| 00:00.12 | coreyfarrell | does res_pjsip "require" res_pjsip_config_wizard? what would happen if I didn't load res_pjsip_config_wizard? |
| 00:00.48 | gtjoseph | nothing except that the user's configs wouldn't be loaded either |
| 00:01.25 | gtjoseph | i did it this way on purpose so there wouldn't be a dependency there |
| 00:03.05 | gtjoseph | iirc there isn't even a real dependency from the config_wizard back to res_pjsip* either. it's just that was so that if someone isn't using chan_pjsip the wizard doesn't get compiled. |
| 00:04.03 | coreyfarrell | config wizard uses ast_sip_get_sorcery which is from res_pjsip. |
| 00:04.59 | *** join/#asterisk-dev CELYA_ (~Thunderbi@2a01cb0583d31d0020e7038e0bd80bc5.ipv6.abo.wanadoo.fr) |
| 00:05.20 | gtjoseph | ah, then that's the only one. it only does that when you want to dump the config from the cli. |
| 00:05.48 | coreyfarrell | This is the only module I'm having trouble fitting into my new module loader. so far the new module loader has found 2 incorrect module priorities - res_stasis and res_speech get loaded earlier due to dependencies. |
| 00:06.31 | gtjoseph | are you revisiting the stuff we talked about last year (or the year before, or the year before that)? :) |
| 00:06.36 | coreyfarrell | when it comes to res_pjsip_config_wizard it loads with the realtime drivers, but then because of .requires = "res_pjsip" it makes res_pjsip load before it, which makes res_pjproject load before that. |
| 00:07.33 | gtjoseph | well, taking the depends out is fine |
| 00:07.55 | gtjoseph | looks like both res_pjsip and res_pjproject |
| 00:08.38 | gtjoseph | we could put a check arounf the ast_sip_get_sorcery call to check that res_pjsip is loaded before trying it. |
| 00:08.39 | coreyfarrell | it uses pjproject, not res_pjproject. |
| 00:08.51 | gtjoseph | oh yeah. no clue why |
| 00:09.35 | coreyfarrell | I'm not sure pjproject isn't included in the symbol import/export scan I did, I was only looking for modules that use other modules. |
| 00:09.52 | gtjoseph | it's certainly not needed by thew wizard |
| 00:10.56 | coreyfarrell | if I remove .requires = "res_pjsip" from res_pjsip_config_wizard it would be the only exception (every other module that uses globals declares the requires). |
| 00:11.27 | gtjoseph | the requires on res_pjsip??? |
| 00:12.04 | coreyfarrell | give me a second |
| 00:13.35 | coreyfarrell | I just pushed to github so I can link.. https://github.com/coreyfarrell/asterisk/blob/loader-work-testing/res/res_pjsip_config_wizard.c#L1349 |
| 00:14.24 | gtjoseph | ok but if it doesn't require it, then it doesn't |
| 00:14.25 | coreyfarrell | this tells the new module load that res_pjsip must be running before res_pjsip_config_wizard can start. |
| 00:14.55 | gtjoseph | so i'm not understanding the issue with removing that line. |
| 00:16.47 | gtjoseph | the only thing the wizard really depends on is core sorcery.c |
| 00:17.31 | gtjoseph | what do the other sorcery wizards .require? |
| 00:18.03 | gtjoseph | not much if anything i'm guessing |
| 00:18.17 | coreyfarrell | the res_pjsip_config_wizard actually uses a res_pjsip symbol though, that's where my issue comes from. |
| 00:18.47 | gtjoseph | only that ast_sip_get_sorcery right? |
| 00:19.06 | *** join/#asterisk-dev CELYA_ (~Thunderbi@2a01cb0583d31d00adf78ba467ab4449.ipv6.abo.wanadoo.fr) |
| 00:19.21 | coreyfarrell | correct. |
| 00:19.41 | gtjoseph | from a code perspective, we can just a check that res_pjsip is loaded before calling it. it's only used for a cli command. |
| 00:20.25 | gtjoseph | if you're doing something automatically based on symbols, maybe oyu need a .optional thing. |
| 00:21.36 | coreyfarrell | well I do have .nonoptreq, but if both modules are loading it still enforces an order. I'm not automatically doing anything based on symbols but I have setup the deps based partly on a manual scan of symbols. |
| 00:22.26 | gtjoseph | ah. not sure i can help you there. |
| 00:23.38 | gtjoseph | btw, .nonoptreq needs a better moniker that actually describes what it does. :) |
| 00:23.57 | coreyfarrell | in all current versions it does nothing. |
| 00:24.20 | gtjoseph | which would make sense because it means nothing to me. :) |
| 00:25.23 | coreyfarrell | I think it stood for non-optional requirements. but yeah I should change that. |
| 00:26.41 | coreyfarrell | so res_pjsip_config_wizard should always load first and unload last? and I'll figure out how to deal with the CLI using ast_sip_get_sorcery later. |
| 00:27.33 | gtjoseph | correct. it's technically a sorcery wizard but one that inserts itself into res_pjsip without res_pjsip knowing or caring. |
| 00:28.41 | gtjoseph | it can only do that at res_pjsip startup which is why it has to be loaded before res_pjsip |
| 00:29.34 | coreyfarrell | ok and anything else that requires res_pjsip_config_wizard will always load after res_pjsip anyways? |
| 00:32.19 | *** join/#asterisk-dev CELYA_ (~Thunderbi@2a01cb0583d31d00f5edd1fe02d9d7ad.ipv6.abo.wanadoo.fr) |
| 00:33.15 | gtjoseph | yep |
| 00:36.37 | coreyfarrell | cool.. so it looks like res_stasis and res_speech are the only modules that the current loader starts in the wrong order. going to post patches for that. |
| 00:36.56 | gtjoseph | cool. looking forward to it. :) |
| 00:45.42 | *** join/#asterisk-dev CELYA_ (~Thunderbi@2a01cb0583d31d00ac8c3ccb25a5acd1.ipv6.abo.wanadoo.fr) |
| 00:48.08 | *** part/#asterisk-dev kharwell (kharwell@nat/digium/x-zegyvjhshbuiapbg) |
| 01:00.42 | *** join/#asterisk-dev CELYA_ (~Thunderbi@2a01cb0583d31d00a06174e5bba30a7b.ipv6.abo.wanadoo.fr) |
| 01:13.23 | *** join/#asterisk-dev CELYA_ (~Thunderbi@2a01cb0583d31d005cb26ad2bb508b1f.ipv6.abo.wanadoo.fr) |
| 01:21.08 | *** join/#asterisk-dev infobot (ibot@rikers.org) |
| 01:21.08 | *** 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 |
| 01:26.38 | *** join/#asterisk-dev CELYA_ (~Thunderbi@2a01cb0583d31d00f54fb5942547f2b0.ipv6.abo.wanadoo.fr) |
| 01:41.28 | *** join/#asterisk-dev CELYA_ (~Thunderbi@2a01cb0583d31d0084f79d7a9ed42cbd.ipv6.abo.wanadoo.fr) |
| 01:57.23 | *** join/#asterisk-dev CELYA_ (~Thunderbi@2a01cb0583d31d0018d3c345741cb507.ipv6.abo.wanadoo.fr) |
| 02:12.14 | *** join/#asterisk-dev CELYA_ (~Thunderbi@2a01cb0583d31d00f91324e03ca296f7.ipv6.abo.wanadoo.fr) |
| 02:35.08 | *** join/#asterisk-dev CELYA_ (~Thunderbi@2a01cb0583d31d00e14960e9f80e6733.ipv6.abo.wanadoo.fr) |
| 02:49.55 | *** join/#asterisk-dev CELYA_ (~Thunderbi@2a01cb0583d31d0025df8d422045ea68.ipv6.abo.wanadoo.fr) |
| 03:05.47 | *** join/#asterisk-dev CELYA_ (~Thunderbi@2a01cb0583d31d00d977257c14af6b7a.ipv6.abo.wanadoo.fr) |
| 03:20.37 | *** join/#asterisk-dev CELYA_ (~Thunderbi@2a01cb0583d31d00305d3f7e1c9f2614.ipv6.abo.wanadoo.fr) |
| 03:35.29 | *** join/#asterisk-dev CELYA_ (~Thunderbi@2a01cb0583d31d005d92613f55d8c285.ipv6.abo.wanadoo.fr) |
| 03:48.22 | *** join/#asterisk-dev CELYA_ (~Thunderbi@2a01cb0583d31d00e1a1030b771647e3.ipv6.abo.wanadoo.fr) |
| 04:01.13 | *** join/#asterisk-dev CELYA_ (~Thunderbi@2a01cb0583d31d007138d07986321ccd.ipv6.abo.wanadoo.fr) |
| 04:16.05 | *** join/#asterisk-dev CELYA_ (~Thunderbi@2a01cb0583d31d0080e91022fc1f8667.ipv6.abo.wanadoo.fr) |
| 04:28.58 | *** join/#asterisk-dev CELYA_ (~Thunderbi@2a01cb0583d31d003cd38ef67965e76b.ipv6.abo.wanadoo.fr) |
| 04:43.46 | *** join/#asterisk-dev CELYA_ (~Thunderbi@2a01cb0583d31d008c82915b733995b4.ipv6.abo.wanadoo.fr) |
| 04:59.42 | *** join/#asterisk-dev CELYA_ (~Thunderbi@2a01cb0583d31d00d9f3fc7ad9592930.ipv6.abo.wanadoo.fr) |
| 05:00.49 | *** join/#asterisk-dev sh_smith (~sh_smith@cpe-76-174-26-91.socal.res.rr.com) |
| 05:17.30 | *** join/#asterisk-dev CELYA_ (~Thunderbi@2a01cb0583d31d00ac32e9c820ca95d0.ipv6.abo.wanadoo.fr) |
| 05:40.26 | *** join/#asterisk-dev CELYA_ (~Thunderbi@2a01cb0583d31d00cdefa74fc647655f.ipv6.abo.wanadoo.fr) |
| 05:55.19 | *** join/#asterisk-dev CELYA_ (~Thunderbi@2a01cb0583d31d00b1f6494a66afa46c.ipv6.abo.wanadoo.fr) |
| 06:10.07 | *** join/#asterisk-dev CELYA_ (~Thunderbi@2a01cb0583d31d00c9f885c0ac529caf.ipv6.abo.wanadoo.fr) |
| 06:22.57 | *** join/#asterisk-dev CELYA_ (~Thunderbi@2a01cb0583d31d0060c626458d1933fb.ipv6.abo.wanadoo.fr) |
| 06:37.59 | *** join/#asterisk-dev CELYA_ (~Thunderbi@2a01cb0583d31d002c3962045c583aed.ipv6.abo.wanadoo.fr) |
| 07:06.40 | *** join/#asterisk-dev DanQuinney (sid18169@gateway/web/irccloud.com/x-pdqsbmcbmjgozvtm) |
| 07:07.45 | *** join/#asterisk-dev X-Rob (sid14615@gateway/web/irccloud.com/x-lximobhtsuarvjww) |
| 07:16.13 | *** join/#asterisk-dev opticron (~opticron@75.76.59.35) |
| 07:16.13 | *** mode/#asterisk-dev [+o opticron] by ChanServ |
| 07:17.41 | *** join/#asterisk-dev CELYA_ (~Thunderbi@2a01:e0a:d:1111:1dd7:b42:3b85:c2f5) |
| 07:30.12 | *** join/#asterisk-dev jkroon (~jkroon@uls-154-73-35-201.wall.uls.co.za) |
| 07:46.22 | *** join/#asterisk-dev tsearle (~tsearle@37.19.10.33) |
| 07:46.22 | *** mode/#asterisk-dev [+o tsearle] by ChanServ |
| 07:50.07 | *** join/#asterisk-dev tsearle (~tsearle@37.19.10.33) |
| 07:50.07 | *** mode/#asterisk-dev [+o tsearle] by ChanServ |
| 08:28.13 | *** join/#asterisk-dev pchero_work (~pchero@109.70.54.56) |
| 09:37.35 | *** join/#asterisk-dev jkroon_ (~jkroon@165.16.204.36) |
| 09:49.27 | *** join/#asterisk-dev jkroon_ (~jkroon@uls-154-73-35-201.wall.uls.co.za) |
| 11:59.28 | *** join/#asterisk-dev bllackjack (67d86345@gateway/web/freenode/ip.103.216.99.69) |
| 12:25.50 | *** join/#asterisk-dev zopsi (~zopsi@dir.ac) |
| 12:50.18 | pchero_work | Hm.. I have a problem with asterisk git push. |
| 12:51.00 | file | what are you attempting to do? |
| 12:51.08 | pchero_work | https://pastebin.com/8GebiNTb |
| 12:51.17 | pchero_work | I'm trying to push it.. |
| 12:51.22 | file | you can't push to a branch |
| 12:51.26 | pchero_work | I expected change-id returns it. |
| 12:51.52 | pchero_work | I thought I used to do it.. |
| 12:52.00 | file | it has to be pushed up for review (not to the main branch), go through review, and then it gets merged |
| 12:52.28 | file | it is easiest to use the git-review utility as it does everything needed |
| 12:52.59 | file | https://wiki.asterisk.org/wiki/display/AST/Gerrit+Usage#GerritUsage-Installgit-review has the details |
| 12:55.28 | pchero_work | Ah! |
| 12:55.37 | pchero_work | I've push it! :) Thanks |
| 12:55.58 | pchero_work | I was forgot to create new branch as you mentioned it. |
| 12:56.06 | pchero_work | file: thank you. :) |
| 12:56.12 | file | yw |
| 13:35.09 | *** join/#asterisk-dev jkroon (~jkroon@41.21.200.174) |
| 14:22.28 | seanbright | happy friday! |
| 14:22.39 | file | yes! |
| 14:23.13 | *** join/#asterisk-dev zopsi (~zopsi@dir.ac) |
| 14:48.39 | *** join/#asterisk-dev bford (d8cff501@gateway/web/freenode/ip.216.207.245.1) |
| 14:48.39 | *** mode/#asterisk-dev [+o bford] by ChanServ |
| 15:23.40 | seanbright | |
| 15:23.43 | seanbright | sorry |
| 15:28.56 | seanbright | waits patiently for kharwell |
| 15:29.45 | bford | he's out today |
| 15:30.36 | seanbright | shakes his fist at the heavens |
| 15:48.14 | seanbright | so if i have a pjsip aor with a fixed contact and a qualify_frequency of 30 and that contact is dead, what should i expect asterisk to send in terms of OPTIONS requests? |
| 15:49.11 | seanbright | should it send 1 or more options requests every 30 seconds? |
| 15:49.26 | seanbright | or does it behave differently if the contact isn't responding? |
| 15:50.49 | file | well, you'd get the normal request retransmission of the SIP stack |
| 15:52.14 | seanbright | ah, right. |
| 15:52.29 | seanbright | what is 4 seconds? |
| 15:52.39 | seanbright | one of the default timers? |
| 15:55.03 | seanbright | ah, T2. |
| 15:55.06 | seanbright | word. |
| 15:56.24 | seanbright | so having a qualify frequency less than 32 seconds is probably dumb |
| 15:57.48 | file | it's probably not the best |
| 16:02.12 | seanbright | ah. ok i think i see the bug. |
| 16:28.08 | seanbright | or maybe not. |
| 16:36.22 | seanbright | well i can't duplicate ASTERISK-26806 so i'll have to wait for kevin to come back |
| 16:36.26 | seanbright | he back on monday? |
| 16:37.58 | file | yes |
| 16:38.02 | seanbright | kk |
| 16:53.37 | *** join/#asterisk-dev Worldexe (~Worldexe@95-107-33-134.dsl.orel.ru) |
| 17:08.04 | *** join/#asterisk-dev rmudgett (rmudgett@nat/digium/x-qykwqobgjzqmqtmh) |
| 17:08.05 | *** mode/#asterisk-dev [+o rmudgett] by ChanServ |
| 17:24.27 | *** join/#asterisk-dev zopsi (~zopsi@dir.ac) |
| 17:52.08 | seanbright | coreyfarrell: how does one use ast_cli_completion_add? |
| 17:52.19 | seanbright | did you update any commands to use it? |
| 18:04.50 | *** join/#asterisk-dev zopsi (~zopsi@dir.ac) |
| 18:15.22 | *** join/#asterisk-dev zopsi (~zopsi@dir.ac) |
| 18:41.09 | coreyfarrell | seanbright: no commands use it yet, but it's easy. just call ast_cli_completion_add repeatedly until you've added all matches, then return NULL from the CLI handler so it doesn't get called again. |
| 18:41.33 | coreyfarrell | I'll be back in like 10 minutes if you need more help. |
| 18:44.35 | seanbright | nope, i should be good. thanks. |
| 18:53.44 | *** join/#asterisk-dev CELYA_ (~Thunderbi@2a01cb0583d31d00d95ae8fc9689de08.ipv6.abo.wanadoo.fr) |
| 18:56.15 | seanbright | coreyfarrell: easy peasy lemon squeezy |
| 18:56.19 | seanbright | very elegant |
| 19:02.46 | coreyfarrell | yeah the first patch required you to build your own vector then turn it over to the CLI core. this way reads threadstorage for each match but I think it's worth the ease of use. |
| 19:07.06 | *** join/#asterisk-dev sh_smith (~sh_smith@cpe-76-174-26-91.socal.res.rr.com) |
| 19:10.02 | *** join/#asterisk-dev sh_smith (~sh_smith@cpe-76-174-26-91.socal.res.rr.com) |
| 19:10.30 | *** join/#asterisk-dev CELYA_ (~Thunderbi@2a01cb0583d31d006054f85cee514720.ipv6.abo.wanadoo.fr) |
| 19:11.34 | coreyfarrell | seanbright: I posted a comment on https://gerrit.asterisk.org/7471 with a +1. feel free to address or ignore the comment.. |
| 19:21.05 | seanbright | coreyfarrell: updated |
| 19:21.16 | seanbright | AND failing the build |
| 19:21.19 | seanbright | so there's that |
| 19:21.24 | coreyfarrell | I saw. sorry |
| 19:21.31 | seanbright | this is why we can't have nice things |
| 19:21.52 | coreyfarrell | BTW I'm looking at your astdb prefix review, can you point me to any document that explains what the hell is going on in your new SQL query? |
| 19:22.04 | seanbright | heh |
| 19:22.13 | seanbright | you mean it's not obvious? |
| 19:22.23 | seanbright | i will add a comment |
| 19:22.33 | coreyfarrell | I sort-of know SQL from MSSQL / mysql, I do not recognize whatever the hell that stuff is |
| 19:22.47 | seanbright | so the concatenation operator in sqlite is || |
| 19:23.19 | seanbright | oh, and ?1 is a numbered place holder |
| 19:23.29 | seanbright | so if you do this: |
| 19:23.31 | coreyfarrell | I figured the ?1 part :) |
| 19:23.33 | seanbright | ok |
| 19:24.06 | coreyfarrell | actually misread it as $1 .. |
| 19:24.11 | *** join/#asterisk-dev CELYA_ (~Thunderbi@2a01cb0583d31d0005adca99e68a88e9.ipv6.abo.wanadoo.fr) |
| 19:24.13 | seanbright | so it basically says "find all of the keys that are gte 'foo' and lt 'foo<high unicode code point>' |
| 19:24.23 | seanbright | which would be everything that started with 'foo' |
| 19:24.43 | seanbright | the REAL question you should be asking is what the fuck is the current gettree_stmt doing |
| 19:24.44 | coreyfarrell | OH. |
| 19:25.55 | seanbright | frankly, i'm not sure why the query planner can't optimize LIKE 'foo%' into something that utilizes indexes |
| 19:26.29 | seanbright | if you look at the gettree_stmt one, it's appending a slash to the key |
| 19:26.35 | seanbright | so that will always do a full table scan |
| 19:26.41 | seanbright | which is awful |
| 19:27.11 | coreyfarrell | remnant of how things were done in dbd? |
| 19:29.20 | seanbright | i guess? |
| 19:30.06 | seanbright | not sure. but we can add to that api all day, so i don't really care what other consumers are doing. |
| 19:30.25 | seanbright | i found that x'ffff' hack on the internet somewhere, let me find a link |
| 19:30.57 | seanbright | coreyfarrell: http://sqlite.1065341.n5.nabble.com/fast-string-prefix-matching-tp10777p10778.html |
| 19:31.27 | gtjoseph | i was wondering about that. you might want to add a comment |
| 19:31.31 | gtjoseph | in the code |
| 19:32.51 | seanbright | coreyfarrell: i'm not following your question about regex |
| 19:33.57 | *** join/#asterisk-dev zopsi (~zopsi@dir.ac) |
| 19:34.33 | coreyfarrell | seanbright: a few lines before the comment we check for regex[0] == '^' then do some special processing so we ast_db_gettree with a prefix.. |
| 19:36.33 | coreyfarrell | seanbright: I'm thinking in the case where we call make_astdb_prefix_pattern we would also want to use ast_db_gettree_prefix instead of ast_db_gettree. |
| 19:37.00 | coreyfarrell | obviously make_astdb_prefix_pattern would need to stop adding '%' to the end for that to work. |
| 19:38.26 | *** join/#asterisk-dev CELYA_ (~Thunderbi@2a01cb0583d31d00992cd0584edb9121.ipv6.abo.wanadoo.fr) |
| 19:41.09 | seanbright | looking |
| 19:42.01 | seanbright | eh |
| 19:42.28 | seanbright | there are things that spring to mind |
| 19:43.09 | seanbright | 1. if you want to do a prefix match, just use ast_sorcery_retrieve_by_prefix() |
| 19:43.46 | seanbright | 2. sqlite3 *does* have a REGEX operator (not built in, but can be installed via package on a lot of distros). so i would much rather push regex processing down into sqlite when we can. |
| 19:43.48 | coreyfarrell | seanbright: maybe dev-mode should print a warning/error message? ast_log(LOG_WARNING, "Using regex for prefix lookup is bad, mkay?\n"); |
| 19:44.30 | seanbright | sure. but does that prevent this patch from going in? |
| 19:45.09 | coreyfarrell | no.. I just -1'ed to make sure you saw it.. if you don't want to mess with the regex prefix thing I understand |
| 19:50.00 | *** join/#asterisk-dev CELYA_ (~Thunderbi@2a01cb0583d31d00f8ed94204d255d5f.ipv6.abo.wanadoo.fr) |
| 20:01.48 | *** join/#asterisk-dev zopsi (~zopsi@dir.ac) |
| 20:04.51 | *** join/#asterisk-dev CELYA_ (~Thunderbi@2a01cb0583d31d00185f5e68d7e714d3.ipv6.abo.wanadoo.fr) |
| 20:11.36 | seanbright | ok, hopefully adequate comments added |
| 20:24.01 | *** join/#asterisk-dev ChanServ (ChanServ@services.) |
| 20:24.01 | *** mode/#asterisk-dev [+o ChanServ] by tolkien.freenode.net |
| 20:25.34 | *** join/#asterisk-dev jkroon (~jkroon@165.16.204.166) |
| 20:28.57 | *** join/#asterisk-dev CELYA_ (~Thunderbi@2a01cb0583d31d009c69f571d6855cc4.ipv6.abo.wanadoo.fr) |
| 20:41.45 | *** join/#asterisk-dev CELYA_ (~Thunderbi@2a01cb0583d31d00510a610902d76946.ipv6.abo.wanadoo.fr) |
| 20:54.44 | *** join/#asterisk-dev CELYA_ (~Thunderbi@2a01cb0583d31d00314d436212bacfe4.ipv6.abo.wanadoo.fr) |
| 21:07.11 | coreyfarrell | does asterisk or digium have anything to create short + permanent url's pointing to a wiki article? something that could be linked from a source with: #error "you must do X, see https://.../" |
| 21:07.36 | *** join/#asterisk-dev CELYA_ (~Thunderbi@2a01cb0583d31d00d586761f152028a1.ipv6.abo.wanadoo.fr) |
| 21:08.36 | seanbright | the wiki has that built in i think |
| 21:09.06 | seanbright | if you go to the "Tools" drop down on the upper right |
| 21:09.18 | coreyfarrell | oh cool. thanks. |
| 21:09.20 | seanbright | there is a "Link to this page..." option which has a "tiny url" |
| 21:09.30 | coreyfarrell | yeah I see it, that'll work. |
| 21:10.59 | coreyfarrell | When I get a chance I'm going to write a quick "How to compile modules from outside the Asterisk source tree" then link to it from asterisk.h. |
| 21:12.49 | seanbright | we need a pcap redactor |
| 21:23.09 | *** join/#asterisk-dev CELYA_ (~Thunderbi@2a01cb0583d31d0010ec87fb6a4d2241.ipv6.abo.wanadoo.fr) |
| 21:34.59 | *** join/#asterisk-dev CELYA_ (~Thunderbi@2a01cb0583d31d00383c56ede9448bee.ipv6.abo.wanadoo.fr) |
| 21:50.14 | *** join/#asterisk-dev CELYA_ (~Thunderbi@2a01cb0583d31d0024931f73350eacb8.ipv6.abo.wanadoo.fr) |
| 22:02.22 | *** join/#asterisk-dev CELYA_ (~Thunderbi@2a01cb0583d31d005d39ca47f0a34631.ipv6.abo.wanadoo.fr) |
| 22:05.43 | *** join/#asterisk-dev zopsi (~zopsi@dir.ac) |
| 22:08.12 | seanbright | coreyfarrell: re: ASTERISK-27383 |
| 22:08.47 | seanbright | i believe the reason there is a global mutex around the astdb code is simply because that is how it has always been |
| 22:09.08 | seanbright | a single sqlite3 database can be used by multiple threads/processes at the same time |
| 22:09.22 | seanbright | writers block readers, but readers can run in parallel |
| 22:09.41 | seanbright | plus there is a shared cache mode and WAL option that reduces contention |
| 22:10.28 | seanbright | in short: i think there is a lot of room for improvement there and it won't take too much effort |
| 22:10.52 | seanbright | sigh... |
| 22:10.56 | seanbright | creates a branch |
| 22:17.35 | *** join/#asterisk-dev CELYA_ (~Thunderbi@2a01cb0583d31d0095fd842243de7938.ipv6.abo.wanadoo.fr) |
| 22:33.29 | *** join/#asterisk-dev CELYA_ (~Thunderbi@2a01cb0583d31d0018a01fd3f35449e3.ipv6.abo.wanadoo.fr) |
| 22:48.16 | *** join/#asterisk-dev CELYA_ (~Thunderbi@2a01cb0583d31d00e4d45aa7fa4c5497.ipv6.abo.wanadoo.fr) |
| 22:54.08 | coreyfarrell | seanbright: I opened that ticket assuming that the performance woes of astdb could not be fixed. it they can that's great. |
| 23:01.20 | *** join/#asterisk-dev CELYA_ (~Thunderbi@2a01cb0583d31d003914a77185762032.ipv6.abo.wanadoo.fr) |
| 23:09.32 | Kobaz | perks up |
| 23:09.34 | Kobaz | astdb |
| 23:10.06 | Kobaz | right |
| 23:10.16 | Kobaz | i've also had performance problems with astdb sqlite |
| 23:10.32 | Kobaz | i had to put it on a ramdisk, i would get serious lock contention that would 'freeze' everything basically |
| 23:10.52 | Kobaz | i converted the locks into trylocks, and abort after 5 attempts, and that's resolved my issues with it |
| 23:11.06 | *** join/#asterisk-dev gtjoseph (~gtjoseph@unaffiliated/gtj) |
| 23:11.06 | *** mode/#asterisk-dev [+o gtjoseph] by ChanServ |
| 23:12.08 | Kobaz | there's also the race condition issues from realtime from the thread |
| 23:15.48 | *** join/#asterisk-dev zopsi (~zopsi@dir.ac) |
| 23:16.41 | *** join/#asterisk-dev CELYA_ (~Thunderbi@2a01cb0583d31d0088618697e59d50d5.ipv6.abo.wanadoo.fr) |
| 23:29.25 | *** join/#asterisk-dev CELYA_ (~Thunderbi@2a01cb0583d31d0038db719b968a247f.ipv6.abo.wanadoo.fr) |
| 23:42.11 | *** join/#asterisk-dev CELYA_ (~Thunderbi@2a01cb0583d31d00c03bc868dfec4d4a.ipv6.abo.wanadoo.fr) |
| 23:56.14 | *** join/#asterisk-dev seanbright (~sean@asterisk/community-developer/seanbright) |
| 23:56.14 | *** mode/#asterisk-dev [+o seanbright] by ChanServ |
| 23:57.01 | *** join/#asterisk-dev CELYA_ (~Thunderbi@2a01cb0583d31d00e159316779436987.ipv6.abo.wanadoo.fr) |