') + ')', 'gi'); if (regex.test(text)) { found = true; var frag = document.createDocumentFragment(); var parts = text.split(regex); parts.forEach(function(part, i) { if (i % 2 === 0) { frag.appendChild(document.createTextNode(part)); } else { var span = document.createElement('span'); span.className = 'userscript-highlight'; span.textContent = part; frag.appendChild(span); } }); node.parentNode.replaceChild(frag, node); } }); } else if (node.nodeType === 1 && node.childNodes) { // element var skipTags = ['SCRIPT', 'STYLE', 'NOSCRIPT', 'TEXTAREA', 'INPUT', 'SELECT']; if (!skipTags.includes(node.tagName)) { Array.from(node.childNodes).forEach(highlight); } } } highlight(document.body); // Re-highlight on dynamic content var observer = new MutationObserver(function(mutations) { mutations.forEach(function(m) { m.addedNodes.forEach(function(node) { if (node.nodeType === 1 || node.nodeType === 3) highlight(node); }); }); }); observer.observe(document.body, { childList: true, subtree: true }); })(); } } catch(__e) { console.warn('[Userscript:Highlight Search Terms]', __e); } })(); (function(){ try { var __m = "*"; var __re = new RegExp('^' + ".*" + ', 'i'); if (__m === '*' || __re.test(location.href)) { // Strip utm_, fbclid, gclid, etc. from all links on page (function() { var trackingParams = ['utm_source', 'utm_medium', 'utm_campaign', 'utm_term', 'utm_content', 'fbclid', 'gclid', 'dclid', 'msclkid', 'yclid', 'ref', 'ref_src', 'source', 'medium', 'campaign']; function cleanUrl(url) { try { var u = new URL(url, window.location.origin); var changed = false; trackingParams.forEach(function(p) { if (u.searchParams.has(p)) { u.searchParams.delete(p); changed = true; } }); return changed ? u.toString() : url; } catch (e) { return url; } } function cleanLinks() { document.querySelectorAll('a[href]').forEach(function(a) { var clean = cleanUrl(a.href); if (clean !== a.href) a.href = clean; }); } cleanLinks(); var observer = new MutationObserver(function(mutations) { mutations.forEach(function(m) { m.addedNodes.forEach(function(node) { if (node.nodeType === 1) { if (node.tagName === 'A') cleanLinks(); node.querySelectorAll('a[href]').forEach(function(a) { var clean = cleanUrl(a.href); if (clean !== a.href) a.href = clean; }); } }); }); }); observer.observe(document.body, { childList: true, subtree: true }); })(); } } catch(__e) { console.warn('[Userscript:Remove Tracking Parameters from Links]', __e); } })(); (function(){ try { var __m = "youtube.com"; var __re = new RegExp('^' + "youtube\\.com" + ', 'i'); if (__m === '*' || __re.test(location.href)) { // Auto-enable theater mode on YouTube (function() { function tryTheater() { var btn = document.querySelector('button[aria-label="Theater mode"], ytd-player #player button[title="Theater mode"]'); if (btn && !btn.classList.contains('activated')) { btn.click(); } } // Try immediately tryTheater(); // Try after navigation (SPA) var lastUrl = location.href; setInterval(function() { if (location.href !== lastUrl) { lastUrl = location.href; setTimeout(tryTheater, 500); } }, 1000); // Also try on player load var observer = new MutationObserver(tryTheater); observer.observe(document.body, { childList: true, subtree: true }); })(); } } catch(__e) { console.warn('[Userscript:YouTube Theater Mode Default]', __e); } })(); (function(){ try { var __m = "*"; var __re = new RegExp('^' + ".*" + ', 'i'); if (__m === '*' || __re.test(location.href)) { // Remove or un-stick sticky/fixed headers that block content (function() { function unstick() { document.querySelectorAll('header, nav, [role="banner"], .header, .navbar, .sticky, .fixed-top, [style*="position: fixed"], [style*="position:sticky"]').forEach(function(el) { if (el.style.position === 'fixed' || el.style.position === 'sticky' || getComputedStyle(el).position === 'fixed' || getComputedStyle(el).position === 'sticky') { el.style.position = 'static'; el.style.top = 'auto'; el.style.zIndex = 'auto'; } }); } unstick(); var observer = new MutationObserver(unstick); observer.observe(document.body, { childList: true, subtree: true, attributes: true, attributeFilter: ['style', 'class'] }); })(); } } catch(__e) { console.warn('[Userscript:Kill Sticky Headers]', __e); } })(); })(); perf(test): shrink storage capacity and legacy Plan fixtures to minimal crossings by GabrielDrapor · Pull Request #2473 · apache/maka · GitHub
Skip to content

perf(test): shrink storage capacity and legacy Plan fixtures to minimal crossings - #2473

Merged
Astro-Han merged 1 commit into
apache:mainfrom
GabrielDrapor:perf/storage-test-amplification
Aug 8, 2026
Merged

perf(test): shrink storage capacity and legacy Plan fixtures to minimal crossings#2473
Astro-Han merged 1 commit into
apache:mainfrom
GabrielDrapor:perf/storage-test-amplification

Conversation

@GabrielDrapor

@GabrielDraporGabrielDrapor commented Aug 8, 2026

Copy link
Copy Markdown
Contributor

Refs #2388 (fixture-reduction PR; the stress-policy change comes separately per the issue).

What changed

Sandbox-boundary capacity test (sqlite-session-metadata-store.test.ts): each expansion request now carries 15 near-MAX_SANDBOX_BOUNDARY_PATH_CHARS paths instead of 32 shorter ones. Per-request serialized size stays near MAX_SANDBOX_BOUNDARY_SERIALIZED_BYTES (the fewest settles that can cross the 1 MiB execution-boundary cap), but each settle's boundary rescan now walks less than half the accumulated entries.

Legacy Plan ledger fixture (sqlite-workflow-store.test.ts): seedLegacyPlanLedger now exceeds each projection bound by just enough to force truncation — 4 steps (enough to push the shared text budget below the asserted 4,000 bytes), 51 files > PLAN_MAX_FILES_PER_STEP, 21 risks > PLAN_MAX_RISKS, 31-char titles > PLAN_STEP_TITLE_MAX_CHARS — instead of ~500KB of near-max fields per ledger event.

What deliberately did not change

  • Assertions: every truncation/atomicity/persisted-state assertion is untouched; the fixtures still cross every asserted boundary.
  • Catalog capacity test: already minimal — each connection sits at CONNECTION_CATALOG_MAX_ENABLED_MODEL_IDS × CONNECTION_MODEL_ID_MAX_LENGTH (~262KB), so ~16 commits is the floor for crossing the 4 MiB document cap.
  • Vault capacity test: already minimal — secrets are at MAX_SECRET_LENGTH (64 KiB), and the provider auth contract (coordinator.ts) requires one connection per stored secret (oauth_token is github-copilot-only), so 32 commits is the floor for the 2 MiB cap.
  • All three capacity loops break on first rejection, so their loop caps were never a cost driver.

Timing (Apple Silicon, node --test per file)

TestBeforeAfter
rejects an expansion atomically before the complete boundary exceeds capacity2,495ms1,171ms
projects a pre-bound legacy Plan ledger into the bounded Host contract517ms118ms
requires a projected legacy proposal to be revised or abandoned114ms41ms

Full @maka/storage suite: 766 tests, 754 pass / 12 skipped / 0 fail before and after. Suite wall time is bounded by the slowest test file, so the win shows up as ~3.8s less per-file CPU rather than lower end-to-end wall time.

Co-Authored-By: Claude noreply@anthropic.com

https://claude.ai/code/session_01Ac5rv6WUKWtMZgQb5QPN16

@GabrielDrapor@claude
perf(test): shrink storage capacity and legacy Plan fixtures to minim…
baab8c9
…al crossings
Refs apache#2388.
The sandbox-boundary capacity test now reaches MAX_EXECUTION_BOUNDARY_SERIALIZED_BYTES
with 15 near-max-length paths per request instead of 32 shorter ones — same
per-request byte size, less than half the accumulated entries each settle has
to rescan (2,495ms -> 1,171ms).
The legacy Plan ledger fixture now exceeds each projection bound by just
enough to force truncation (4 steps, 51 files, 21 risks, 31-char step titles)
instead of ~500KB of near-max fields per event (517ms -> 118ms, 114ms -> 41ms).
The catalog and vault capacity tests are unchanged: their per-item sizes
already sit at CONNECTION_CATALOG_MAX_ENABLED_MODEL_IDS x
CONNECTION_MODEL_ID_MAX_LENGTH and MAX_SECRET_LENGTH, and the provider auth
contract requires one connection per stored secret, so their commit counts are
already the minimum that crosses the aggregate byte limits.
Co-Authored-By: Claude <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Ac5rv6WUKWtMZgQb5QPN16
This was referenced Aug 8, 2026
Merged
Closed
Astro-Han
Astro-Han approved these changes Aug 8, 2026
View reviewed changes

@Astro-HanAstro-Han left a comment

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Thanks — this keeps the important capacity, atomicity, persistence, and truncation assertions intact while materially reducing fixture amplification. The direction is sound and CI is green.

Two non-blocking P3 notes:

  • The legacy Plan fixture can be reduced from four steps to three and still cross the shared text budget while preserving the existing files, risks, title, and truncation assertions.
  • The sandbox-boundary comment overstates “fewest settles”: 16 near-max paths still fit below the per-request cap and cross the aggregate boundary one settle earlier. Since #2388 originally marked this capacity test as already minimal, it may be worth either tightening the fixture/wording or leaving that hunk unchanged in a follow-up.

Neither point weakens the current assertions or blocks this improvement. Approved.

简体中文

感谢这次优化——容量、原子拒绝、持久状态和截断断言都保留了，同时明显降低了 fixture 放大成本。方向正确，CI 也已全绿。

两个不阻塞合并的 P3 建议：

  • Legacy Plan fixture 从 4 个 steps 改为 3 个，仍然足以越过共享文本预算，并保持现有 files、risks、title 和 truncation 断言。
  • Sandbox boundary 注释中的 “fewest settles” 不够准确：每轮 16 个接近上限的路径仍低于单请求容量，并能少一次 settle 越过总容量。考虑到 #2388 原先把这项容量测试列为已经最小，可以在后续进一步收紧 fixture/描述，或者保持该 hunk 不变。

这两点都没有削弱当前断言，也不阻塞本次优化。Approve。

@Astro-Han
Astro-Han merged commit c878316 into apache:mainAug 8, 2026
11 checks passed
@UncertaintyDeterminesYou4ndMeUncertaintyDeterminesYou4ndMe mentioned this pull request Aug 8, 2026
Merged
Sign up for freeto join this conversation on GitHub. Already have an account? Sign in to comment

Reviewers

@Astro-HanAstro-HanAstro-Han approved these changes

Assignees

No one assigned

Labels

None yet

Projects

None yet

Milestone

No milestone

Development

Successfully merging this pull request may close these issues.

2 participants

@GabrielDrapor@Astro-Han