 3; timeout + empty: 4	—
Observed duration	19.88 s	19.87 s	-0.01 s (-0.03%)
Estimated CPU	0.45 s	0.42 s	-0.03 s (-6.21%)
Peak PSS	66.85 MiB	66.95 MiB	+0.11 MiB (+0.16%)
Peak RSS	69.88 MiB	70.11 MiB	+0.23 MiB (+0.32%)
Peak threads	18	18	0 (0.00%)
Mismatched sites	0	0	0
Timeout with items	0	0	0

Full HTML, JSON, CSV, logs, and page snapshots: workflow run and full spider-bench-results artifact.

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.

[lanyue-llk]

Closing after the consolidation audit. This only repairs regressions introduced by the earlier #1337 visibility stack, which is not being adopted. It has no independent role in the minimal #1501 fix.