Skip to content

Model image-backed object fallback - #96

Closed
ldm0 wants to merge 1 commit into
mainfrom
layout-object-image-content
Closed

Model image-backed object fallback#96
ldm0 wants to merge 1 commit into
mainfrom
layout-object-image-content

Conversation

@ldm0

@ldm0 ldm0 commented Aug 16, 2026

Copy link
Copy Markdown
Member

No description provided.

@github-actions

github-actions Bot commented Aug 16, 2026

Copy link
Copy Markdown

Spider Bench A/B

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

Public HEAD: 179 / 240 rows; 37 / 48 sites produced rows; 11 unexpected empty sites.

✅ Deterministic fixture diagnostic is clean.

Common ancestor 2692c5ce7663 → HEAD 4524e23e5e02; benchmark order: base-first; workflow: success.

Public 48-site run · informational

Metric Base HEAD Δ
Run status success success —
Extracted rows / contract 179 / 240 179 / 240 0 (0.00%)
Contract fill 74.58% 74.58% 0.00 pp (0.00%)
Sites with rows 37 / 48 37 / 48 0 (0.00%)
Sites meeting row contract 35 / 48 35 / 48 0 (0.00%)
Unexpected empty sites 11 11 0 (0.00%)
Partial-row sites 2 2 0 (0.00%)
Site outcomes extracted: 36; HTTP error + rows: 1; empty: 7; HTTP error + empty: 1; snapshot error + empty: 3 extracted: 36; HTTP error + rows: 1; empty: 7; HTTP error + empty: 1; snapshot error + empty: 3 —
Observed duration 40.30 s 37.17 s -3.12 s (-7.75%)
Estimated CPU 36.61 s 34.36 s -2.25 s (-6.15%)
Peak PSS 498.87 MiB 478.32 MiB -20.54 MiB (-4.12%)
Peak RSS 501.52 MiB 474.81 MiB -26.71 MiB (-5.33%)
Peak threads 27 27 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 27 / 40 0
games 37 / 40 37 / 40 0
life 25 / 40 25 / 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" [1.043, 2.545, 3.547, 4.049, 4.549, 5.551, 7.053, 8.054, 9.056, 10.057, 11.559, 12.06, 12.56, 13.56, 14.561, 16.064, 17.064, 18.064, 19.067, 20.567, 21.57, 22.57, 23.571, 24.572, 25.073, 26.074, 27.076, 28.077, 29.579, 30.581, 31.582, 32.583, 34.085, 35.085, 36.086, 37.089, 38.591, 39.593]
    y-axis "CPU percent" 0 --> 200
    line [49.9, 117.81, 115.8, 171.58, 135.78, 139.84, 155.75, 43.93, 97.85, 129.81, 72.01, 0, 25.98, 83.98, 43.92, 45.96, 99.93, 69.93, 125.82, 109.91, 131.7, 105.92, 101.9, 120.07, 131.73, 57.93, 133.6, 119.81, 75.9, 25.97, 61.94, 123.93, 135.9, 77.97, 51.91, 23.95, 47.95, 115.83]
Loading 
xychart-beta
    title "Base memory: RSS then PSS"
    x-axis "Elapsed seconds" [0.042, 1.043, 2.545, 3.547, 4.049, 4.549, 5.551, 7.053, 8.054, 9.056, 10.057, 11.559, 12.06, 12.56, 13.56, 14.561, 16.064, 17.064, 18.064, 19.067, 20.567, 21.57, 22.57, 23.571, 24.572, 25.073, 26.074, 27.076, 28.077, 29.579, 30.581, 31.582, 32.583, 34.085, 35.085, 36.086, 37.089, 38.591, 39.593, 40.298]
    y-axis "Memory MiB" 0 --> 1000
    line [23.3, 120.15, 174.61, 191.93, 219.84, 246.47, 232.2, 249.93, 230.69, 216.87, 251.17, 208.6, 206.73, 182.38, 181.63, 191.67, 190.94, 221, 207.44, 219.54, 368.21, 334.27, 326.71, 364.66, 501.52, 342.86, 213.18, 237.28, 238.88, 209.07, 213.19, 221.46, 307.37, 271.21, 257.72, 255.7, 240, 235.08, 266.2, 0]
    line [20.96, 117.33, 171.79, 189.22, 217.14, 244.34, 229.53, 247.32, 228.03, 214.21, 249.78, 205.95, 204.08, 179.72, 178.99, 189.04, 188.28, 218.36, 205.12, 225.39, 365.71, 331.61, 324.06, 362.03, 498.87, 333.21, 210.52, 234.63, 236.23, 206.42, 210.52, 218.79, 304.72, 268.54, 255.06, 253.1, 237.52, 232.37, 263.48, 0]
Loading

HEAD

xychart-beta
    title "HEAD CPU"
    x-axis "Elapsed seconds" [1.035, 2.036, 3.037, 4.04, 5.041, 6.042, 6.543, 7.544, 8.545, 9.546, 10.546, 11.547, 12.548, 13.55, 14.551, 15.553, 16.554, 17.555, 19.057, 20.057, 21.058, 22.059, 22.56, 23.06, 24.061, 25.061, 26.062, 27.063, 28.064, 29.067, 30.068, 31.57, 32.572, 33.072, 33.573, 34.574, 35.575, 36.577]
    y-axis "CPU percent" 0 --> 200
    line [53.97, 105.9, 107.97, 139.64, 129.83, 107.85, 159.76, 79.89, 110.02, 117.81, 109.85, 68.03, 45.93, 4, 53.91, 67.8, 103.89, 101.85, 125.94, 125.95, 113.73, 98.02, 121.79, 135.95, 61.92, 129.89, 112.02, 45.95, 79.79, 7.98, 73.9, 87.84, 18, 2, 65.92, 155.82, 53.95, 129.82]
Loading 
xychart-beta
    title "HEAD memory: RSS then PSS"
    x-axis "Elapsed seconds" [0.035, 1.035, 2.036, 3.037, 4.04, 5.041, 6.042, 6.543, 7.544, 8.545, 9.546, 10.546, 11.547, 12.548, 13.55, 14.551, 15.553, 16.554, 17.555, 19.057, 20.057, 21.058, 22.059, 22.56, 23.06, 24.061, 25.061, 26.062, 27.063, 28.064, 29.067, 30.068, 31.57, 32.572, 33.072, 33.573, 34.574, 35.575, 36.577, 37.175]
    y-axis "Memory MiB" 0 --> 1000
    line [23.39, 122.37, 149.1, 190.3, 223.51, 233.95, 227.9, 245.32, 222.62, 246.07, 252.79, 218.57, 208.52, 190.39, 189.98, 181.95, 210.95, 204.84, 257.98, 313.89, 321.21, 333.24, 391.78, 474.81, 373.48, 219.91, 211.95, 227.47, 205.76, 206.67, 204.96, 218.4, 248.79, 212.09, 211.87, 222.68, 245.89, 216.16, 255.85, 0]
    line [21.21, 119.84, 146.64, 187.84, 220.92, 231.48, 225.81, 242.85, 220.17, 243.61, 250.3, 215.89, 206.03, 187.89, 187.49, 179.48, 208.49, 202.38, 255.52, 311.43, 318.64, 330.89, 389.32, 478.32, 371, 217.42, 209.47, 225.01, 203.27, 204.47, 202.47, 215.91, 246.48, 209.6, 209.38, 220.19, 243.41, 213.62, 253.16, 0]
Loading

Deterministic fixture · diagnostic

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 20.10 s 20.04 s -0.06 s (-0.32%)
Estimated CPU 0.47 s 0.48 s +0.01 s (+1.90%)
Peak PSS 85.11 MiB 86.73 MiB +1.62 MiB (+1.91%)
Peak RSS 87.92 MiB 89.46 MiB +1.54 MiB (+1.75%)
Peak threads 20 18 -2 (-10.00%)
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.

Spider Bench is informational: site failures, missing benchmark reports, failed browser service runs, fixture diagnostics, timing, and performance deltas do not fail the workflow. Build, runner, and artifact infrastructure errors can still fail it.

@github-actions

github-actions Bot commented Aug 16, 2026 •
edited
Loading

Copy link
Copy Markdown

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 3cbc3c2b49 → HEAD e4bfdeec66; benchmark order: base-first; workflow: success.

Session resilience

Metric Base HEAD
Attempted / planned 200 / 200 200 / 200
Direct observable pass 200 / 200 200 / 200
Failures / recovered 0 / 0 0 / 0
Recovery failures 0 0
Lifecycle order violations 0 0
Network order violations 0 0
Wall time 255.3 s 246.0 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 258.0 MiB 256.8 MiB -1.2 MiB
Periodic peak PSS 255.2 MiB 253.8 MiB -1.4 MiB
RSS · first 10 avg 216.2 MiB 204.7 MiB -11.5 MiB
RSS · last 10 avg 219.0 MiB 228.3 MiB +9.3 MiB
RSS · last − first window +2.8 MiB +23.7 MiB +20.8 MiB
RSS · warm slope / 100 nav +2.4 MiB +4.8 MiB +2.4 MiB
PSS · first 10 avg 213.4 MiB 202.5 MiB -11.0 MiB
PSS · last 10 avg 216.8 MiB 226.0 MiB +9.2 MiB
PSS · last − first window +3.4 MiB +23.6 MiB +20.2 MiB
PSS · warm slope / 100 nav +2.8 MiB +4.8 MiB +2.0 MiB
Peak file descriptors 53 54 1
Peak threads 28 28 0

HEAD memory by 50-navigation quarter

HEAD range Avg RSS Final RSS Peak RSS Avg PSS Final PSS Peak FD
1–50 215.7 MiB 195.3 MiB 255.5 MiB 213.4 MiB 192.8 MiB 44
51–100 219.3 MiB 212.2 MiB 254.8 MiB 216.9 MiB 209.8 MiB 44
101–150 222.1 MiB 197.3 MiB 255.8 MiB 219.6 MiB 194.8 MiB 44
151–200 225.3 MiB 218.0 MiB 255.6 MiB 223.0 MiB 215.7 MiB 42

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.

@ldm0
ldm0 force-pushed the layout-object-image-content branch from 4524e23 to ecfa299 Compare August 25, 2026 13:23
@ldm0
ldm0 force-pushed the layout-object-image-content branch from ecfa299 to e4bfdee Compare August 25, 2026 15:24
@github-actions

github-actions Bot commented Aug 25, 2026

Copy link
Copy Markdown

Spider Bench A/B

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

Public HEAD: 179 / 240 rows; 37 / 48 sites produced rows; 11 unexpected empty sites.

✅ Deterministic fixture diagnostic is clean.

Common ancestor 3cbc3c2b495a → HEAD e4bfdeec6607; benchmark order: base-first; workflow: success.

Public 48-site run · informational

Metric Base HEAD Δ
Run status success success —
Extracted rows / contract 179 / 240 179 / 240 0 (0.00%)
Contract fill 74.58% 74.58% 0.00 pp (0.00%)
Sites with rows 37 / 48 37 / 48 0 (0.00%)
Sites meeting row contract 35 / 48 35 / 48 0 (0.00%)
Unexpected empty sites 11 11 0 (0.00%)
Partial-row sites 2 2 0 (0.00%)
Site outcomes extracted: 35; HTTP error + rows: 2; empty: 7; HTTP error + empty: 1; snapshot error + empty: 3 extracted: 36; HTTP error + rows: 1; empty: 8; HTTP error + empty: 1; snapshot error + empty: 2 —
Observed duration 55.69 s 36.83 s -18.86 s (-33.87%)
Estimated CPU 37.23 s 31.99 s -5.24 s (-14.07%)
Peak PSS 288.57 MiB 314.18 MiB +25.61 MiB (+8.87%)
Peak RSS 288.13 MiB 316.46 MiB +28.33 MiB (+9.83%)
Peak threads 29 26 -3 (-10.34%)
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 27 / 40 0
games 37 / 40 37 / 40 0
life 25 / 40 25 / 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" [1.538, 3.039, 4.545, 5.546, 6.046, 8.053, 9.553, 10.054, 11.057, 12.559, 14.06, 15.562, 17.063, 18.566, 20.067, 22.07, 23.572, 25.073, 26.575, 28.077, 29.579, 31.08, 32.081, 32.581, 34.083, 36.085, 37.586, 39.087, 40.589, 42.091, 43.591, 45.093, 46.595, 48.097, 50.099, 51.606, 53.111, 54.614]
    y-axis "CPU percent" 0 --> 200
    line [111.65, 101.94, 157.67, 93.85, 109.96, 83.9, 146.02, 189.74, 127.64, 135.86, 0, 2, 11.99, 15.95, 49.95, 51.88, 11.98, 17.98, 9.99, 15.98, 89.96, 141.94, 107.81, 130.01, 43.87, 101.9, 39.96, 29.95, 0, 95.78, 35.98, 69.91, 7.99, 111.83, 10, 2, 71.43, 113.83]
Loading 
xychart-beta
    title "Base memory: RSS then PSS"
    x-axis "Elapsed seconds" [0.036, 1.538, 3.039, 4.545, 5.546, 6.046, 8.053, 9.553, 10.054, 11.057, 12.559, 14.06, 15.562, 17.063, 18.566, 20.067, 22.07, 23.572, 25.073, 26.575, 28.077, 29.579, 31.08, 32.081, 32.581, 34.083, 36.085, 37.586, 39.087, 40.589, 42.091, 43.591, 45.093, 46.595, 48.097, 50.099, 51.606, 53.111, 54.614, 55.694]
    y-axis "Memory MiB" 0 --> 500
    line [23.48, 161.05, 179.18, 223.69, 225.83, 243.25, 201.06, 219.71, 206.81, 225.24, 228.04, 197.91, 195.11, 196.73, 195.48, 206.68, 185.44, 217.57, 219.09, 219.21, 214.71, 202.6, 235.58, 288.13, 220.55, 200.76, 231.81, 244.91, 237.14, 234.13, 238.52, 209.93, 218.46, 204.97, 272.21, 207.39, 204.37, 198.52, 238.88, 0]
    line [21.29, 158.58, 176.57, 221.07, 223.25, 240.67, 198.52, 217.3, 204.32, 222.69, 228.1, 195.59, 192.79, 194.41, 193.16, 204.35, 183.12, 215.25, 216.76, 216.88, 212.39, 200.28, 233.4, 288.57, 218.47, 198.38, 229.69, 242.59, 234.82, 231.81, 236.2, 207.61, 216.13, 202.66, 269.75, 205.06, 202.01, 196.23, 236.64, 0]
Loading

HEAD

xychart-beta
    title "HEAD CPU"
    x-axis "Elapsed seconds" [1.036, 2.038, 3.039, 4.04, 4.542, 5.044, 5.545, 6.046, 7.047, 8.048, 9.552, 10.553, 11.554, 12.56, 13.563, 14.563, 15.565, 16.566, 17.567, 18.568, 19.57, 20.571, 21.572, 22.576, 23.578, 24.58, 25.584, 26.588, 28.09, 29.091, 29.592, 30.093, 31.093, 32.095, 33.097, 34.097, 35.1, 36.101]
    y-axis "CPU percent" 0 --> 200
    line [53.98, 105.82, 105.91, 159.74, 119.63, 161.23, 119.69, 119.86, 149.84, 81.86, 125.33, 101.91, 67.92, 65.35, 45.88, 12, 49.89, 43.98, 109.92, 115.93, 101.97, 119.87, 62, 111.43, 113.77, 13.96, 69.64, 77.58, 63.95, 105.91, 153.8, 103.81, 107.85, 5.99, 2, 143.86, 77.77, 123.8]
Loading 
xychart-beta
    title "HEAD memory: RSS then PSS"
    x-axis "Elapsed seconds" [0.036, 1.036, 2.038, 3.039, 4.04, 4.542, 5.044, 5.545, 6.046, 7.047, 8.048, 9.552, 10.553, 11.554, 12.56, 13.563, 14.563, 15.565, 16.566, 17.567, 18.568, 19.57, 20.571, 21.572, 22.576, 23.578, 24.58, 25.584, 26.588, 28.09, 29.091, 29.592, 30.093, 31.093, 32.095, 33.097, 34.097, 35.1, 36.101, 36.831]
    y-axis "Memory MiB" 0 --> 500
    line [22.92, 120.77, 144.93, 192.9, 226.99, 238.1, 236.75, 229.91, 218.91, 245.5, 219.23, 214.19, 210.59, 207.39, 175.82, 182.26, 182.94, 172.86, 187.44, 205.77, 213.61, 251.3, 222.29, 197.52, 243.75, 220.73, 195.52, 185.68, 188.88, 190.23, 297.84, 316.46, 315.07, 224.22, 203.76, 218.43, 221.41, 210.67, 231.51, 0]
    line [20.92, 118.52, 142.69, 185.79, 224.69, 235.83, 234.42, 227.65, 216.61, 242.61, 216.94, 211.91, 208.3, 205.1, 173.53, 179.97, 180.65, 170.57, 184.13, 203.48, 211.7, 249.01, 219.87, 195.23, 241.45, 218.44, 193.23, 183.4, 186.59, 187.95, 295.56, 314.18, 312.79, 221.94, 201.45, 216.09, 219.08, 208.38, 229.22, 0]
Loading

Deterministic fixture · diagnostic

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 19.94 s 19.99 s +0.05 s (+0.25%)
Estimated CPU 0.46 s 0.54 s +0.08 s (+17.10%)
Peak PSS 88.73 MiB 82.22 MiB -6.51 MiB (-7.34%)
Peak RSS 91.21 MiB 84.71 MiB -6.50 MiB (-7.13%)
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.

Spider Bench is informational: site failures, missing benchmark reports, failed browser service runs, fixture diagnostics, timing, and performance deltas do not fail the workflow. Build, runner, and artifact infrastructure errors can still fail it.

@ldm0 ldm0 closed this Aug 26, 2026
@ldm0
ldm0 deleted the layout-object-image-content branch August 26, 2026 19:13
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Reviewers

No reviews

Assignees

No one assigned

Labels

None yet

Projects

None yet

Milestone

No milestone

Development

Successfully merging this pull request may close these issues.

1 participant

@ldm0