[Bug c/36255] -fstrict-aliasing breaks __builtin_va_args

2008-05-17 Thread rguenth at gcc dot gnu dot org


--- Comment #1 from rguenth at gcc dot gnu dot org  2008-05-17 12:22 ---
Can you provide an executable testcase that abort()s if the miscompilation
happens?


-- 

rguenth at gcc dot gnu dot org changed:

   What|Removed |Added

 CC||rguenth at gcc dot gnu dot
   ||org


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=36255



[Bug c/36255] -fstrict-aliasing breaks __builtin_va_args

2008-05-17 Thread juanjose dot garciaripoll at gmail dot com


--- Comment #2 from juanjose dot garciaripoll at gmail dot com  2008-05-17 
13:27 ---
Created an attachment (id=15649)
 -- (http://gcc.gnu.org/bugzilla/attachment.cgi?id=15649action=view)
Files for test case (simple.c 1/3)


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=36255



[Bug c/36255] -fstrict-aliasing breaks __builtin_va_args

2008-05-17 Thread juanjose dot garciaripoll at gmail dot com


--- Comment #3 from juanjose dot garciaripoll at gmail dot com  2008-05-17 
13:28 ---
Created an attachment (id=15650)
 -- (http://gcc.gnu.org/bugzilla/attachment.cgi?id=15650action=view)
Files for test case (simple.h 2/3)


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=36255



[Bug c/36255] -fstrict-aliasing breaks __builtin_va_args

2008-05-17 Thread juanjose dot garciaripoll at gmail dot com


--- Comment #4 from juanjose dot garciaripoll at gmail dot com  2008-05-17 
13:29 ---
Created an attachment (id=15651)
 -- (http://gcc.gnu.org/bugzilla/attachment.cgi?id=15651action=view)
Files for test case (main.c 3/3)


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=36255



[Bug c/36255] -fstrict-aliasing breaks __builtin_va_args

2008-05-17 Thread juanjose dot garciaripoll at gmail dot com


--- Comment #5 from juanjose dot garciaripoll at gmail dot com  2008-05-17 
13:32 ---
I uploaded the program that fails if the GCC miscompiles the file simple.c
However, I had to split the test case into two different files, for otherwise
GCC inlines the offending function and the error is not produced.

I compiled the files separately, with and without -fstrict-aliasing, and it
only fails in the first case.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=36255