| 03:27.14 | *** join/#asterisk-doc jsmith (n=jsmith@h46055150.area3.spcsdns.net) |
| 03:27.15 | *** mode/#asterisk-doc [+o jsmith] by ChanServ |
| 06:12.39 | *** join/#asterisk-doc file (n=file@asterisk/developer-and-muffin-lover/file) |
| 06:17.03 | *** join/#asterisk-doc file (n=file@asterisk/developer-and-muffin-lover/file) |
| 06:39.42 | *** join/#asterisk-doc file (n=file@asterisk/developer-and-muffin-lover/file) |
| 12:59.25 | *** join/#asterisk-doc Alric (n=nbowyer@masq.hyperusa.com) |
| 13:26.08 | *** join/#asterisk-doc blitzrage (n=blitzrag@asterisk/documenteur-extraordinaire/blitzrage) |
| 13:26.08 | *** mode/#asterisk-doc [+o blitzrage] by ChanServ |
| 14:11.11 | *** join/#asterisk-doc fakhir (n=Fakhir@unaffiliated/fakhir) |
| 15:20.47 | *** join/#asterisk-doc jsmith (n=jsmith@000-185-916.area3.spcsdns.net) |
| 15:20.47 | *** mode/#asterisk-doc [+o jsmith] by ChanServ |
| 15:26.07 | file | jsmith: ! |
| 15:26.50 | file | eep |
| 15:30.24 | jsmith | That reminds me... why did the elephant paint it's toenails red? |
| 15:30.37 | file | unknown |
| 15:36.17 | jsmith | To hide in the strawberry patch. |
| 15:46.22 | file | jsmith = CRM |
| 15:47.14 | jsmith | Have you ever seen an elephant hide in a strawberry patch? |
| 15:48.07 | file | I won't answer that! |
| 16:07.25 | jsmith | Must work pretty well then ;-) |
| 16:13.22 | *** join/#asterisk-doc fakhir (n=Fakhir@unaffiliated/fakhir) |
| 17:33.31 | jsmith | file: Yes? |
| 17:33.55 | file | nevermind |
| 17:34.33 | Juggie | file, do i need to destroy a list after i'm done with it? |
| 17:34.50 | file | depends on how it is allocated and whether it has a lock or not |
| 17:34.56 | file | if it has a lock, then yes - it is proper to destroy it |
| 17:34.58 | file | and initialize it |
| 17:38.48 | *** join/#asterisk-doc Juggie (n=Juggie@CPE00c049d9f271-CM000a73a18a20.cpe.net.cable.rogers.com) |
| 17:38.57 | *** join/#asterisk-doc fakhir (n=Fakhir@unaffiliated/fakhir) |
| 17:39.39 | Juggie | file, as i was saying, i have a AST_RWLIST_HEAD_STATIC, then in my unload to clean up, i lock it, then safe_traverse, in the loop i free and then remove each entry, then unlock. |
| 17:39.57 | Juggie | i'm just asking if i need to use AST_RWLIST_HEAD_DESTROY |
| 17:40.00 | Juggie | i guess it coudnt hurt. |
| 17:42.31 | file | I can't answer that without seeing the code and whether you are doing things properly or not ... |
| 17:42.54 | file | if I suspect you are, then you shouldn't have to |
| 17:43.40 | file | as the static stuff automatically initializes/deinitializes as needed |
| 17:45.03 | Juggie | ic, well i'll put it there for good mesure :) |
| 17:45.09 | Juggie | can allways remove it later |
| 17:46.32 | Juggie | file, whats the weather like today |
| 17:46.37 | Juggie | its been pouring here since yesterday |
| 17:46.44 | file | it is the same. |
| 17:46.56 | Juggie | was going to laronde (six flags montreal) yesterday |
| 17:47.02 | Juggie | and coudnt cuz of the weather |
| 18:01.24 | Juggie | file, still there? |
| 18:01.30 | file | yes |
| 18:02.20 | Juggie | i'm getting, res_agi.c:1647: warning: assignment from incompatible pointer type on my uses of AST_RWLIST_TRAVERSE(&agi_commands, temp_command, list) { but it appears to be working fine. |
| 18:02.35 | Juggie | any idea what that could be. |
| 18:02.44 | file | without seeing the code I can't answer |
| 18:03.16 | Juggie | do you have time to take a quick look? |
| 18:03.41 | Juggie | i suspect its how i have my struct defnied |
| 18:03.56 | file | pastebin the relevant parts and I will look |
| 18:04.14 | Juggie | *defined, but i did it that way so i could use the struct in the list, and also outside just as an array. |
| 18:05.08 | file | you could have also just left the array stuff alone and added a static rwlock |
| 18:06.43 | Juggie | file, http://pastebin.ca/628027 |
| 18:06.58 | file | using an array is also better on the CPU cache... |
| 18:07.49 | file | AST_LIST_ENTRY(agi_command_ll) list; |
| 18:07.56 | file | AST_LIST_ENTRY(agi_command) list; |
| 18:07.59 | Juggie | oops |
| 18:08.08 | Juggie | left over from something i was previously trying :) |
| 18:08.45 | Juggie | file i could, but then its quite hard to add/remove commands |
| 18:08.55 | Juggie | eg, if another module wanted to register agi commands |
| 18:10.36 | file | not particularly, we aren't nowhere near running out of array slots... nobody registers/unregisters (well, I know someone who does... but anyway)... and it's more efficient for this scenario imo |
| 18:12.54 | Juggie | i doubt those ast_agi_register/unregister functions even work |
| 18:13.09 | Juggie | they arnt referenced anywhere |
| 18:13.26 | file | I know they do |
| 18:13.42 | file | :D |
| 18:15.02 | Juggie | i'm sure they work, but they dont work properly i dont think |
| 18:15.09 | Juggie | unless i'm reading the code wrong |
| 18:15.44 | Juggie | they dont appear to handle commands with more then one part. |
| 18:15.49 | Juggie | so if i registered hey world |
| 18:16.02 | Juggie | then hey josh would probally fail |
| 18:16.19 | Juggie | it only appears to be checking the first command element |
| 18:16.29 | Juggie | secondly, why would you even rewrite that code when you already have find_command |
| 18:18.33 | file | the code that checks to see if the command is already registered is mehish aye |
| 18:18.49 | Juggie | its useless :) |
| 18:18.55 | Juggie | find_command should just be used |
| 18:20.35 | Juggie | find_command(agi->cmda,1) would tell you |
| 18:23.59 | Juggie | hmmmm |
| 18:29.10 | file | if the AGI commands were spread out over the tree more I could see using a linked list, but as they are all in the same file... and not much registers, I find an array more efficient |
| 18:30.43 | Juggie | i would tend to agree, but they should be spread out though. |
| 18:31.50 | Juggie | eg, say should be registering its own agi commands |
| 18:32.46 | Juggie | we woudnt have the cli registering commands for applications, why should we have agi registering for outside apps |
| 18:33.54 | file | then you get into the dependency game where say would depend on having res_agi loaded |
| 18:34.33 | Juggie | theres no is_loaded(res_agi) ? |
| 18:35.15 | file | not in 1.4 |
| 18:35.23 | Juggie | trunk? |
| 18:35.48 | file | trunk has ast_module_check |
| 18:37.17 | Juggie | ya, this is for trunk not 1.4 |
| 18:37.20 | file | but you can't use that because say stuff gets initialized before module loading :) |
| 18:37.45 | file | isn't this just fun? :D |
| 18:37.48 | Juggie | i guess :) |
| 18:37.55 | Juggie | coudn't app_say register it? |
| 18:38.11 | Juggie | er, nm, i must of dreamed that app |
| 18:38.15 | Juggie | who provides SAY() |
| 18:38.33 | file | I don't remember |
| 18:38.51 | Juggie | well, a better example would be record |
| 18:39.06 | Juggie | app_record should register the record application |
| 18:39.45 | file | if you go down the route of moving the AGI commands around, then yes... linked list |
| 18:42.36 | Juggie | should they be moved around |
| 18:43.14 | Juggie | i'm asking for an opinion |
| 18:49.37 | *** join/#asterisk-doc fakhir (n=Fakhir@unaffiliated/fakhir) |
| 20:20.21 | *** join/#asterisk-doc Juggie (n=Juggie@CPE00c049d9f271-CM000a73a18a20.cpe.net.cable.rogers.com) |
| 20:22.16 | Juggie | file, you still around? |
| 20:23.06 | file | yes |
| 20:24.30 | Juggie | i'm not sure how to solve this given my newb ness :) |
| 20:24.57 | Juggie | ast_agi_register is just defined as an int, where as find_command (i mentioned this earlier) is defined as static agi_command (a struct) |
| 20:25.09 | Juggie | is it possible to call a static from a non static function? |
| 20:26.41 | Juggie | hmm nm |
| 20:26.42 | Juggie | i'm dumb |
| 20:26.57 | Juggie | it was just unhappy because find_command wasnt defined until after ast_agi_register |
| 21:32.14 | Juggie | file can i but you got 30 seconds again? |
| 21:32.32 | Juggie | *bug |
| 21:37.47 | Juggie | i cant quite get this memory to free properly |
| 21:43.04 | file | wazzat? |
| 21:51.22 | Juggie | you saw in #dev, i was trying to do a free on a pointer which was pointing to a static array |