Hi T00n3,
I did not see this thread since I do not check the forum very often anymore, however, kickvb pointed it out to me. Some people read the forum primarily, and some read only the mailing lists, so it can be difficult to know where to post. I usually follow only the mailing lists and the issue tracker. Also, there are several Mac OS X users subscribed to the crystal-main and crystal-develop mailing lists, so it might be useful posting there. At least one of those persons is actively extending CS (for GSoC), so he presumably has a working installation of CS on Snow Leopard. (He filed several of the bug reports mentioned below, which we have since fixed).
The
CS::Threading::ThreadID problem should be addressed by commit r33845 in trunk (
http://trac.crystalspace3d.org/trac/CS/changeset/33845). Unfortunately, this fix slipped through the cracks and has not yet been merged into the 1.4 branch, although I just added it to the list of patches which should be applied:
http://trac.crystalspace3d.org/trac/CS/wiki/Version14Plan#Candidatesfor1.4.1The
-bundle issue has been raised several times (most recently in this thread
http://thread.gmane.org/gmane.comp.graphics.crystalspace.devel/3003/). My understanding is that this flag merely needs to be dropped when building with Snow Leopard. In each case I suggested that we should automate detection of Snow Leopard and conditionally remove the
-bundle flag, although there was never any follow-up to my inquiries. If you would like to work with me to resolve the issue, that would be ideal.
Setting
MACOSX_DEPLOYMENT_TARGET probably should not be necessary since the project itself should still be deployable on older Mac OS X releases, so no need to restrict deployment only to Snow Leopard.
The
NSOpenGLContext issue was the subject of ticket #810 (
http://trac.crystalspace3d.org/trac/CS/ticket/810) and resolved in trunk by r33867 (
http://trac.crystalspace3d.org/trac/CS/changeset/33867). The fix was merged into the 1.4 branch with r33868.
The
size_t vs
Uint32 compilation failure in
driver_coreaudio.cpp was the subject of ticket #811 (
http://trac.crystalspace3d.org/trac/CS/ticket/811) and resolved in trunk by r33872 (
http://trac.crystalspace3d.org/trac/CS/changeset/33872). The fix was merged into the 1.4 branch with r33873.
Did you ever resolve the problem with
vfs plugin not being found? Typically, this would occur if
zlib was not installed or detected at CS
configure time.
Regarding the build system,
jam and
make are equivalent in CS. They both invoke the same build commands, so there is no functional difference between the two. Whether you use "
./configure; make; make install" or "
./configure; jam; jam install" is a matter of personal preference.
Was the
shaderweaver crash from 1.4 or pre-2.0 CS? Perhaps this problem also has been fixed in the trunk.
Generally speaking, unless there are constraints which force you to use the 1.4 release (or branch), I would recommend instead using the pre-2.0 (trunk) version directly from Subversion since it has advanced a lot from when 1.4 was released, and many bugs have been fixed in the trunk.