Auto PR from release/0.6 to release/0.7 - #1536
Closed
github-actions[bot] wants to merge 37 commits into
Closed
Conversation
Signed-off-by: liulanzheng <lanzheng.liulz@alibaba-inc.com>
…t empty/full (#504) * FIX: lockfree MPMC queue should not fail to pop/push when queue is not empty/full Signed-off-by: Coldwings <coldwings@me.com> * Old CI image not able to access repo, change to preset image Signed-off-by: Coldwings <coldwings@me.com> --------- Signed-off-by: Coldwings <coldwings@me.com>
…oding (#1113) * fix(http): Fix stale buffer data being parsed in chunked transfer encoding Signed-off-by: yuchen.cc <yuchen.cc@alibaba-inc.com> * fix(http): prevent duplicate stream read in chunked body when parsing final chunk Signed-off-by: yuchen.cc <yuchen.cc@alibaba-inc.com> --------- Signed-off-by: yuchen.cc <yuchen.cc@alibaba-inc.com>
…1165) Signed-off-by: zhuangbowei.zbw <zhuangbowei.zbw@alibaba-inc.com>
…heck on e2fsprogs >= 1.46 (#1237) Signed-off-by: zhuangbowei.zbw <zhuangbowei.zbw@alibaba-inc.com>
…1407) (#1440) * Cap HTTP client retry backoff at 5 seconds The exponential backoff formula (sleep_interval + 500) * 2 grows unbounded, reaching 30+ seconds after a few retries. Cap at 5s to prevent excessive tail latency on transient failures. * Address review: bound retry backoff by remaining timeout Per maintainer feedback: instead of an arbitrary 5s cap, clamp each backoff sleep to the remaining operation timeout (tmo.timeout()). This prevents a single backoff from overshooting the deadline regardless of how the exponential interval grows, and removes the magic constant. * Address review: use microseconds for sleep_interval directly Per maintainer feedback: define sleep_interval in microseconds from the start, eliminating the * 1000 conversion. Both thread_usleep() and tmo.timeout() use microseconds, so this is cleaner. --------- Co-authored-by: Jiangtian Feng <fengjiangtian.fjt@alibaba-inc.com> Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>
… writev cap (#1288) (#1370) (#1380) (#1416) (#1448) * Fix HTTP range parsing, cookie jar, and body writev cap (#1288) (#1370) * Fix HTTP Range parsing, cookie header corruption, and writev cap * Address review: move wc calculation into branches - Only clone/shrink iovec in overflow path - Calculate wc separately in each branch to avoid extra branching - Fixes sign comparison warning --------- * Update cookie_jar.cpp --------- Co-authored-by: Jiangtian Feng <fengjiangtian.fjt@alibaba-inc.com> Co-authored-by: Claude Opus 4.6 (1M context) <noreply@anthropic.com> Co-authored-by: Huiba Li <huiba.lhb@alibaba-inc.com>
… fini() (#1268) (#1318) (#1384) (#1420) (#1445) * Fix iouring eventfd double-close on repeated fini() (#1268) (#1318) (#1384) (#1420) Co-authored-by: Jiangtian Feng <fengjiangtian.fjt@alibaba-inc.com> Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com> * Update iouring-wrapper.cpp --------- Co-authored-by: Jiangtian Feng <fengjiangtian.fjt@alibaba-inc.com> Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com> Co-authored-by: Huiba Li <huiba.lhb@alibaba-inc.com>
#1306) (#1334) (#1375) (#1404) (#1452) * Fix macOS sendfile source/dest argument order (#1306) (#1334) (#1375) * Update basic_socket.cpp --------- Co-authored-by: Jiangtian Feng <fengjiangtian.fjt@alibaba-inc.com> Co-authored-by: Claude Opus 4.6 (1M context) <noreply@anthropic.com> Co-authored-by: Huiba Li <huiba.lhb@alibaba-inc.com>
… failure (#1252) (#1328) (#1393) (#1428) (#1459) * Fix crash in threads_create_join on partial failure (#1252) * Fix crash in threads_create_join on partial failure When thread_create() fails mid-loop, only join threads that were actually created. Previously the join loop iterated all n slots, calling thread_join on uninitialized pointers (stack array case) or nullptr (vector case), causing a segfault. * Address review: return # of threads created, use created++ index Per maintainer feedback: - threads_create_join now returns the actual number of threads created (< n on partial failure), so callers can detect partial failure. - Use pthreads[created++] = th instead of a separate created++ line. --------- * Update thread.h * Update thread.h --------- Co-authored-by: Jiangtian Feng <fengjiangtian.fjt@alibaba-inc.com> Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com> Co-authored-by: Coldwings <coldwings@me.com> Co-authored-by: Huiba Li <huiba.lhb@alibaba-inc.com>
…le switch and fix extfs_close (#1276) (#1346) (#1364) (#1397) (#1460) * extfs: add resize_extfs interface with compile switch and fix extfs_close (#1276) (#1346) (#1364) * Update CMakeLists.txt * Update CMakeLists.txt --------- Co-authored-by: OsentryO <75356449+OsentryO@users.noreply.github.com> Co-authored-by: Huiba Li <huiba.lhb@alibaba-inc.com> Co-authored-by: Coldwings <coldwings@me.com>
…ce/iterator bugs (#1283) (#1382) (#1421) (#1444) (#1463) * Fix alog snprintf overflow and iovector slice/iterator bugs (#1283) * Fix log buffer overflow and iovector slice/iterator bugs * Address review: refactor slice to process first element before loop - Skip loop only skips complete iovec elements - Process first (potentially partial) element separately - Main loop only handles full elements, no branching --------- * Update iovector.cpp --------- Co-authored-by: Jiangtian Feng <fengjiangtian.fjt@alibaba-inc.com> Co-authored-by: Claude Opus 4.6 (1M context) <noreply@anthropic.com> Co-authored-by: Huiba Li <huiba.lhb@alibaba-inc.com>
…#1265) (#1322) (#1387) (#1423) (#1457) * Fix OOB access in event engines when fd is 0 (#1265) (#1322) (#1387) (#1423) Co-authored-by: Jiangtian Feng <fengjiangtian.fjt@alibaba-inc.com> Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com> * Update epoll.cpp --------- Co-authored-by: Jiangtian Feng <fengjiangtian.fjt@alibaba-inc.com> Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com> Co-authored-by: Huiba Li <huiba.lhb@alibaba-inc.com>
…url init (#1285) (#1371) (#1379) (#1408) (#1455) * Fix TLS null ctx crash, OpenSSL leaks, and curl init (#1285) (#1371) (#1379) (#1408) * Fix TLS null ctx crash, OpenSSL resource leaks, and curl init * Fix TLS null ctx crash, OpenSSL resource leaks, and curl init * Address review: use LOG_ERROR_RETURN and DEFER - tls-stream.cpp: Replace LOG_ERROR + return with LOG_ERROR_RETURN - curl.cpp: Use DEFER for cleanup on init failure instead of separate function --------- Co-authored-by: Jiangtian Feng <fengjiangtian.fjt@alibaba-inc.com> Co-authored-by: Claude Opus 4.6 (1M context) <noreply@anthropic.com> * Update curl.cpp * Update curl.cpp --------- Co-authored-by: Jiangtian Feng <fengjiangtian.fjt@alibaba-inc.com> Co-authored-by: Claude Opus 4.6 (1M context) <noreply@anthropic.com> Co-authored-by: Huiba Li <huiba.lhb@alibaba-inc.com>
…ead join and skip_read static buffer race (#1287) (#1337) (#1372) (#1402) (#1472) * [Backport][0.8 to 0.7] | Fix terminate() thread join and skip_read static buffer race (#1287) (#1337) (#1372) (#1402) * Fix terminate() thread join and skip_read static buffer race (#1287) (#1337) (#1372) Co-authored-by: Jiangtian Feng <fengjiangtian.fjt@alibaba-inc.com> Co-authored-by: Claude Opus 4.6 (1M context) <noreply@anthropic.com> * Update kernel_socket.cpp * Update kernel_socket.cpp --------- Co-authored-by: Jiangtian Feng <fengjiangtian.fjt@alibaba-inc.com> Co-authored-by: Claude Opus 4.6 (1M context) <noreply@anthropic.com> Co-authored-by: Huiba Li <huiba.lhb@alibaba-inc.com> * Update kernel_socket.cpp --------- Co-authored-by: Jiangtian Feng <fengjiangtian.fjt@alibaba-inc.com> Co-authored-by: Claude Opus 4.6 (1M context) <noreply@anthropic.com> Co-authored-by: Huiba Li <huiba.lhb@alibaba-inc.com>
… init and ObjectCache move/lock bugs (#1284) (#1339) (#1413) (#1446) (#1480) * [Backport][main to 0.9] | Fix lockfree queue init and ObjectCache move/lock bugs (#1284) (#1339) (#1413) * Fix lockfree queue init and ObjectCache move/lock bugs (#1284) * Refactor release method in ObjectCacheBase * Update lockfree_queue.h --------- * Delete common/objectcachev2.h * fix conflict * Update expirecontainer.cpp * Update expirecontainer.cpp --------- Co-authored-by: Jiangtian Feng <fengjiangtian.fjt@alibaba-inc.com> Co-authored-by: Coldwings <coldwings@me.com> Co-authored-by: Huiba Li <huiba.lhb@alibaba-inc.com>
… arguments in Content-Range parsing (#1262) (#1325) (#1412) (#1454) (#1481) * [Backport][main to 0.9] | Fix swapped sscanf arguments in Content-Range parsing (#1262) (#1325) (#1412) * Fix swapped sscanf arguments in Content-Range parsing (#1262) * Update message.cpp --------- * Update message.cpp * Update message.cpp --------- Co-authored-by: Jiangtian Feng <fengjiangtian.fjt@alibaba-inc.com> Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com> Co-authored-by: Huiba Li <huiba.lhb@alibaba-inc.com>
…1476) (#1483) (#1489) (#1492) * Fix timer thread leak, stop() hang, and thread-key constant * Fix Timer.OneShot hang: drop stop() from ~Timer After removing the stub's _th=nullptr (the leak fix), ~Timer no longer short-circuits at if(!_th) return, so it runs stop() first. For an already-fired one-shot the stub thread has exited, so _waiting stays false and it never notifies _wait_ready again; stop()'s while(cancel()) _wait_ready.wait_no_lock() then blocks forever (lost wakeup) -> Timer.OneShot 3600s hang. The destructor's own _repeating=false + interrupt(ECANCELED) + join already tears the stub down (and joins the DONE one-shot immediately, fixing the leak), so stop() is redundant and harmful here. --------- Co-authored-by: Jiangtian Feng <fengjiangtian.fjt@alibaba-inc.com> Co-authored-by: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
… label gate, conflict handling, and CI skip (#1410) (#1471) (#1507) (#1520) (#1523) * [Backport][0.8 to 0.7] | | | ci: harden backport autopr with bugfix label gate, conflict handling, and CI skip (#1410) (#1471) (#1507) (#1520) * [Backport][0.9 to 0.8] | | ci: harden backport autopr with bugfix label gate, conflict handling, and CI skip (#1410) (#1471) (#1507) * [Backport][main to 0.9] | ci: harden backport autopr with bugfix label gate, conflict handling, and CI skip (#1410) (#1471) * ci: harden backport autopr with bugfix label gate, conflict handling, and CI skip (#1410) - Replace commit message parsing with GitHub API (listPullRequestsAssociatedWithCommit) to reliably detect source PR head ref regardless of merge strategy (rebase/squash/merge) - Add bugfix label gate: only backport commits from PRs with 'bugfix' label - Cherry-pick conflict handling: mark conflicts, create draft PR with 'needs-manual-merge' label - Auto-drop new files introduced by cherry-pick that don't exist in target branch - Add 'bugfix' label to created backport PRs for consistency - CI workflows skip jobs on PRs with 'needs-manual-merge' label to avoid wasting runners - Support 'unlabeled' event to re-trigger CI after label removal without empty commits - Symmetric fix for auto-pr-precise.yml using same API-based detection * resolve conflicts * Update ci.linux.x86_64.yml 0.9 do not test with GCC13/14 * fix(arm): add explicit +crc+crypto target for aarch64 SIMD section GCC 9 cannot correctly detect crypto features via -march=native on newer ARM CPUs (e.g. Graviton3/4) used by GitHub-hosted ubuntu-24.04-arm runners. This causes vmull_p64 (PMULL) and CRC32 instructions to fail compilation when MinSizeRel build type sets -march=native, which prevents the fallback -mcpu=generic+crc+crypto+lse from being applied. Add #pragma GCC target ("+crc+crypto") (and equivalent clang attribute) to explicitly enable these extensions at the function level, matching the approach already used on the main branch. --------- Co-authored-by: Coldwings <coldwings@me.com> * chore: drop files not present in 0.8 * Update ci.linux.arm.yml * Update ci.linux.x86_64.yml --------- Co-authored-by: Coldwings <coldwings@me.com> Co-authored-by: lihuiba <282919+lihuiba@users.noreply.github.com> * Update ci.linux.x86_64.yml --------- Co-authored-by: Coldwings <coldwings@me.com> Co-authored-by: lihuiba <282919+lihuiba@users.noreply.github.com> * chore: drop files not present in 0.6 * Update ci.linux.arm.yml * Update ci.macos.arm.yml * Update ci.linux.arm.yml * Modify ulimit command to include sudo Updated CI configuration to use 'sudo' for setting unlimited memory lock. * Update ci.macos.arm.yml --------- Co-authored-by: Coldwings <coldwings@me.com> Co-authored-by: lihuiba <282919+lihuiba@users.noreply.github.com>
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
8 participants
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.
Created by Github action