test(release): Keep current version checks from going stale

The repo is already on 2.3.0, but the release truth test and manual publish fallback were still pinned to 2.2.0. Derive the assertions from package metadata and normalize macOS temp paths so the full suite checks the real contract instead of local path spelling.

Author: aolin480

.github/workflows/publish-npm-on-release.yml

@@ -8,9 +8,9 @@ on:
   workflow_dispatch:
     inputs:
       tag:
-        description: 'Tag to publish (e.g. v2.2.0)'
+        description: 'Tag to publish (e.g. v2.3.0)'
         required: true
-        default: 'v2.2.0'
+        default: 'v2.3.0'
 
 permissions:
   contents: read

tests/contextbench-baseline-schema-gate.test.ts

@@ -1,5 +1,12 @@
 import { execFileSync } from 'node:child_process';
-import { chmodSync, mkdtempSync, readFileSync, rmSync, writeFileSync } from 'node:fs';
+import {
+  chmodSync,
+  mkdtempSync,
+  readFileSync,
+  realpathSync,
+  rmSync,
+  writeFileSync
+} from 'node:fs';
 import { tmpdir } from 'node:os';
 import path from 'node:path';
 import { describe, expect, it, vi } from 'vitest';
@@ -55,6 +62,10 @@ function cleanupSessionRoot(sessionRoot: string): void {
   }
 }
 
+function normalizeFilesystemPath(filePath: string): string {
+  return realpathSync(filePath);
+}
+
 function tempSessionRoot(phase: 'phase40' | 'phase41' = 'phase40'): string {
   return path.join(
     mkdtempSync(path.join(tmpdir(), `contextbench-${phase}-schema-gate-`)),
@@ -398,7 +409,9 @@ describe('ContextBench Phase 40 schema gate', () => {
         repoCheckoutPath: repoPath,
         verificationStrict: false
       });
-      expect(readFileSync(cwdCapturePath, 'utf8')).toBe(repoPath);
+      expect(normalizeFilesystemPath(readFileSync(cwdCapturePath, 'utf8'))).toBe(
+        normalizeFilesystemPath(repoPath)
+      );
       const stdin = readFileSync(stdinCapturePath, 'utf8');
       expect(stdin).toContain('Problem statement:');
       expect(stdin).toContain('Fix the failing ContextBench task');
@@ -584,7 +597,9 @@ describe('ContextBench Phase 40 schema gate', () => {
           ],
           { encoding: 'utf8', env }
         );
-        expect(readFileSync(cwdPath, 'utf8')).toBe(repoPath);
+        expect(normalizeFilesystemPath(readFileSync(cwdPath, 'utf8'))).toBe(
+          normalizeFilesystemPath(repoPath)
+        );
       }
       execFileSync(
         'node',
@@ -681,7 +696,9 @@ describe('ContextBench Phase 40 schema gate', () => {
         (candidate) => candidate.scoring && 'baselineArmId' in candidate.scoring
       );
       expect(row).toMatchObject({ status: 'completed' });
-      expect(readFileSync(cwdCapturePath, 'utf8')).toBe(repoPath);
+      expect(normalizeFilesystemPath(readFileSync(cwdCapturePath, 'utf8'))).toBe(
+        normalizeFilesystemPath(repoPath)
+      );
       const stdin = readFileSync(stdinCapturePath, 'utf8');
       expect(stdin).toContain('Problem statement:');
       expect(stdin).toContain('Run the diagnostic arm with materialized task text.');

tests/release-truth-surfaces.test.ts

@@ -73,17 +73,17 @@ function extractMarkdownLinks(markdown: string): string[] {
 describe('release truth surfaces', () => {
   const packageJson = readJson<PackageJson>('package.json');
   const releaseManifest = readJson<ReleaseManifest>('.release-please-manifest.json');
+  const expectedVersion = packageJson.version;
   const changelog = readText('CHANGELOG.md');
   const readme = readText('README.md');
   const workflow = readText('.github/workflows/publish-npm-on-release.yml');
   const todoDoc = readOptionalText('docs/TODO.md');
   const visualsDoc = readOptionalText('docs/visuals.md');
   const packagedPaths = ['README.md', 'LICENSE', ...(packageJson.files ?? [])];
 
-  it('keeps package metadata, release manifest, and changelog on 2.2.0', () => {
-    expect(packageJson.version).toBe('2.2.0');
-    expect(releaseManifest['.']).toBe('2.2.0');
-    expect(changelog).toContain('## [2.2.0]');
+  it('keeps package metadata, release manifest, and changelog aligned', () => {
+    expect(releaseManifest['.']).toBe(expectedVersion);
+    expect(changelog).toContain(`## [${expectedVersion}]`);
     expect(changelog).not.toContain('## Unreleased');
   });
 
@@ -114,8 +114,8 @@ describe('release truth surfaces', () => {
     expect(visualsDoc).toContain('Historical snapshot');
   });
 
-  it('keeps the manual publish fallback aligned to v2.2.0', () => {
-    expect(workflow).toContain("description: 'Tag to publish (e.g. v2.2.0)'");
-    expect(workflow).toContain("default: 'v2.2.0'");
+  it('keeps the manual publish fallback aligned to the package version', () => {
+    expect(workflow).toContain(`description: 'Tag to publish (e.g. v${expectedVersion})'`);
+    expect(workflow).toContain(`default: 'v${expectedVersion}'`);
   });
 });