dbtx

That's unexpected and interesting. Of course Yoshimi has nothing to do with GTK but XFCE does. In your coredumps you can see the only mentions of Yoshimi are in what look exactly like window titles including Firefox viewing the wiki. I also see some GL shader source which gives me an idea-- are you using XFCE's desktop compositor? If yes, does it help to switch it off?

It was X which dumped core, not Yoshimi. I thought Yoshimi never did anything at all directly to an X server (that's crazy!) -- so my first guess is that arch's X or FLTK package has a problem and Yoshimi is just the one finding it. OP, there aren't many options but can you try messing around with something else like Rakarrack? What version of FLTK is installed? Are you able to try with older versions of that and (separately) another version of Yoshimi?

Are you trying to use ALSA directly? In any case, do you have a USB audio adapter? I ask because the last time I tried to run a synth on a Pi3B+, the builtin audio device was just about useless.

Hello, the Pi family's CPUs don't have SSE instructions, so that build of GCC naturally doesn't even have the options to indicate they should be used. You can try to remove those hardcoded optimization flags from bristol/Makefile.{in,am} possibly enabling auto-detection of the current CPU with -march=native. Here's a patch that does that, and I hope that GCC already complained about all the things it didn't like. Those flags also appear in libbristol/Makefile.{in,am} but the lines appear to be commented...

Reposition FLTK_LIBRARIES, unbreak linker command line (we hope)

Merge branch 'master' of github.com:Yoshimi/yoshimi into bugfix

Initialize and freshen MasterUI->activePart

Prelim. attempt at tunneling through EffUI for updates to dynfilter subwindow