unix,pipe: fix handling null buffer in uv_pipe_get{sock,peer}name

Fixes: https://github.com/libuv/libuv/issues/4610

build(deps): bump actions/checkout from 3 to 4 (#4490)

Requires updating the android builder, since the arm emulator is
deprecated and unavailable now. Switch to using a Github Action pl

build(deps): bump actions/checkout from 3 to 4 (#4490)

Requires updating the android builder, since the arm emulator is
deprecated and unavailable now. Switch to using a Github Action plugin
instead of a container, so that hopefully future updates will be
delivered via that channel instead.

Changed the idna test since printf returns EILSEQ for some byte
sequences in the format on Android in glibc. We don't fully understand
the cause, but we can avoid that by not asking it to reencode the bytes
in the current locale settings.

show more ...

doc: move cjihrig back to active maintainers (#4615)

kqueue: lower overhead in uv__io_check_fd (#4617)

Merge kevent calls along with the improvement of code simplicity.

Signed-off-by: Andy Pan <i@andypan.me>

unix: fix build breakage on haiku, openbsd, etc (#4618)

The compile-time detection check from commit 7b75935 ("kqueue: use
EVFILT_USER for async if available") was not being used, breaki

unix: fix build breakage on haiku, openbsd, etc (#4618)

The compile-time detection check from commit 7b75935 ("kqueue: use
EVFILT_USER for async if available") was not being used, breaking
numerous operating systems. This commit hopefully unbreaks them.

Fixes; https://github.com/libuv/libuv/issues/4608
Signed-off-by: Jeffrey H. Johnson <trnsz@pobox.com>

show more ...

doc: add scala-native-loop to LINKS.md (#4613)

unix,win: fix off-by-one in uv_wtf8_to_utf16() (#4609)

uv_wtf8_length_as_utf16() checks if codepoints are > 0xFFFF (to see if
it should be encoded as a surrogate pair), therefore uv_wtf8

unix,win: fix off-by-one in uv_wtf8_to_utf16() (#4609)

uv_wtf8_length_as_utf16() checks if codepoints are > 0xFFFF (to see if
it should be encoded as a surrogate pair), therefore uv_wtf8_to_utf16()
should too. Instead it checked > 0x1000. Harmonize the checks.

Fixes: https://github.com/nodejs/node/issues/55914

show more ...

kqueue: use EVFILT_USER for async if available (#4588)

Establishes a user event for kqueue to eliminate the overhead
of the pipe and the system call read(2) per wakeup event.

Re

kqueue: use EVFILT_USER for async if available (#4588)

Establishes a user event for kqueue to eliminate the overhead
of the pipe and the system call read(2) per wakeup event.

Relands commit 27134547ff using VSCode merge, since it shows the
conflict is just on the order of #ifdef calls.

Co-authored-by: Andy Pan <panjf2000@gmail.com>

show more ...

build: update minimum cmake to 3.10 (#4604)

Fixes a deprecation warning with new cmake versions.

Changing the minimum from 3.9 to 3.10 should be safe because there isn't
even a

build: update minimum cmake to 3.10 (#4604)

Fixes a deprecation warning with new cmake versions.

Changing the minimum from 3.9 to 3.10 should be safe because there isn't
even a non-EoL'd distro left that ships 3.10, let alone 3.9.

Fixes: https://github.com/libuv/libuv/issues/4603

show more ...

darwin,test: squelch EBUSY error on multicast join

The firewall was suspected to be the culprit, but the test
intermittently fails in the CI, not locally.

ci: make sure the macOS firewall is disabled

It seems to be disabled by default, for now, but let's log its status
and disable it just in case.

test: skip multicast join test on ENOEXEC

It happens due to the default firewall configuration on macOS >= 13.

Note: GH action runners have their firewall disabled, and yet, the tes

test: skip multicast join test on ENOEXEC

It happens due to the default firewall configuration on macOS >= 13.

Note: GH action runners have their firewall disabled, and yet, the test
fails all the same. Oh well...

Closes: https://github.com/libuv/libuv/issues/4263

show more ...

unix,win: map ENOEXEC errno

ci: run macOS and iOS tests also on macOS 14

The macOS 14 runners are ARM64 (in the non "large" version) whereas
macOS 13 runners are still x64, so keep that one around too.

Now working on version 1.49.3

Add SHA to ChangeLog

2024.10.18, Version 1.49.2 (Stable)

Changes since version 1.49.1:

* win,fs: remove trailing slash in junctions (Hüseyin Açacak)

* Revert "linux: eliminate a read on eventfd

2024.10.18, Version 1.49.2 (Stable)

Changes since version 1.49.1:

* win,fs: remove trailing slash in junctions (Hüseyin Açacak)

* Revert "linux: eliminate a read on eventfd per wakeup" (Ben Noordhuis)

* win: Fix linked list logic in getaddrinfo (Thad House)

* win: fix compilation against Windows 24H2 SDK (Thad House)

* win: remap ERROR_NOACCESS and ERROR_BUFFER_OVERFLOW (Jameson Nash)

* win,fs: match trailing slash presence in junctions to user input
(Jameson Nash)

show more ...

win,fs: match trailing slash presence in junctions to user input (#4590)

Refs: #4582

win: remap ERROR_NOACCESS and ERROR_BUFFER_OVERFLOW (#4567)

It seemed incorrect to map a segfault to EACCES, since posix would typically
map this to EFAULT. The ERROR_BUFFER_OVERFLOW is

win: remap ERROR_NOACCESS and ERROR_BUFFER_OVERFLOW (#4567)

It seemed incorrect to map a segfault to EACCES, since posix would typically
map this to EFAULT. The ERROR_BUFFER_OVERFLOW is literally "the filename is too
long", and is not typically an invalid parameter in posix.

Test originally added in #1060 to test the API, not the value.

show more ...

win: fix compilation against Windows 24H2 SDK (#4576)

Compilation against the 24H2 SDK is broken by #4327.

Fix that issue by only conditionally defining the new values.

Clo

win: fix compilation against Windows 24H2 SDK (#4576)

Compilation against the 24H2 SDK is broken by #4327.

Fix that issue by only conditionally defining the new values.

Closes #4575

show more ...

win: Fix linked list logic in getaddrinfo (#4578)

The logic in #4254 is incorrect, and results in the addrinfo linked
list only having a single result. Fix this by correcting the logic.

win: Fix linked list logic in getaddrinfo (#4578)

The logic in #4254 is incorrect, and results in the addrinfo linked
list only having a single result. Fix this by correcting the logic.

Closes #4577

show more ...

Revert "linux: eliminate a read on eventfd per wakeup (#4400)" (#4585)

This reverts commit e5cb1d3d3d4ab3178ac567fb6a7f0f4b5eef3083.

Reason: bisecting says it breaks dnstap.

Revert "linux: eliminate a read on eventfd per wakeup (#4400)" (#4585)

This reverts commit e5cb1d3d3d4ab3178ac567fb6a7f0f4b5eef3083.

Reason: bisecting says it breaks dnstap.

Also revert commit 27134547ff ("kqueue: use EVFILT_USER for async if
available") because otherwise the first commit doesn't revert cleanly,
with enough conflicts in src/unix/async.c that I'm not comfortable
fixing those up manually.

Fixes: https://github.com/libuv/libuv/issues/4584

show more ...

win,fs: remove trailing slash in junctions

Fixes: https://github.com/libuv/libuv/issues/3329

Now working on version 1.49.2

Add SHA to ChangeLog