unix: simplify atomic op in uv_tty_reset_mode() (#3773)

Compare-and-exchange is only useful in a loop. Replace it with a simple
exchange.

doc: update LINKS.md (#3772)

win: fix uv__pipe_accept memory leak (#3767)

build: restore qemu download logic (#3768)

I recently changed it to download a fixed .deb but seems it's updated
more frequently than I anticipated because the dfsg-7ubuntu1_package is

build: restore qemu download logic (#3768)

I recently changed it to download a fixed .deb but seems it's updated
more frequently than I anticipated because the dfsg-7ubuntu1_package is
already gone, replaced with dfsg-7ubuntu2.

Bring back the downloader logic that fetches the filename from the
directory listing.

show more ...

unix,win: fix memory leak in uv_fs_scandir() (#3760)

uv_fs_scandir() leaked an entry when you called it on a directory with
a single entry _and_ you didn't run the iterator until UV_EOF.

unix,win: fix memory leak in uv_fs_scandir() (#3760)

uv_fs_scandir() leaked an entry when you called it on a directory with
a single entry _and_ you didn't run the iterator until UV_EOF.

Fixes: https://github.com/libuv/libuv/issues/3748

show more ...

unix: remove pre-macos 10.8 compatibility hack (#3759)

linux: move epoll.c back into linux-core.c

epoll.c is only used on Android and Linux after commit 5fe59726 ("sunos:
restore use of event ports") so move it back into linux-core.c

linux: move epoll.c back into linux-core.c

epoll.c is only used on Android and Linux after commit 5fe59726 ("sunos:
restore use of event ports") so move it back into linux-core.c

This commit removes a workaround for pre-2.6.27 kernels that don't have
the epoll_create1() system call.

show more ...

build: upgrade qemu-user-static package

4.2 is three years old by now. The linux/alpha emulator in that version
doesn't support the epoll_create1() system call.

linux: teach uv_get_constrained_memory() cgroupsv2 (#3744)

Fixes: https://github.com/libuv/libuv/issues/2315

kqueue: DRY file descriptor deletion logic (#3746)

Remove the TODO that says to batch up kevent(EV_DELETE) system calls.
It's unduly complicated in case of errors.

If you pass a

kqueue: DRY file descriptor deletion logic (#3746)

Remove the TODO that says to batch up kevent(EV_DELETE) system calls.
It's unduly complicated in case of errors.

If you pass an out array to kevent(), it will store error records that
tell you which file descriptor caused the error but it also stores new
events in the array, complicating state management.

If you don't pass an out array, it simply doesn't tell you anything
except that _something_ failed. Optimistically trying batch deletion
and falling back to one-by-one deletion is probably possible but is
something of a deoptimization and also feels somewhat dangerous.

macOS has a mostly-undocumented kevent_qos() system call that accepts a
KEVENT_FLAG_ERROR_EVENTS flag that fixes the first problem but that of
course isn't portable.

Long story short, it seems like too much hassle for too little payoff.
Libuv has been doing one-by-one deletion for over a decade now and no
one complained about performance so far so let's just stick with that.

show more ...

test: fix musl libc.a dlerror() test expectation (#3735)

Dynamic loading with musl only works with libc.so because it is
is implemented in musl's dynamic linker, not in libc proper.

test: fix musl libc.a dlerror() test expectation (#3735)

Dynamic loading with musl only works with libc.so because it is
is implemented in musl's dynamic linker, not in libc proper.

libc.a contains just a stub dlopen() that always fails with a "Dynamic
loading not supported" error message. Update the test to handle that.

Fix: https://github.com/libuv/libuv/issues/3706

show more ...

doc: add note about offset -1 in uv_fs_read/write (#3677)

unix: fix uv_getrusage() ru_maxrss reporting (#3721)

Most platforms report it in kilobytes but macOS and Solaris report it in
bytes and pages respectively.

Fixes: https://github

unix: fix uv_getrusage() ru_maxrss reporting (#3721)

Most platforms report it in kilobytes but macOS and Solaris report it in
bytes and pages respectively.

Fixes: https://github.com/nodejs/node/issues/44332

show more ...

win: support Windows 11 in uv_os_uname() (#3718)

Fixes: https://github.com/libuv/libuv/issues/3381
Fixes: https://github.com/nodejs/node/issues/40862

unix,win: remove UV_HANDLE_SHUTTING flag (#3705)

Replace it with a `uv__is_stream_shutting()` macro that checks the
`shutdown_req` field.

It partially fixes: https://github.com/

unix,win: remove UV_HANDLE_SHUTTING flag (#3705)

Replace it with a `uv__is_stream_shutting()` macro that checks the
`shutdown_req` field.

It partially fixes: https://github.com/libuv/libuv/issues/3663.

show more ...

win: fix thread race in uv_cwd() and uv_chdir() (#3708)

Another thread can change the working directory between calls to
GetCurrentDirectoryW().

The first call was used to deter

win: fix thread race in uv_cwd() and uv_chdir() (#3708)

Another thread can change the working directory between calls to
GetCurrentDirectoryW().

The first call was used to determine the size of the buffer to use in
the second call. Retry if the reported size does not match the expected
size because the buffer's contents is undefined in that case.

show more ...

build: switch ci from macos-10.15 to macos-11 (#3707)

The former will be removed by GitHub in August.

Fixes: https://github.com/libuv/libuv/issues/3697

win: fix off-by-1 buffer overrun in uv_exepath() (#3695)

uv_exepath() wrote the nul byte *after* the end of the buffer. It's not
necessary to write said nul byte in the first place becau

win: fix off-by-1 buffer overrun in uv_exepath() (#3695)

uv_exepath() wrote the nul byte *after* the end of the buffer. It's not
necessary to write said nul byte in the first place because that was a
workaround for a bug in the Windows XP version of GetModuleFileName().

Fix uv_cwd() in the same fashion, it doesn't need the nul byte either.

Fixes: https://github.com/libuv/libuv/issues/3691

show more ...

unix: don't accept() connections in a loop (#3696)

After analysis of many real-world programs I've come to conclude that
accepting in a loop is nearly always suboptimal.

1. 99.9

unix: don't accept() connections in a loop (#3696)

After analysis of many real-world programs I've come to conclude that
accepting in a loop is nearly always suboptimal.

1. 99.9% of the time the second accept() call fails with EAGAIN, meaning
there are no additional connections to accept. Not super expensive
in isolation but it adds up.

2. When there are more connections to accept but the listen socket is
shared between multiple processes (ex. the Node.js cluster module),
libuv's greedy behavior necessitated the UV_TCP_SINGLE_ACCEPT hack
to slow it down in order to give other processes a chance.

Accepting a single connection and relying on level-triggered polling to
get notified on the next incoming connection both simplifies the code
and optimizes for the common case.

show more ...

unix: switch to c11 atomics (#3688)

Fixes: https://github.com/libuv/libuv/issues/3683

build: enable MSVC_RUNTIME_LIBRARY setting (#3624)

android: remove pthread-fixes.c (#3674)

This was a kludge for a bug in old versions (API level <= 16) of the
Android SDK.

The os390 port had a build dependency on the file but d

android: remove pthread-fixes.c (#3674)

This was a kludge for a bug in old versions (API level <= 16) of the
Android SDK.

The os390 port had a build dependency on the file but does not actually
use it so that too has been removed.

show more ...

win: remove stdint-msvc2008.h (#3673)

Support for VS 2008 was removed two years ago in commit a779fccfd so
this file can safely be deleted.

Now working on version 1.44.3

Fixes: https://github.com/libuv/libuv/issues/3586

Add SHA to ChangeLog