feat: guard h3 + honest ADR update on the flare reactor driver gap - #38
Merged
Conversation
While wiring HTTP/3 into opengateway (PR #36's flare pin bump pulled in bind_with_http3 + QuicServerConfig), the integration ran into a real gap: at the pinned flare commit (edf3f22) the reactor's main ``serve()`` does not call ``tick_http3_once`` / ``pump_http3_handler_once``, which the source itself labels "Test-only entry point". Binding via ``bind_with_http3`` accepts the UDP socket but never processes datagrams. So the work to land a working h3 path is a real reactor-loop integration (select on the UDP fd, feed into recv_datagram, dispatch via pump_http3_handler_once, drain egress via send_to). That's upstream flare work, not opengateway's scope. Rather than ship a broken-looking path that binds a UDP socket and accepts nothing, this commit: - Reverts the bind_with_http3 wiring from main.mojo (it didn't drive any packets). - Adds a clear guard: if ``HTTP3_ENABLED=1`` is set, raise an Error explaining the reactor driver isn't integrated and pointing at the pinned commit's serve() for the gap. - Updates ADR-003 #6 and #4 to reflect the actual status: TLS is fully shipped via the reactor; h3 is blocked on the missing reactor integration (not the rustls build or cert path, both of which landed). Verified: 42 pytest, 8 Mojo router tests, ruff + mypy clean. Co-authored-by: opencode-agent <noreply@opencode.ai>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
Investigating HTTP/3 wiring in opengateway (the natural completion of ADR-003 #6 once #35 landed the rustls cdylib build and #36 pulled in
bind_with_http3via the flare pin bump) hit a real gap: at the pinned flare commit (edf3f22), the reactor's mainserve()does not calltick_http3_once/pump_http3_handler_once, which the source itself labels "Test-only entry point". Binding viabind_with_http3accepts the UDP socket but never processes datagrams, so the reactor-integration step is the actual missing piece.Rather than ship a broken-looking path (bind UDP, accept nothing, look like an h3 server that just isn't getting traffic), this commit:
bind_with_http3wiring fromopengateway/mojo/main.mojo.HTTP3_ENABLED=1raises anErrorexplaining the gap and pointing at the pinned commit'sserve()for where the integration belongs.The honest engineering move: don't ship a half-broken h3 path. The fix is either upstream in flare (the reactor-loop work that consumes
recvfrom→recv_datagramand emitssend_to) or a self-contained UDP-loop driver in opengateway that gets upstreamed. Either way, it's real work and belongs as a follow-up PR, not a quiet bug.Consequences
bind_with_http3API + rustls cdylib + TLS) are landed. The remaining work is the reactor-side driver.flare#7issue was a cert hostname verification failure masquerading as a stall).Testing
uv run pytest— 42 passedmake mojo-test— 8 passed