fix(ci): paginate report artifacts (#3797)

Co-authored-by: Gergo Magyar <gergomagyar0@gmail.com>

.github/workflows/ci-report.yml

@@ -43,14 +43,16 @@ jobs:
             const path = require('path');
             const runId = context.payload.workflow_run.id;
 
-            const allArtifacts = await github.rest.actions.listWorkflowRunArtifacts({
+            // Sharded CI runs can upload more than one page of artifacts.
+            const allArtifacts = await github.paginate(github.rest.actions.listWorkflowRunArtifacts, {
               owner: context.repo.owner,
               repo: context.repo.repo,
               run_id: runId,
+              per_page: 100,
             });
 
             async function downloadArtifact(name, dest) {
-              const match = allArtifacts.data.artifacts.find(a => a.name === name);
+              const match = allArtifacts.find(a => a.name === name);
               if (!match) {
                 core.warning(`Artifact "${name}" not found`);
                 return false;
@@ -74,6 +76,8 @@ jobs:
       - name: Extract artifacts
         shell: bash
         run: |
+          # With no artifacts, let the metadata step skip the report cleanly.
+          mkdir -p "$RUNNER_TEMP/dl"
           cd "$RUNNER_TEMP/dl"
           # Extract each artifact into its own directory to avoid filename collisions
           for z in *.zip; do
@@ -160,13 +164,14 @@ jobs:
 
             // Try each run until we find a downloadable test-reports artifact
             for (const run of runs.data.workflow_runs) {
-              const artifacts = await github.rest.actions.listWorkflowRunArtifacts({
+              const artifacts = await github.paginate(github.rest.actions.listWorkflowRunArtifacts, {
                 owner: context.repo.owner,
                 repo: context.repo.repo,
                 run_id: run.id,
+                per_page: 100,
               });
 
-              const testReports = artifacts.data.artifacts.find(a => a.name === 'test-reports');
+              const testReports = artifacts.find(a => a.name === 'test-reports');
               if (!testReports) {
                 core.info(`Run ${run.id}: no test-reports artifact, trying next`);
                 continue;

gitnexus/test/unit/ci-report-workflow.test.ts

@@ -0,0 +1,127 @@
+import { execFileSync } from 'node:child_process';
+import { mkdtempSync, readFileSync, readdirSync, rmSync } from 'node:fs';
+import { createRequire } from 'node:module';
+import { tmpdir } from 'node:os';
+import path from 'node:path';
+import { load } from 'js-yaml';
+import { afterEach, beforeEach, describe, expect, it, vi } from 'vitest';
+
+const workflow = load(
+  readFileSync(path.resolve(__dirname, '../../../.github/workflows/ci-report.yml'), 'utf8'),
+) as {
+  jobs: Record<string, { steps: Array<{ name: string; run?: string; with?: { script: string } }> }>;
+};
+const steps = workflow.jobs['pr-report'].steps;
+const requireCjs = createRequire(import.meta.url);
+const AsyncFunction = Object.getPrototypeOf(async function () {}).constructor as new (
+  ...arguments_: string[]
+) => (...arguments_: unknown[]) => Promise<void>;
+
+type Artifact = { id: number; name: string };
+let temp: string;
+
+beforeEach(() => {
+  temp = mkdtempSync(path.join(tmpdir(), 'gitnexus-ci-report-'));
+});
+
+afterEach(() => {
+  rmSync(temp, { recursive: true, force: true });
+});
+
+function fixture(artifacts: Artifact[]) {
+  const listWorkflowRunArtifacts = vi.fn(async ({ per_page = 30 }: { per_page?: number }) => ({
+    data: { total_count: artifacts.length, artifacts: artifacts.slice(0, per_page) },
+  }));
+  const downloadArtifact = vi.fn(async ({ artifact_id }: { artifact_id: number }) => ({
+    data: Buffer.from(`archive-${artifact_id}`),
+  }));
+  const github = {
+    rest: {
+      actions: {
+        listWorkflowRunArtifacts,
+        downloadArtifact,
+        listWorkflowRuns: vi.fn(async () => ({ data: { workflow_runs: [{ id: 123 }] } })),
+      },
+    },
+    // Octokit's paginate returns the flattened artifact array, unlike the REST response.
+    paginate: vi.fn(async () => artifacts),
+  };
+  const core = { warning: vi.fn(), info: vi.fn(), setOutput: vi.fn() };
+  const context = {
+    repo: { owner: 'owner', repo: 'repo' },
+    payload: { workflow_run: { id: 456 } },
+  };
+  return {
+    github,
+    core,
+    downloadArtifact,
+    async run(name: string) {
+      const script = steps.find((step) => step.name === name)?.with?.script;
+      if (!script) throw new Error(`Missing script: ${name}`);
+      const execute = new AsyncFunction('github', 'context', 'core', 'require', 'process', script);
+      await execute(github, context, core, requireCjs, { env: { RUNNER_TEMP: temp } });
+    },
+  };
+}
+
+function precedingArtifacts(count: number): Artifact[] {
+  return Array.from({ length: count }, (_, id) => ({ id, name: `shard-${id}` }));
+}
+
+describe('CI report artifact discovery', () => {
+  it.each([0, 130])('downloads report artifacts after %i unrelated artifacts', async (count) => {
+    const names = ['pr-meta', 'test-reports', 'test-dupes'];
+    const api = fixture([
+      ...precedingArtifacts(count),
+      ...names.map((name, index) => ({ id: 1000 + index, name })),
+    ]);
+
+    await api.run('Download artifacts');
+
+    expect(readdirSync(path.join(temp, 'dl')).sort()).toEqual(
+      names.map((name) => `${name}.zip`).sort(),
+    );
+    for (const [index, name] of names.entries()) {
+      expect(readFileSync(path.join(temp, 'dl', `${name}.zip`), 'utf8')).toBe(
+        `archive-${1000 + index}`,
+      );
+    }
+    expect(api.core.warning).not.toHaveBeenCalled();
+  });
+
+  it.each([0, 130])('finds base coverage after %i unrelated artifacts', async (count) => {
+    const api = fixture([...precedingArtifacts(count), { id: 1000, name: 'test-reports' }]);
+
+    await api.run('Fetch base branch coverage');
+
+    expect(readFileSync(path.join(temp, 'base-coverage', 'base.zip'), 'utf8')).toBe('archive-1000');
+    expect(api.core.setOutput).toHaveBeenCalledWith('found', 'true');
+  });
+
+  it.skipIf(process.platform === 'win32')(
+    'skips reporting cleanly when no artifacts exist',
+    async () => {
+      const api = fixture([]);
+      await api.run('Download artifacts');
+
+      const output = path.join(temp, 'outputs');
+      for (const name of ['Extract artifacts', 'Read PR metadata']) {
+        const script = steps.find((step) => step.name === name)?.run;
+        if (!script) throw new Error(`Missing script: ${name}`);
+        execFileSync('bash', ['--noprofile', '--norc', '-e', '-o', 'pipefail', '-c', script], {
+          env: { ...process.env, RUNNER_TEMP: temp, GITHUB_OUTPUT: output },
+        });
+      }
+
+      expect(readFileSync(output, 'utf8').trim()).toBe('skip=true');
+      expect(api.downloadArtifact).not.toHaveBeenCalled();
+    },
+  );
+
+  it('still fails when downloading an existing report artifact fails', async () => {
+    const api = fixture([{ id: 1000, name: 'pr-meta' }]);
+    api.downloadArtifact.mockRejectedValueOnce(new Error('download failed'));
+
+    await expect(api.run('Download artifacts')).rejects.toThrow('download failed');
+  });
+});