| 00:20.00 | *** join/#asterisk-dev galeras (n=galeras@186.80.192.102) |
| 00:21.38 | galeras | any idea?: somtimes res_monitor.c chan->monitor->filename_base has duplicated path (version 1.4.27) |
| 00:35.00 | *** join/#asterisk-dev snuff-work (n=aaa@120.17.144.110) |
| 00:35.01 | *** mode/#asterisk-dev [+o snuff-work] by ChanServ |
| 01:50.20 | *** join/#asterisk-dev ticoit (n=ticoit@201.191.187.58) |
| 02:15.16 | *** join/#asterisk-dev kpfleming (n=kpflemin@asterisk/digium-director-of-software-tech/kpfleming) |
| 02:15.16 | *** mode/#asterisk-dev [+o kpfleming] by ChanServ |
| 02:15.32 | Corydon76-dig | Evening, Kevin |
| 02:17.04 | kpfleming | howdy |
| 02:19.16 | Corydon76-dig | Did that explanation on ffsll make sense? |
| 02:20.32 | kpfleming | yep |
| 02:20.36 | kpfleming | we'll need to test it |
| 02:20.50 | kpfleming | the __builtin_ffsll version does compile on freebsd, at least |
| 02:20.59 | Corydon76-dig | Which version? |
| 02:21.12 | Corydon76-dig | ffsll was introduced in FreeBSD 7, I think |
| 02:21.36 | kpfleming | whichever one buildbot is on... freebsd 6 i think |
| 02:22.37 | kpfleming | the comments in the gcc manual about using the 'library function' just means it gets the implementation from libgcc if it's not inlined |
| 02:22.45 | kpfleming | in either case, though, it's provided by gcc |
| 02:30.24 | *** join/#asterisk-dev seanbright (n=sean@asterisk/contributor-and-bug-marshal/seanbright) |
| 02:30.24 | *** mode/#asterisk-dev [+o seanbright] by ChanServ |
| 02:33.44 | *** join/#asterisk-dev ticoit (n=ticoit@201.191.187.58) |
| 02:36.33 | *** part/#asterisk-dev alecdavis (n=alec@202.36.76.21) |
| 02:45.59 | *** join/#asterisk-dev xpot-mobile (n=xpot@70-91-210-233-BusName-Utah.hfc.comcastbusiness.net) |
| 03:56.35 | *** join/#asterisk-dev paravoid (n=paravoid@zebra.cslab.ece.ntua.gr) |
| 04:42.20 | *** join/#asterisk-dev felipe_ (n=felipe@my.nada.kth.se) |
| 04:46.53 | *** join/#asterisk-dev moy (n=moy@CPE001cdfec4cee-CM00080dab8485.cpe.net.cable.rogers.com) |
| 05:51.28 | *** join/#asterisk-dev ticoit (n=ticoit@201.191.187.58) |
| 05:52.32 | *** join/#asterisk-dev ticoit (n=ticoit@201.191.187.58) |
| 06:50.18 | *** join/#asterisk-dev |Cybex| (n=John@atwork-20.r-212.178.82.atwork.nl) |
| 07:09.20 | *** join/#asterisk-dev snuff-work (n=aaa@210.9.82.197) |
| 07:09.21 | *** mode/#asterisk-dev [+o snuff-work] by ChanServ |
| 07:29.01 | *** join/#asterisk-dev tzafrir_laptop (n=tzafrir@local.xorcom.com) |
| 07:29.01 | *** mode/#asterisk-dev [+o tzafrir_laptop] by ChanServ |
| 07:30.37 | *** join/#asterisk-dev ticoit (n=ticoit@201.191.187.58) |
| 07:48.30 | *** join/#asterisk-dev kazaa_lite (n=msaleem@94-193-98-124.zone7.bethere.co.uk) |
| 08:03.30 | *** join/#asterisk-dev mpe (n=mpe@gate.ipvision.dk) |
| 08:18.23 | *** join/#asterisk-dev tamiel (n=tamiel@213.30.183.226) |
| 08:51.30 | *** join/#asterisk-dev fish9370 (n=Miranda@79.120.100.132) |
| 10:26.30 | *** join/#asterisk-dev snuff-home (n=snuffy@203-219-29-182-bri-ts6-2600.tpgi.com.au) |
| 10:26.31 | *** mode/#asterisk-dev [+o snuff-home] by ChanServ |
| 11:09.36 | *** join/#asterisk-dev caio1982 (i=caio1982@CAcert-br/caio1982) |
| 11:28.14 | *** join/#asterisk-dev sgimeno (n=sgimeno@226.Red-80-33-64.staticIP.rima-tde.net) |
| 11:40.33 | *** join/#asterisk-dev MAbbas (i=Jinbaba@115.186.16.74) |
| 11:42.25 | MAbbas | Hi people .. |
| 11:42.30 | mvanbaak | hi |
| 11:43.22 | *** join/#asterisk-dev snuff-home (n=snuffy@203-219-29-182-bri-ts6-2600.tpgi.com.au) |
| 11:43.23 | *** mode/#asterisk-dev [+o snuff-home] by ChanServ |
| 11:44.00 | MAbbas | I want to modify queue behaviour, so that whenever a calls comes in .. instead of queue using policy like round robin ... |
| 11:44.36 | MAbbas | ask my application on a seperate server .. which agent to transfer call |
| 11:45.10 | MAbbas | I have modified code for astersik ver 1.4.26 |
| 11:45.32 | MAbbas | Will my modified code work on every asterisk version .. |
| 11:45.46 | mvanbaak | probably not |
| 11:58.33 | *** join/#asterisk-dev file (n=file@asterisk/developer-and-muffin-lover/file) |
| 11:58.33 | *** join/#asterisk-dev mnicholson_ (n=mnichols@nat/digium/x-papvixpbbkjgeobr) |
| 11:58.33 | *** join/#asterisk-dev angler (n=angler@pdpc/sponsor/digium/angler) |
| 11:58.33 | *** join/#asterisk-dev malcolmd (n=malcolmd@pdpc/sponsor/digium/malcolmd) |
| 11:58.33 | *** join/#asterisk-dev Corydon76-dig (i=blush@c-69-137-80-31.hsd1.tn.comcast.net) |
| 11:58.34 | *** join/#asterisk-dev jamesh1 (n=jhenders@xob.neospire.net) |
| 11:58.34 | *** join/#asterisk-dev mog (n=mog@c-68-62-169-247.hsd1.al.comcast.net) |
| 11:58.34 | *** mode/#asterisk-dev [+oooo file malcolmd Corydon76-dig mog] by irc.freenode.net |
| 11:58.34 | *** join/#asterisk-dev jsmith-away (n=jsmith@asterisk/training-and-documentation-guru/jsmith) |
| 11:58.34 | *** join/#asterisk-dev Mw3 (i=mw3@mw3.hu) |
| 11:58.34 | *** join/#asterisk-dev tzanger (n=tzanger@gromit.mixdown.ca) [NETSPLIT VICTIM] |
| 11:58.34 | *** join/#asterisk-dev corramor (n=corramor@nat/digium/x-rkkzxkyswyktomqg) [NETSPLIT VICTIM] |
| 11:58.34 | *** join/#asterisk-dev russellb (n=russellb@asterisk/digium-open-source-team-lead/russellb) |
| 11:58.34 | *** join/#asterisk-dev snuff-work (n=aaa@210.9.82.197) [NETSPLIT VICTIM] |
| 11:58.34 | *** join/#asterisk-dev snuff-home (n=snuffy@203-219-29-182-bri-ts6-2600.tpgi.com.au) [NETSPLIT VICTIM] |
| 11:58.34 | *** mode/#asterisk-dev [+oooo jsmith-away russellb snuff-work snuff-home] by irc.freenode.net |
| 11:58.34 | *** join/#asterisk-dev MAbbas (i=Jinbaba@115.186.16.74) |
| 11:58.34 | *** join/#asterisk-dev caio1982 (i=caio1982@CAcert-br/caio1982) |
| 11:58.34 | *** join/#asterisk-dev fish9370 (n=Miranda@79.120.100.132) |
| 11:58.34 | *** join/#asterisk-dev tamiel (n=tamiel@213.30.183.226) |
| 11:58.34 | *** join/#asterisk-dev mpe (n=mpe@gate.ipvision.dk) |
| 11:58.34 | *** join/#asterisk-dev kazaa_lite (n=msaleem@94-193-98-124.zone7.bethere.co.uk) [NETSPLIT VICTIM] |
| 11:58.34 | *** join/#asterisk-dev tzafrir_laptop (n=tzafrir@local.xorcom.com) |
| 11:58.34 | *** join/#asterisk-dev |Cybex| (n=John@atwork-20.r-212.178.82.atwork.nl) |
| 11:58.35 | *** join/#asterisk-dev felipe_ (n=felipe@my.nada.kth.se) |
| 11:58.35 | *** join/#asterisk-dev paravoid (n=paravoid@zebra.cslab.ece.ntua.gr) [NETSPLIT VICTIM] |
| 11:58.35 | *** join/#asterisk-dev seanbright (n=sean@asterisk/contributor-and-bug-marshal/seanbright) |
| 11:58.35 | *** join/#asterisk-dev philipp64 (n=chatzill@63.224.43.239) |
| 11:58.35 | *** join/#asterisk-dev JunK-Y (n=junky@modemcable206.69-20-96.mc.videotron.ca) |
| 11:58.35 | *** join/#asterisk-dev TheDavidFactor (n=chatzill@fw1.safedataisp.net) |
| 11:58.35 | *** join/#asterisk-dev DocAwesome (n=Leif@asterisk/documenteur-extraordinaire/blitzrage) |
| 11:58.35 | *** join/#asterisk-dev Marquis42 (n=bfhbmw0@65-127-126-34.dia.static.qwest.net) |
| 11:58.35 | *** join/#asterisk-dev ApOgEE (i=apogee@elite.bshellz.net) [NETSPLIT VICTIM] |
| 11:58.35 | *** join/#asterisk-dev Chicago (n=Chicago@c-98-223-75-214.hsd1.in.comcast.net) |
| 11:58.35 | *** join/#asterisk-dev korihor (n=korihor@201.210.226.98) |
| 11:58.35 | *** join/#asterisk-dev Kobaz (n=kobaz@its.kobaz.net) |
| 11:58.35 | *** join/#asterisk-dev pyite (n=dschreib@unaffiliated/pyite) [NETSPLIT VICTIM] |
| 11:58.35 | *** join/#asterisk-dev hugorebelo (n=hugorebe@200-171-132-124.completo.com.br) |
| 11:58.35 | *** join/#asterisk-dev xpot (n=xpot@70-91-210-233-BusName-Utah.hfc.comcastbusiness.net) |
| 11:58.35 | *** join/#asterisk-dev jtodd (i=j5s3feix@asterisk/community-director-and-tie-dye-shirt-lover/jtodd) |
| 11:58.35 | *** mode/#asterisk-dev [+oooo tzafrir_laptop seanbright DocAwesome jtodd] by irc.freenode.net |
| 11:58.35 | *** join/#asterisk-dev ccesario (n=ccesario@189-19-6-236.dsl.telesp.net.br) |
| 11:58.36 | *** join/#asterisk-dev micols (n=mio@rlogin.dk) |
| 11:58.36 | *** join/#asterisk-dev denon (n=denon@66.207.128.103) |
| 11:58.36 | *** join/#asterisk-dev qdk (n=qdk@0x573d8df9.bynqu1.dynamic.dsl.tele.dk) |
| 11:58.36 | *** join/#asterisk-dev kaii (n=kai@ciphron.de) [NETSPLIT VICTIM] |
| 11:58.36 | *** join/#asterisk-dev mvanbaak (n=michiel@asterisk/contributor-and-bug-marshal/mvanbaak) |
| 11:58.36 | *** join/#asterisk-dev Katty (n=asterisk@mail.copi-rite.com) |
| 11:58.36 | *** join/#asterisk-dev ecrane (n=ecrane@o1-69-19-166-10.static.o1.com) |
| 11:58.36 | *** join/#asterisk-dev ChanServ (ChanServ@services.) |
| 11:58.36 | *** mode/#asterisk-dev [+oo mvanbaak ChanServ] by irc.freenode.net |
| 11:59.11 | *** part/#asterisk-dev MAbbas (i=Jinbaba@115.186.16.74) |
| 12:32.46 | *** join/#asterisk-dev kpfleming (n=kpflemin@asterisk/digium-director-of-software-tech/kpfleming) |
| 12:32.47 | *** mode/#asterisk-dev [+o kpfleming] by ChanServ |
| 12:39.19 | *** join/#asterisk-dev fiddur (n=fiddur@dhcp01.textalk.com) |
| 12:48.41 | *** join/#asterisk-dev michael-i (n=michael-@141.41.40.153) |
| 12:51.30 | michael-i | I'm updating some packages in AskoziaPBX from 1.4 and decided to add packages to build 1.6.0, 1.6.1 and 1.6.2 as well as trunk. Everything's going fine except I believe some changes to autoconf in trunk have broken cross-compilation. Does anyone have experience cross-compiling trunk? |
| 12:54.24 | michael-i | Here's a log of my last attempt (http://pastebin.ca/1700041). It fails building menuselect. In my build scripts, I specify that to be built natively...it seems to be being built a second time and failing. |
| 12:55.04 | michael-i | I'd hugely appreciate a second pair of eyes on this. My brain's fried after getting the others working. |
| 12:59.25 | kpfleming | there aren't many people with experience cross compiling asterisk trunk, unfortunately |
| 12:59.36 | kpfleming | in addition, it's trunk, which means it could undergo more changes before branching for release, so to some extent this is not a terribly useful exercise |
| 13:02.04 | michael-i | kpfleming: unfortunately, I'm left with the option of getting trunk in shape or backporting bri_net_ptmp support in chan_dahdi to 1.6.2. Diff'ing those two versions, however, left me with a sinking feeling about its feasibility. |
| 13:02.23 | kpfleming | it's not worth doing |
| 13:06.24 | michael-i | My cross compiling error seems trivial (autoconf's updated --host --build policy/processing) so I will continue searching. I'd just hoped someone had dealt with that before. I'm cross compiling for cleanliness right now 686 host to 486 toolchain |
| 13:35.33 | michael-i | over that hurdle... the menuselect directory was being cleaned after being initially built by the native compiler. Let the workarounds begin ;) If I come up actual issues I'll make sure to get them in mantis |
| 13:43.40 | michael-i | made it all the way through...now I just have the same error as with 1.6.2 when linking (undefined reference to ast_complete_source_filename). Having the same error as a release is acceptable I guess! |
| 13:44.42 | *** join/#asterisk-dev ticoit (n=ticoit@201.191.187.58) |
| 13:45.35 | *** join/#asterisk-dev atis_work (n=atis_wor@193.238.212.171) |
| 13:53.59 | michael-i | looks like 'ast_complete_source_filename' and 'ast_rtp_destroy' are used but not declared/visible when LOW_MEMORY is enabled (1.6.2+) |
| 13:57.04 | *** join/#asterisk-dev Victor_Yure (n=victor@unaffiliated/victoryure/x-837844) |
| 13:57.04 | TheDavidFactor | mornin' all |
| 13:58.55 | TheDavidFactor | I'm trying to mergetrunk6 on the 6 branches. due to changes in the common parts of the unified diffs it won't complete the merge. Is it correct to simply modify the file that has the conflicts and then commit? Will that leave all the properties and everything as they should be? |
| 13:59.07 | kpfleming | michael-i: sounds like a bug for sure |
| 13:59.19 | kpfleming | TheDavidFactor: yes |
| 13:59.30 | kpfleming | just let the merge tool try to do it, then clean up the result and commit it |
| 13:59.40 | michael-i | kpfleming: ah the reward of messing when I shouldn't mess :) I'll put something in mantis |
| 14:00.24 | TheDavidFactor | thanks |
| 14:18.39 | *** join/#asterisk-dev Amorsen (n=Amorsen@94.127.50.7) |
| 14:28.18 | michael-i | it's in : https://issues.asterisk.org/view.php?id=16381 |
| 14:35.17 | *** join/#asterisk-dev TheDavidFactor (n=chatzill@fw1.safedataisp.net) |
| 14:37.50 | *** join/#asterisk-dev mohawk (n=ross@host217-40-110-153.in-addr.btopenworld.com) |
| 14:57.58 | michael-i | phone working in bri_net_ptmp mode :) I get dropped straight to the 's' extension but I'm guessing that's just a result of still using the conf generators for 1.4 on FreeBSD! |
| 14:59.14 | tzafrir_laptop | michael-i, you call from an ISDN phone? |
| 14:59.31 | tzafrir_laptop | If so, be sure to set overlapdial=yes |
| 15:02.44 | michael-i | tzafrir_laptop: will do. I'm just happy that my phone's alive after going through 1.4, 1.6.1-2 and now finally on trunk |
| 15:03.51 | *** join/#asterisk-dev MAbbas (i=Jinbaba@115.186.1.114) |
| 15:10.30 | michael-i | with overlap enabled, now I keep getting hung up on by the system after dialing. |
| 15:13.27 | *** join/#asterisk-dev The_Boy_Wonder (n=vossel@asterisk/batman-developer/dvossel) |
| 15:13.27 | *** mode/#asterisk-dev [+o The_Boy_Wonder] by ChanServ |
| 15:13.33 | tzafrir_laptop | can you provide a trace? Do you have the dialplan set to handle it? |
| 15:15.14 | michael-i | tzafrir_laptop: I don't have any 'pri' related commands in the cli |
| 15:15.48 | tzafrir_laptop | dahdi restart ? |
| 15:16.38 | *** join/#asterisk-dev DocAwesome (n=Leif@asterisk/documenteur-extraordinaire/blitzrage) |
| 15:16.38 | *** mode/#asterisk-dev [+o DocAwesome] by ChanServ |
| 15:18.01 | michael-i | now I have them |
| 15:18.38 | *** join/#asterisk-dev sgimeno (n=sgimeno@226.Red-80-33-64.staticIP.rima-tde.net) |
| 15:19.54 | michael-i | pri set debug on == no such command?! |
| 15:20.13 | michael-i | (sorry but this really is my first venture into pri/bri debugging) |
| 15:20.20 | *** join/#asterisk-dev martyndev (i=be18869a@gateway/web/freenode/x-aqcwjvytnibxkmet) |
| 15:20.56 | martyndev | Hi Asterisk developers.. |
| 15:23.13 | *** join/#asterisk-dev mvanbaak (n=michiel@asterisk/contributor-and-bug-marshal/mvanbaak) |
| 15:23.13 | *** mode/#asterisk-dev [+o mvanbaak] by ChanServ |
| 15:23.25 | michael-i | tzafrir_laptop: got it..sorry for the stupiditiy (http://pastebin.ca/1700237) |
| 15:24.31 | martyndev | Ehy, someone has an implementation with Loquendo MRCP ? |
| 15:27.27 | *** join/#asterisk-dev moy (n=moy@bas1-unionville55-1177733516.dsl.bell.ca) |
| 15:29.40 | *** join/#asterisk-dev putnopvut (n=putnopvu@asterisk/master-of-queues/mmichelson) |
| 15:29.40 | *** mode/#asterisk-dev [+o putnopvut] by ChanServ |
| 15:33.33 | michael-i | just fyi, trunk cross compiles fine for blackfin :) |
| 15:35.32 | TheDavidFactor | does an if terminate on a false before continuing evaluation of the condition? i.e. for a condition like (x && (y = func())) if x is false will the func still be executed? |
| 15:36.59 | *** join/#asterisk-dev anthm (n=anthm@freeswitch/developer/anthm) |
| 15:37.15 | atis_work | TheDavidFactor: generally no |
| 15:37.32 | atis_work | TheDavidFactor: i don't think that in c it will be executed |
| 15:38.01 | atis_work | TheDavidFactor: however i've seen some weird optimization in delphi that executed something in reverse |
| 15:38.25 | TheDavidFactor | ok, using strchr on a NULL will cause a segfault, so I wanted to test the null before the strchr |
| 15:39.01 | atis_work | actually i'm not pro at C, so it could be executed on some compilers.. better to test in separate if |
| 15:39.28 | atis_work | also coding guidelines should say something about this |
| 15:39.34 | *** join/#asterisk-dev mog (n=mog@c-68-62-169-247.hsd1.al.comcast.net) |
| 15:39.34 | *** mode/#asterisk-dev [+o mog] by ChanServ |
| 15:42.43 | TheDavidFactor | I don't see anything in the coding guidelines, if it's in there I missed it, but I'll go with the separate if. That's how I coded it, but I wasn't sure if it was the best way. |
| 15:42.43 | atis_work | umm, i can't find anything in guidelines |
| 15:43.08 | atis_work | it's probably somewhere in ISO C rules |
| 15:43.25 | TheDavidFactor | thanks |
| 15:43.39 | atis_work | i.e. everything is evaluated left-to-right until first failrue |
| 15:44.17 | michael-i | enough bri_net_ptmp fun for today... take care everyone |
| 15:45.34 | atis_work | TheDavidFactor: http://en.wikipedia.org/wiki/Short-circuit_evaluation |
| 15:46.01 | atis_work | so you can write that |
| 15:47.17 | TheDavidFactor | thanks! |
| 15:50.54 | *** join/#asterisk-dev Deeewayne (n=dwayne@75.76.254.162) |
| 15:50.55 | *** mode/#asterisk-dev [+o Deeewayne] by ChanServ |
| 15:54.13 | *** join/#asterisk-dev |Cybex| (n=John@atwork-20.r-212.178.82.atwork.nl) |
| 15:57.01 | *** join/#asterisk-dev jpeeler (n=jpeeler@asterisk/digium-software-dev/jpeeler) |
| 15:57.01 | *** mode/#asterisk-dev [+o jpeeler] by ChanServ |
| 15:57.33 | *** join/#asterisk-dev kpfleming (n=kpflemin@asterisk/digium-director-of-software-tech/kpfleming) |
| 15:57.33 | *** mode/#asterisk-dev [+o kpfleming] by ChanServ |
| 16:06.34 | *** join/#asterisk-dev Victor_Yure (n=victor@unaffiliated/victoryure/x-837844) |
| 16:21.26 | martyndev | .. |
| 16:21.30 | *** part/#asterisk-dev martyndev (i=be18869a@gateway/web/freenode/x-aqcwjvytnibxkmet) |
| 16:25.15 | *** join/#asterisk-dev marv[work] (n=timr@router.asteriasgi.com) |
| 16:25.15 | atis_work | huh, any way to force symbol table.. i have SSL_accept() where should be mysql_reconnect and mysql_update |
| 16:25.34 | atis_work | trying to debug res_config_mysql |
| 16:27.27 | atis_work | gdb says it's loading symbols: Reading symbols from /opt/asterisk-1.6.1.6-svn-iqlabs/lib/asterisk/modules/res_config_mysql.so...done. |
| 16:28.47 | kpfleming | your question does not make sense |
| 16:29.00 | atis_work | well, my gdb output doesn't make sense :p |
| 16:29.15 | atis_work | maybe i messed up memory too much |
| 16:31.00 | kpfleming | or you built with optimization |
| 16:31.48 | atis_work | well, it should be on.. i see everything else except res_config_mysql in bactrace |
| 16:34.46 | *** join/#asterisk-dev Victor_Yure (n=victor@unaffiliated/victoryure/x-837844) |
| 16:35.50 | tzafrir_laptop | atis_work, is it trunk or earlier? (with earlier res_config_mysql was built separately) |
| 16:36.11 | tzafrir_laptop | ah, earlier |
| 16:36.11 | atis_work | tzafrir_laptop: earlier.. however it should grab settings from installed asterisk |
| 16:36.24 | atis_work | http://pastebin.ca/1700331 |
| 16:36.34 | kpfleming | have you confirmed that it was built with debugging symbols at all? |
| 16:36.38 | atis_work | this was forced crash without any changes |
| 16:36.44 | tzafrir_laptop | file /opt/asterisk-1.6.1.6-svn-iqlabs/lib/asterisk/modules/res_config_mysql.so # has debug symbols? |
| 16:37.51 | atis_work | tzafrir_laptop: should it output something more than just "Reading symbols... done" |
| 16:37.58 | kpfleming | gdb won't tell you |
| 16:38.00 | atis_work | that's supposed to be gdb command? |
| 16:38.09 | kpfleming | you have to use nm or objdump or something |
| 16:38.12 | kpfleming | or 'file' |
| 16:39.44 | atis_work | nm tells there is something, but not mysql_reconnect |
| 16:40.04 | atis_work | <PROTECTED> |
| 16:40.07 | kpfleming | mysql_reconnect is not a symbol in res_config_mysql, it is in the mysql library itself, isn't it? |
| 16:40.24 | atis_work | kpfleming: nope it's function in res_config_mysql |
| 16:40.26 | tzafrir_laptop | atis_work, the code/symbol must be there, or things would exploded at load time |
| 16:40.33 | tzafrir_laptop | ah, ok. not stripped |
| 16:40.40 | kpfleming | are you doing this on Linux? |
| 16:40.44 | atis_work | sure |
| 16:40.49 | atis_work | is there RedHat BSD? :D |
| 16:41.11 | kpfleming | does 'nm -a res_config_mysql.so' list any symbols with 'N' tags? |
| 16:41.18 | tzafrir_laptop | No, but there's Debian GNU/kFreeBSD :-p |
| 16:41.53 | atis_work | kpfleming: nope, nothing with N |
| 16:42.02 | kpfleming | then it wasn't built with debugging symbols |
| 16:42.07 | kpfleming | so they aren't there |
| 16:42.09 | atis_work | i have --prefix=/opt/asterisk in both asterisk and addons |
| 16:42.16 | kpfleming | that is irrelevant |
| 16:42.19 | atis_work | perhaps it's that what's messing up |
| 16:42.31 | kpfleming | when it was compiled, the compile needs to have included -g |
| 16:43.03 | atis_work | kpfleming: how do i check that |
| 16:43.11 | kpfleming | watch the compile |
| 16:43.23 | kpfleming | look at the command line that runs the compiler |
| 16:43.24 | kpfleming | etc. |
| 16:43.41 | atis_work | yeah, but how did i configure to see full command line? |
| 16:43.46 | kpfleming | you don't |
| 16:43.49 | kpfleming | there is no configure for that |
| 16:43.53 | kpfleming | what version of asterisk is this? |
| 16:43.59 | atis_work | 1.6.1.6 |
| 16:44.05 | atis_work | but it's addons 1.6.1.1 |
| 16:44.11 | kpfleming | right, hang on |
| 16:44.28 | kpfleming | # make NOISY_BUILD=yes |
| 16:44.37 | kpfleming | that will display the compiler command lines |
| 16:45.22 | atis_work | well, it displays only for menuselect |
| 16:45.31 | kpfleming | did you 'make clean' first? |
| 16:45.36 | atis_work | yeah |
| 16:45.43 | kpfleming | i don't see anything in addons 1.6.1 that would cause it to build with debug symbols |
| 16:45.47 | atis_work | i shouldn't need dist-clean? |
| 16:46.20 | kpfleming | no |
| 16:46.26 | kpfleming | make NOISY_BUILD=yes works for me |
| 16:46.30 | atis_work | ah, dist-clean helped |
| 16:46.36 | kpfleming | i see the compiler trying to compile chan_mobile |
| 16:46.47 | kpfleming | but... the addons build system does nto generate debug symbols at all |
| 16:46.47 | atis_work | gcc -o res_config_mysql.o -c res_config_mysql.c -I/opt/asterisk/include -DDONT_OPTIMIZE -fPIC -DAST_MODULE=\"res_config_mysql\" -MD -MT res_config_mysql.o -MF .res_config_mysql.o.d -MP |
| 16:47.29 | kpfleming | try this: |
| 16:47.39 | kpfleming | # ASTCFLAGS=-g3 make NOISY_BUILD=yes |
| 16:47.48 | kpfleming | that will generate debug symbols |
| 16:48.08 | atis_work | g3 is there |
| 16:48.17 | tzafrir_laptop | aren't debug symbols enabled by default? |
| 16:48.21 | kpfleming | no |
| 16:48.36 | kpfleming | atis_work: what does 'g3 is there' mean |
| 16:48.55 | atis_work | yeah, debug symbols are on :) |
| 16:48.57 | atis_work | thanks |
| 16:49.01 | kpfleming | ok good |
| 16:49.08 | atis_work | well, i meant - g3 is in gcc line |
| 16:49.12 | kpfleming | now |
| 16:49.14 | kpfleming | it was not before |
| 16:49.55 | atis_work | is going back to write master-hostname-switching-detection |
| 16:50.04 | *** join/#asterisk-dev caio1982 (i=caio1982@CAcert-br/caio1982) |
| 16:58.23 | *** join/#asterisk-dev DocAwesome (n=Leif@asterisk/documenteur-extraordinaire/blitzrage) |
| 16:58.23 | *** mode/#asterisk-dev [+o DocAwesome] by ChanServ |
| 17:06.51 | *** join/#asterisk-dev DocAwesome (n=Leif@asterisk/documenteur-extraordinaire/blitzrage) |
| 17:06.51 | *** mode/#asterisk-dev [+o DocAwesome] by ChanServ |
| 17:07.02 | *** join/#asterisk-dev errotan (n=errotan@81.0.115.3) |
| 17:08.02 | *** join/#asterisk-dev nd- (n=nd-@xes-mad.com) |
| 17:10.12 | nd- | i'm writing a custom channel driver and i have a separate monitor thread that queues up incoming frames using ast_queue_frame() rather than using the tech->read() callback. the problem is that my monitor thread gets stuck waiting for the channel lock which is being held by __ast_read() |
| 17:11.12 | nd- | i'm not holding any other locks at this point in the monitor thread either. it looks like __ast_read() gets called over and over again in this condition and my monitor thread never wins out to get the channel lock |
| 17:11.35 | kpfleming | ast_read() should not be geting called repeatedly. something else is wrong. |
| 17:11.57 | nd- | i thought that was fishy... any common causes for that? |
| 17:12.32 | nd- | i would think the generic bridge function would always see the queued up frames and never have to call __ast_read() |
| 17:12.45 | nd- | sorry, would never have to call tech->read() |
| 17:13.11 | kpfleming | that is correct |
| 17:13.24 | kpfleming | when it calls ast_read(), a frame will be pulled off the queue and returned |
| 17:13.36 | kpfleming | when the queue is empty, the bridge will wait for it to become non-empty |
| 17:14.40 | nd- | i don't have my tech->fds set for it to be able to poll on file descriptors... i thought this wouldn't be a problem since i'm queuing frames directly in the channel driver |
| 17:15.01 | kpfleming | that is fine |
| 17:15.15 | kpfleming | there is an alertpipe on the channel that is used to tell it when frames have been queued |
| 17:15.40 | nd- | ast_read() would be called repeatedly in a loop though from the generic bridge function right? so I assumed that's why ast_read() was getting called repeatedly |
| 17:16.23 | kpfleming | ast_read() will be called by the bridge when the channel has a frame available to read |
| 17:16.33 | kpfleming | whether that is 'repeatedly' or not depends on when frames arrive and other stuff |
| 17:16.47 | kpfleming | it will not call it over and over again even without frames available, unless something is broken |
| 17:17.23 | nd- | hmm... i guess something is broken then. i had to make my tech->read() callback return &ast_null_frame and also usleep() so that the CPU wouldn't get pegged. it sounds like i shouldn't have to do this |
| 17:18.06 | kpfleming | you should not |
| 17:18.18 | kpfleming | and so you are holding the lock while you are in usleep |
| 17:18.34 | *** join/#asterisk-dev mpe (n=mpe@0x4dd624b2.adsl.cybercity.dk) |
| 17:18.35 | nd- | right, which is making it harder for my monitor thread competing for that lock |
| 17:19.27 | nd- | it sounds like i should focus on making sure my tech->read() never gets called in the first place |
| 17:19.41 | nd- | and get rid of the usleep() |
| 17:20.22 | kpfleming | that is mandatory |
| 17:20.56 | nd- | alright, thanks for the advice. much appreciated |
| 17:44.38 | *** join/#asterisk-dev xpot-mobile (n=xpot@173-14-232-121-Utah.hfc.comcastbusiness.net) |
| 17:47.16 | TheDavidFactor | kpfleming, read your FAX post, I appreciate the detail and the background. Would the TDM-T.38 FAX relay allow: |
| 17:47.18 | TheDavidFactor | fax -> ata -> T.38 -> * -> PRI/T1 -> POTS -> analog fax? |
| 17:48.06 | kpfleming | yes |
| 17:48.59 | TheDavidFactor | That is something we need so I will be glad to be in on testing/development for it. |
| 17:56.14 | *** join/#asterisk-dev Victor_Yure (n=victor@unaffiliated/victoryure/x-837844) |
| 17:56.52 | atis_work | fortunately we don't use analog fax machines :) |
| 17:57.35 | kpfleming | i'd be happy to send you some |
| 17:58.01 | atis_work | ah, you want me to cover posting? :D |
| 17:58.23 | kpfleming | i'll just send them freight collect |
| 17:59.57 | atis_work | btw, anyone up for some fax-testing-group? i'm really curious about my provider's quality as lots of faxes are failing |
| 18:01.33 | atis_work | i can share my fax number, so if anyone needs to test - i wouldn't mind receiving FEW faxes |
| 18:03.15 | atis_work | just don't make it as public as 888-hp-fax-me |
| 18:03.19 | atis_work | TheDavidFactor: ? |
| 18:04.36 | TheDavidFactor | yes sir? |
| 18:05.02 | atis_work | TheDavidFactor: we can exchange our fax testing numbers? |
| 18:05.51 | kpfleming | TheDavidFactor: you need to learn how trim your mailing list replies and not top-post :-) |
| 18:06.45 | TheDavidFactor | yes sir, I meant to delete all that and then didn't think about it before I it send. |
| 18:15.58 | *** join/#asterisk-dev atis_work (n=atis_wor@193.238.212.171) |
| 18:48.10 | *** join/#asterisk-dev CunningPike (n=CunningP@204.239.8.157) |
| 18:54.21 | Kobaz | is there an override for the /var/lib/asterisk/documentation path, in ./configure? |
| 18:54.38 | Kobaz | i don't want to set the entire /var/lib/asterisk directory, just documentation/ |
| 18:59.28 | *** join/#asterisk-dev outtolunc (n=me@c-98-248-96-110.hsd1.ca.comcast.net) |
| 19:08.34 | *** join/#asterisk-dev jtodd (i=u07hqgeu@asterisk/community-director-and-tie-dye-shirt-lover/jtodd) |
| 19:08.34 | *** mode/#asterisk-dev [+o jtodd] by ChanServ |
| 19:13.41 | nd- | Kobaz: I don't think there is one. it looks to be forced to $datadir/documentation |
| 19:14.16 | nd- | but you could always create a symlink i guess for $datadir/documentation |
| 19:17.19 | tzafrir_laptop | Kobaz, what do you mean? Why do you have a problem with it? |
| 19:18.17 | Kobaz | tzafrir_laptop: when i build asteriskeseses, i do a custom --prefix |
| 19:18.27 | Kobaz | so i can change asterisk versions with a single symlink |
| 19:18.48 | Kobaz | but then now with 1.6.2.x the documentation doesn't go with it |
| 19:19.13 | tzafrir_laptop | If you use a non-matching documentation directory, you may get an incorrect on-line help |
| 19:19.22 | tzafrir_laptop | Which is rather funny |
| 19:19.44 | Kobaz | yeah |
| 19:19.47 | Kobaz | exactly |
| 19:19.56 | Kobaz | ./configure --prefix=/apps/asterisk-1.6.2.0-rc7 --sysconfdir=/etc --localstatedir=/var |
| 19:20.09 | Kobaz | docs go into /var... everything else goes into /apps/asterisk.. |
| 19:22.50 | tzafrir_laptop | that's odd. It goes with datadir here (rc6, though) |
| 19:23.15 | tzafrir_laptop | time to upgrade to my freshly-built rc7 anyway |
| 19:24.01 | Kobaz | this is with all the rc's for 1.6.2 |
| 19:24.04 | Kobaz | rc7 too |
| 19:24.44 | Kobaz | /usr/bin/install -c -m 644 doc/appdocsxml.dtd /var/lib/asterisk/documentation |
| 19:33.32 | Kobaz | --datadir doesn't do anything |
| 19:34.21 | Kobaz | --datarootdir is the man pages |
| 19:36.16 | tzafrir_laptop | BUILDFLAGS += ASTDATADIR=/usr/share/asterisk |
| 19:36.28 | tzafrir_laptop | that's what the Debian package has |
| 19:37.23 | Kobaz | yeah, that's the man page location |
| 20:15.59 | *** join/#asterisk-dev jtodd (i=ua1kv7tr@asterisk/community-director-and-tie-dye-shirt-lover/jtodd) |
| 20:15.59 | *** mode/#asterisk-dev [+o jtodd] by ChanServ |
| 20:23.23 | tzafrir_laptop | asterisk-addons still uses the old config.sub that breaks avr32 |
| 20:39.37 | *** join/#asterisk-dev mpe (n=mpe@x1-6-00-1e-2a-2a-b3-a2.k758.webspeed.dk) |
| 20:43.24 | *** join/#asterisk-dev lmsteffan (n=laurent@reef.ac-noumea.nc) |
| 21:02.23 | *** part/#asterisk-dev |Cybex| (n=John@atwork-20.r-212.178.82.atwork.nl) |
| 21:02.35 | nd- | kpfleming: I got rid of my usleep() in tech->read() and the locking problems seem fixed, but tech->read() is still getting hammered quite a bit (>2000 calls/sec). i investigated and found that ast_waitfor_nandfds() is returning right away when called by the generic_bridge function and doesn't seem to wait for any real activity. do you have any ideas on what might cause this? |
| 21:06.01 | putnopvut | If frames are getting queued at that rate, that could be the cause. |
| 21:08.42 | nd- | well, in this case i tracked all the cases where my channel drive can queue frames, and it's happening at a pretty steady rate of 50/second |
| 21:08.50 | nd- | s/drive/driver |
| 21:09.49 | nd- | it's hard for me to follow exactly what ast_waitfor_nandfds() is doing -- it's not clear how it could know about the queued frames since it only seems to be looking at the file descriptors for polling? is the alertpipe that kpfleming mentioned automatically included in this list? |
| 21:10.30 | putnopvut | I believe so. |
| 21:10.33 | putnopvut | Let me double-check |
| 21:11.36 | putnopvut | nd-: yes, the read end of the alertpipe is set as the AST_ALERT_FD on the channel. |
| 21:11.53 | putnopvut | Which is AST_MAX_FDS-1 |
| 21:12.03 | putnopvut | So it's towards the end of the array of FDs on a channel. |
| 21:12.24 | nd- | i see, that makes sense. strange then because it sounds like it should work fine |
| 21:13.07 | putnopvut | ast_read will only read a single byte off the alertpipe per read. I wonder if more than 1 byte is being written to it at a time or some such... |
| 21:13.13 | putnopvut | actually not true. |
| 21:13.19 | putnopvut | not a single byte, a single int. |
| 21:15.07 | nd- | hmm.. my queuing is done via ast_queue_frame(). let me check to see how many ints it ends up writing to the alertpipe |
| 21:18.47 | putnopvut | ast_queue_frame will write a number of ints equal to the number of frames being queued at the time that ast_queue_frame is called, it appears. |
| 21:26.23 | nd- | i put some debug in to see how many new frames ast_queue_frame() thinks are new, and noticed that it doesn't look like it's even writing to the alertpipe (the if (chan->alertpipe[1] > -1) check is failing) |
| 21:28.08 | file | did you tell ast_channel_alloc you needed a queue? |
| 21:30.31 | nd- | i didn't realize I needed to, but I suspect I did not file... d'oh. checking now |
| 21:31.15 | file | unless you are REALLY creative in the design of your channel driver and can drive frames into the core from your own file descriptor then you need one |
| 21:31.22 | nd- | well, the first argument i passed to ast_channel_alloc() was 1 so needqueue=1 |
| 21:31.28 | file | k |
| 21:38.36 | nd- | yeah it looks like __ast_queue_frame() is seeing that chan->alertpipe[1] == -1 so it's not writing to it |
| 21:38.40 | nd- | hmm |
| 21:40.06 | putnopvut | So I wonder what's triggering ast_waitfor to say that the channel is ready for reading... |
| 21:40.49 | nd- | not sure about that |
| 21:41.13 | nd- | __ast_queue_frame() is using the this case instead of writing to the alertpipe: |
| 21:41.20 | nd- | <PROTECTED> |
| 21:41.26 | nd- | <PROTECTED> |
| 21:41.42 | putnopvut | ohhhhhhh |
| 21:42.12 | putnopvut | So when the timer is in continuous mode, that means that ast_waitfor will immediately return every time saying that the channel can be read. |
| 21:42.34 | nd- | ah, so that makes some sense.. now i just need to figure out how to get it to stop using this mode and get the alertpipe working |
| 21:43.28 | nd- | it looks like it should create it properly with pipe() if i pass needqueue=1 like file was saying |
| 21:49.36 | nd- | hmm, right after ast_channel_alloc() is called: |
| 21:49.39 | nd- | chan->alertpipe[0] = -1 |
| 21:49.39 | nd- | chan->alertpipe[1] = -1 |
| 21:49.40 | nd- | chan->alertpipe[2] = 0 |
| 21:49.59 | file | no pipe foo you |
| 21:50.07 | nd- | pipe() gets called and even returns 0 but i get -1 for those pipes |
| 21:50.31 | nd- | i'm on a powerpc platform with eglibc so maybe there's something weird going on outside of asterisk |
| 22:15.59 | nd- | oh i think i see what's happening |
| 22:16.21 | nd- | when ast_channel_alloc() runs, the call to ast_timer_open() fails and so it forces needqueue=0 |
| 22:17.12 | nd- | or maybe not fails... but tmp->timer is non-zero and then needqueue gets set to 0 |
| 22:18.04 | *** join/#asterisk-dev mpe (n=mpe@0x4dd624b2.adsl.cybercity.dk) |
| 22:18.53 | nd- | it actually looks like ast_timer_open() succeeds (returns non-NULL) and so it does this.. |
| 22:19.54 | nd- | it looks to me like asterisk (at least this 1.6.2 snapshot I'm using) favors chan->timingfd usage over chan->alertpipe |
| 22:26.52 | kpfleming | yes, that uses less resources than alertpipes |
| 22:26.59 | kpfleming | but the timers have to work properly |
| 22:28.45 | nd- | i don't know anything about how these timers work, but is it expected to work with my method of queuing frames directly from a monitor thread? |
| 22:30.32 | kpfleming | yes |
| 22:30.47 | kpfleming | the timers are provided by the generic timing API, and some res_timer_<foo> module you have loaded |
| 22:31.01 | kpfleming | there is also a timer test module in the test directory you can enable, load and run |
| 22:33.11 | nd- | hmm, at a glance i don't see any res_timer_* mentioned in the console log, maybe i'll look into this |
| 22:33.25 | kpfleming | what version of asterisk are you using? |
| 22:33.59 | nd- | 1.6.2 from svn, r220531 |
| 22:34.34 | kpfleming | yeah, there are two or three timing modules available. you probably don't have dahdi, so you'd be using res_timer_pthread |
| 22:34.44 | kpfleming | or res_timing_pthread |
| 22:34.46 | nd- | right, i don't have dahdi |
| 22:35.04 | kpfleming | running on linux? |
| 22:35.15 | nd- | yep |
| 22:35.23 | nd- | i see that res_timing_pthread.so was built at least |
| 22:36.20 | kpfleming | what kernel version? |
| 22:37.08 | nd- | 2.6.30 + some patches for the single-board-computer i'm using |
| 22:37.18 | nd- | (which i wrote myself but they're not very invasive) |
| 22:37.29 | kpfleming | then you should have res_timing_timerfd as well, which is a better choice because it uses kernel timers and is lower cost |
| 22:38.23 | nd- | i see from my config.log that one of the tests failed for timerfd support |
| 22:38.38 | kpfleming | fix that if you can... you'll be happier |
| 22:38.44 | nd- | will do, thanks :) |
| 22:39.46 | kpfleming | if it turns out that res_timing_pthread is causing this problem, then either POSIX timers are broken on your system, or something is wrong with the module |
| 22:45.46 | nd- | ok, it will take me a while to find out for sure since it looks like I need to recompile my glibc |
| 22:46.24 | nd- | for timerfd support that is |
| 22:47.48 | nd- | meanwhile, I don't see this timer test module you were referring to? |
| 22:49.07 | putnopvut | nd-: yeah, it's not a module, actually, it's part of the core. It's in main/timing.c |
| 22:49.24 | putnopvut | (it probably should be made into a module and moved to the tests/ directory though, fwiw) |
| 22:49.41 | nd- | ah |
| 22:49.53 | nd- | *CLI> timing test |
| 22:49.53 | nd- | Attempting to test a timer with 50 ticks per second. |
| 22:49.53 | nd- | Using the 'pthread' timing module for this test. |
| 22:49.53 | nd- | It has been 1001 milliseconds, and we got 50 timer ticks |
| 22:49.56 | nd- | looks OK to me :) |
| 22:49.58 | putnopvut | yep |
| 22:50.17 | putnopvut | The problem may be that the timer isn't being taken out of continuous mode when it should be. |
| 22:50.45 | kpfleming | right, that's what it sounds like |
| 22:59.52 | nd- | is this something that is supposed to be more automatic or is ast_timer_disable_continuous() what you're referring to? because I only see one case where ast_timer_disable_continuous() is called and it's when __ast_read() gets a special AST_TIMING_EVENT_CONTINUOUS timing event |
| 23:00.05 | russellb | what version of asterisk? |
| 23:00.09 | nd- | 1.6.2 |
| 23:00.19 | russellb | 1.6.2.recent? |
| 23:00.24 | russellb | or 1.6.2.superoldbeta |
| 23:00.56 | putnopvut | He said rev 220531 earlier. |
| 23:00.58 | putnopvut | so recent |
| 23:01.01 | russellb | Okay. |
| 23:01.01 | nd- | yeah |
| 23:01.10 | russellb | I had fixed a res_timing_pthread continuous mode bug back in May |
| 23:01.13 | russellb | rev 198148 |
| 23:01.38 | *** join/#asterisk-dev puzzled (n=foobar@puzzled.xs4all.nl) |
| 23:01.51 | nd- | ok, I verified in my svn log that that fix is in there already russellb |
| 23:02.00 | russellb | nods |
| 23:05.23 | *** join/#asterisk-dev atis_home (n=atis@193.238.213.215) |
| 23:05.42 | atis_home | jpeeler: ping |
| 23:05.51 | jpeeler | atis_home: pong |
| 23:06.18 | atis_home | jpeeler: Hold: __ast_queue_control_data channel.c:1172 |
| 23:06.29 | atis_home | unhold: __ast_queue_control channel.c:1157 |
| 23:06.40 | russellb | debug fail :-( |
| 23:06.44 | atis_home | i'm not sure is this what you wanted to see, as.. |
| 23:06.45 | atis_home | yes :D |
| 23:06.56 | atis_home | you forgot to pass function and line |
| 23:07.14 | jpeeler | man... |
| 23:07.20 | atis_home | and it was early, so i didn't read the code much, just applied |
| 23:07.29 | kpfleming | well, you have line, but not function, right? |
| 23:08.01 | atis_home | well, i meant __ast_queue_control_data has passed function name, line, etc but it doesn't that to src |
| 23:08.08 | kpfleming | oh |
| 23:08.12 | atis_home | yeah :D |
| 23:08.23 | atis_home | ok, i'll fix that tomorrow |
| 23:08.47 | atis_home | so some debug either tomorrow late evening or monday |
| 23:13.59 | jpeeler | so yeah, the non macro version of ast_queue_frame needs to be called in all the modified functions |
| 23:15.15 | atis_home | jpeeler: doesn't this tell where the problem is? http://pastebin.ca/1700988 |
| 23:17.16 | jpeeler | depends where that is in the file, but i doubt it. i suspect that is the location inside ast_queue_control_data calling ast_queue_frame instead of __ast_queue_frame |
| 23:17.42 | jpeeler | actually of course it isn't |
| 23:18.35 | putnopvut | grates jpeeler up with some pine nuts, olive oil, and parmesan cheese. |
| 23:18.36 | jpeeler | i'm sorry i messed it up, i'll fix it tonight if you want me to |
| 23:18.47 | nd- | well, i verified that ast_timer_disable_continuous() is getting called regularly at least |
| 23:19.18 | jpeeler | putnopvut: ! |
| 23:19.27 | kpfleming | but it's not actually disabling the timer then |
| 23:19.33 | putnopvut | jpeeler: oh sorry, I thought you were a sprig of basil there for a second |
| 23:19.53 | *** join/#asterisk-dev tzafrir_laptop (n=tzafrir@bzq-218-155-145.cablep.bezeqint.net) |
| 23:19.53 | *** mode/#asterisk-dev [+o tzafrir_laptop] by ChanServ |
| 23:23.25 | atis_home | jpeeler: so it should be return __ast_queue_frame(chan, &f, file, line, function); |
| 23:23.28 | atis_home | in both places |
| 23:30.42 | atis_home | jpeeler: what confused me is that this "jpeeler: inserting frame %p into frame list with datalen %d and data at %p..." already says datalen=0 |
| 23:32.40 | jpeeler | yes to above and what? |
| 23:33.06 | atis_home | ah, nothing.. it's supposed to set src already |