| 03:48.47 | jkroon | cresl1n, is right, function pointer calls are indirect, meaning the target address needs to be loaded first, it's always a "far" call (relative calls are an awesome thing to minimize code size, so eg, if two functions are close in memory then shorter instructions can be used in terms of code size). Interestingly most cross-module functions are similar to function pointers anyway, so if I've got a call to a dynamically linked function that goes |
| 03:48.47 | jkroon | through a jmp instruction to a placeholder for the call instruction, which by default calls a linker function upon first call, which then updates that code segment with the call to the right address. |
| 06:03.56 | *** join/#asterisk-dev CELYA_ (~Thunderbi@2a01:e0a:d:1111:b078:cb6c:125:933b) |
| 09:41.26 | *** join/#asterisk-dev dadrc (~quassel@unaffiliated/dadrc) |
| 11:01.26 | *** join/#asterisk-dev jkroon (~jkroon@165.16.203.126) |
| 12:24.00 | jkroon | file, i note you abandoned https://gerrit.asterisk.org/#/c/asterisk/+/10047/ ?? |
| 12:24.18 | file | I moved it to the 13 branch |
| 12:24.29 | file | it was put up against a release branch, which is only done when a regression release is to be done |
| 12:24.30 | jkroon | will look. |
| 12:25.09 | jkroon | will track one thanks. |
| 12:26.27 | jkroon | file, i'm sorry for being a pain, it just seems you're very responsive in general here, and i'm starting to feel like i'm abusing that. |
| 12:26.40 | file | I'm generally responsive everywhere |
| 12:26.57 | jkroon | you recall we discussed the whole qualify issue where pjsip extensions keeps doing the unavailable thing? sounded like you were involved in that code? |
| 12:27.08 | file | I rewrote qualify support |
| 12:28.49 | jkroon | the reports i'm getting is that phones are going unavailable and reachable type of thing, but i'm not seeing anything odd anywhere. upon startup things are very shoddy until I at least issue pjsip reload |
| 12:29.13 | file | the current code will tell you what it is doing with a sufficiently high debug level, and why it has chosen such a thing |
| 12:29.15 | jkroon | i'm not sure how to approach it, what would you need from me to be able to assist? |
| 12:29.32 | file | otherwise a precise scenario and details to reproduce the scenario |
| 12:29.46 | jkroon | what's your timezone and availability? |
| 12:29.58 | file | I'm not offering one on one help like that, I'm busy :P |
| 12:30.04 | file | if you file an issue then it would go into the queue |
| 12:30.32 | file | I can answer specific questions but debugging something live like that is not something I can do while keeping up my normal responsibilities |
| 12:31.08 | jkroon | we all are :) |
| 12:31.36 | jkroon | cool, understood, and appreciate the honesty :). really do. |
| 12:32.10 | file | in particular with the run up to Astricon, things are going to get even moreeeeeee busy |
| 12:32.31 | file | plus we were acquired... |
| 12:32.36 | file | :D |
| 12:33.13 | jkroon | hehehe, i can imagine things are spinning at the moment, would have loved to be at astricon but i'm presenting the day before in ZA ... and cheaps aren't cheap right at the moment. |
| 12:40.28 | jkroon | file, don't want to abuse, actually I'd just like contribute. I've got some code-writing skills I'm trying to offer, so I think the reality is more that I need pointers in many cases more so than outright someone to write code on my behalf to fix my problems. |
| 12:41.10 | file | the debug code will tell you what it is doing and why, otherwise you really just have to dig in and follow the flow and execution |
| 12:41.20 | file | I can't really give any specific hints without labbing it up and digging in myself |
| 12:41.56 | jkroon | thank you, i'll re-install with al the debug stuff as soon as the client closes shop. Obviously you'd prefer that to happen on 13.23 rather than 13.22 :). |
| 12:42.35 | file | the options code I don't think changed between those |
| 12:43.05 | jkroon | but something else that the code interact with might've, which could already fix it. |
| 12:43.14 | file | I added the debug messages for this precise reason though, so it's not an absolute guess |
| 12:44.08 | jkroon | file, i know i'm keeping you from other work, thanks, i'll report later with feedback, you've helped a tremendous lot already. |
| 12:45.20 | file | this is an excuse to not be in chan_sip for something |
| 12:48.40 | jkroon | if i can help, please let me know. |
| 12:48.50 | jkroon | got about half an hour now. |
| 12:49.42 | jkroon | otherwise I'll be online again ~3 hours debugging the PJSIP availability issue. |
| 13:01.34 | jkroon | file, on which file should I set the debug level for the pjsip qualify? |
| 13:02.17 | file | um res_pjsip |
| 13:02.32 | jkroon | was just wondering if it's one of the add-on modules, thanks for confirming it's not. |
| 13:11.48 | jkroon | file, could it be about module load order? |
| 13:11.59 | file | it shouldn't care. |
| 13:58.16 | off-nominal | so i may want to move these conversion routines... :-) |
| 13:59.03 | off-nominal | into res_pjproject |
| 14:24.57 | *** join/#asterisk-dev kharwell (kharwell@nat/digium/x-nymerbrsihzyfbxz) |
| 14:24.57 | *** mode/#asterisk-dev [+o kharwell] by ChanServ |
| 14:32.38 | *** join/#asterisk-dev rmudgett (rmudgett@nat/digium/x-podrvthoywfidqqp) |
| 14:32.50 | *** mode/#asterisk-dev [+o rmudgett] by ChanServ |
| 14:40.22 | *** join/#asterisk-dev bford (uid283514@gateway/web/irccloud.com/x-hklqocmcjmhqcjou) |
| 14:40.22 | *** mode/#asterisk-dev [+o bford] by ChanServ |
| 16:45.19 | *** join/#asterisk-dev jkroon (~jkroon@165.16.204.173) |
| 17:25.47 | *** join/#asterisk-dev jpsharp (~jsharp@linode.fivecats.org) |
| 17:27.02 | jpsharp | If I'm doing an ast_read(chan,100) on a channel and nothing comes in on that 100ms, will that also affect an ast_stream playing on that channel? Cause it to stutter? |
| 17:28.06 | cresl1n | IIRC yes |
| 17:28.12 | cresl1n | goes back to look at some code |
| 17:28.51 | cresl1n | although ast_read doesn't have a blocking timeout |
| 17:28.53 | cresl1n | so no |
| 17:28.54 | cresl1n | :-) |
| 17:33.34 | jpsharp | Hmm. Odd. I'm trying to handle ast_read on some channels tht don't have continuous RTP. |
| 17:33.39 | jpsharp | damn silence suppression. |
| 17:44.51 | jpsharp | win 3 |
| 17:50.50 | cresl1n | If ast_read is blocking internally for 100 ms, it's quite likely that it will cause nothing to be written |
| 17:51.18 | cresl1n | At least historically, you only have one thread writing/reading audio from a channel at a time |
| 17:51.34 | jpsharp | Its definitely history, I'm still dealing with 11.25 |
| 17:51.38 | file | there's a timer file descriptor on it, so it should wake up every 20ms if you have a working timer |
| 17:51.45 | file | oh, I have no idea for that old then |
| 17:51.45 | cresl1n | I was only saying that there wasn't a version of ast_read() with a 100 ms timeout option |
| 17:52.35 | cresl1n | You'd have to do a ast_waitfor(chan, 100) to wait for a frame |
| 17:52.58 | jpsharp | Er, wait. Thats what I'm doing. Darnit. |
| 17:53.06 | jpsharp | ast_waitfor() not ast_read |
| 17:53.07 | cresl1n | Usually you don't want to wait on things though |
| 17:53.12 | cresl1n | unless you know what you're doing |
| 17:53.16 | jpsharp | Not enough coffee yet. |
| 17:53.19 | cresl1n | blocking is bad for real time media |
| 17:53.57 | rmudgett | You should only call ast_read() when you know there is something to be read. That's what ast_waitfor_xxx is supposed to do. |
| 17:54.31 | cresl1n | But even ast_waitfor_xxx won't automatically "write media", if you wait too long |
| 17:54.34 | jpsharp | But if I'm waiting in ast_waitfor(), will that stop stream playback? |
| 17:54.36 | jpsharp | Okay. |
| 17:55.08 | cresl1n | those are just the "right functions" to use to wait for media on a channel |
| 17:58.56 | *** join/#asterisk-dev jameswf (uid27319@gateway/web/irccloud.com/x-jglcpmvrxrqhmdme) |
| 18:00.13 | rmudgett | There are plenty of code examples using the ast_waitfor functions to handle read and write. Mostly the write is going to be handled by a generator run by a timer to know when to send the next frame. |
| 18:02.43 | cresl1n | I guess it depends on the application |
| 18:48.35 | jpsharp | can I put in a request for someone to look at doxygen.asterisk.org? Returns 503 for everythign. |
| 18:49.53 | file | points at cresl1n |
| 18:50.14 | cresl1n | Didn't we decide to kill it? |
| 18:50.20 | cresl1n | tries to remember |
| 18:50.26 | cresl1n | it hasn't been updated in forever, IIRC |
| 18:50.35 | cresl1n | and you can generate the doxygen docs yourself |
| 18:50.43 | file | cresl1n: I think it was on your list |
| 18:50.48 | cresl1n | crap. |
| 18:50.56 | file | like, months ago |
| 18:50.58 | jpsharp | Just that google hits return doxygen.asterisk fir the first hit. |
| 18:51.02 | cresl1n | jpsharp - that's what I remember about it |
| 18:51.15 | cresl1n | apparently I dropped the ball on either killing it or fixign it |
| 18:51.42 | cresl1n | jpsharp: it's better to just generate them yourself |
| 18:51.59 | jpsharp | that's what I'm doing. And doing a local search engine on it. |
| 19:20.56 | *** join/#asterisk-dev libardi (~libardi@177.194.124.58) |
| 19:23.46 | jpsharp | so i should do a ast_sched_wait() and then an ast_waitfor based on that that result? |
| 19:25.41 | rmudgett | ast_sched_wait() is dealing with the scheduler. That is something entirely different. |
| 19:26.21 | jpsharp | Then I guess I'm really confused then. |
| 19:35.12 | rmudgett | jpsharp: Are you trying to write your own dialplan application? |
| 19:36.01 | jpsharp | Yes. I'm building an application that does background recording and silence/talk detection to work with our speech recognition IVR. |
| 19:36.49 | jpsharp | It mostly works, but it burns a whole lot of CPU spinning in a while loop. |
| 19:38.46 | rmudgett | So you basically want to combine the playback and AMD applications. |
| 19:38.58 | jpsharp | yes. |
| 19:39.10 | jpsharp | And waitforsilence/noise |
| 19:39.53 | jpsharp | And some custom code that uses some ast_dsp stuff to do voicemail/answering machine beep detection. |
| 19:49.31 | *** join/#asterisk-dev sduthil (~sduthil@modemcable094.94-70-69.static.videotron.ca) |
| 19:52.44 | rmudgett | jsmith: I see why you were calling ast_sched_wait() now. Yeah. You'll need to use that to run any channel scheduled jobs. |
| 19:53.27 | jpsharp | Which I wasn't. |
| 19:53.47 | rmudgett | There is a loop in app_followme which opens a stream to playback a file but doesn't then subsequently call ast_waitstream() that you should study for ideas. |
| 19:54.17 | rmudgett | The followme loop is in wait_for_winner() |
| 19:54.28 | jpsharp | Ah, okay. I'll go check that. |
| 21:00.30 | off-nominal | file: i have a question that is so bad i refuse to ask it on gerrit |
| 21:00.44 | file | why do birds suddenly appear? |
| 21:01.24 | off-nominal | re: generator patch - could that frametype check be second in that if statement rather than third? |
| 21:01.30 | off-nominal | without causing a problem? |
| 21:01.40 | file | it could be first if it wanted to |
| 21:01.51 | off-nominal | zeroth? |
| 21:02.09 | file | as long as it exists somewhere in space and time |
| 21:02.31 | off-nominal | ok. could you make it second then? |
| 21:02.34 | off-nominal | or do i have to request that? |
| 21:02.35 | off-nominal | :-) |
| 21:02.37 | off-nominal | on gerrit. |
| 21:02.47 | file | such a Richard request |
| 21:02.50 | off-nominal | heh |
| 21:02.55 | off-nominal | HOW DARE YOU |
| 21:02.56 | file | I'm done work for the day, leave a comment and I'll do it tomorrow |
| 21:02.59 | off-nominal | k |
| 21:03.05 | file | I don't really care either way |
| 21:03.20 | off-nominal | nano-optimization |
| 21:03.24 | off-nominal | premature |
| 21:03.29 | *** join/#asterisk-dev GameGamer43 (sid5533@gateway/web/irccloud.com/x-uiybnolnaemizros) |
| 21:04.18 | off-nominal | note left. i refuse to -1 it. |
| 21:31.49 | rmudgett | is now known as nano-optimization |
| 21:32.40 | off-nominal | there you go. |
| 23:00.00 | *** part/#asterisk-dev kharwell (kharwell@nat/digium/x-nymerbrsihzyfbxz) |