Run the horizon-client parity probes from one binary

81 client examples that harnesses run become modules of examples/probe:
harnesses run `probe <name> ...` (or `cargo run --example probe -- <name>`),
and the shared helpers (audio_common, hud_fixture, chat_screen_fixture)
move with them. The benchmarks and manual tools stay standalone examples.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
This commit is contained in:
Ludovic Marie
2026-10-23 14:46:08 +02:00
parent 4a922a5772
commit 858f701ee8
145 changed files with 551 additions and 340 deletions
+3 -3
View File
@@ -64,9 +64,9 @@ def main():
registries = json.loads((build / 'registries.json').read_text())
corpus = corpus_module.corpus(tree, registries)
(build / 'corpus.txt').write_text(''.join(line + '\n' for line in corpus), encoding='utf-8')
run(['cargo', 'build', '--release', '-p', 'horizon-client', '--example', 'command_parse_fixture'], 'compile-rust')
run(['cargo', 'build', '--release', '-p', 'horizon-client', '--example', 'probe'], 'compile-rust')
metadata = json.loads(subprocess.check_output(['cargo', 'metadata', '--no-deps', '--format-version', '1'], cwd=lab.ROOT, text=True))
binary = Path(metadata['target_directory']) / 'release/examples' / ('command_parse_fixture.exe' if os.name == 'nt' else 'command_parse_fixture')
binary = Path(metadata['target_directory']) / 'release/examples' / ('probe.exe' if os.name=='nt' else 'probe')
resources = lab.RESOURCES
total_failures = 0
summary = []
@@ -79,7 +79,7 @@ def main():
if not (stamp.exists() and stamp.read_text() == digest and (level_dir / 'vanilla.jsonl').exists()):
java('run', str(level_dir), str(build / 'corpus.txt'), str(permission))
stamp.write_text(digest)
run([binary, resources, build / 'registries.json', build / 'corpus.txt', str(permission), level_dir / 'rust.jsonl'], f'rust{permission}')
run([binary, 'command_parse_fixture',resources, build / 'registries.json', build / 'corpus.txt', str(permission), level_dir / 'rust.jsonl'], f'rust{permission}')
failures = compare(level_dir / 'vanilla.jsonl', level_dir / 'rust.jsonl', args.limit)
count = sum(1 for _ in (level_dir / 'vanilla.jsonl').open())
print(f'permission {permission}: {count - failures}/{count} commands match', flush=True)