| *** trebmuh has quit IRC | 00:39 | |
| *** Yruama_Lairba has quit IRC | 02:30 | |
| *** ColaEuphoria_ has quit IRC | 05:16 | |
| *** ColaEuphoria_ has joined #lv2 | 05:31 | |
| *** grejppi has quit IRC | 05:55 | |
| *** grejppi has joined #lv2 | 06:31 | |
| *** trebmuh has joined #lv2 | 09:09 | |
| *** sigma6 has joined #lv2 | 09:55 | |
| *** Yruama_Lairba has joined #lv2 | 10:44 | |
| *** edogawa has joined #lv2 | 11:50 | |
| *** SianaGearz has quit IRC | 11:56 | |
| *** SianaGearz has joined #lv2 | 11:56 | |
| *** SianaGearz has joined #lv2 | 11:56 | |
| *** ricardocrudo has joined #lv2 | 12:28 | |
| ventosus | rgareus: does ardour still run 'state_restore' and 'run' concurrently? | 12:48 |
|---|---|---|
| ventosus | I'm investigating a race condition in my plug in 'state_restore' | 12:49 |
| *** son0p has joined #lv2 | 12:50 | |
| falktx | something was added to lv2 because of this | 12:57 |
| drobilla | Yeah, state:threadSafeRestore is the proper way | 13:22 |
| drobilla | I don't remember if the normal case was fixed in ardour | 13:22 |
| ventosus | seems not, and it does not explicitely check for the feature either | 13:27 |
| rgareus | drobilla: I thought you fixed it when adding state:threadSafeRestore | 13:27 |
| rgareus | ventosus: loading presets pushes a different worker-thread (state_worker) | 13:29 |
| drobilla | rgareus: I don't think I figured out the "stop processing while this happens" bit | 13:32 |
| rgareus | ventosus: which plugin(s) are affected by that, anyway? | 13:33 |
| drobilla | Also github search apparently doesn't search commit messages, which is impressively useless | 13:35 |
| drobilla | A smart handle would be nice, RAII "if you have this, the plugin isn't being run" | 13:37 |
| drobilla | Waiting for the first/next cycle to arrive, but also not blocking forever if it never comes, is the annoying part | 13:37 |
| ventosus | rgareus: moony. I can easily switch to state:threadSafeRestore, though, no big deal | 13:40 |
| rgareus | drobilla: smells like TRY_LOCK to me. | 13:41 |
| rgareus | drobilla: is that actually in the spec, that run() and state save/restore must not be called concurrently? | 13:43 |
| rgareus | it's indirectly mentioned in http://lv2plug.in/ns/ext/state/state.html#threadSafeRestore | 13:44 |
| rgareus | http://lv2plug.in/doc/html/group__state.html#ac6d423f49c55a6df24ef687450d3cf1a has it. good | 13:46 |
| rgareus | drobilla: RAII will be more work. at the end of LV2Plugin::set_state() there's a latency_compute_run() which needs to run. | 13:48 |
| drobilla | rgareus: I guess that will do | 14:00 |
| drobilla | rgareus: Not really more difficult in that respect, if (locked) vs if (should_run_flag). Acquiring the mutex is much simpler though | 14:02 |
| *** ricardocrudo has quit IRC | 14:22 | |
| ssj71 | reMID.lv2 does a bunch of allocation in the state restore funciton fwiw | 15:46 |
| rgareus | that should be fine. state load/save is not meant to be rt-safe, is it? | 15:47 |
| rgareus | it's instantiation threading class. | 15:48 |
| ssj71 | oh duh, thread safe and rt safe are independent | 15:49 |
| ssj71 | my bad | 15:50 |
| rgareus | well, a lot of plugins don't care if state save/restore is called concurrently with run(). but for those that do care it's usually fatal. | 15:50 |
| drobilla | For some value of "don't care" anyway. Almost all can cause an audible disaster | 15:53 |
| drobilla | If not a crashy one | 15:53 |
| rgareus | if it's only GUI state.. no. | 15:57 |
| *** ssj71 has quit IRC | 16:35 | |
| drobilla | Well, sure | 16:40 |
| *** ssj71 has joined #lv2 | 16:40 | |
| *** son0p has quit IRC | 16:52 | |
| *** sigma6 has quit IRC | 17:18 | |
| *** deva has joined #lv2 | 18:51 | |
| *** ricardocrudo has joined #lv2 | 19:26 | |
| *** oofus has quit IRC | 20:03 | |
| *** oofus has joined #lv2 | 20:08 | |
| *** HarryHaaren has joined #lv2 | 20:30 | |
| HarryHaaren | drobilla: ping, you have an interesting nanogui repo fork - I'm trying to compile but its failing to link - any advice? | 20:31 |
| *** ricardocrudo has quit IRC | 20:36 | |
| drobilla | HarryHaaren: link harder? :P | 20:38 |
| drobilla | (gonna need more information than that) | 20:38 |
| HarryHaaren | drobilla: heh good point. http://pastebin.com/vMnq1zVQ that's already build the library, just linking the examples | 20:39 |
| HarryHaaren | VERBOSE=1 output http://pastebin.com/wLKvqED9 | 20:39 |
| HarryHaaren | and cmake .. output http://pastebin.com/QAWaJpzv | 20:41 |
| drobilla | HarryHaaren: You appear to have no pugl lib, or an out of date one, or some such | 20:42 |
| HarryHaaren | yep, but its the latest git cloned, and installed in /usr | 20:42 |
| drobilla | HarryHaaren: Old one in /usr/local? | 20:43 |
| drobilla | ... hm, though I get a compile error with most recent | 20:46 |
| HarryHaaren | drobilla: yeah I "hacked" around that puglGetModifier one | 20:46 |
| HarryHaaren | but then hit the linking error which I haven't been able to solve yet.. | 20:47 |
| HarryHaaren | no old versions as far as I can see here - nuked all of them and reinstalled but no joy | 20:47 |
| * HarryHaaren tries clean cmake .. && recompile | 20:47 | |
| HarryHaaren | it is *finding* libpugl-0, because if i manually change the command it fails differently | 20:49 |
| drobilla | I guess your pugl build could be botched somehow and not exporting symbols.. | 20:49 |
| drobilla | I'd exhaustively look around for libpugls and objdump them for said symbols etc | 20:50 |
| HarryHaaren | good plan, i'll get to it | 20:50 |
| drobilla | Fixed the compile issue | 20:51 |
| drobilla | For some value of "fixed" that means it compiles at least | 20:51 |
| HarryHaaren | i'll do a pull | 20:54 |
| HarryHaaren | drobilla: did you push? | 20:54 |
| drobilla | mmhmm | 20:54 |
| HarryHaaren | do you use Github or your own Git? I don't see the commit on either.. | 20:55 |
| *** deva has quit IRC | 20:57 | |
| * HarryHaaren pulls commit - took a while to show up on Github. I'm a bit impatient that way :) | 21:02 | |
| HarryHaaren | strange - only the examples fail to link | 21:04 |
| HarryHaaren | the python + nanogui.so link fine, and ldd shows them to be fine too | 21:04 |
| JaVelDa | http://www.premierguitar.com/articles/24993 Nice article, but links to non informative subscribe-page about MOD | 21:31 |
| ssj71 | interesting. IMO it would have been a good idea to launch the new website before a press release | 21:33 |
| ssj71 | but what do I know | 21:33 |
| JaVelDa | falktx ; there must be more than this by now. | 21:41 |
| ssj71 | JaVelDa: forums.moddevices.com | 21:42 |
| *** oofus has quit IRC | 21:42 | |
| ssj71 | oops, forum.moddevices.com | 21:42 |
| *** HarryHaaren has quit IRC | 21:42 | |
| *** oofus has joined #lv2 | 21:45 | |
| JaVelDa | ssj71 : That more like it, but http://moddevices.com/ should also link to this. | 21:46 |
| ssj71 | well, were it my company, I wouldn't link to it, but I sure wouldn't have just a subscribe box either | 21:48 |
| ssj71 | they had a decent front-page before | 21:48 |
| ssj71 | JaVelDa: there's #moddevices too, if you want to talk to more of the mod team | 21:49 |
| JaVelDa | ssj71 I think falktx is on the mod team so all I wanted to do was to warn him of the "dead link" in the fresh new article I found outside Linux/Foss usual places. | 21:57 |
| JaVelDa | In best wishes for them | 21:58 |
| rgareus | yeah it's an odd choice to remove all links from the mod front-page | 22:03 |
| *** edogawa has quit IRC | 23:00 | |
| *** JaVelDa has left #lv2 | 23:30 | |
| *** JaVelDa has joined #lv2 | 23:31 | |
| *** grejppi has quit IRC | 23:31 | |
| *** grejppi has joined #lv2 | 23:46 | |
Generated by irclog2html.py 2.13.0 by Marius Gedminas - find it at mg.pov.lt!