Created attachment 41583 [details] zynaddsubfx-2.1.1.ebuild
Created attachment 41584 [details] zynaddsubfx-2.1.1.patch
i did my own 2.1.1 ebuild and patches independently, currently trying to iron out the Bug #65752 problems.
So why is this an issue when gentoo uses gcc 3.3.4 at the moment?
no mikael, the problem is with Zyn. bug #65752. Frank is hard to reach (I hear he is quite busy). Chris White is also quite busy, but has found enough time to commit sound-related ebuilds. you should re-assign to chriswhite@gentoo.org because he is a gentoo homeboy and rocks the house. also, please change the keywords to "ebuild" when the bug is ebuild-related, and add "inclusion" to the keywords when it is ready for inclusion. also, when attaching text files, please mark the content type NOT as appl/octet-stream, instead mark them as text. i appreciate the effor mikael, but i'm already hip to the new release and am working on a full patch set with a new ebuild and possible work-around / hack for the gcc 3.4.x users. i would appreciate if you could mark the ebuild and patch you have up there as "obsolete" to prevent any confusion. be patient, i will post my attachments soon.
Created attachment 41777 [details, diff] v1, zynaddsubfx-2.1.1-gcc34_mem_hack.patch it's bizarre, doesn't make sense, but it makes Zyn run happy on x86 against gcc 3.4.2-r2
Created attachment 42010 [details, diff] v1, fltk_flags.patch corrects build scripts so they use proper fltk flags
Created attachment 42011 [details, diff] v1, fltk_version.patch comment out the (unnecessary) version string in *.fl
Created attachment 42012 [details, diff] v2, gcc34_mem_hack.patch hacks for both Synth/OscilGen.C and DSP/AnalogSynth.C not sure why it works, but what i did is allocate some memory, use sprintf to write values to the buffer, and then free the allocated memory. whatever mis-alignement happens due to gcc 3.4.x+ is then not crashing ZynAddSubFX, and after some limited testing I found no ill effects.
Created attachment 42013 [details, diff] v1, gcc_subscript_warning.patch douse gcc warnings
Created attachment 42014 [details, diff] v1, sanitize_makefile.patch 1-liner, gets rid of silly 'yes "" | head' -ism
Created attachment 42015 [details] v2, media-sound/zynaddsubfx/zynadsubfx-2.1.1.ebuild fixed a whole lot of brokeness. i hope. no, i'm pretty sure of it. brokeness be gone!
chris, have a look at this. i've done as much testing as i can do, and ebuild i have attached to the bug should be much more complete than the last one (for v2.0.0). cheers. -E
Closed and in cvs, wee :D.