include: uv_spawn takes const uv_process_options_t*

Passing or returning structs as values makes life hard for people that
work with libuv through a foreign function interface. Switch to

include: uv_spawn takes const uv_process_options_t*

Passing or returning structs as values makes life hard for people that
work with libuv through a foreign function interface. Switch to a
pointer-based approach.

Fixes #684.

show more ...

include: uv_udp_send{6} now takes sockaddr_in*

Passing or returning structs as values makes life hard for people that
work with libuv through a foreign function interface. Switch to a

include: uv_udp_send{6} now takes sockaddr_in*

Passing or returning structs as values makes life hard for people that
work with libuv through a foreign function interface. Switch to a
pointer-based approach.

Fixes #684.

show more ...

include: uv_udp_bind{6} now takes sockaddr_in*

Passing or returning structs as values makes life hard for people that
work with libuv through a foreign function interface. Switch to a

include: uv_udp_bind{6} now takes sockaddr_in*

Passing or returning structs as values makes life hard for people that
work with libuv through a foreign function interface. Switch to a
pointer-based approach.

Fixes #684.

show more ...

include: uv_udp_recv_cb now takes const uv_buf_t*

Passing or returning structs as values makes life hard for people that
work with libuv through a foreign function interface. Switch to a

include: uv_udp_recv_cb now takes const uv_buf_t*

Passing or returning structs as values makes life hard for people that
work with libuv through a foreign function interface. Switch to a
pointer-based approach.

Fixes #684.

show more ...

include: uv_tcp_connect{6} now takes sockaddr_in*

Passing or returning structs as values makes life hard for people that
work with libuv through a foreign function interface. Switch to a

include: uv_tcp_connect{6} now takes sockaddr_in*

Passing or returning structs as values makes life hard for people that
work with libuv through a foreign function interface. Switch to a
pointer-based approach.

Fixes #684.

show more ...

include: uv_tcp_bind{6} now takes sockaddr_in*

Passing or returning structs as values makes life hard for people that
work with libuv through a foreign function interface. Switch to a

include: uv_tcp_bind{6} now takes sockaddr_in*

Passing or returning structs as values makes life hard for people that
work with libuv through a foreign function interface. Switch to a
pointer-based approach.

Fixes #684.

show more ...

include: uv_ip[46]_addr now takes sockaddr_in*

Passing or returning structs as values makes life hard for people that
work with libuv through a foreign function interface. Switch to a

include: uv_ip[46]_addr now takes sockaddr_in*

Passing or returning structs as values makes life hard for people that
work with libuv through a foreign function interface. Switch to a
pointer-based approach.

Fixes #684.

show more ...

include: uv_read{2}_cb now takes const uv_buf_t*

Passing or returning structs as values makes life hard for people that
work with libuv through a foreign function interface. Switch to a

include: uv_read{2}_cb now takes const uv_buf_t*

Passing or returning structs as values makes life hard for people that
work with libuv through a foreign function interface. Switch to a
pointer-based approach.

Fixes #684.

show more ...

include: uv_alloc_cb now takes uv_buf_t*

Passing or returning structs as values makes life hard for people that
work with libuv through a foreign function interface. Switch to a
poin

include: uv_alloc_cb now takes uv_buf_t*

Passing or returning structs as values makes life hard for people that
work with libuv through a foreign function interface. Switch to a
pointer-based approach.

Fixes #684.

show more ...

test: fix epoll_wait() usage in test-embed.c

Passing 0 as the `maxevents` argument fails with EINVAL, passing NULL as
the `events` argument fails with EFAULT.

Fixes #915.

Now working on v0.11.12

2013.08.30, Version 0.11.11 (Unstable)

Changes since version 0.11.10:

* unix, windows: add thread-local storage API (Ben Noordhuis)

* linux: don't turn on SO_REUSEPORT sock

2013.08.30, Version 0.11.11 (Unstable)

Changes since version 0.11.10:

* unix, windows: add thread-local storage API (Ben Noordhuis)

* linux: don't turn on SO_REUSEPORT socket option (Ben Noordhuis)

* darwin: fix 10.6 build error in fsevents.c (Ben Noordhuis)

* windows: make uv_shutdown() for write-only pipes work (Bert Belder)

* include: update uv_udp_open() / uv_udp_bind() docs (Ben Noordhuis)

* unix: req queue must be empty when destroying loop (Ben Noordhuis)

* unix: move loop functions from core.c to loop.c (Ben Noordhuis)

* darwin: remove CoreFoundation dependency (Ben Noordhuis)

* windows: make autotools build system work with mingw (Keno Fischer)

* windows: fix mingw build (Alex Crichton)

* windows: tweak Makefile.mingw for easier usage (Alex Crichton)

* build: remove _GNU_SOURCE macro definition (Ben Noordhuis)

show more ...

windows: make uv_shutdown() for write-only pipes work

A couple of issues prevented uv_shutdown() from working correctly with
write-only pipes.

* The pipe handle wasn't opened

windows: make uv_shutdown() for write-only pipes work

A couple of issues prevented uv_shutdown() from working correctly with
write-only pipes.

* The pipe handle wasn't opened with the right permissions, so an
attempt to probe the state of the write buffer would fail with
ERROR_ACCESS_DENIED.

* The pipe flags for child process stdio pipes were always set to
UV_HANDLE_READABLE and UV_HANDLE_WRITABLE, even in cases where it
was actually half-duplex.

* There was no code path that lead to closing the pipe handle if the
pipe was write-only.

show more ...

windows: make uv_shutdown() for write-only pipes work

A couple of issues prevented uv_shutdown() from working correctly with
write-only pipes.

* The pipe handle wasn't opened

windows: make uv_shutdown() for write-only pipes work

A couple of issues prevented uv_shutdown() from working correctly with
write-only pipes.

* The pipe handle wasn't opened with the right permissions, so an
attempt to probe the state of the write buffer would fail with
ERROR_ACCESS_DENIED.

* The pipe flags for child process stdio pipes were always set to
UV_HANDLE_READABLE and UV_HANDLE_WRITABLE, even in cases where it
was actually half-duplex.

* There was no code path that lead to closing the pipe handle if the
pipe was write-only.

show more ...

darwin: fix 10.6 build error in fsevents.c

Work around an 'initializer element is not constant' build error in
src/unix/fsevents.c by turning the const int flags into #defines.

darwin: fix 10.6 build error in fsevents.c

Work around an 'initializer element is not constant' build error in
src/unix/fsevents.c by turning the const int flags into #defines.

Only an issue on OS X 10.6 due to the old compiler it uses.

Fixes #908.

show more ...

darwin: remove CoreFoundation dependency

Load the required symbols at run-time rather than linking against the
CoreFoundation (and CoreServices and ApplicationServices) frameworks
at

darwin: remove CoreFoundation dependency

Load the required symbols at run-time rather than linking against the
CoreFoundation (and CoreServices and ApplicationServices) frameworks
at build time.

Should make integration easier for people that bundle libuv with their
own projects because they no longer have to replicate magic -framework
incantations in their top-level build system.

show more ...

unix: move loop functions from core.c to loop.c

Move uv_default_loop(), uv_loop_new() and uv_loop_delete() to loop.c.

unix: req queue must be empty when destroying loop

Assert that the request queue is empty when destroying the event loop.

Should catch errors where people call uv_loop_delete() when

unix: req queue must be empty when destroying loop

Assert that the request queue is empty when destroying the event loop.

Should catch errors where people call uv_loop_delete() when there are
still in-progress work requests (for example.)

show more ...

windows: fix mingw build

On older mingw installations, DEVICE_TYPE isn't defined. This was
redefined in windows.h, but it was done after it was already attempted
to be used. This mov

windows: fix mingw build

On older mingw installations, DEVICE_TYPE isn't defined. This was
redefined in windows.h, but it was done after it was already attempted
to be used. This moves the #define for DEVICE_TYPE above the usage of
it in windows.h to prevent compilation failures.

show more ...

windows: tweak Makefile.mingw for easier usage

Default CC to `gcc` instead of `cc` because at least on the mingw
installation I had there was no `cc` executable. This avoids having
t

windows: tweak Makefile.mingw for easier usage

Default CC to `gcc` instead of `cc` because at least on the mingw
installation I had there was no `cc` executable. This avoids having
to type `CC=gcc` in all the `make` commands.

Also append to `CFLAGS` instead of defining if not previously defined.
These flags are required to build libuv, so they should not be
overridden if other extra CFLAGS are supplied via the command
line as well.

show more ...

unix, windows: add thread-local storage API

Uses the pthread_key_{create,delete} and pthread_{get,set}specific
functions on UNIX platforms, Tls{Alloc,Free} and Tls{Get,Set}Value
on W

unix, windows: add thread-local storage API

Uses the pthread_key_{create,delete} and pthread_{get,set}specific
functions on UNIX platforms, Tls{Alloc,Free} and Tls{Get,Set}Value
on Windows.

Fixes #904.

show more ...

Revert "build: remove mingw makefile"

As requested by the Rust people.

This reverts commit 8f3621e4dd2d55befd0f0b8635cf8de17f849aff.

include: update uv_udp_open() / uv_udp_bind() docs

Mention that:

* these functions set the SO_REUSEADDR and SO_REUSEPORT socket flags,
* what the effect of those flags is, and

include: update uv_udp_open() / uv_udp_bind() docs

Mention that:

* these functions set the SO_REUSEADDR and SO_REUSEPORT socket flags,
* what the effect of those flags is, and
* that we may remove it someday

show more ...

linux: don't turn on SO_REUSEPORT socket option

On the BSDs, SO_REUSEPORT is pretty much SO_REUSEADDR with some special
casing for IP multicast. When two processes (that don't do multic

linux: don't turn on SO_REUSEPORT socket option

On the BSDs, SO_REUSEPORT is pretty much SO_REUSEADDR with some special
casing for IP multicast. When two processes (that don't do multicast)
bind to the same address, only the last one receives traffic. It allows
one to "steal" the bound address from another process. (Both processes
have to enable SO_REUSEPORT though, so it only works in a cooperative
setting.)

On Linux however, it enables port sharing, not stealing - both processes
receive a share of the traffic. This is a desirable trait but pre-3.9
kernels don't support the socket option and a libuv program therefore
behaves differently with older kernels or on another platform.

The difference in behavior (sharing vs. stealing) is, in my opinion,
big enough and confusing enough that it merits a rollback. People
that want this kind of functionality can prepare the socket manually
and hand it off to uv_udp_open().

This commit effectively reverts commit 17452cd.

show more ...

build: remove _GNU_SOURCE macro definition

Compile libuv without -D_GNU_SOURCE, remove mention from the README.

The only place where it's still used is in the test suite and only

build: remove _GNU_SOURCE macro definition

Compile libuv without -D_GNU_SOURCE, remove mention from the README.

The only place where it's still used is in the test suite and only
because test/test-fs.c uses struct stat directly.

show more ...