fix(evidence): bind HTTP validators to final URL

This commit is contained in:
2026-07-12 03:38:11 +02:00
parent 81ff1810d1
commit d1fdf7d9f5
3 changed files with 110 additions and 21 deletions
@@ -11,6 +11,9 @@ from tht.ports.evidence import EvidenceSourceError
class Handler(BaseHTTPRequestHandler):
etag_requests = 0
etag_body_responses = 0
redirect_target = "/redirected-v1"
redirect_request_validators = []
final_request_validators = []
def do_GET(self):
if self.path.startswith("/etag"):
@@ -46,6 +49,24 @@ class Handler(BaseHTTPRequestHandler):
"Location", f"http://user:password@127.0.0.1:{self.server.server_port}/etag"
)
self.end_headers()
elif self.path == "/stable-redirect":
type(self).redirect_request_validators.append(self.headers.get("If-None-Match"))
self.send_response(302)
self.send_header("Location", type(self).redirect_target)
self.end_headers()
elif self.path in {"/redirected-v1", "/redirected-v2"}:
type(self).final_request_validators.append(
(self.path, self.headers.get("If-None-Match"))
)
etag = '"v1"' if self.path.endswith("v1") else '"v2"'
if self.headers.get("If-None-Match") == etag:
self.send_response(304)
self.end_headers()
return
self.send_response(200)
self.send_header("ETag", etag)
self.end_headers()
self.wfile.write(self.path.encode())
else:
self.send_response(200)
self.send_header("Last-Modified", "Wed, 21 Oct 2015 07:28:00 GMT")
@@ -60,6 +81,9 @@ class Handler(BaseHTTPRequestHandler):
def server_url():
Handler.etag_requests = 0
Handler.etag_body_responses = 0
Handler.redirect_target = "/redirected-v1"
Handler.redirect_request_validators = []
Handler.final_request_validators = []
server = ThreadingHTTPServer(("127.0.0.1", 0), Handler)
thread = threading.Thread(target=server.serve_forever, daemon=True)
thread.start()
@@ -246,3 +270,62 @@ def test_http_closes_response_when_streaming_fails():
with pytest.raises(EvidenceSourceError):
list(source.discover())
assert response.closed
def test_http_binds_validators_to_exact_final_redirect_url(server_url):
source = HttpManifestEvidenceSource(
[server_url + "/stable-redirect"], allow_private_hosts=True
)
first = next(iter(source.discover()))
second = next(iter(source.discover()))
assert first == second
assert Handler.redirect_request_validators == [None, None]
assert Handler.final_request_validators == [
("/redirected-v1", None),
("/redirected-v1", '"v1"'),
]
def test_http_redirect_path_change_fetches_and_replaces_body(server_url):
source = HttpManifestEvidenceSource(
[server_url + "/stable-redirect"], allow_private_hosts=True
)
first = next(iter(source.discover()))
assert source.acquire(first).content == b"/redirected-v1"
Handler.redirect_target = "/redirected-v2"
with pytest.raises(EvidenceSourceError):
source.acquire(first)
current = next(iter(source.discover()))
assert source.acquire(current).content == b"/redirected-v2"
assert ("/redirected-v2", None) in Handler.final_request_validators
def test_http_rejects_unsolicited_304_without_bound_validator(monkeypatch):
monkeypatch.setattr(socket, "getaddrinfo", lambda *args, **kwargs: [
(socket.AF_INET, socket.SOCK_STREAM, 6, "", ("93.184.216.34", 443)),
])
source = HttpManifestEvidenceSource(["https://example.test/doc"])
response = FakeResponse(peer="93.184.216.34")
response.status_code = 304
source._session = FakeSession(response)
with pytest.raises(EvidenceSourceError) as caught:
list(source.discover())
assert not caught.value.retryable
assert response.closed
def test_http_rejects_cross_origin_304_for_cached_provenance(server_url):
source = HttpManifestEvidenceSource([server_url + "/etag"], allow_private_hosts=True)
item = next(iter(source.discover()))
response = FakeResponse()
response.status_code = 304
source._session = FakeSession(response)
with pytest.raises(EvidenceSourceError) as caught:
source._download("https://other.example/doc", item.uri)
assert not caught.value.retryable
assert response.closed