| 00:18.24 | *** join/#asterisk-dev infobot (ibot@rikers.org) |
| 00:18.24 | *** topic/#asterisk-dev is Asterisk Development Discussion -=- http://www.asterisk.org/developers -=- Tier 2 and 3.14159265 support is in #asterisk -=- Check out our blog! blogs.asterisk.org -=- Follow on Twitter at @AsteriskDev |
| 01:17.37 | *** join/#asterisk-dev leedm777 (~textual@173-16-52-28.client.mchsi.com) |
| 01:17.37 | *** mode/#asterisk-dev [+o leedm777] by ChanServ |
| 01:21.21 | *** join/#asterisk-dev leedm777 (~textual@173-16-52-28.client.mchsi.com) |
| 01:21.21 | *** mode/#asterisk-dev [+o leedm777] by ChanServ |
| 09:18.59 | *** join/#asterisk-dev file (uid178970@asterisk/developer-and-muffin-lover/file) |
| 09:18.59 | *** mode/#asterisk-dev [+o file] by ChanServ |
| 09:57.21 | *** join/#asterisk-dev jkroon (~jkroon@uls-154-73-32-14.wall.uls.co.za) |
| 12:44.47 | *** join/#asterisk-dev scgm11 (~scgm11@r186-52-191-4.dialup.adsl.anteldata.net.uy) |
| 12:57.49 | *** join/#asterisk-dev scgm11__ (~scgm11@r186-52-208-162.dialup.adsl.anteldata.net.uy) |
| 13:54.05 | *** join/#asterisk-dev newtonr (RustyNewto@nat/digium/x-toqasdomemxlqwfo) |
| 13:54.05 | *** mode/#asterisk-dev [+o newtonr] by ChanServ |
| 14:00.25 | *** join/#asterisk-dev putnopvut (putnopvut@asterisk/master-of-queues/mmichelson) |
| 14:00.25 | *** mode/#asterisk-dev [+o putnopvut] by ChanServ |
| 14:01.43 | *** join/#asterisk-dev scgm11 (~scgm11@r186-52-253-177.dialup.adsl.anteldata.net.uy) |
| 14:07.35 | *** join/#asterisk-dev kharwell (kharwell@nat/digium/x-emavsxpzfqjvmxrs) |
| 14:07.35 | *** mode/#asterisk-dev [+o kharwell] by ChanServ |
| 14:07.53 | *** join/#asterisk-dev scgm11 (~scgm11@r186-52-253-177.dialup.adsl.anteldata.net.uy) |
| 14:07.54 | seanbright | MONDAY |
| 14:08.04 | seanbright | WHY HAVE YOU FORSAKEN ME!? |
| 14:14.51 | *** join/#asterisk-dev rmudgett (rmudgett@nat/digium/x-pvjsammokdbhdold) |
| 14:14.51 | *** mode/#asterisk-dev [+o rmudgett] by ChanServ |
| 14:36.08 | *** join/#asterisk-dev cresl1n (Adium@asterisk/libpri-and-libss7-expert/Cresl1n) |
| 14:36.08 | *** mode/#asterisk-dev [+o cresl1n] by ChanServ |
| 15:24.52 | seanbright | in regards to ASTERISK-26940 |
| 15:25.30 | seanbright | aren't there standard patterns for non-blocking reads? |
| 15:25.37 | seanbright | is the session lock the problem? |
| 15:27.37 | *** join/#asterisk-dev leedm777 (textual@nat/digium/x-xvnfpyauujuhwcyt) |
| 15:27.37 | *** mode/#asterisk-dev [+o leedm777] by ChanServ |
| 15:34.24 | *** join/#asterisk-dev leedm777 (textual@nat/digium/x-ohvfagjwxxarvdrt) |
| 15:34.24 | *** mode/#asterisk-dev [+o leedm777] by ChanServ |
| 15:38.44 | rmudgett | seanbright: The memory consumption is because there is a log message in such a tight loop. |
| 15:39.35 | *** join/#asterisk-dev leedm777 (textual@nat/digium/x-orfvrqhzerulvmdg) |
| 15:39.35 | *** mode/#asterisk-dev [+o leedm777] by ChanServ |
| 15:43.17 | seanbright | i just think the function is poorly written |
| 15:44.34 | seanbright | i assume the poll() without the timeout is there to prevent taking the session lock unnecessarily |
| 15:45.46 | seanbright | but you could also just do a blocking read() into a temporary buffer until you have enough data and then lock/copy/unlock |
| 16:03.42 | rmudgett | seanbright: There is no question of how elegant it is at giving heart burn if you are prone to such a thing. :) |
| 16:19.09 | *** join/#asterisk-dev leedm777 (textual@nat/digium/x-mdsucqyoszopgwdb) |
| 16:19.09 | *** mode/#asterisk-dev [+o leedm777] by ChanServ |
| 19:02.02 | *** join/#asterisk-dev zopsi (~zopsi@dir.ac) |
| 19:22.51 | *** join/#asterisk-dev cresl1n (Adium@asterisk/libpri-and-libss7-expert/Cresl1n) |
| 19:22.51 | *** mode/#asterisk-dev [+o cresl1n] by ChanServ |
| 19:29.21 | *** join/#asterisk-dev leedm777 (textual@nat/digium/x-plxukdajfenzdlby) |
| 19:29.21 | *** mode/#asterisk-dev [+o leedm777] by ChanServ |
| 19:48.31 | *** join/#asterisk-dev leedm777 (textual@nat/digium/x-afitblsobkdgdgko) |
| 19:48.32 | *** mode/#asterisk-dev [+o leedm777] by ChanServ |
| 20:46.31 | seanbright | for the alert pipe on a channel |
| 20:47.00 | seanbright | is there an expectation that the reads & writes and lock stepped? |
| 20:49.09 | rmudgett | There should be an alert pipe write for every frame put in the read queue. There should be an alert pipe read for every queue read. The alert pipe is what signals the channel thread to read the queue. |
| 20:50.22 | seanbright | right |
| 20:50.33 | seanbright | so the only exception is flush which consumes everything |
| 20:50.39 | seanbright | otherwise its a one-for-one |
| 20:51.44 | rmudgett | yes. An imbalance may have issues depending upon how it is imbalanced. |
| 20:52.23 | rmudgett | The alert pipe is also only used when timerfd is being used. |
| 20:52.37 | seanbright | ahhh. interesting. |
| 20:53.32 | rmudgett | s/used/needed |
| 20:58.48 | *** join/#asterisk-dev malcolmd (malcolmd@pdpc/sponsor/digium/malcolmd) |
| 20:58.48 | *** mode/#asterisk-dev [+o malcolmd] by ChanServ |
| 21:09.30 | seanbright | rmudgett: only timerfd? not pthread? |
| 21:09.48 | seanbright | or kqueue |
| 21:10.55 | rmudgett | Don't know anything about kqueue but only timerfd uses/needs the alert pipe |
| 21:11.54 | seanbright | ok, cool |
| 21:11.55 | seanbright | thanks |
| 21:15.37 | *** join/#asterisk-dev scgm11 (~scgm11@r186-52-208-162.dialup.adsl.anteldata.net.uy) |
| 21:34.59 | *** join/#asterisk-dev elguero (~miguel323@74-95-21-41-Connecticut.hfc.comcastbusiness.net) |
| 22:01.42 | *** join/#asterisk-dev scgm11 (~scgm11@r186-52-208-162.dialup.adsl.anteldata.net.uy) |
| 22:45.33 | scgm11 | @rmudgett patchset 6 more than 45K no issue till now |
| 23:31.14 | *** part/#asterisk-dev kharwell (kharwell@nat/digium/x-emavsxpzfqjvmxrs) |