On Feb 4, 2009, at 4:02 PM, Thomas Treichl wrote:
Ben Abbott schrieb:
On Feb 3, 2009, at 3:58 AM, Thomas Treichl wrote:
Hi Ben, just to make sure once again:
The second one gives an error.
$ g++ -framework OpenGL testc.cc
testc.cc: In function ‘int main()’:
testc.cc:12: error: invalid conversion from ‘GLvoid (*)(...)’ to
‘GLvoid (*)()’
testc.cc:12: error: initializing argument 3 of ‘void
gluTessCallback(GLUtesselator*, GLenum, GLvoid (*)())’
... means that the first one does not give an error?
Correct, the first one gave no error.
Ok, from what I've seen until now:
- the problem appears on the current 10.5.x system
- framework OpenGL API between 10.4 and 10.5 has not been changed
- compiling on 10.5 against SDK 10.4 doesn't change anything
- the example program can be compiled if not "(...)" is used
Ergo: it's the compiler. I don't know if it is a bug or if the
compiler has been changed (ie. it is a feature?). I have not been
able to build a small example program that illustrates the problem
in short. So I think there is nothing more to do than to create the
configure.in test now and to set up framework OpenGL for both of us
in a different way...
I'll continue replying to the our thread on the list tomorrow (or
the day after, or the day after, or the day after ;). Meanwhile best
regards,
Thomas
PS. I found this reference that is exactly our problem
http://www.openframeworks.cc/forum/viewtopic.php?t=269
Thomas,
Looking at that link stirs a memory of another problem I had.
Unfortunately, I do no recall the context. However, I have some
recollection that David fixed the problem. I'd estimate this occurred
in Oct-Dec. The part I recall is an issue regarding
"void(<<somethinghere>>)(...)" vs "void(<<somethinghere>>)()" ... or
vice versa.
I don't know if testing David's memory would help, but if you think
it
would help I can ask him.