Jean-François Mertens wrote:
On 30 Jul 2005, at 01:23, Jens Nöckel wrote:
OK, one more thing to finish this off:
My earlier success report was for a 10.3.9 system. On Tiger, I just
saw that the build still doesn't work unless one sets gcc_select 3.3
first.
Please be more specific in the future in your error msgs...
(I guess it was a linking error)
Here it did build with gcc4 _ but then, about everything does here -).
But since it contains c++ code, I added the appropriate field.
..which of course prevents one from using gcc_select 3.3 and thus
building matplotlib. I get the same error as Jens.
If I rip out the GCC field and set gcc_select 3.3, it builds OK.
It also builds OK with gcc-4.0 and g++-4.0. Only Fink's chimaera of
gcc-4.0 and g++-3.3 doesn't work.
Actually, matplotlib outsmarts Fink's logic here: It compiles its C++
sources using "gcc" and links them with "g++-3.3". Using "gcc" for
compiling C++ code means using g++-4.0 in the current state of affairs,
so that an ABI clash is unavoidable.
Unfortunately, matplotlib uses the python build system which I don't
understand sufficiently to see where it is defined that "gcc" or
"g++-3.3" should be used for compiling and linking, respectively.
Someone else?
--
Martin
-------------------------------------------------------
SF.Net email is sponsored by: Discover Easy Linux Migration Strategies
from IBM. Find simple to follow Roadmaps, straightforward articles,
informative Webcasts and more! Get everything you need to get up to
speed, fast. http://ads.osdn.com/?ad_idt77&alloc_id492&op=click
_______________________________________________
Fink-users mailing list
Fink-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/fink-users
