A checker for a bridge that is actually deployed #1

Merged
paulhitt merged 1 commits from tool/bridgecheck into main 2026-09-06 23:24:13 +00:00
paulhitt commented 2026-09-06 23:19:14 +00:00 (Migrated from gitlab.com)

The interop suite proves the protocol against a bridge the test starts
itself. Nothing proved the deployed one, and the gap is not academic: a
container reporting "Up" establishes that a process is running, not that
its published line is dialable from outside, not that the handshake
completes against the Tor Project's own client, and not that whatever it
fronts still answers.

bridgecheck takes a bridge line exactly as an operator receives it,
dials with real lyrebird, and speaks HTTP through the tunnel. Used to
verify the first deployed bridge today:

handshake: OK against real lyrebird
upstream:  HTTP 200 OK via 135.148.121.177:8443
server:    flamenet-relay

The Server: header is the part that matters — that is the relay
answering from behind the bridge, so the whole path is proven rather than
the socket being open.

⚠️ Deliberately not wired into any test suite. It needs a live bridge and
the network, so putting it in CI would produce exactly the flaky, ignored
job that teaches people to stop reading test results.

Co-Authored-By: Claude Opus 5 noreply@anthropic.com

The interop suite proves the protocol against a bridge the test starts itself. Nothing proved the deployed one, and the gap is not academic: a container reporting "Up" establishes that a process is running, not that its published line is dialable from outside, not that the handshake completes against the Tor Project's own client, and not that whatever it fronts still answers. `bridgecheck` takes a bridge line exactly as an operator receives it, dials with real lyrebird, and speaks HTTP through the tunnel. Used to verify the first deployed bridge today: handshake: OK against real lyrebird upstream: HTTP 200 OK via 135.148.121.177:8443 server: flamenet-relay The `Server:` header is the part that matters — that is the relay answering from behind the bridge, so the whole path is proven rather than the socket being open. ⚠️ Deliberately not wired into any test suite. It needs a live bridge and the network, so putting it in CI would produce exactly the flaky, ignored job that teaches people to stop reading test results. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
paulhitt commented 2026-09-06 23:19:14 +00:00 (Migrated from gitlab.com)

assigned to @paulhitt

assigned to @paulhitt
paulhitt commented 2026-09-06 23:24:13 +00:00 (Migrated from gitlab.com)

mentioned in commit 8ad485465a

mentioned in commit 8ad485465a9e1db09760918770872a3d5707bd04
paulhitt (Migrated from gitlab.com) merged commit 8ad485465a into main 2026-09-06 23:24:13 +00:00
Sign in to join this conversation.
No Reviewers
No labels
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: extant2000/flamenet-elligator2#1