| 10:04.59 | *** join/#asterisk-dev infobot (ibot@c-174-52-60-165.hsd1.ut.comcast.net) |
| 10:04.59 | *** topic/#asterisk-dev is Take the March 2019 Asterisk User Survey! https://goo.gl/forms/xL1VUHRsf95saly13 -- 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 |
| 10:23.29 | *** join/#asterisk-dev jkroon_ (~jkroon@41.113.104.109) |
| 11:36.44 | *** join/#asterisk-dev DanJenkins[m] (dannimblea@gateway/shell/matrix.org/x-lcmzxllfxqosywyj) |
| 11:36.44 | *** mode/#asterisk-dev [+o DanJenkins[m]] by ChanServ |
| 11:37.32 | DanJenkins[m] | Hello! |
| 12:46.42 | *** join/#asterisk-dev jkroon (~jkroon@41.113.104.109) |
| 12:48.03 | jkroon | so ... with mysql-connector-c my_bool is no longer exposed ... i can implement a fix, but I'm afraid it'll break backwards compatibility with older versions - someone that can help test on that front? |
| 12:48.43 | jkroon | also, fix compile error separately from warnings or all the related things in a single patch? |
| 12:48.52 | file | what is exposed? |
| 12:49.01 | file | there is configure logic to check for "bool" and "my_bool" |
| 12:50.09 | file | any change in that regard would need to do similar to determine the correct choice |
| 12:50.22 | file | and all related is fine |
| 12:51.10 | jkroon | i note there is a fix in master, sorry for the noise, so just warnings about my_ulonglong should be using %lu instead of %llu ... go figure. |
| 12:51.29 | jkroon | checks 13 branch specifically. |
| 12:51.42 | file | the configure change went into all applicable supported branches |
| 12:52.37 | jkroon | thanks. |
| 12:53.50 | jkroon | Does this make sense: https://gerrit.asterisk.org/c/asterisk/+/13816 |
| 12:54.34 | file | don't know, would have to dig in deeper |
| 13:15.32 | *** join/#asterisk-dev jkroon (~jkroon@41.113.95.26) |
| 13:22.53 | seanbright | they dropped my_ulonglong in favor of uint64_t but kept a #define there for backwards compat |
| 13:23.15 | seanbright | i think they even recommend the cast that you put there |
| 13:23.20 | seanbright | or suggest it, at least |
| 13:24.08 | seanbright | // Legacy definition for the benefit of old code. Use uint64_t in new code. |
| 13:24.11 | seanbright | // If you get warnings from printf, use the PRIu64 macro, or, if you need |
| 13:24.13 | seanbright | // compatibility with older versions of the client library, cast |
| 13:24.16 | seanbright | // before printing. |
| 13:28.03 | seanbright | so... i think that PRIu64 will work regardless of the mysql version, but that would need to be tested |
| 13:28.18 | seanbright | that might be a cleaner patch. i don't like casts. |
| 13:28.50 | seanbright | up until about a month ago i thought if you cast a larger integer to a smaller one that you would saturate to the upper/lower bound of the smaller type |
| 13:29.45 | seanbright | but that is so blatantly wrong it's embarassing |
| 13:35.36 | coreyfarrell | does that consistently get the LSB's of the larger int or does it depend on endian? |
| 13:36.03 | *** join/#asterisk-dev jkroon_ (~jkroon@41.113.22.253) |
| 13:39.32 | seanbright | that's a good question that i don't know the answer to |
| 13:40.22 | seanbright | on LE it just takes the LSBs and lops off the rest |
| 14:14.38 | seanbright | i don't have a BE machine to test with. anyone have a sparc? |
| 14:19.14 | *** join/#asterisk-dev tmoore (~tmoore@2600:1702:4110:3d10:e9dc:6920:c1b0:9136) |
| 14:20.38 | gtjoseph | I do but it's fairly dusty and probably rusted solid. |
| 14:20.53 | gtjoseph | qemu? |
| 14:22.57 | gtjoseph | seanbright: isn't "memcpy(ext, ".WAV", sizeof(".WAV"))" going to change the extension from wav49 to WAV49? Was that what was intended? |
| 14:26.53 | *** join/#asterisk-dev spatel (~spatel@static-71-174-102-210.bstnma.fios.verizon.net) |
| 14:28.05 | seanbright | it's not |
| 14:28.22 | seanbright | sizeof(".WAV") == 5 |
| 14:29.14 | seanbright | sorry, the "it's not" was an answer to your first question, not the "was that intended?" part |
| 14:30.45 | gtjoseph | Oh I gotcha. the null byte is included in the memcpy. |
| 14:30.52 | seanbright | indeed. |
| 14:30.55 | gtjoseph | didn't catch that |
| 14:31.05 | seanbright | if it's too "clever" i can just change it to a strcpy() |
| 14:31.34 | gtjoseph | nah. i'm fine with it. |
| 14:31.38 | seanbright | alrighty |
| 14:34.49 | seanbright | i also tested, which is unusual for me, but it did work :P |
| 14:35.10 | seanbright | the more i think about that particular patch the more i want to make it master only |
| 14:35.39 | seanbright | since X-Rob opened it, i'm thinking there might be freepbx code that relies on the broken behavior? |
| 14:36.42 | gtjoseph | hmmmm |
| 14:37.05 | seanbright | https://github.com/FreePBX/core/commit/fba815ba2ea111165716be3059bf39ddd55403fb |
| 14:37.17 | gtjoseph | if you want ot make it master only, i need to kill the jenkins jobs quickly. |
| 14:37.34 | seanbright | oh |
| 14:37.39 | seanbright | didn't realize you +2d it |
| 14:37.43 | seanbright | i think it will be fine |
| 14:37.49 | seanbright | it's technically a bug fix |
| 14:38.24 | gtjoseph | kk |
| 14:58.22 | *** join/#asterisk-dev kharwell (uid358942@gateway/web/irccloud.com/x-fmnbgiwndiynlvyg) |
| 14:58.22 | *** mode/#asterisk-dev [+o kharwell] by ChanServ |
| 15:01.17 | *** join/#asterisk-dev bford (uid283514@gateway/web/irccloud.com/x-ldppdaaxlzeaxjpu) |
| 15:01.17 | *** mode/#asterisk-dev [+o bford] by ChanServ |
| 17:06.54 | DanJenkins[m] | if i make an external media issue with 16.2 version of asterisk am i just going to get told to upgrade ? |
| 17:10.06 | cresl1n | I think we all know the answer to that question |
| 17:10.07 | cresl1n | :-P |
| 17:10.44 | file | 16.2 didn't have external media |
| 17:10.49 | DanJenkins[m] | say it aint so |
| 17:10.56 | DanJenkins[m] | yes it does |
| 17:11.12 | DanJenkins[m] | sorry |
| 17:11.13 | DanJenkins[m] | im tired |
| 17:11.14 | DanJenkins[m] | i meant 16.6.2 |
| 17:11.24 | DanJenkins[m] | too many 6's |
| 17:11.47 | file | I don't think anything was touched since then |
| 17:12.04 | DanJenkins[m] | i take back my "yes it does" short response an apologise |
| 17:12.07 | DanJenkins[m] | ok :) |
| 17:12.11 | DanJenkins[m] | thanks file |
| 17:25.03 | DanJenkins[m] | ok issue submitted |
| 17:25.09 | DanJenkins[m] | hopefully with pcaps that worked |
| 17:26.48 | file | oh god, slin |
| 17:28.42 | DanJenkins[m] | google can take it in quite a few formats but pretty much only gives it back in slin |
| 17:29.16 | DanJenkins[m] | Opus encoded audio wrapped in an ogg container., MP3 audio at 32kbps, Uncompressed 16-bit signed little-endian samples (Linear PCM). Audio content returned as LINEAR16 also contains a WAV header. |
| 17:32.13 | DanJenkins[m] | whereas they can take ulaw, flac etc in - https://cloud.google.com/dialogflow/docs/reference/rest/v2beta1/QueryInput#audioencoding |
| 17:39.22 | seanbright | 118 is dynamic |
| 17:39.33 | seanbright | hard-coded or not |
| 17:39.47 | seanbright | well, i should speak to it, because i'm not sure |
| 17:39.51 | seanbright | shouldn't* |
| 17:40.23 | DanJenkins[m] | so i'd agree with you based on the specs |
| 17:40.31 | DanJenkins[m] | but asterisk sends me type 118 |
| 17:40.58 | DanJenkins[m] | if i cant send back 118...... of the same type of audio... thats very counterintuitive |
| 17:41.09 | DanJenkins[m] | theres no sdp setup because its external media |
| 17:41.46 | seanbright | i have theories but i don't want to throw them out there because i will inevitably be proven wrong |
| 17:42.13 | DanJenkins[m] | and if its dynamic.... why are any of these static payloads in rtp_engine.c ? https://github.com/asterisk/asterisk/blob/master/main/rtp_engine.c#L3675 |
| 17:42.24 | DanJenkins[m] | hahaha seanbright |
| 17:42.49 | DanJenkins[m] | im 100% out of my comfort zone - ive no idea what ive done is right etc |
| 17:43.02 | seanbright | the static assignments are a fallback (see comment about 25 lines up) |
| 17:43.50 | DanJenkins[m] | soooo it should fallback to the list and say, look its slin16 16k |
| 17:43.53 | DanJenkins[m] | right? |
| 17:44.01 | seanbright | sure? |
| 17:44.05 | DanJenkins[m] | because there was no sdp setting up the session etc |
| 17:44.10 | DanJenkins[m] | haha |
| 17:44.17 | seanbright | that might work for stuff going out of asterisk but not coming in |
| 17:44.27 | seanbright | again, these are just my knee-jerk thoughts |
| 17:44.33 | DanJenkins[m] | oh i hadnt thought of that |
| 17:44.49 | seanbright | luckily this code is trivial and we can just trace right throu... |
| 17:44.52 | seanbright | wait |
| 17:44.55 | seanbright | heh |
| 17:45.02 | DanJenkins[m] | yeah 100% |
| 17:45.07 | DanJenkins[m] | you know who i blame..... |
| 17:45.16 | seanbright | mark spencer |
| 17:45.17 | seanbright | me too |
| 17:45.21 | DanJenkins[m] | that amazing guy gtjoseph who gave me just enough info to hang from :D |
| 17:45.28 | DanJenkins[m] | lol |
| 17:45.41 | seanbright | it is his fault we are all in this position |
| 17:46.10 | DanJenkins[m] | wait mark spencer or gtjoseph ? |
| 17:46.13 | gtjoseph | uhm who? |
| 17:46.18 | seanbright | mspencer |
| 17:47.19 | DanJenkins[m] | that is indeed fair. |
| 17:47.21 | DanJenkins[m] | the most important thing though.... did i get the components right in the jira? |
| 17:48.04 | gtjoseph | Close enough :) |
| 17:48.32 | DanJenkins[m] | omg |
| 17:48.37 | DanJenkins[m] | thats an achievement in itself |
| 17:48.49 | *** join/#asterisk-dev kharwell (uid358942@gateway/web/irccloud.com/x-sqbsahoagcewvmhv) |
| 17:48.49 | *** mode/#asterisk-dev [+o kharwell] by ChanServ |
| 17:50.20 | gtjoseph | DanJenkins[m]: I should be able to work my way through your scenario in the next day or so. |
| 17:51.34 | DanJenkins[m] | sweet, thanks gtjoseph |
| 17:51.36 | DanJenkins[m] | i hope ive given you everything you need in those two repos |
| 17:51.46 | DanJenkins[m] | if not, shout! |
| 17:52.16 | DanJenkins[m] | this one im less bothered by - because i can at least get it working with 8k audio |
| 17:52.38 | DanJenkins[m] | the non made issue because it'll be something my end..... the lack of audio from linux |
| 17:52.45 | DanJenkins[m] | although thats going to be an issue for you isnt it! |
| 17:53.04 | gtjoseph | heh |
| 17:53.09 | DanJenkins[m] | so you may have to fix my main issue in my code before you can figure out whats wrong with 16k :D |
| 17:53.16 | DanJenkins[m] | double win |
| 17:53.21 | seanbright | so... the slin formats have a smoother on them |
| 17:53.39 | seanbright | which should be doing the byte order stuff already |
| 17:54.03 | gtjoseph | goes off to have a liquid lunch |
| 17:54.10 | DanJenkins[m] | youre talking dutch to me seanbright ;) |
| 17:54.21 | seanbright | you said you swapped byte orders |
| 17:54.34 | DanJenkins[m] | you think i shouldnt have to do little to big endian |
| 17:54.35 | seanbright | the endians |
| 17:54.46 | seanbright | i'm thinking |
| 17:55.01 | DanJenkins[m] | right - I wouldnt have even known to do that but gtjoseph did that on the way out of asterisk in his original demo code |
| 17:55.12 | seanbright | interesting |
| 17:55.46 | DanJenkins[m] | google wants it in a particular way and i can only assume from george's code..... asterisk does the opposite |
| 17:55.46 | gtjoseph | yeah had to |
| 17:56.13 | DanJenkins[m] | so i can only assume that i have to swap it before sending it back |
| 17:56.18 | DanJenkins[m] | and i can get 8k slin16 working |
| 17:56.26 | DanJenkins[m] | so the swapping must work |
| 17:56.48 | seanbright | https://github.com/asterisk/asterisk/blob/16/res/res_rtp_asterisk.c#L7240-L7242 |
| 17:56.53 | DanJenkins[m] | it wouldnt suprise me if the endians had something to do with not getting any audio from linux |
| 17:57.22 | seanbright | so on the way in we do the swap |
| 17:57.31 | seanbright | on the way out, the smoother should be doing the swap |
| 17:57.32 | DanJenkins[m] | oohhh |
| 17:57.54 | seanbright | try not changing the endianess of the packets going _to_ asterisk |
| 17:57.55 | DanJenkins[m] | i seem to recall trying not swapping and it made no difference |
| 17:58.01 | seanbright | just as a silly guess |
| 17:58.03 | seanbright | ok |
| 17:58.08 | DanJenkins[m] | on 8k or 16? |
| 17:58.13 | seanbright | the one that doesn't work |
| 17:58.21 | DanJenkins[m] | no ill give it a go now :) |
| 17:59.14 | DanJenkins[m] | oh actually its not easy to do before i clock off |
| 17:59.23 | DanJenkins[m] | i just relised... i dont swap the endians in my basic test |
| 17:59.41 | DanJenkins[m] | because im using a sound file from Asterisk's lib |
| 17:59.47 | DanJenkins[m] | so i presumed that was "in the correct format" |
| 18:00.07 | seanbright | ok... maybe ast_frame_byteswap_be is broken |
| 18:00.07 | DanJenkins[m] | although saying that.... im swapping the endians on the wav file |
| 18:00.28 | DanJenkins[m] | it gives me a headache |
| 18:01.11 | seanbright | it's going to end up being a byte order issue |
| 18:01.17 | seanbright | i am officially putting $11 on it |
| 18:01.40 | seanbright | which i think is like £1 |
| 18:01.41 | DanJenkins[m] | a whole $11?! |
| 18:01.52 | gtjoseph | I'm betting that the source of the audio makes a difference. I tested with live audio transcoded by asterisk |
| 18:02.16 | gtjoseph | playing from a sound file takes a different path |
| 18:02.33 | seanbright | i always bet on gtjoseph, so whatever he says |
| 18:03.59 | gtjoseph | a fool and his money are soon parted |
| 18:04.27 | seanbright | i'm COUNTING ON IT |
| 18:04.29 | seanbright | no wait |
| 18:05.08 | DanJenkins[m] | Woot! |
| 18:05.17 | DanJenkins[m] | It'll 100% be my code other than the 118 thing |
| 18:05.22 | DanJenkins[m] | Maybe |
| 18:06.38 | seanbright | nice, closed as WONTFIX |
| 18:08.01 | DanJenkins[m] | Hahahaha |
| 18:10.20 | seanbright | you'd think being in the business of phones that it wouldn't drive me absolute bonkers to hear a phone ring |
| 18:10.30 | seanbright | yet here we are |
| 18:10.43 | file | seanbright: or... hear we are |
| 18:11.03 | seanbright | touchâé |
| 18:11.06 | seanbright | doh |
| 18:11.11 | seanbright | you know what i mean |
| 18:19.20 | gtjoseph | just for reference... signed linear over RTP is always network byte order (BE) and google expects LE hence the byte swap in my sample. |
| 18:20.37 | gtjoseph | goes back to lunch |
| 18:30.49 | DanJenkins[m] | ah ha! |
| 18:31.29 | seanbright | because L16 <> signed linear |
| 18:32.00 | seanbright | DanJenkins[m]: eureka? |
| 18:32.22 | DanJenkins[m] | which makes my wav slin16 8k audio read -> send even weirder |
| 18:32.34 | DanJenkins[m] | nope |
| 18:33.04 | DanJenkins[m] | because my actual slin16 16k from dialogflow doesnt play |
| 18:33.33 | DanJenkins[m] | my demo repo for not needing dialogflow might be a little skewiff |
| 18:33.40 | DanJenkins[m] | but the actual problem is still a problem |
| 18:34.26 | DanJenkins[m] | unless my mory is failing and i didnt actually try not swapping |
| 18:34.36 | DanJenkins[m] | but im 99% certain i did |
| 18:36.32 | DanJenkins[m] | but more curious... i want to know whats in gtjoseph 's liquid lunch |
| 18:36.55 | seanbright | well, he'll probably have to talk to his sponsor after lunch, so... |
| 18:37.44 | DanJenkins[m] | so i was thinking about submitting a talk about dana to astricon |
| 18:41.11 | seanbright | i approve |
| 18:41.15 | seanbright | i also have no say |
| 18:42.29 | DanJenkins[m] | would mean having to fix it ;) |
| 18:42.43 | seanbright | well... you're always there when you need us |
| 18:42.48 | seanbright | <3 |
| 18:43.06 | DanJenkins[m] | not sure what else to talk about other than maybe a detailed dialogflow one with external media |
| 18:43.35 | DanJenkins[m] | oh yioure too kind seanbright ;) |
| 18:55.13 | *** join/#asterisk-dev coreyfarrell (~coreyfarr@68-114-57-249.dhcp.davl.vt.charter.com) |
| 18:55.13 | *** mode/#asterisk-dev [+o coreyfarrell] by ChanServ |
| 19:04.04 | seanbright | DanJenkins[m]: just so i'm clear... |
| 19:04.19 | seanbright | i downloaded 16k.pcap and extracted the audio streams |
| 19:04.30 | seanbright | both should be slin16? |
| 19:04.48 | DanJenkins[m] | yes i think so |
| 19:04.59 | seanbright | ok, the allison side sounds great |
| 19:05.04 | DanJenkins[m] | this is where you say its not? |
| 19:05.07 | seanbright | the other sid... does not sound great |
| 19:05.10 | seanbright | side* |
| 19:05.44 | DanJenkins[m] | so the non allison side is the audio coming out of asterisk |
| 19:06.01 | seanbright | ok, then it must not be slin16 |
| 19:06.27 | DanJenkins[m] | but it is..... |
| 19:06.43 | DanJenkins[m] | or at least thats what i asked for |
| 19:06.44 | DanJenkins[m] | :D |
| 19:07.41 | DanJenkins[m] | and thats what i tell google it is... and google does speech to text on it fine |
| 19:07.50 | seanbright | interesting |
| 19:08.05 | DanJenkins[m] | and its interesting that you hear zombies fine |
| 19:08.17 | seanbright | you're swapping the byte order before sending to google though, yes? |
| 19:08.22 | DanJenkins[m] | because i hear nothing on my phone |
| 19:08.36 | DanJenkins[m] | yes |
| 19:08.42 | seanbright | ok, let me try to do that |
| 19:10.09 | seanbright | the smoother flags on the asterisk side should swap the byte order for you, so that part of the code must not be firing |
| 19:10.18 | seanbright | are you in a position to add some debug code and re-run your test? |
| 19:11.33 | DanJenkins[m] | not right now, about to put my son to bed but could after |
| 19:12.04 | seanbright | "should've just heard 16k zombies... but... ya didn't" |
| 19:16.34 | seanbright | https://pastebin.com/raw/BRKnNHBC |
| 19:17.23 | seanbright | if you never see "DJ wants a smoother" then asterisk is not creating a smoother and therefore not swapping byte order |
| 19:19.07 | DanJenkins[m] | ð |
| 19:19.21 | seanbright | my local branch name is dj-smooth |
| 19:19.32 | seanbright | which you are free to use on your business cards if you wish |
| 19:19.38 | DanJenkins[m] | Ofc |
| 19:21.59 | seanbright | but that is really a secondary issue. i'll go back to my previous guess, it doesn't like 118 because it's not static and there is not SDP. |
| 19:22.25 | seanbright | ... which is why i brought up a configurable payload code when external media was going in but no one listens to me |
| 19:23.18 | DanJenkins[m] | I mean all this wouldn't have been an issue if I could just write slin16 to a websocket, no headers needed ;) |
| 19:23.24 | DanJenkins[m] | Right? |
| 19:23.25 | DanJenkins[m] | Hahaha |
| 19:23.26 | DanJenkins[m] | I jest |
| 19:23.39 | seanbright | stop trying to make media over TCP happen |
| 19:24.23 | seanbright | if someone implemented such a thing, i don't see many people putting up a fight against it going in |
| 19:24.35 | seanbright | i just have 0 interest in working on it |
| 19:24.58 | DanJenkins[m] | Haha, I'll never stop going on about it |
| 19:25.05 | DanJenkins[m] | And that is your perogative ;) |
| 19:26.32 | seanbright | actually, sean bright industries will get it implemented for $15k |
| 19:26.39 | seanbright | reach out to my people |
| 19:27.04 | file | who let you have people? |
| 19:27.30 | DanJenkins[m] | sounds fair |
| 20:18.22 | *** join/#asterisk-dev electronic_eel (~quassel@HSI-KBW-46-223-65-185.hsi.kabel-badenwuerttemberg.de) |
| 20:18.45 | DanJenkins[m] | so seanbright |
| 20:18.56 | DanJenkins[m] | with how thngs were I get "[2020-02-20 20:17:37] ERROR[22401]: res_rtp_asterisk.c:5028 ast_rtp_write: DJ see smoother run!" |
| 20:21.04 | DanJenkins[m] | if i swap endianness I also see "DJ see smoother run!" |
| 20:21.06 | DanJenkins[m] | neither give me audio |
| 20:28.40 | kharwell | who thinks it would be a lost cause to start const'ing things |
| 20:29.15 | file | try, try, try? |
| 20:49.51 | seanbright | maybe not a lost cause, but other than warm and fuzzies i don't think it buys you much |
| 20:50.29 | seanbright | i guess the compiler could get more aggressive with optimizations? i dunno |
| 20:51.44 | seanbright | DanJenkins[m]: interesting |
| 20:52.32 | DanJenkins[m] | so it could well be that i need to swap the endianness etc but it doesnt understand 118 |
| 20:52.39 | DanJenkins[m] | shrug |
| 20:53.12 | seanbright | i think whether it is right or wrong to swap the endianess, you should still hear _something_, even if is just noise/garbage |
| 20:53.34 | seanbright | and my "doesn't understand the payload code" theory is just that, a theory |
| 20:53.43 | seanbright | gtjoseph is going to get to the bottom of this if it is the last thing that he does |
| 20:54.14 | DanJenkins[m] | i shall comment |
| 20:54.17 | DanJenkins[m] | on the ticket |
| 21:01.26 | DanJenkins[m] | thanks so much for checking it out seanbright ! |
| 21:01.30 | seanbright | sure |
| 21:02.06 | DanJenkins[m] | and confirming the audio im sending back is indeed "good" |
| 21:02.26 | seanbright | well, that is arguable |
| 21:02.41 | DanJenkins[m] | you should see a recording of what dialogflow actually sends bck in their 8k version |
| 21:02.42 | seanbright | when you get asterisk to play it and it sounds correct, then it is "good" |
| 21:03.12 | DanJenkins[m] | if you just play it back its full of static |
| 21:03.19 | DanJenkins[m] | but if you just play back one of the 2 channels.... its fine |
| 21:03.26 | DanJenkins[m] | i just send it to asterisk and its fine |
| 21:03.48 | DanJenkins[m] | (hence why im using "known" good audio from asterisk's own audio files) |
| 21:04.33 | seanbright | it's just a confusing set of circumstances because i find the notion of 'signed linear' be me a nebulous thing |
| 21:04.40 | seanbright | to be* |
| 21:05.11 | DanJenkins[m] | ya |
| 21:05.22 | seanbright | either way, i'm done talking about it. i look forward to gtjoseph's diagnosis. |
| 21:05.34 | DanJenkins[m] | if i understood enough about how to remove an ogg container from opus I'd have tried that |
| 21:05.38 | DanJenkins[m] | but alas im an idiot |
| 21:07.10 | seanbright | don't be so hard on yourself |
| 21:07.26 | seanbright | you can't compare yourself to people like... oh i dunno... me |
| 21:07.34 | seanbright | i'm a pretty special person, dj. |
| 21:08.12 | kharwell | seanbright: I agree, and maybe constness might get us some optimization. I mostly like it though as a way to keep the potential side effects down on an object. Or modifying said object outside its scope |
| 21:08.25 | seanbright | just think about how dumb an average person is and then remember that half of people are dumber than that |
| 21:08.59 | DanJenkins[m] | but youre my idol seanbright |
| 21:09.04 | seanbright | bro i know |
| 21:09.42 | seanbright | kharwell: i'm all about immutability |
| 21:09.56 | kharwell | exactly |
| 21:10.01 | seanbright | so rather than modifying a member of a struct, you duplicate the entire struct and return it |
| 21:10.05 | seanbright | heh |
| 21:10.15 | kharwell | well there are some drawbacks I suppose |
| 21:10.46 | kharwell | if only the compiler implemented copy on write |
| 21:11.08 | seanbright | i started looking at ast_channel a while ago and was like, ok, let's find some members that never change and then we won't have to lock around them. oh! the channel name doesn't change, so that's a good candidate! oh wait, you can change it. ok, well i give up. |
| 21:12.07 | seanbright | well, then you have to think about deep copying, because everything has a pointer to something else |
| 21:12.10 | seanbright | gets very hairy |
| 21:12.22 | kharwell | yeah ao2'ing everything makes it hard too. The ref count is kinda embedded in the object now so if passing around and inc/dec'ing the ref then blargh |
| 21:15.14 | kharwell | I guess it's a problem that will remain...constant even :-P |
| 21:22.47 | seanbright | i see what you did there |
| 22:08.34 | *** join/#asterisk-dev spatel (~spatel@pool-96-237-230-175.bstnma.fios.verizon.net) |
| 23:32.51 | *** join/#asterisk-dev spatel (~spatel@pool-96-237-230-175.bstnma.fios.verizon.net) |