We avoid the problem by using poll(2).
On systems without poll(2) (older bsd-ish systems, and win32), we emulate
poll(2) using select(2) and check for valid descriptors before attempting
to access them via the descriptor sets.
If an out-of-range descriptor is detected, an E_WARNING is raised suggesting
that PHP should be recompiled with a larger FD_SETSIZE (and also with a
suggested value).
Most uses of select(2) in the source are to poll a single descriptor, so
a couple of handy wrapper functions have been added to make this easier.
A configure option --enable-fd-setsize has been added to both the unix and
win32 builds; on unix we default to 16384 and on windows we default to 256.
Windows FD_SETSIZE imposes a limit on the maximum number of descriptors that
can be select()ed at once, whereas the unix FD_SETSIZE limit is based on the
highest numbered descriptor; 256 should be plenty for PHP scripts under windows
(the default OS setting is 64).
The win32 specific parts are untested; will do that now.
Register filters as resources when
instantiated by stream_filter_(ap|pre)pend().
Export php_stream_filter_flush() internal function to wind buffered data
out of a particular filter until consumed by a later filter or sent to
stream->readbuffer or stream->ops->write()
Disables a wrapper (user-defined or built-in) for the life of the request.
Add stream_wrapper_restore()
Restores the wrapper originally defined at the time the request started
to the protocol name mentioned.
used to return "" and not bool(false). It's not worth keeping it because
STR_FREE() and zval_dtor() always have to check for it and it slows down
the general case. In addition, it seems that empty_string has been abused
quite a lot, and was used not only for setting zval's but generally in
PHP code instead of "", which wasn't the intention. Last but not least,
nuking empty_string should improve stability as I doubt every place
correctly checked if they are not mistakenly erealloc()'ing it or
calling efree() on it.
NOTE: Some code is probably broken. Each extension maintainer should
check and see that my changes are OK. Also, I haven't had time to touch
PECL yet. Will try and do it tomorrow.
Userdefined wrappers were being registered into a global wrapper hash
which can cross threads. Termination of once instance then has the
potential to leave an active stream in another instance with no wrapper
leading to segfault.
