fix(daemon): restore blocking I/O before client authentication - #258
Merged
Merged
Conversation
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
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Accepted daemon client sockets can inherit the listener’s nonblocking mode on macOS. A captured failure rejected authentication about88µs after handler entry, before the client finished sending it; the resulting reset was presented as “timeout after auth.” The dedicated handler now restores blocking mode before creating its read/write halves, allowing the existing deadlines to govern I/O. Authentication, timeout values, and retry policy are unchanged.
The regression explicitly starts with a nonblocking socket and delayed authentication: it fails before the fix and passes afterward, verifies the greeting, and joins the handler. Actual failed connection timing and successful socket-mode controls are retained. Older timeout/EOF/reset observations are preserved without claiming universal attribution; iteration268 remains open.
Validation: focused auth regressions; final-source original160/recurring165 real-Firefox tests; ordered stable-update/fmt/strict workspace clippy/workspace tests; all nine xtask checks (dogfood explicitly skipped: no script). Independent Astra review found no product-code/security/substantive-evidence issues; one carry-over reference corrected. Main’s independent265 change was reconciled with hashes unchanged and combined-tree ordered gates verified.
The dual-gate sweep (
FF_RDP_LIVE_TESTS=1 FF_RDP_LIVE_NETWORK_TESTS=1) retained this actual summary on the267 baseline:CLI334passed/1failed. The raw browser exited before sweep startup; the exact nine omitted core tests subsequently passed in a separate dual-gate run. All344 unique names reconcile as343passes/1failure across those separate runs. Main’s two265 Watcher tests have their independently reviewed live evidence inPR257; no invented combined sweep summary.
Carry-over: external-CSS cascade test returned
no element matching selector h1; an unchanged-source isolation passed2.66s and does not explain the failure. Filed plan276 retains the missing document/route evidence; it is outside the selected execution queue. No full-green sweep claim. Iteration267 fulfills4/4 original criteria for its demonstrated occurrence.