docs: remove stale 2.0.0 benchmark snapshot from llms.txt and README by developerz-ai[bot] · Pull Request #27 · sebyx07/ruby2html

developerz-ai · 2026-11-10T13:48:11Z

What

The quoted ratios describe the 2.0.0 release (README snapshot dated
2026-10-31, same as CHANGELOG) while version.rb is 2.1.0; AGENTS.md's
docs rule forbids hardcoded benchmark numbers. Replaced the snapshot
tables with pointers to the CI benchmark workflow (.github/workflows/
benchmark.yml) and bundle exec rake bench, kept the code-level
"How it is fast" claims, left CHANGELOG.md's dated release notes alone.

Acceptance

llms.txt has no benchmark figures labelled with a release version: the 'Performance at 2.0.0' sentence is gone, replaced by a pointer to CI/README Performance
README.md has no 'Snapshot at 2.0.0' section and no 'ruby2html 2.0' table rows; grep for '2.0.0|1.35x|2.4–16x' in README.md finds no benchmark prose
Both files name the benchmark workflow / bundle exec rake bench as the source of truth: README keeps its .github/workflows/benchmark.yml link, llms.txt points at CI + README Performance
Every path named in the rewritten sections exists: .github/workflows/benchmark.yml, bench/render_bench.rb, bench/escape_bench.rb, spec/benchmarks/requests_spec.rb (all confirmed present in this survey)
Remaining factual claims in the touched sections match code: SSE2/NEON escape scan (ext/ruby2html/html_escape.c:28,62), attributes written straight into the buffer with no cache (ext/ruby2html/attributes.c:2,14), pure-Ruby fallback and lazy loading per AGENTS.md code map
CHANGELOG.md and version.rb untouched; README.md and llms.txt remain valid markdown/plaintext with the Performance anchor heading intact

Changes

llms.txt (edit) — Line 3 intro blockquote: delete the stale 'Performance at 2.0.0: ... 1.35x ... 2.4-16x ... require costs about 0.4 MB' sentence; replace with one pointer sentence: benchmarks run in CI on every push/PR and README "Performance" holds the current numbers; run locally with bundle exec rake bench. Keep all other lines byte-identical (file is 132 lines).
README.md (edit) — Line 23 'Why ruby2html' bullet: drop hardcoded ratios (1.35x Erubi, 2.4-16x Phlex, 40-90% allocations); keep 'Fastest Ruby HTML renderer we've measured' and point at #performance. Line 24: drop 'require costs about 0.4 MB and 19 ms' clause. Performance section: keep intro paragraph (line 199, already links .github/workflows/benchmark.yml and rake bench); delete the '### Snapshot at 2.0.0 release (2026-10-31)' subsection (lines 201-238: page-render table, Rails-request table, 5.9x escaping line, footprint table, Phlex footnote) and replace with a short 'What is measured' list naming b [trimmed]

Verification

./bin/check passed (exit 0) in 26s.

What the editors ran to see the change working, and what they saw:

grep README.md for '2\.0\.0|1\.35x|2\.4.16x|0\.4 MB|Snapshot|19 ms|2026-10-31' (docs): 0 hits — no stale figures or snapshot section remain in README.md
grep llms.txt for '2\.0\.0|1\.35x|0\.4 MB' (docs): 0 hits; line 3 now reads "Performance: benchmarks run in CI on every push and pull request, and README "Performance" holds the current numbers; run them locally with bundle exec rake bench"
code_read README.md:20-29 and 196-211 after the edits (docs): Why bullet keeps 'Fastest Ruby HTML renderer we've measured' + Performance link, require-cost clause gone; Performance section = intro (benchmark.yml link) + '### What is measured' list of bench/render_bench.rb, bench/tag_bench.rb, bench/escape_bench.rb, spec/benchmarks/requests_spec.rb + 'How it is fast' list intact; single blank lines
code_glob bench/*.rb and code_read of each script's header + spec/benchmarks/requests_spec.rb (docs): all four cited benchmark paths exist and each bullet's description matches the script's own header comments

🤖 Opened by developerz.ai for task tsk_1e0d885c9b903cf9f46031138834403b on sebyx07/ruby2html.

Raised automatically by the raw_task_triage lane.

🤖 developerz.ai, an automated agent, not a human — this PR is one slice of a stacked plan.
Dispatched automatically by a developerz.ai maintainer lane
Part 1/1

The quoted ratios describe the 2.0.0 release (README snapshot dated 2026-10-31, same as CHANGELOG) while version.rb is 2.1.0; AGENTS.md's docs rule forbids hardcoded benchmark numbers. Replaced the snapshot tables with pointers to the CI benchmark workflow (.github/workflows/ benchmark.yml) and `bundle exec rake bench`, kept the code-level "How it is fast" claims, left CHANGELOG.md's dated release notes alone. Dz-Task-Id: tsk_1e0d885c9b903cf9f46031138834403b

github-actions · 2026-11-10T13:50:54Z

Benchmarks: aarch64 (linux)

ruby 4.0.9 (2026-10-23 revision 3e93b23afc) +PRISM [aarch64-linux] · commit 6e8b67ee6c · 2.0s per report
Δ is against main b69cba8f67 (same runner type; expect a few % noise).

escape_bench (C ext) · peak RSS 120 MB

Benchmark	i/s	ms / call	allocs / call	vs fastest	Δ vs main
escape_html short clean (10B)	617,829 ± 0.2%	0.0016	0.00	fastest	-0.1%
escape_html medium clean (200B)	336,366 ± 0.2%	0.0030	0.00	1.84x slower	-0.9%
append short clean (10B)	208,939 ± 0.2%	0.0048	0.02	2.96x slower	+0.8%
append short dirty (10B)	161,433 ± 0.2%	0.0062	0.00	3.83x slower	+0.9%
append medium clean (200B)	147,781 ± 0.2%	0.0068	0.00	4.18x slower	+0.3%
escape_html short dirty (10B)	108,721 ± 3.4%	0.0092	100	5.68x slower	-2.1%
append medium dirty (200B)	73,534 ± 0.2%	0.0136	0.00	8.40x slower	-0.3%
escape_html medium dirty (200B)	58,717 ± 0.4%	0.0170	100	10.52x slower	+0.3%
escape_html long clean (10KB)	11,478 ± 0.2%	0.0871	0.00	53.83x slower	-0.2%
append long clean (10KB)	9,087 ± 0.5%	0.1101	0.00	67.99x slower	+0.1%
append long dirty (10KB)	1,205 ± 0.3%	0.8296	0.00	512.54x slower	+0.3%
escape_html long dirty (10KB)	1,133 ± 3.2%	0.8829	100	545.47x slower	+0.0%

escape_bench (pure Ruby) · peak RSS 150 MB

Benchmark	i/s	ms / call	allocs / call	vs fastest	Δ vs main
escape_html short clean (10B)	164,643 ± 0.2%	0.0061	0.00	fastest	+0.0%
append short clean (10B)	104,622 ± 0.2%	0.0096	0.02	1.57x slower	+0.2%
escape_html short dirty (10B)	69,162 ± 0.4%	0.0145	100	2.38x slower	-0.2%
append short dirty (10B)	57,421 ± 0.4%	0.0174	100	2.87x slower	-0.6%
escape_html medium clean (200B)	56,877 ± 0.2%	0.0176	0.00	2.89x slower	+0.0%
append medium clean (200B)	45,560 ± 0.2%	0.0219	0.00	3.61x slower	+0.1%
escape_html medium dirty (200B)	24,769 ± 1.0%	0.0404	200	6.65x slower	-0.9%
append medium dirty (200B)	22,102 ± 0.8%	0.0452	200	7.45x slower	-0.5%
escape_html long clean (10KB)	1,632 ± 0.3%	0.6126	0.00	100.86x slower	+0.1%
append long clean (10KB)	1,557 ± 0.3%	0.6421	0.00	105.71x slower	+0.1%
escape_html long dirty (10KB)	1,086 ± 0.6%	0.9205	200	151.55x slower	+1.3%
append long dirty (10KB)	948 ± 4.2%	1.0544	200	173.59x slower	-0.2%

render_bench (C ext) · peak RSS 131 MB

Benchmark	i/s	ms / call	allocs / call	vs fastest	Δ vs main
erubi	7,446 ± 1.0%	0.1343	713	fastest	-2.1%
ruby2html	5,750 ± 0.7%	0.1739	419	1.29x slower	-0.6%
phlex	555 ± 0.9%	1.8015	6,374	13.41x slower	-1.3%

render_bench (pure Ruby) · peak RSS 131 MB

Benchmark	i/s	ms / call	allocs / call	vs fastest	Δ vs main
erubi	7,468 ± 0.8%	0.1339	713	fastest	-1.1%
ruby2html	2,310 ± 0.5%	0.4330	1,029	3.23x slower	-0.7%
phlex	562 ± 1.1%	1.7779	6,374	13.28x slower	-0.6%

tag_bench (C ext) · peak RSS 39 MB

Benchmark	i/s	ms / call	allocs / call	vs fastest	Δ vs main
br	214,582 ± 1.6%	0.0047	7.00	fastest	-1.7%
li(text)	158,192 ± 0.6%	0.0063	7.02	1.36x slower	-0.4%
li(dirty text)	125,323 ± 0.6%	0.0080	7.00	1.71x slower	+0.8%
div { block }	98,495 ± 0.6%	0.0102	7.00	2.18x slower	-0.1%
span(text, class:)	30,927 ± 1.7%	0.0323	207	6.94x slower	-0.3%
div(class:) { block }	29,298 ± 0.4%	0.0341	207	7.32x slower	-0.3%
a(text, href:, class:)	23,513 ± 1.1%	0.0425	207	9.13x slower	-1.0%

tag_bench (pure Ruby) · peak RSS 26 MB

Benchmark	i/s	ms / call	allocs / call	vs fastest	Δ vs main
br	81,100 ± 2.3%	0.0123	107	fastest	-1.0%
div { block }	63,753 ± 0.4%	0.0157	107	1.27x slower	+1.0%
li(text)	54,809 ± 0.4%	0.0182	107	1.48x slower	+1.3%
li(dirty text)	41,810 ± 0.3%	0.0239	207	1.94x slower	+0.9%
div(class:) { block }	15,746 ± 0.4%	0.0635	207	5.15x slower	+0.6%
span(text, class:)	15,458 ± 0.5%	0.0647	207	5.25x slower	+0.0%
a(text, href:, class:)	9,801 ± 0.7%	0.1020	207	8.27x slower	+0.5%

Rails request (C ext) · peak RSS 243 MB

Benchmark	i/s	ms / call	allocs / call	vs fastest	Δ vs main
Ruby2html::Component	493 ± 2.6%	2.0276	4,452	fastest	+4.3%
ruby2html .html.rb templates	487 ± 2.3%	2.0526	4,588	1.01x slower	+3.2%
Slim	483 ± 5.8%	2.0696	4,126	1.02x slower	+4.0%
ViewComponent + ruby2html	461 ± 3.0%	2.1676	4,309	1.07x slower	+3.7%
ERB	422 ± 16.8%	2.3675	2,946	1.17x slower	-5.7%
Phlex	420 ± 1.2%	2.3786	5,807	1.17x slower	+1.6%

github-actions · 2026-11-10T13:50:58Z

Benchmarks: x86_64 (linux)

ruby 4.0.9 (2026-10-23 revision 3e93b23afc) +PRISM [x86_64-linux] · commit 6e8b67ee6c · 2.0s per report
Δ is against main b69cba8f67 (same runner type; expect a few % noise).

escape_bench (C ext) · peak RSS 120 MB

Benchmark	i/s	ms / call	allocs / call	vs fastest	Δ vs main
escape_html short clean (10B)	874,444 ± 2.9%	0.0011	0.00	fastest	+24.0%
escape_html medium clean (200B)	533,690 ± 0.2%	0.0019	0.00	1.64x slower	+27.4%
append short clean (10B)	335,227 ± 0.5%	0.0030	0.02	2.61x slower	+32.6%
append short dirty (10B)	248,897 ± 0.7%	0.0040	0.00	3.51x slower	+34.8%
append medium clean (200B)	236,110 ± 0.2%	0.0042	0.00	3.70x slower	+36.9%
escape_html short dirty (10B)	173,911 ± 0.7%	0.0058	100	5.03x slower	+35.6%
append medium dirty (200B)	117,233 ± 0.6%	0.0085	0.00	7.46x slower	+28.1%
escape_html medium dirty (200B)	89,987 ± 1.0%	0.0111	100	9.72x slower	+22.3%
escape_html long clean (10KB)	18,486 ± 0.8%	0.0541	0.00	47.30x slower	+13.9%
append long clean (10KB)	14,303 ± 2.8%	0.0699	0.00	61.14x slower	+25.2%
append long dirty (10KB)	1,901 ± 0.7%	0.5259	0.00	459.91x slower	+25.4%
escape_html long dirty (10KB)	1,822 ± 3.1%	0.5490	100	480.06x slower	+30.4%

escape_bench (pure Ruby) · peak RSS 146 MB

Benchmark	i/s	ms / call	allocs / call	vs fastest	Δ vs main
escape_html short clean (10B)	262,464 ± 1.9%	0.0038	0.00	fastest	+30.6%
append short clean (10B)	193,858 ± 5.4%	0.0052	0.02	1.35x slower	+27.0%
escape_html short dirty (10B)	113,563 ± 3.3%	0.0088	100	2.31x slower	+28.8%
append short dirty (10B)	96,403 ± 1.9%	0.0104	100	2.72x slower	+44.2%
escape_html medium clean (200B)	69,688 ± 2.2%	0.0143	0.00	3.77x slower	+21.3%
append medium clean (200B)	61,180 ± 1.1%	0.0163	0.00	4.29x slower	+22.9%
escape_html medium dirty (200B)	36,381 ± 2.2%	0.0275	200	7.21x slower	+42.9%
append medium dirty (200B)	32,540 ± 2.2%	0.0307	200	8.07x slower	+41.7%
escape_html long clean (10KB)	1,766 ± 0.5%	0.5662	0.00	148.60x slower	+14.1%
append long clean (10KB)	1,706 ± 0.3%	0.5862	0.00	153.85x slower	+14.4%
escape_html long dirty (10KB)	1,181 ± 0.3%	0.8469	200	222.27x slower	+15.2%
append long dirty (10KB)	1,011 ± 8.4%	0.9893	200	259.66x slower	+10.4%

render_bench (C ext) · peak RSS 132 MB

Benchmark	i/s	ms / call	allocs / call	vs fastest	Δ vs main
erubi	10,351 ± 4.2%	0.0966	713	fastest	+28.5%
ruby2html	7,794 ± 0.8%	0.1283	419	1.33x slower	+44.9%
phlex	654 ± 3.8%	1.5292	6,374	15.83x slower	+32.4%

render_bench (pure Ruby) · peak RSS 132 MB

Benchmark	i/s	ms / call	allocs / call	vs fastest	Δ vs main
erubi	10,385 ± 1.2%	0.0963	713	fastest	+28.2%
ruby2html	3,012 ± 0.9%	0.3320	1,029	3.45x slower	+39.0%
phlex	660 ± 4.2%	1.5149	6,374	15.73x slower	+33.5%

tag_bench (C ext) · peak RSS 39 MB

Benchmark	i/s	ms / call	allocs / call	vs fastest	Δ vs main
br	312,416 ± 1.9%	0.0032	7.00	fastest	+21.3%
li(text)	211,501 ± 1.9%	0.0047	7.02	1.48x slower	+24.3%
li(dirty text)	179,557 ± 5.8%	0.0056	7.00	1.74x slower	+23.7%
div { block }	148,308 ± 1.1%	0.0067	7.00	2.11x slower	+33.0%
span(text, class:)	47,671 ± 1.5%	0.0210	207	6.55x slower	+46.6%
div(class:) { block }	42,167 ± 5.0%	0.0237	207	7.41x slower	+52.1%
a(text, href:, class:)	34,030 ± 0.7%	0.0294	207	9.18x slower	+45.4%

tag_bench (pure Ruby) · peak RSS 27 MB

Benchmark	i/s	ms / call	allocs / call	vs fastest	Δ vs main
br	133,714 ± 1.3%	0.0075	107	fastest	+34.7%
div { block }	106,891 ± 2.8%	0.0094	107	1.25x slower	+36.6%
li(text)	94,238 ± 0.7%	0.0106	107	1.42x slower	+30.6%
li(dirty text)	65,716 ± 5.3%	0.0152	207	2.03x slower	+30.5%
span(text, class:)	20,989 ± 0.6%	0.0476	207	6.37x slower	+45.0%
div(class:) { block }	19,534 ± 0.9%	0.0512	207	6.85x slower	+47.7%
a(text, href:, class:)	13,995 ± 0.6%	0.0715	207	9.55x slower	+46.8%

Rails request (C ext) · peak RSS 232 MB

Benchmark	i/s	ms / call	allocs / call	vs fastest	Δ vs main
Slim	724 ± 8.8%	1.3821	3,996	fastest	+34.4%
Ruby2html::Component	692 ± 1.3%	1.4458	4,312	1.05x slower	+31.3%
ruby2html .html.rb templates	679 ± 2.2%	1.4719	4,443	1.06x slower	+33.3%
ViewComponent + ruby2html	653 ± 0.8%	1.5309	4,169	1.11x slower	+34.2%
ERB	650 ± 11.4%	1.5391	2,858	1.11x slower	+36.9%
Phlex	575 ± 2.6%	1.7384	5,609	1.26x slower	+26.0%

developerz-ai · 2026-11-10T14:07:55Z

✅ Reviewed — nothing blocking

0 actionable comment(s) · grounded on your code

_{⚠️ 1 config note(s): test files were not shown to the reviewer — this diff touched no path recognised as a test, so no assertion was available to read the change against}

_{⏱ 27s wall clock · glm-5.3-flash via router · 1 model call(s) · 119 output token(s)}

_{🤖 developerz.ai — automated review, running on your model and your box.}

developerz-ai

Review summary — 2 file(s), 0 finding(s).

Docs-only edit removing stale 2.0.0 benchmark figures from README and llms.txt, replaced by pointers to CI benchmarks and rake bench; the shown files match the acceptance criteria and no contradictions remain.

test files were not shown to the reviewer — this diff touched no path recognised as a test, so no assertion was available to read the change against

🤖 developerz.ai review — automated, running on your model and your box. What is this? · Reviewed by router/glm-5.3-flash — 16.3k in / 119 out over 1 call, on your flat-rate plan: no per-token charge (~<$0.01 at list price).

developerz-ai Bot added dz/fleet-pr Opened by developerz.ai, an automated agent, from a fleet run. dz/model:glm-5.3-flash developerz.ai: model that wrote this bot PR. dz/provider:router.ai.developerz.ai developerz.ai: provider that served this bot PR. labels Nov 10, 2026

developerz-ai Bot commented Nov 10, 2026

View reviewed changes

developerz-ai Bot merged commit c566c6e into main Nov 10, 2026
30 checks passed

developerz-ai Bot deleted the dz/remove-stale-hardcoded-benchmark-numbers-from--1e0d885c branch November 10, 2026 14:10

Navigation Menu

Repository navigation

docs: remove stale 2.0.0 benchmark snapshot from llms.txt and README - #27

docs: remove stale 2.0.0 benchmark snapshot from llms.txt and README#27
developerz-ai[bot] merged 1 commit into
mainfrom
dz/remove-stale-hardcoded-benchmark-numbers-from--1e0d885c

developerz-ai Bot commented Nov 10, 2026 •

edited

Loading

Uh oh!

github-actions Bot commented Nov 10, 2026

Uh oh!

github-actions Bot commented Nov 10, 2026

Uh oh!

developerz-ai Bot commented Nov 10, 2026 •

edited

Loading

Uh oh!

developerz-ai Bot left a comment

Uh oh!

Uh oh!

Reviewers

Uh oh!

Assignees

Labels

Projects

Milestone

Development

Uh oh!

0 participants

Conversation

developerz-ai Bot commented Nov 10, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

What

Acceptance

Changes

Verification

Uh oh!

github-actions Bot commented Nov 10, 2026

Benchmarks: aarch64 (linux)

escape_bench (C ext) · peak RSS 120 MB

escape_bench (pure Ruby) · peak RSS 150 MB

render_bench (C ext) · peak RSS 131 MB

render_bench (pure Ruby) · peak RSS 131 MB

tag_bench (C ext) · peak RSS 39 MB

tag_bench (pure Ruby) · peak RSS 26 MB

Rails request (C ext) · peak RSS 243 MB

Uh oh!

github-actions Bot commented Nov 10, 2026

Benchmarks: x86_64 (linux)

escape_bench (C ext) · peak RSS 120 MB

escape_bench (pure Ruby) · peak RSS 146 MB

render_bench (C ext) · peak RSS 132 MB

render_bench (pure Ruby) · peak RSS 132 MB

tag_bench (C ext) · peak RSS 39 MB

tag_bench (pure Ruby) · peak RSS 27 MB

Rails request (C ext) · peak RSS 232 MB

Uh oh!

developerz-ai Bot commented Nov 10, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

✅ Reviewed — nothing blocking

Uh oh!

developerz-ai Bot left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Reviewers

Uh oh!

Assignees

Labels

Projects

Milestone

Development

Uh oh!

0 participants

developerz-ai Bot commented Nov 10, 2026 •

edited

Loading

developerz-ai Bot commented Nov 10, 2026 •

edited

Loading