| *** edogawa has quit IRC | 00:02 | |
| *** tytel has joined #lv2 | 02:06 | |
| *** uncle-j_j has quit IRC | 02:32 | |
| *** Magnus_RM has quit IRC | 03:19 | |
| *** Bernhard_L has joined #lv2 | 06:41 | |
| Bernhard_L | Hello, looking for a scatter plugin. Tried granular scatter processor, but don't get any result out of it. | 06:42 |
|---|---|---|
| *** tytel has quit IRC | 07:59 | |
| *** edogawa has joined #lv2 | 08:19 | |
| *** tytel has joined #lv2 | 08:32 | |
| *** tytel has quit IRC | 08:37 | |
| *** falktx has joined #lv2 | 08:49 | |
| *** tytel has joined #lv2 | 09:27 | |
| *** tytel has quit IRC | 09:31 | |
| *** uncle-j_j has joined #lv2 | 09:39 | |
| *** tytel has joined #lv2 | 10:21 | |
| *** tytel has quit IRC | 10:25 | |
| *** uncle-j_j has quit IRC | 10:36 | |
| *** uncle-j_j has joined #lv2 | 10:38 | |
| *** rncbc has joined #lv2 | 10:40 | |
| *** Mark__ has joined #lv2 | 10:51 | |
| *** Mark__ has left #lv2 | 10:51 | |
| *** uncle-j_j has quit IRC | 10:53 | |
| *** tytel has joined #lv2 | 11:15 | |
| *** tytel has quit IRC | 11:19 | |
| *** rgareus|afk is now known as rgareus | 11:24 | |
| rgareus | drobilla: nice. | 11:25 |
| rgareus | drobilla: still fails when double-clicking (not embedd) (_instance is NULL) | 11:25 |
| rgareus | drobilla: http://paste.debian.net/317996/ | 11:25 |
| rgareus | mmh there still seems to be a race-condition. Sisco.lv2 (x11_in_gtk2). embedd/close/embedd-again/... the UI signal sent during instantiate does not always make it though. | 11:31 |
| *** Bernhard_L has quit IRC | 11:37 | |
| rgareus | ssh in sisco's case. the issue is now the other way 'round. the first message(s) from DSP to UI may not be delivered | 11:42 |
| rgareus | UI to DSP works now properly. but the other way round not yet. in this case: GUI settings are not always restored. | 11:43 |
| rgareus | s/restored/initialized/ | 11:44 |
| *** tytel has joined #lv2 | 11:59 | |
| *** tytel has quit IRC | 12:04 | |
| *** rncbc has quit IRC | 12:31 | |
| *** uncle-j_j has joined #lv2 | 12:45 | |
| rgareus | drobilla: found it. there can be N DSP run() before PortImpl::monitor() | 14:17 |
| rgareus | drobilla: I added a printf() to the plugin's run() and a printf() to ingen's PortImpl::monitor() | 14:17 |
| rgareus | I get 2 runs for one PortImpl::monitor() | 14:17 |
| drobilla | Hm. Yes, I guess the explicitly monitored case needs to work differently | 14:18 |
| rgareus | atom-messages send in the first run() just go nowhere. the 2nd run() overwrites the buffer | 14:18 |
| drobilla | (The usual is e.g. audio/activity monitoring where the goal is to not go insane with traffic) | 14:18 |
| rgareus | drobilla: it's odd. PortImpl::monitor() is supposed to be called in RT-context, isn't it? | 14:19 |
| rgareus | from LV2Block::post_process() | 14:19 |
| drobilla | rgareus: yes | 14:21 |
| drobilla | ... Though that parameter isn't ProcessContext& for some reason I can't remember | 14:22 |
| rgareus | hang on. LV2Block::post_process() is called every cycle. but context.notify() in PortImpl::monitor() is not | 14:25 |
| drobilla | Correct | 14:26 |
| drobilla | At least in general. The _monitored check in the early return there is supposed to make it happen every time for explicitly monitored ports | 14:27 |
| rgareus | I get a regular pattern | 14:28 |
| rgareus | run(); LV2Block::post_process(); run(); LV2Block::post_process() -> context.notify(); // repeat | 14:28 |
| rgareus | give me a min. I'll check where in in PortImpl::monitor() it stops every 2nd time | 14:29 |
| *** sigma6 has joined #lv2 | 14:31 | |
| rgareus | right at the top !context.must_notify(this) | 14:31 |
| rgareus | if I comment that out in PortImpl::monitor() all works well | 14:32 |
| rgareus | scratch that. | 14:37 |
| rgareus | that does not make a difference. | 14:37 |
| rgareus | drobilla: ok. now the definitive answer: time_to_send is false every 2 time | 14:43 |
| drobilla | rgareus: Yes, but _monitored should be true | 14:47 |
| rgareus | ingen/src/server/PortImpl.cpp:429 if (!time_to_send && (!is_sequence || _monitored || buffer(0)->value())) { return ; } | 14:47 |
| rgareus | drobilla: what's the logic here? | 14:47 |
| rgareus | I think that should be changed: monitored sequences should ignore time_to_send. | 14:48 |
| rgareus | if (!time_to_send && !(is_sequence && _monitored) && (!is_sequence && buffer(0)->value())) | 14:51 |
| rgareus | ^^ seems right to me | 14:51 |
| drobilla | Mmm, yes, seems odd | 14:52 |
| drobilla | Seems reasonable | 14:52 |
| rgareus | ie always notify the GUI if it's a monitored sequence. but buffer-value updates depend on the time | 14:52 |
| rgareus | it seems that there's some issue with _frames_since_monitor | 14:54 |
| rgareus | but still atom-sequences, should always be forwarded regardless. | 14:55 |
| drobilla | Shouldn't matter in that case, it's only used for value things | 14:56 |
| drobilla | The monitored sequence branch just calls notify in the loop directly | 14:57 |
| rgareus | _monitored is always true. and time_to_send is false every 2nd time (here 1024fpp, jack dummy) | 15:06 |
| rgareus | monitor rate is 25Hz, so 1920 frames. (@48KSPS) | 15:07 |
| rgareus | so that's good too | 15:07 |
| rgareus | drobilla: http://paste.debian.net/318028/ fixes all my issues | 15:14 |
| rgareus | drobilla: can you commit that? | 15:16 |
| drobilla | rgareus: Sure. Thanks | 15:25 |
| *** edogawa has quit IRC | 15:30 | |
| *** deva has joined #lv2 | 17:14 | |
| *** rgareus is now known as rgareus|afk | 17:49 | |
| drobilla | rgareus|afk: http://dev.drobilla.net/changeset/5786 | 18:17 |
| *** rncbc has joined #lv2 | 18:21 | |
| *** NickSB2 has quit IRC | 19:18 | |
| *** uncle-j_j has quit IRC | 19:38 | |
| *** deva has quit IRC | 19:40 | |
| *** drobilla` has joined #lv2 | 19:48 | |
| *** drobilla has quit IRC | 19:51 | |
| *** drobilla` has quit IRC | 20:30 | |
| *** uncle-j_j has joined #lv2 | 20:44 | |
| *** uncle-j_j has quit IRC | 20:49 | |
| *** uncle-j_j has joined #lv2 | 20:54 | |
| *** uncle-j_j has quit IRC | 21:34 | |
| *** aombk has joined #lv2 | 21:51 | |
| *** tytel has joined #lv2 | 21:57 | |
| *** tytel has quit IRC | 22:07 | |
| *** tytel has joined #lv2 | 22:09 | |
| *** edogawa has joined #lv2 | 22:11 | |
| *** tytel has quit IRC | 22:37 | |
| *** falktx has quit IRC | 22:58 | |
| *** rncbc has quit IRC | 23:03 | |
| *** edogawa has quit IRC | 23:36 | |
Generated by irclog2html.py 2.13.0 by Marius Gedminas - find it at mg.pov.lt!