Atte Andr=E9 Jensen writes:
> * The installation should be documented. A simple *not autogenerated*
I agree. I just want to concentrate on code until it is more complete. I
accept any help, feedback, documentaiton, code. Your feedback in
particular is very valuable.
> * For some reason zyn complains that I haven't set LV2_PATH (although
Current Makefile hackery installs the plugin into LV2_PATH. Have you
"export"-ed the LV2_PATH variable?
I plan to switch zyn and zynjacku projects to autotools or scons in
> * I installed the latest phat + pyphat from source, but (after
Yes, this is best to do (I guess you mean SVN). I should talk with Loki
to release newer version of PHAT. This will help jack_mixer too. Loki
> I get this:
Recently someone in IRC (maybe you) had problem with those constants
failing to autogenerate. This causes window without controls. So I added
check to detect this at start. However the guy left and I still don't
know what causes it. Can you please run this command from within
./gen_py_constants.py `pkg-config --variable=3Dincludedir lv2dynparamhost-1=
Do you get any output?
gen_py_constants.py is executed several times, with different regex,
From=20Makefile. Result is collected in init_py_constants.c
I guess that file is empty for some strange reason?
Also, once I got report that gen_py_constants.py is not executable. That
was strage to me because AFAIK subversion client should put +x to the
file when it is checked out. Please check if gen_py_constants.py is
executable and if not, make it executable. And tell me :)
> A few questions about await me once the thingy is running:
ATM, no. This feature is in my plan however, as "import" in zynjacku. Of
course it will load only adsynth-zynadd settings until other synths are
ported too. But for this to happen I need to implement project save/load
in zynjacku. With LASH support. And maybe publishing ready-for-use
patches/presets for zynjacku, including ones converted from original
zynaddsubfx patches, is good idea.
> * Are the realtime problems of zynaddsubfx still there?
I continue to dig the code and it is still work in progress. Howver, you
should feel the progress I've made right now. I particular zynadd has
no lock (mutex) at all. I removed lots of memory allocations in audio
processing context, but this is not complete yet. The results I get
here are quite amazing, I never got any "clicks" or JACK
crashes. They are caused by sleeping in realtime context. However JACK
crashing because of an app seems like bug in JACK itself to me, but I
may be wrong.