[Bug c/40172] [4.5 Regression] Revision 147596 breaks bootstrap

[ddaney at caviumnetworks dot com]

------- Comment #27 from ddaney at caviumnetworks dot com  2009-05-19 20:58 -------
Subject: Re:  [4.5 Regression]  Revision 147596 breaks bootstrap

manu at gcc dot gnu dot org wrote:
> ------- Comment #26 from manu at gcc dot gnu dot org  2009-05-19 20:29 -------
> The case in toplev.c cannot be fixed without tracking macro expansions somehow,
> but I wonder why it warns (multiple times!) for this case:
> 
>> ../../trunk/gcc/config/mips/sb1.md:159: error: logical �or� of collectively
>> exhaustive tests is always true
>> ../../trunk/gcc/config/mips/sb1.md:159: error: logical �or� of collectively
>> exhaustive tests is always true
>> ../../trunk/gcc/config/mips/sb1.md:159: error: logical �or� of collectively
>> exhaustive tests is always true
>> ../../trunk/gcc/config/mips/sb1.md:159: error: logical �or� of collectively
>> exhaustive tests is always true
> 
> David, could you produce a testcase?
> 

It is in insn-attrtab.c, which is machine generated.

I since the fatal warning is now disabled, it should be fine.

The problem with insn-attrtab.c is that it is generated from the .md 
files and then includes all the target macros.  So for this file you 
should probably never use -Wlogical-ops as filters that try to eliminate 
things in macros will fail.  The whole file is conceptually one big macro.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=40172