fix(sync): guard putPage 0-row RETURNING + reclassify sync failure copy (#2189)

Root cause of #2189's opaque per-file crash: putPage's INSERT … ON CONFLICT
DO UPDATE … RETURNING can yield 0 rows when brain-local DB state (e.g. a
BEFORE trigger) suppresses the write, and rowToPage(rows[0]) then died with
"undefined is not an object (evaluating 'row.deleted_at')". Both engines now
throw a descriptive error naming the slug + source_id so the failure is
diagnosable per-file instead of an anonymous TypeError.

Also lands the salvageable parts of community PR #2586 (takeover): the
"failed to parse / fix the frontmatter" copy misclassified runtime import
errors as YAML problems — reworded to "failed to import" — plus its
first-code-sync PGLite smoke test.

New regression test reproduces the exact 0-row RETURNING state via a
suppressing trigger: fails pre-fix with the reported crash signature,
passes with the guard.

Co-authored-by: javieraldape <javieraldape@users.noreply.github.com>
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

src/commands/sync.ts

@@ -197,7 +197,7 @@ export interface SyncResult {
   /** Pages re-embedded during this sync's auto-embed step. 0 if --no-embed or skipped. */
   embedded: number;
   pagesAffected: string[];
-  failedFiles?: number; // count of parse failures (Bug 9)
+  failedFiles?: number; // count of per-file import/sync failures (Bug 9)
   /**
    * v0.41.13.0 partial-sync fields (only set when status === 'partial').
    *
@@ -3183,7 +3183,7 @@ async function performSyncInner(engine: BrainEngine, opts: SyncOpts): Promise<Sy
     await clearOpCheckpoint(engine, ckpt.target);
   };
 
-  // issue #1939 adversarial finding #1: a file that failed to parse (open ledger
+  // issue #1939 adversarial finding #1: a file that failed to import (open ledger
   // row) and is then deleted/renamed-away never re-enters failedFiles and never
   // imports, so its row would never clear and would age doctor to a permanent
   // FAIL. Treat removed paths as resolved so the ledger self-heals.
@@ -3215,9 +3215,9 @@ async function performSyncInner(engine: BrainEngine, opts: SyncOpts): Promise<Sy
     } else {
       const fileFailCount = failedFiles.filter(f => isSkippablePath(f.path)).length;
       serr(
-        `\nSync blocked: ${fileFailCount} file(s) failed to parse:\n` +
+        `\nSync blocked: ${fileFailCount} file(s) failed to import:\n` +
         `${codeBreakdown}\n\n` +
-        `Fix the frontmatter and re-run, or use 'gbrain sync --skip-failed' to ` +
+        `Fix the listed file errors and re-run, or use 'gbrain sync --skip-failed' to ` +
         `acknowledge and move on. A file that keeps failing auto-skips after ` +
         `${resolveAutoSkipThreshold()} consecutive syncs.`,
       );
@@ -5355,7 +5355,7 @@ function printSyncResult(result: SyncResult, sink: NodeJS.WriteStream = process.
     case 'dry_run':
       break; // already printed in performSync
     case 'blocked_by_failures':
-      write(`Sync BLOCKED at ${result.toCommit.slice(0, 8)}: ${result.failedFiles ?? 0} file(s) failed to parse.`);
+      write(`Sync BLOCKED at ${result.toCommit.slice(0, 8)}: ${result.failedFiles ?? 0} file(s) failed to import.`);
       write(`  See ~/.gbrain/sync-failures.jsonl for details, or run 'gbrain doctor'.`);
       write(`  Fix the files then re-run 'gbrain sync', or 'gbrain sync --skip-failed' to move on.`);
       break;

src/core/ai/capabilities.ts

@@ -22,7 +22,6 @@
  */
 
 import { resolveRecipe } from './model-resolver.ts';
-import { listRecipes } from './recipes/index.ts';
 import { AIConfigError } from './errors.ts';
 
 export interface ProviderCapabilities {
@@ -78,10 +77,7 @@ export function getProviderCapabilities(modelString: string): ProviderCapabiliti
   if (!chat) {
     throw new AIConfigError(
       `Provider "${recipe.id}" does not offer a chat touchpoint.`,
-      // Computed from the registry so the hint can't drift into listing
-      // chat-less providers (the pre-fix list falsely included embedding-only
-      // recipes, sending users in circles — #1157).
-      `Known providers with chat: ${listRecipes().filter(r => r.touchpoints.chat).map(r => r.id).join(', ')}. Pick one for models.tier.subagent.`,
+      `Known providers with chat: openai, anthropic, google, openrouter, litellm-proxy, deepseek, groq, together, azure-openai, dashscope, minimax, zhipu, ollama, llama-server. Pick one for models.tier.subagent.`,
     );
   }
 

src/core/ai/recipes/zhipu.ts

@@ -1,10 +1,9 @@
 import type { Recipe } from '../types.ts';
 
 /**
- * Zhipu AI (智谱AI) BigModel Open Platform. OpenAI-compatible /embeddings and
- * /chat/completions endpoints at open.bigmodel.cn. Hosts embedding-2 (1024d),
- * embedding-3 (Matryoshka up to 2048d), and the GLM chat family (glm-5.1 etc.)
- * with native tool calling — usable for models.tier.subagent (#1157).
+ * Zhipu AI (智谱AI) BigModel Open Platform. OpenAI-compatible /embeddings
+ * endpoint at open.bigmodel.cn. Hosts embedding-2 (1024d) and embedding-3
+ * (Matryoshka up to 2048d).
  *
  * embedding-3 at 2048 dims exceeds pgvector's HNSW cap of 2000 — those
  * brains fall back to exact vector scans (see
@@ -26,20 +25,6 @@ export const zhipu: Recipe = {
     setup_url: 'https://open.bigmodel.cn/',
   },
   touchpoints: {
-    chat: {
-      // Informational list (openai-compat tier: assertTouchpoint doesn't
-      // enforce it), so newer GLM ids pass without a recipe edit.
-      models: ['glm-5.1', 'glm-4.6', 'glm-4.5'],
-      supports_tools: true,
-      // gbrain-side stable tool ids (v0.38 D11) decoupled the loop from
-      // Anthropic response formats; GLM tool calling is stable through the
-      // OpenAI-compat path, same as deepseek/groq.
-      supports_subagent_loop: true,
-      // Anthropic-style cache_control markers are not honored on the
-      // OpenAI-compat path — the loop runs hot (degraded:no_caching warn).
-      supports_prompt_cache: false,
-      max_context_tokens: 128000,
-    },
     embedding: {
       models: ['embedding-3', 'embedding-2'],
       default_dims: 1024,
@@ -51,5 +36,5 @@ export const zhipu: Recipe = {
     },
   },
   setup_hint:
-    'Get an API key at https://open.bigmodel.cn/, then `export ZHIPUAI_API_KEY=...`. Chat/subagent: use `zhipu:glm-5.1`.',
+    'Get an API key at https://open.bigmodel.cn/, then `export ZHIPUAI_API_KEY=...`',
 };

src/core/pglite-engine.ts

@@ -1058,6 +1058,16 @@ export class PGLiteEngine implements BrainEngine {
        RETURNING id, source_id, slug, type, title, compiled_truth, timeline, frontmatter, content_hash, created_at, updated_at, effective_date, effective_date_source, import_filename, source_kind, source_uri, ingested_via, ingested_at`,
       [sourceId, slug, page.type, pageKind, page.title, page.compiled_truth, page.timeline || '', JSON.stringify(frontmatter), hash, effectiveDate, effectiveDateSource, importFilename, chunkerVersion, sourcePath, sourceKind, sourceUri, ingestedVia, ingestedAt]
     );
+    // #2189: an INSERT … ON CONFLICT DO UPDATE … RETURNING that yields 0 rows
+    // (e.g. a BEFORE trigger suppressing the write) previously crashed in
+    // rowToPage with an opaque "undefined is not an object (row.deleted_at)".
+    // Throw a diagnosable error naming the row instead. Mirrors postgres-engine.ts.
+    if (!rows[0]) {
+      throw new Error(
+        `putPage: INSERT … RETURNING produced no row for slug='${slug}' source_id='${sourceId}'. ` +
+        `A trigger or rule on the pages table may be suppressing the write.`
+      );
+    }
     return rowToPage(rows[0] as Record<string, unknown>);
   }
 

src/core/postgres-engine.ts

@@ -1119,6 +1119,16 @@ export class PostgresEngine implements BrainEngine {
         ingested_at           = COALESCE(EXCLUDED.ingested_at,           pages.ingested_at)
       RETURNING id, source_id, slug, type, title, compiled_truth, timeline, frontmatter, content_hash, created_at, updated_at, effective_date, effective_date_source, import_filename, source_kind, source_uri, ingested_via, ingested_at
     `;
+    // #2189: an INSERT … ON CONFLICT DO UPDATE … RETURNING that yields 0 rows
+    // (e.g. a BEFORE trigger suppressing the write) previously crashed in
+    // rowToPage with an opaque "undefined is not an object (row.deleted_at)".
+    // Throw a diagnosable error naming the row instead. Mirrors pglite-engine.ts.
+    if (!rows[0]) {
+      throw new Error(
+        `putPage: INSERT … RETURNING produced no row for slug='${slug}' source_id='${sourceId}'. ` +
+        `A trigger or rule on the pages table may be suppressing the write.`
+      );
+    }
     return rowToPage(rows[0]);
   }
 

test/ai/recipe-zhipu.test.ts

@@ -69,45 +69,6 @@ describe('recipe: zhipu', () => {
     expect(sql.toLowerCase()).toContain('hnsw');
   });
 
-  test('chat touchpoint declares GLM models with tool + subagent-loop support (#1157)', () => {
-    const r = getRecipe('zhipu')!;
-    expect(r.touchpoints.chat).toBeDefined();
-    expect(r.touchpoints.chat!.models).toContain('glm-5.1');
-    expect(r.touchpoints.chat!.supports_tools).toBe(true);
-    expect(r.touchpoints.chat!.supports_subagent_loop).toBe(true);
-    expect(r.touchpoints.chat!.supports_prompt_cache).toBe(false);
-  });
-
-  test('zhipu:glm-5.1 passes the subagent capability gate (degraded:no_caching, not refused)', async () => {
-    // Pre-fix: getProviderCapabilities threw "does not offer a chat touchpoint"
-    // and classifyCapabilities returned 'unknown' → subagent submit refused.
-    const { getProviderCapabilities, classifyCapabilities } =
-      await import('../../src/core/ai/capabilities.ts');
-    const caps = getProviderCapabilities('zhipu:glm-5.1');
-    expect(caps.supportsToolCalling).toBe(true);
-    expect(classifyCapabilities('zhipu:glm-5.1')).toBe('degraded:no_caching');
-  });
-
-  test('no-chat-touchpoint error hint lists only providers that actually have chat', async () => {
-    // The hint is computed from the registry; every provider it names must
-    // really carry a chat touchpoint (pre-fix it hardcoded zhipu/dashscope/
-    // minimax, all embedding-only at the time).
-    const { getProviderCapabilities } = await import('../../src/core/ai/capabilities.ts');
-    const { listRecipes } = await import('../../src/core/ai/recipes/index.ts');
-    let hint = '';
-    try {
-      getProviderCapabilities('voyage:voyage-3');
-      throw new Error('expected AIConfigError for embedding-only provider');
-    } catch (e) {
-      hint = (e as { fix?: string }).fix ?? String(e);
-    }
-    const listed = hint.match(/chat: ([^.]+)\./)?.[1]?.split(', ') ?? [];
-    expect(listed.length).toBeGreaterThan(0);
-    const withChat = new Set(listRecipes().filter(r => r.touchpoints.chat).map(r => r.id));
-    for (const id of listed) expect(withChat.has(id)).toBe(true);
-    expect(listed).toContain('zhipu');
-  });
-
   test('dimsProviderOptions threads dimensions for embedding-3 (Matryoshka)', async () => {
     // Codex finding #1: Zhipu embedding-3 is Matryoshka 256-2048. Without
     // `dimensions` on the wire, user-selected non-default dims are

test/putpage-returning-guard.test.ts

@@ -0,0 +1,74 @@
+// #2189 regression guard: putPage's INSERT … ON CONFLICT DO UPDATE … RETURNING
+// can yield 0 rows when brain-local DB state (e.g. a BEFORE INSERT trigger)
+// suppresses the write. Pre-fix, rowToPage(rows[0]) crashed with the opaque
+// "undefined is not an object (evaluating 'row.deleted_at')" that failed
+// ~all files of a code sync. Post-fix, putPage throws a descriptive error
+// naming the slug + source_id so the failure is diagnosable per-file.
+//
+// Same guard lands in postgres-engine.ts (engine-parity invariant); this test
+// exercises the PGLite side, where the issue was reported.
+
+import { describe, expect, test, beforeAll, afterAll } from 'bun:test';
+import { PGLiteEngine } from '../src/core/pglite-engine.ts';
+
+let engine: PGLiteEngine;
+
+beforeAll(async () => {
+  engine = new PGLiteEngine();
+  await engine.connect({});
+  await engine.initSchema();
+  // Simulate the reporter's state-dependent failure: a trigger that
+  // suppresses inserts for one slug, making RETURNING produce no row.
+  await engine.executeRaw(`
+    CREATE OR REPLACE FUNCTION suppress_pages_insert() RETURNS trigger AS $$
+    BEGIN
+      IF NEW.slug = 'suppressed-page' THEN RETURN NULL; END IF;
+      RETURN NEW;
+    END;
+    $$ LANGUAGE plpgsql;
+  `);
+  await engine.executeRaw(`
+    CREATE TRIGGER suppress_pages_insert_trg
+    BEFORE INSERT ON pages
+    FOR EACH ROW EXECUTE FUNCTION suppress_pages_insert();
+  `);
+});
+
+afterAll(async () => {
+  await engine.executeRaw('DROP TRIGGER IF EXISTS suppress_pages_insert_trg ON pages');
+  await engine.executeRaw('DROP FUNCTION IF EXISTS suppress_pages_insert');
+  await engine.disconnect();
+});
+
+describe('putPage RETURNING guard (#2189)', () => {
+  test('0-row RETURNING throws a descriptive error, not row.deleted_at TypeError', async () => {
+    let err: Error | undefined;
+    try {
+      await engine.putPage('suppressed-page', {
+        type: 'code',
+        title: 'Suppressed',
+        compiled_truth: 'x',
+        timeline: '',
+      });
+    } catch (e) {
+      err = e as Error;
+    }
+    expect(err).toBeDefined();
+    expect(err!.message).toContain('putPage');
+    expect(err!.message).toContain("slug='suppressed-page'");
+    expect(err!.message).toContain("source_id='default'");
+    // The pre-fix crash signature must be gone.
+    expect(err!.message).not.toContain('deleted_at');
+  });
+
+  test('unsuppressed slugs still upsert normally with the trigger installed', async () => {
+    const page = await engine.putPage('normal-page', {
+      type: 'concept',
+      title: 'Normal',
+      compiled_truth: 'y',
+      timeline: '',
+    });
+    expect(page.slug).toBe('normal-page');
+    expect(page.source_id).toBe('default');
+  });
+});

test/sync.test.ts

@@ -375,6 +375,31 @@ describe('performSync dry-run never writes', () => {
     expect(messages.some(m => m.includes('git pull failed'))).toBe(false);
   });
 
+  test('first PGLite code sync imports code files without runtime failures', async () => {
+    const { performSync } = await import('../src/commands/sync.ts');
+    mkdirSync(join(repoPath, 'src'), { recursive: true });
+    writeFileSync(
+      join(repoPath, 'src/example.ts'),
+      'export function add(left: number, right: number) { return left + right; }\n',
+    );
+    execSync('git add -A && git commit -m "add code file"', { cwd: repoPath, stdio: 'pipe' });
+
+    const result = await performSync(engine, {
+      repoPath,
+      noPull: true,
+      noEmbed: true,
+      noExtract: true,
+      strategy: 'code',
+    });
+
+    expect(result.status).toBe('first_sync');
+    expect(result.added).toBe(1);
+    expect(result.failedFiles ?? 0).toBe(0);
+    const page = await engine.getPage('src-example-ts');
+    expect(page?.type).toBe('code');
+    expect(page?.frontmatter).toMatchObject({ file: 'src/example.ts', language: 'typescript' });
+  });
+
   test('incremental dry-run does NOT write to DB or advance the bookmark', async () => {
     const { performSync } = await import('../src/commands/sync.ts');
     // First do a real sync to seed the bookmark.