fix(search): preserve Gmail reads when replies disappear (#8388)

* fix(search): preserve Gmail reads when replies disappear

* fix(search): handle replies missing during Gmail verification

Author: waleedlatif1

apps/sim/lib/sim-search/live/google.ts

@@ -435,20 +435,30 @@ export async function readGmail(
   for (const candidate of latest) {
     options.signal.throwIfAborted()
     const candidateId = string(candidate.id)
-    if (
-      !(await options.verify({
-        id: candidateId,
-        accessMetadata: { id: candidateId, labelIds: candidate.labelIds },
-      }))
-    ) {
-      warnings.add('messages outside the source search scope were omitted')
-      continue
+    let row: Record<string, unknown>
+    try {
+      if (
+        !(await options.verify({
+          id: candidateId,
+          accessMetadata: { id: candidateId, labelIds: candidate.labelIds },
+        }))
+      ) {
+        warnings.add('messages outside the source search scope were omitted')
+        continue
+      }
+      row = object(
+        await client.json(`/gmail/v1/users/me/messages/${segment(candidateId)}`, {
+          query: { format: 'full' },
+        })
+      )
+    } catch (error) {
+      options.signal.throwIfAborted()
+      if (error instanceof NativeSearchError && error.httpStatus === 404) {
+        warnings.add('some messages are no longer available')
+        continue
+      }
+      throw error
     }
-    const row = object(
-      await client.json(`/gmail/v1/users/me/messages/${segment(candidateId)}`, {
-        query: { format: 'full' },
-      })
-    )
     if (row.id !== candidateId || row.threadId !== threadId)
       throw new NativeSearchError(
         'unavailable',

apps/sim/scripts/test-search-google-content-e2e.ts

@@ -79,6 +79,7 @@ const mailbox: GmailMessage[] = Array.from({ length: 12 }, (_, index) => ({
     body: { data: Buffer.from(`Synthetic evidence ${index}.`).toString('base64url') },
   },
 }))
+const messageFailures = new Map<string, { status: number; format: string }>()
 const mediaStarted = createDeferred<void>()
 const mediaClosed = createDeferred<void>()
 const server = http.createServer((request, response) => {
@@ -92,6 +93,14 @@ const server = http.createServer((request, response) => {
     const row = mailbox.find((message) => message.id === id)
     const isThread = url.pathname.startsWith('/gmail/v1/users/me/threads/')
     const isAttachment = url.pathname.includes('/attachments/')
+    const failure = !isThread ? messageFailures.get(id) : undefined
+    if (failure && url.searchParams.get('format') === failure.format) {
+      requests.push({ id, resource: 'message', status: failure.status })
+      response
+        .writeHead(failure.status, { 'Content-Type': 'application/json' })
+        .end(JSON.stringify({ error: { code: failure.status } }))
+      return
+    }
     const threadMessages = isThread ? mailbox.filter((message) => message.threadId === id) : []
     const metadata = (message: (typeof mailbox)[number]) => ({
       ...message,
@@ -284,6 +293,81 @@ try {
   server.listen(0, '127.0.0.1', () => started.resolve())
   await started.promise
   origin = `http://127.0.0.1:${(server.address() as AddressInfo).port}`
+  for (const { name, failedId, status, expectedStatus } of [
+    { name: 'healthy conversation', failedId: '', status: 200 },
+    { name: 'missing sibling', failedId: 'reply', status: 404 },
+    { name: 'missing anchor', failedId: 'anchor', status: 404, expectedStatus: 'unavailable' },
+    { name: 'expired credentials', failedId: 'reply', status: 401, expectedStatus: 'reconnect' },
+    { name: 'denied credentials', failedId: 'reply', status: 403, expectedStatus: 'reconnect' },
+    { name: 'rate limit', failedId: 'reply', status: 429, expectedStatus: 'rate_limited' },
+    { name: 'server failure', failedId: 'reply', status: 500, expectedStatus: 'unavailable' },
+  ]) {
+    for (const format of failedId === 'reply' ? ['metadata', 'full'] : ['full']) {
+      await check(
+        `Gmail conversation handles ${name} during ${format} reads without hiding unrelated failures`,
+        async () => {
+          const before = mailbox.length
+          mailbox.push(
+            ...['anchor', 'retained', 'reply'].map((id, index) => ({
+              id,
+              threadId: 'failure-conversation',
+              labelIds: ['INBOX'],
+              internalDate: String(Date.UTC(2026, 0, 1, 12, index)),
+              payload: {
+                mimeType: 'text/plain',
+                body: { data: Buffer.from(`Synthetic ${id} evidence.`).toString('base64url') },
+              },
+            }))
+          )
+          if (failedId) messageFailures.set(failedId, { status, format })
+          try {
+            const signal = AbortSignal.timeout(5000)
+            const api = createNativeClient({ origin, accessToken: TOKEN, signal })
+            const policy = {
+              ...defaultLiveSearchPolicy('gmail'),
+              mode: 'selected' as const,
+              included: ['INBOX'],
+            }
+            const verify = createPolicyVerifier('gmail', policy, api, origin)
+            assert.ok(await verify({ id: 'anchor' }))
+            const reading = readGmail(api, 'anchor', {
+              policy,
+              signal,
+              verify: (reference) =>
+                verify(reference, format === 'full' ? reference.accessMetadata : undefined),
+            })
+            if (expectedStatus) {
+              await assert.rejects(
+                reading,
+                (error: unknown) =>
+                  error instanceof NativeSearchError && error.status === expectedStatus
+              )
+              return
+            }
+            const document = await reading
+            assert.equal(document.id, 'anchor')
+            assert.ok(document.content.includes('Synthetic anchor evidence.'))
+            assert.ok(document.content.includes('Synthetic retained evidence.'))
+            assert.equal(document.content.includes('Synthetic reply evidence.'), status === 200)
+            assert.equal(document.content.includes('Coverage incomplete:'), status === 404)
+            assert.deepEqual(
+              document.accessDependencies,
+              status === 404 ? [{ id: 'retained' }] : [{ id: 'retained' }, { id: 'reply' }]
+            )
+            const current = createPolicyVerifier('gmail', policy, api, origin, undefined, {
+              fresh: true,
+            })
+            assert.ok(await current({ id: document.id }))
+            for (const dependency of document.accessDependencies ?? [])
+              assert.ok(await current(dependency))
+          } finally {
+            mailbox.splice(before)
+            messageFailures.clear()
+          }
+        }
+      )
+    }
+  }
   await check(
     'Gmail conversation and fresh selected-label checks fit one native request budget',
     async () => {