| rgareus | ColaEuphoria: a host-provided one | 00:00 |
|---|---|---|
| ColaEuphoria | ah | 00:00 |
| rgareus | ie control-ports and properties (atom patch:get/set) | 00:00 |
| ColaEuphoria | it's really a double edged sword | 00:00 |
| ColaEuphoria | doesn't work too well for plugins with matrix-like controls, e.g., fm synthesis | 00:01 |
| *** edogawa has quit IRC | 01:20 | |
| *** oofus has quit IRC | 05:48 | |
| jmagnusj | Is there any way of printing an atom type for debugging? | 07:33 |
| drobilla | urid unmap | 07:35 |
| jmagnusj | ah yes, thanks | 07:36 |
| *** edogawa has joined #lv2 | 07:46 | |
| jmagnusj | hmm, jalv doesn't work for me, it tries to start a jack daemon and the daemon fails. This happens both if I have qjackctl running and not | 07:50 |
| *** oofus_lt has joined #lv2 | 07:51 | |
| drobilla | PEBKAC. Jalv is just a jack client like any other | 07:51 |
| drobilla | e.g. JACK_DEFAULT_SERVER is set, default jack server isn't actually running, etc | 07:53 |
| jmagnusj | JACK_DEFAULT_SERVER is not set | 07:53 |
| jmagnusj | indeed, PEBKAC, was running it as the wrong user | 07:54 |
| jmagnusj | thanks | 07:54 |
| * jmagnusj goes to bed | 07:54 | |
| *** JaVelDa has left #lv2 | 08:41 | |
| *** JaVelDa has joined #lv2 | 08:42 | |
| *** Yruama_Lairba has joined #lv2 | 11:29 | |
| *** GK_1wm___SU has joined #lv2 | 13:26 | |
| *** GK_1wm___SU has left #lv2 | 13:26 | |
| *** sigma6 has quit IRC | 14:40 | |
| jmagnusj | x42-plugins causes lv2bench to crash | 14:43 |
| rgareus | atom ports, I think | 14:44 |
| jmagnusj | crash as in segfault | 14:44 |
| rgareus | last I looked lv2bench barfs for all plugins with atom i/o. | 14:44 |
| rgareus | yep | 14:44 |
| rgareus | my current workaround is LV2PATH=/subset/to/check lv2bench | 14:45 |
| rgareus | LV2_PATH (not LV2PATH) | 14:45 |
| rgareus | also segfaults with the eg-plugins from lv2plug.in | 14:46 |
| rgareus | and fomp.lv2 ... | 14:47 |
| jmagnusj | :) | 14:48 |
| rgareus | in case of x42-plugins it's probably one of the midi-filters or midigen.lv2 | 14:48 |
| rgareus | IIRC the atom output buffer is not set by lv2bench | 14:50 |
| rgareus | so any plugin writing something there -> crash | 14:50 |
| *** sigma6 has joined #lv2 | 14:50 | |
| rgareus | yep, http://pastebin.com/gd73Cf03 (and yes, the plugin does check if the atom-port buffer is not NULL, and check the capacity) | 14:53 |
| *** deva has joined #lv2 | 15:02 | |
| rgareus | drobilla: lv2bench use a single LV2_Atom_Sequence seq for _all_ atom ports, and it's not reset between runs, either. | 15:04 |
| jmagnusj | ardour has an std:exception if I add a patch to my ttl, the only thing that gets called in my plugin is instantiate | 15:10 |
| rgareus | drobilla, jmagnusj: proposed fix http://paste.debian.net/918383/ | 15:11 |
| rgareus | jmagnusj: can you get a backtrace? | 15:13 |
| jmagnusj | yes, 1 sec | 15:13 |
| rgareus | best guess, some name is missing or duplicate | 15:13 |
| jmagnusj | rgareus: that fix looks fine to me, with my superficial understanding of atom sequences | 15:15 |
| jmagnusj | http://pastebin.com/cg3sRjWr | 15:17 |
| rgareus | arguably lv2bench should inject some noise. into audio-input ports, and maybe even modulate or ramp control-ports | 15:17 |
| jmagnusj | Does look like some null somewhere | 15:17 |
| jmagnusj | "what: basic_string::_M_construct null not valid" | 15:17 |
| rgareus | some c-string char* is NULL -> C++ std::string -> error | 15:19 |
| rgareus | odd though, ardour is not directly involved here. | 15:19 |
| jmagnusj | I guess ardour passes null on to gtk | 15:19 |
| jmagnusj | and it crashes later | 15:19 |
| jmagnusj | patch:writable [ | 15:20 |
| jmagnusj | a lv2:Parameter ; | 15:20 |
| jmagnusj | rdfs:label "Scala .scl file" ; | 15:20 |
| jmagnusj | rdfs:range atom:Path ; | 15:20 |
| jmagnusj | ] ; | 15:20 |
| jmagnusj | I guess I need to pass a named entity, not [ ... ] | 15:22 |
| jmagnusj | Yes!! that was it. | 15:24 |
| jmagnusj | Pretty stupid mistake, I guess my mental model of RDF isn't quite right yet | 15:25 |
| drobilla | Yeeeeeeeah, lv2bench could use some work | 15:50 |
| drobilla | That shouldn't crash (of course, nothing should crash), but it's a pointless declaration anyway | 15:51 |
| drobilla | "I have a writable parameter with no name and therefore you can't possibly write to it" | 15:51 |
| drobilla | rgareus: Seems reasonable | 15:53 |
| * drobilla wonders if there's a way to say "these things must have URIs" | 15:55 | |
| drobilla | Probably not | 15:56 |
| rgareus | drobilla: can you apply that patch from here or should I file a ticket on your tracker? | 16:10 |
| drobilla | rgareus: No worries, I can apply it | 16:23 |
| *** trebmuh has joined #lv2 | 16:25 | |
| rgareus | drobilla: resetting the seq_in.atom.size and seq_in.atom.type should not be needed, actually. | 16:41 |
| rgareus | and it'll probably still fail for plugin with multiple atom-outputs (some lsp-plugins) | 16:42 |
| rgareus | those would forge a sequence head on both atom-ports and the forge may get messed up if writes are interleaved. | 16:42 |
| drobilla | Probably, but I think just garbage output sort of messed up | 16:44 |
| * drobilla fixes a bunch of stuff and ends up at Ladspa_CS_chorus1::runproc again | 16:46 | |
| *** sigma6 has quit IRC | 17:16 | |
| *** dsheeler has quit IRC | 17:35 | |
| *** oofus_lt has quit IRC | 17:41 | |
| *** rncbc has joined #lv2 | 18:04 | |
| *** badosu has joined #lv2 | 18:45 | |
| *** oofus has joined #lv2 | 19:06 | |
| *** deva has quit IRC | 19:39 | |
| *** NickSB has quit IRC | 19:41 | |
| *** NickSB has joined #lv2 | 19:44 | |
| *** NickSB has quit IRC | 20:59 | |
| *** edogawa has quit IRC | 22:23 | |
| *** NickSB has joined #lv2 | 23:04 | |
| *** jmagnusj has quit IRC | 23:43 | |
Generated by irclog2html.py 2.13.0 by Marius Gedminas - find it at mg.pov.lt!