Repository navigation
Conversation
Retrying a truncated reply over TCP connected with no timeout, so a nameserver that never completed the handshake held the query until the kernel gave up, regardless of `timeouts=`, and the resulting error escaped instead of moving on to the next nameserver. The connect now spends from the same interval as the request that follows it. `Addrinfo#connect` is used because `TCPSocket.new` accepts `connect_timeout:` only since Ruby 3.0, and it raises `Errno::ETIMEDOUT` on every supported version. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
This branch has not been deployed
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
When a UDP reply comes back truncated, resolv retries the same nameserver over TCP, but that connect had no timeout. Against a nameserver that never completes the handshake,
getaddresswithtimeouts = 1waited about 135 seconds on Linux and then raisedIO::TimeoutError(Ruby 4.0.6) instead of trying the next nameserver.The connect now spends from the same interval as the request that follows it, and a timed out connect is handled like a refused one. I used
Addrinfo#connect(timeout:)becauseTCPSocket.newacceptsconnect_timeout:only since Ruby 3.0.The new tests fill the listener's accept queue so that Linux drops the SYN. Where the kernel completes or refuses such a connect instead, as macOS does, the tests are omitted.
Generated with Claude Code