Mercurial > pidgin.yaz
view doc/core-signals.dox @ 14960:0c80979077cf
[gaim-migrate @ 17738]
* Patch from Christopher "cuberoot" to fix the repeated, CPU-intensive call loop on url_fetch_recv_cb() which was discussed on gaim-devl with the subject "Yahoo (and maybe others): gaim_util_fetch_url_request() at 100% CPU until complete with repeated EAGAIN". He writes:
"When read() returns 0, that means eof... period. Don't check errno, don't collect $200.
The loop was caused by eof condition when errno had previously been EAGAIN/EWOULDBLOCK. The CPU-bound loop stopped when errno happened to be overwritten.
I also believe trying to use some/part of an HTTP request on ETIMEDOUT to be a mistake so I removed it. Funny enough the read(2) manpage doesn't document ETIMEDOUT as a possible errno, but it is. (see tcp_timer.c and tcp_subr.c in src/sys/netinet)."
This closes Adium Trac ticket #5685.
* Added some passing-NULL-to-printf() safety which has been sitting around in this file locally for a while, waiting for a worthwhile commit to accompany.
committer: Tailor Script <tailor@pidgin.im>
author | Evan Schoenberg <evan.s@dreskin.net> |
---|---|
date | Sat, 11 Nov 2006 20:29:35 +0000 |
parents | 1ec93dd2f359 |
children | 1414e0e01dc5 |
line wrap: on
line source
/** @page core-signals Core Signals @signals @signal quitting @endsignals <hr> @signaldef quitting @signalproto void (*quitting)(); @endsignalproto @signaldesc Emitted when gaim is quitting. @endsignaldef */ // vim: syntax=c tw=75 et