Prakash Punnoor wrote:
Am Sonntag 21 Januar 2007 16:32 schrieb Bill Hoffman:
Prakash Punnoor wrote:
ome soon to be released 2.4.4. version has the fix. :-)
2.4.6 segfaults as well. Wasn't the fix from cvs incorporated?
Well, current cvs segfaults as well. Smells like regression...
Problem is, I don't think the cause was ever identified for this. It
just happened to be fixed in cvs. If you have a cvs build that shows
the problem, can you build with debug, and post the call stack?
Sorry, I don't know very much about debugging binaries in windows. Could you
provide me with more infos what exactly to do?
Also,
creating a small test case that shows the problem would be good as well,
so we can add it to the regression tests.
It is attached. Taking out the include directive doesn't cause the segfault
for me. Perhaps to reproduce exactly, here the tools I used:
- visual c++ express sp1 - debug x86 target
- latest windows (vista) platform sdk - windows xp debug x86 target
cmake with NMake Makefiles generator
Even the cvs from October 2006 segfaults here, whereas it worked on another
machine (which has another sdk installed though...)
OK, found it...
You need to add this:
SET(CMAKE_INCLUDE_FLAG_ASM "-I")
To CMakeASMInformation.cmake.
I will also fix the seg fault, but this will get around the bug.
-Bill
_______________________________________________
CMake mailing list
CMake@cmake.org
http://www.cmake.org/mailman/listinfo/cmake
