Tuesday, 2015-06-23

ssj71I figured since half of my object I was trying to pass came out as garbage00:02
rgareusardour/jack uses   4 *  jack_port_type_get_buffer_size (, JACK_DEFAULT_MIDI_TYPE);    for the worker buffer00:02
rgareus on can't set a minimumSize for the worker-thread buffer. I think falkTX mixed it up with Atom Ports.00:04
rgareusssj71: you can mutex in the worker thread if needed.  it's not realtime and can block.00:05
rgareusbut a mutex is not needed. there is only one worker thread00:05
ssj71ok. I wasn't sure.00:06
ssj71thanks00:06
rgareusplugin_run() {    if (self->should_do_some_work  && !self->already_queued_work) { self->already_queued_work = true; ... schedule_work(); }00:07
rgareusyou could catch    if (self->should_do_some_work  && self->already_queued_work) {  .. remember last work that should be done and do it later .. }00:07
rgareuswork_response   is always called _after_  run() once it's ready   (and before the next run)00:09
ssj71worker in this case only has 1 task. It only loads files and the response only assigns the returned data to the plugin00:09
rgareusand work_response needs to be RT-safe00:10
ssj71yes I'm aware of that00:10
rgareusok. please take it as confirmation of existing knowlege then :)00:11
ssj71I think if a second call to work() happened before the work_response() completed I could have threading issues though00:11
ssj71so I should probably mutex00:11
ssj71its unlikely but may as well be safe00:12
rgareusthe work will be done in correct order (there is only one worker thread).00:22
rgareusbut you can simply not call work() if you know that you already scheduled work and am waiting for it to complete.  a simple flag (see above) will do the trick.00:23
*** NickSB2 has joined #lv201:26
*** aombk has joined #lv202:47
*** aombk2 has quit IRC02:47
*** NickSB2 has quit IRC03:51
*** edogawa has joined #lv206:11
*** sigma6 has joined #lv208:27
*** ricardocrudo has joined #lv208:31
*** NickSB2 has joined #lv208:37
*** yann-kaelig has joined #lv208:50
*** falktx has joined #lv208:51
*** ricardocrudo has quit IRC09:09
*** sigma6 has quit IRC09:15
*** sigma6 has joined #lv209:15
*** ricardocrudo has joined #lv209:24
*** NickSB2 has quit IRC11:12
*** NickSB2 has joined #lv211:18
*** NickSB2 has quit IRC13:14
*** rncbc has joined #lv215:35
*** ricardocrudo has quit IRC16:20
*** ricardocrudo has joined #lv216:34
*** ricardocrudo has quit IRC16:38
*** ricardocrudo has joined #lv216:50
*** sigma6 has quit IRC16:56
*** ricardocrudo has quit IRC17:17
*** rncbc has quit IRC17:30
*** ricardocrudo has joined #lv217:31
*** ricardocrudo has quit IRC17:57
*** artfwo has joined #lv218:38
falktxthe more I post in the mailing lists, the more people ask things back18:38
falktxerr18:48
rgareuslol18:56
rgareuswell, so far I think it was a good discussion. a bit repetative but sti clearing thins up18:57
*** flexus has joined #lv219:07
ssj71I've been thinking about osc control and the arguments are all about the same.19:39
*** ricardocrudo has joined #lv219:48
*** Haskellfant has left #lv220:07
*** NickSB2 has joined #lv220:12
*** ricardocrudo_ has joined #lv220:19
*** ricardocrudo has quit IRC20:22
*** EntropyS1nk has quit IRC20:34
*** flexus has quit IRC20:50
*** ricardocrudo__ has joined #lv221:02
*** EntropySink has joined #lv221:04
*** ricardocrudo has joined #lv221:05
*** ricardocrudo_ has quit IRC21:06
*** falktx has quit IRC21:07
*** ricardocrudo__ has quit IRC21:08
*** ricardocrudo has quit IRC21:10
*** edogawa has quit IRC21:30
*** falktx__ has joined #lv221:44
*** Anchakor has joined #lv222:25
*** falktx__ has quit IRC23:28
*** NickSB2 has quit IRC23:32

Generated by irclog2html.py 2.13.0 by Marius Gedminas - find it at mg.pov.lt!