This is the mail archive of the gcc-patches@gcc.gnu.org mailing list for the GCC project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: [lno] Re: Your change to lno branch



On Mar 9, 2004, at 06:22, Andrew Pinski wrote:



On Mar 9, 2004, at 02:02, Zdenek Dvorak wrote:


... causes the following bootstrap failure on i686; could you please
either fix it or revert the change?

For some reason I thought that invert_truthvalue only returned a non-gimple
which only had a cast added, well I was wrong so I disabled the extra check
for now.


Thanks,
Andrew Pinski


Well that patch was not enough. This patch just disables the use of the new
temporary variables in conditional_replacement.


-Andrew


Index: tree-ssa-phiopt.c =================================================================== RCS file: /cvs/gcc/gcc/gcc/Attic/tree-ssa-phiopt.c,v retrieving revision 1.1.4.3 diff -u -p -r1.1.4.3 tree-ssa-phiopt.c --- tree-ssa-phiopt.c 9 Mar 2004 14:25:28 -0000 1.1.4.3 +++ tree-ssa-phiopt.c 9 Mar 2004 14:53:36 -0000 @@ -258,9 +258,14 @@ conditional_replacement (basic_block bb, && (TYPE_MAIN_VARIANT (TREE_TYPE (cond)) != TYPE_MAIN_VARIANT (TREE_TYPE (result)))) { + /* FIXME: disabled for now as the check for non-gimple is too + weak and will not find some cases. */ + return false; +#if 0 new_var = make_temp (TREE_TYPE (cond)); old_result = cond; cond = new_var; +#endif }

/* If the condition was a naked SSA_NAME and the type is not the


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]