unix: fix close-before-connect bug

Closing the TCP socket before the SYN/ACK handshake completed hung the event
loop, the connect req didn't get unregistered.

unix: streamline shutdown req cleanup logic

test: enable tcp_connect_error_fault test

test: add tcp connect timeout test

unix: back out new idle watcher for now

Its semantics don't quite match what node.js expects. This breaks the
stdio_over_pipes and shutdown_close_pipe tests but that can't be helped.

windows: support junctions with uv_fs_synlink
based on @piscisaureus implementation

unix, windows: rework reference counting scheme

This commit changes how the event loop determines if it needs to stay alive.

Previously, an internal counter was increased whenever a

unix, windows: rework reference counting scheme

This commit changes how the event loop determines if it needs to stay alive.

Previously, an internal counter was increased whenever a handle got created
and decreased again when the handle was closed.

While conceptually simple, it turned out hard to work with: you often want
to keep the event loop alive only if the handle is actually doing something.
Stopped or inactive handles were a frequent source of hanging event loops.

That's why this commit changes the reference counting scheme to a model where
a handle only references the event loop when it's active. 'Active' means
different things for different handle types, e.g.:

* timers: ticking
* sockets: reading, writing or listening
* processes: always active (for now, subject to change)
* idle, check, prepare: only active when started

This commit also changes how the uv_ref() and uv_unref() functions work: they
now operate on the level of individual handles, not the whole event loop.

The Windows implementation was done by Bert Belder.

show more ...

unix: change #if FIONBIO to #ifdef FIONBIO

Fixes compilation on Debian/ppc 6.0.4.

test: prevent accidental division by zero

Put POLL into the handle_map, so handle_size returns the size of it correctly.

sunos: make fs.watch scale better

Instead of using one port per watch, use one port for all the watches.
This is a cherry-pick of commit 7326962 from v0.6 into master.

Conflicts

sunos: make fs.watch scale better

Instead of using one port per watch, use one port for all the watches.
This is a cherry-pick of commit 7326962 from v0.6 into master.

Conflicts:

include/uv-private/uv-unix.h
src/unix/core.c
src/unix/sunos.c

show more ...

Windows: fix MinGW

MinGW's headers are missing a definition for SIO_BASE_HANDLE

Windows: skip GetFileAttributes call when opening a file

It wasn't working, and everything seemed to work fine nonetheless. Removing it just saves a syscall.

Windows: fix iocp emulation bugs for tcp handles

Makes ipc_tcp_connection pass on Windows XP

Tests: make test-tcp-write-error pass on XP

unix: zero out new loop struct

Fixes spurious segfaults in pyuv.

windows: remove unused variables

unix: don't use setlocale()

setlocale() is not inherently thread-safe. We'll have to live with the fact
that test/test-dlerror.c fails on systems with localized error messages.

unix, win: rework uv_dlopen() API

unix: change uv_dl*() error code

Return UV_ENOENT instead of UV_EINVAL. UV_EINVAL was arbitrarily chosen and
turns out to be inconsistent with the Windows implementation.

Fixes

unix: change uv_dl*() error code

Return UV_ENOENT instead of UV_EINVAL. UV_EINVAL was arbitrarily chosen and
turns out to be inconsistent with the Windows implementation.

Fixes #395.

show more ...

Rename uv_platform_socket_t to uv_os_sock_t

Benchmarks: add size of uv_poll_t to benchmark-sizes

Test: add tests for uv_poll

Unix: implement uv_poll

Windows: implement uv_poll