Cap-go · riderx · May 2, 2026 · May 1, 2026 · May 1, 2026 · May 1, 2026
diff --git a/.github/workflows/deploy-translation.yml b/.github/workflows/deploy-translation.yml
@@ -53,7 +53,7 @@ jobs:
             if [ "$status" -eq 0 ]; then
               return 0
             fi
-            printf '%s\n' "$output" | grep -Eiq "already exists|already.*${queue}"
+            printf '%s\n' "$output" | grep -Eiq "already exists|already taken|already.*${queue}"
           }
 
           ensure_queue capgo-translation-refresh

diff --git a/apps/translation-worker/package.json b/apps/translation-worker/package.json
@@ -3,7 +3,8 @@
   "private": true,
   "type": "module",
   "scripts": {
-    "check": "tsc --noEmit",
+    "check": "tsc --noEmit && bun run test:parser",
+    "test:parser": "bun run scripts/verify-parser.ts",
     "test": "bun run test:real",
     "test:real": "bun run scripts/verify-real-ai.ts",
     "dev": "wrangler dev -c wrangler.jsonc -c ../web/wrangler.jsonc -c ../docs/wrangler.jsonc",

diff --git a/apps/translation-worker/scripts/verify-parser.ts b/apps/translation-worker/scripts/verify-parser.ts
@@ -0,0 +1,56 @@
+import { __translationWorkerTest } from '../src/index'
+
+function assert(condition: unknown, message: string): void {
+  if (!condition) throw new Error(message)
+}
+
+const html = `<!doctype html>
+<html lang="en">
+  <head>
+    <title>Capgo - Live Updates for Capacitor Apps</title>
+    <script>
+      !function(){for(var n=0;n<o.length;n++)g(o[n]);}();
+    </script>
+  </head>
+  <body>
+    <a href="#main-content">Skip to main content</a>
+    <h1>Ship mobile updates instantly to every user</h1>
+    <svg><svg><text>Do not collect nested SVG text</text></svg><text>Do not collect outer SVG text</text></svg>
+    <p>Translate the paragraph after a nested skipped SVG.</p>
+    <script>if (current < total) console.log(current)</script>
+    <p>Deploy fixes and features without waiting for app store review delays.</p>
+  </body>
+</html>`
+
+const { parts, segments } = __translationWorkerTest.collectSegments(html)
+const bodySegments = segments.filter((segment) => segment.inBody).map((segment) => segment.text)
+
+assert(
+  bodySegments.some((text) => text.includes('Skip to main content')),
+  'Parser did not collect body text after a script with a less-than operator',
+)
+assert(
+  bodySegments.some((text) => text.includes('Ship mobile updates instantly')),
+  'Parser did not collect the body heading',
+)
+assert(
+  bodySegments.every((text) => !text.includes('Do not collect')),
+  'Parser collected text from a nested skipped SVG',
+)
+assert(
+  bodySegments.some((text) => text.includes('paragraph after a nested skipped SVG')),
+  'Parser did not resume body text after a nested skipped SVG',
+)
+assert(
+  bodySegments.some((text) => text.includes('Deploy fixes and features')),
+  'Parser did not collect the body paragraph after a skipped body script',
+)
+
+const translations = segments.map((segment) => (segment.inBody ? `FR: ${segment.text}` : segment.text))
+const stats = __translationWorkerTest.bodyTranslationStats(segments, translations)
+assert(stats.candidateCount > 0, 'Body translation validator found no body candidates')
+assert(stats.changedCount > 0, 'Body translation validator did not detect changed body text')
+
+const rendered = __translationWorkerTest.renderTranslatedHtml(parts, segments, translations)
+assert(rendered.includes('FR: Ship mobile updates instantly to every user'), 'Renderer did not write translated body text')
+assert(rendered.includes('current < total'), 'Renderer changed skipped script content')
diff --git a/apps/translation-worker/scripts/verify-real-ai.ts b/apps/translation-worker/scripts/verify-real-ai.ts
@@ -9,18 +9,31 @@ type ProbePayload = {
     cache?: boolean
     r2?: boolean
   }
+  page?: {
+    path?: string
+    locale?: string
+    segmentCount?: number
+    bodySegmentCount?: number
+    batchCount?: number
+    translatedBatchCount?: number
+    translatedSegmentCount?: number
+    changedCount?: number
+    bodyChecks?: unknown
+    samples?: unknown
+  }
   translations?: unknown
   error?: string
 }
 
 const WORKER_DIR = resolve(dirname(fileURLToPath(import.meta.url)), '..')
 const MODEL = process.env.TRANSLATION_REAL_TEST_MODEL || '@cf/meta/llama-3.1-8b-instruct-fast'
-const TIMEOUT_MS = Number.parseInt(process.env.TRANSLATION_REAL_TEST_TIMEOUT_MS || '180000', 10)
-const REQUEST_TIMEOUT_MS = Math.min(10_000, TIMEOUT_MS)
+const TIMEOUT_MS = Number.parseInt(process.env.TRANSLATION_REAL_TEST_TIMEOUT_MS || '240000', 10)
+const REQUEST_TIMEOUT_MS = Math.min(60_000, TIMEOUT_MS)
 const LOG_LIMIT = 16_000
 const WRANGLER_CONFIG = 'wrangler.real-test.jsonc'
 const DEVELOPMENT_R2_BUCKET = 'capgo-translation-cache-development'
 const SOURCE_TEXTS = ['Ship updates instantly', 'Pricing', 'Keep Capgo, Capacitor, code, API, SDK, CLI, npm, bun, GitHub, and Cloudflare unchanged.']
+const REAL_PAGE_PROBES = ['/', '/docs/'] as const
 
 let wranglerLog = ''
 
@@ -127,7 +140,7 @@ function assertProbePayload(payload: ProbePayload): void {
   }
 }
 
-async function fetchProbe(url: string): Promise<ProbePayload> {
+async function fetchJsonProbe(url: string): Promise<ProbePayload> {
   const controller = new AbortController()
   const timeout = setTimeout(() => controller.abort(), REQUEST_TIMEOUT_MS)
   let response: Response
@@ -151,18 +164,49 @@ async function fetchProbe(url: string): Promise<ProbePayload> {
   }
 
   if (!response.ok) throw new Error(payload.error || `Probe returned HTTP ${response.status}`)
+  return payload
+}
+
+async function fetchRuntimeProbe(url: string): Promise<ProbePayload> {
+  const payload = await fetchJsonProbe(url)
   assertProbePayload(payload)
   return payload
 }
 
+async function fetchRealPageProbe(url: string, path: string): Promise<ProbePayload> {
+  const payload = await fetchJsonProbe(url)
+  if (!payload.ok) throw new Error(payload.error || `Real page probe failed for ${path}`)
+  if (payload.model !== MODEL) throw new Error(`Real page probe used ${payload.model || 'unknown model'} instead of ${MODEL}`)
+
+  const page = payload.page
+  if (!page) throw new Error(`Real page probe returned no page result for ${path}`)
+  if (page.path !== path) throw new Error(`Real page probe returned ${page.path || 'unknown path'} instead of ${path}`)
+  if (page.locale !== 'es') throw new Error(`Real page probe returned ${page.locale || 'unknown locale'} instead of es`)
+  if (!page.segmentCount || page.segmentCount < 1) throw new Error(`Real page probe found no segments for ${path}`)
+  if (!page.bodySegmentCount || page.bodySegmentCount < 1) throw new Error(`Real page probe found no body segments for ${path}`)
+  if (!page.batchCount || page.batchCount < 1) throw new Error(`Real page probe found no batches for ${path}`)
+  if (!page.translatedBatchCount || page.translatedBatchCount < 1) throw new Error(`Real page probe translated no batches for ${path}`)
+  if (!page.translatedSegmentCount || page.translatedSegmentCount < 1) throw new Error(`Real page probe translated no segments for ${path}`)
+  if (!page.changedCount || page.changedCount < 1) throw new Error(`Real page probe left ${path} untranslated`)
+  if (!Array.isArray(page.bodyChecks) || page.bodyChecks.length < 1) throw new Error(`Real page probe returned no translated body checks for ${path}`)
+  if (!Array.isArray(page.samples) || page.samples.length < 1) throw new Error(`Real page probe returned no translated samples for ${path}`)
+
+  return payload
+}
+
 async function exitedCode(process: Bun.Subprocess<'pipe', 'pipe', 'inherit'>): Promise<number | null> {
   return await Promise.race([process.exited, sleep(0).then(() => null)])
 }
 
 await ensureDevelopmentBucket()
 
 const port = await getFreePort()
-const probeUrl = `http://127.0.0.1:${port}/__translation-test__/real-runtime`
+const probeBaseUrl = `http://127.0.0.1:${port}`
+const runtimeProbeUrl = `${probeBaseUrl}/__translation-test__/real-runtime`
+const realPageProbeUrls = REAL_PAGE_PROBES.map((path) => ({
+  path,
+  url: `${probeBaseUrl}/__translation-test__/real-page?path=${encodeURIComponent(path)}&locale=es&batches=2`,
+}))
 const wrangler = Bun.spawn(
   [
     'bunx',
@@ -209,8 +253,11 @@ try {
     if (code !== null) throw new Error(`wrangler dev exited early with code ${code}`)
 
     try {
-      const payload = await fetchProbe(probeUrl)
-      console.log(`Real translation worker probe passed with ${payload.model}`)
+      const payload = await fetchRuntimeProbe(runtimeProbeUrl)
+      for (const probe of realPageProbeUrls) {
+        await fetchRealPageProbe(probe.url, probe.path)
+      }
+      console.log(`Real translation worker probe passed with ${payload.model} on ${REAL_PAGE_PROBES.join(', ')}`)
       passed = true
       break
     } catch (error) {