implement cancellation handling in CaptureService and Database, update job states accordingly

This commit is contained in:
kstyagi23
2026-09-05 06:54:10 +05:30
parent 9c1a301c1a
commit cab7a68b8b
7 changed files with 293 additions and 38 deletions
+93
View File
@@ -37,6 +37,38 @@ class MinimalSiteHandler(BaseHTTPRequestHandler):
return
class SlowResponseHandler(BaseHTTPRequestHandler):
slow_request_started = threading.Event()
release_slow_request = threading.Event()
def do_GET(self) -> None: # noqa: N802
if self.path == "/":
body = b"<html><body><img src='/slow.bin'></body></html>"
self.send_response(200)
self.send_header("Content-Type", "text/html")
self.send_header("Content-Length", str(len(body)))
self.end_headers()
self.wfile.write(body)
return
if self.path == "/slow.bin":
self.send_response(200)
self.send_header("Content-Type", "application/octet-stream")
self.send_header("Content-Length", "1")
self.end_headers()
self.slow_request_started.set()
self.release_slow_request.wait(timeout=30)
try:
self.wfile.write(b"x")
except BrokenPipeError:
pass
return
self.send_response(404)
self.end_headers()
def log_message(self, format: str, *args) -> None: # type: ignore[override]
return
def test_capture_service_creates_private_archive(tmp_path) -> None:
server = ThreadingHTTPServer(("127.0.0.1", 0), MinimalSiteHandler)
thread = threading.Thread(target=server.serve_forever, daemon=True)
@@ -89,3 +121,64 @@ def test_capture_service_creates_private_archive(tmp_path) -> None:
names = archive.namelist()
assert any(name.endswith("siteharbor-manifest.json") for name in names)
assert any(name.endswith("site.css") for name in names)
def test_capture_cancels_while_a_resource_is_stalled(tmp_path) -> None:
SlowResponseHandler.slow_request_started.clear()
SlowResponseHandler.release_slow_request.clear()
server = ThreadingHTTPServer(("127.0.0.1", 0), SlowResponseHandler)
server_thread = threading.Thread(target=server.serve_forever, daemon=True)
server_thread.start()
port = server.server_address[1]
test_settings = replace(
settings,
data_dir=tmp_path,
database_path=tmp_path / "siteharbor.sqlite3",
work_dir=tmp_path / "work",
artifacts_dir=tmp_path / "artifacts",
reports_dir=tmp_path / "reports",
allow_private_networks=True,
allow_nonstandard_ports=True,
respect_robots=False,
)
test_settings.ensure_directories()
database = Database(test_settings.database_path)
database.initialize()
source_url = f"http://127.0.0.1:{port}/"
job = database.create_job(
source_url,
source_url,
{
"include_external_assets": False,
"max_pages": 10,
"max_depth": 2,
"max_bytes": 2_000_000,
"max_duration_seconds": 60,
"retention_days": 1,
},
)
claimed = database.claim_next("test-worker", lease_seconds=120)
assert claimed is not None
capture_thread = threading.Thread(
target=CaptureService(database, test_settings, "test-worker").run,
args=(claimed,),
daemon=True,
)
capture_thread.start()
try:
assert SlowResponseHandler.slow_request_started.wait(timeout=2)
cancelled = database.request_cancel(job["id"])
assert cancelled is not None
capture_thread.join(timeout=3)
assert not capture_thread.is_alive()
finally:
SlowResponseHandler.release_slow_request.set()
capture_thread.join(timeout=2)
server.shutdown()
server.server_close()
cancelled_job = database.get_job(job["id"])
assert cancelled_job is not None
assert cancelled_job["state"] == "cancelled"