Closed orbea closed 7 years ago
Ugh, seems the buildbot is using cmake for some retarded reason and I didn't bother to test that, only the static Makefile. If I really have to I can look into that tomorrow...
Actually I think its because cmake is using glew.c
instead of glew_libretro.c
. I'll have to see how to tell cmake to do that tomorrow, if someone who actually likes cmake would like to help here it'd be very much appreciated.
It can also be worked around if glew is installed on the build OS, but that may not be portable for a buildbot?
It should really be using glew_libretro instead.
I pushed another commit to use glew_libretro.c. I'm not sure about how to force it instead of a natively installed glew, but I think this should make the buildbot happy.
Unfortunately that didn't seem to help the cmake build issues even if it works locally?
I pushed another commit to hopefully fix the android build issues, but I would need the buildbot to test it...
I'm not sure why travis insists on using glew.c
insted of glew_libretro.c
with cmake. It works here?!
Also for android this is the current issue.
jni/../../Common/Vulkan/VulkanContext.cpp:917:2: error: use of undeclared identifier 'init_resources'
init_resources(Resources);
^
Any ideas?
Seems travis is now happy with android even though I didn't change anything....
Let's merge it and then we can go from there.
This attempts to merge the upstream ppsspp changes into the libretro core. The few merge conflicts were fixed with entirely upstream code. I also encountered these problems which I attempted to fix in the last three commits.
bSeparateCPUThread
was removed upstream.Note that I have only tested this on Linux.