[Bug other/54701] [4.8 Regression] double_int conversion breaks avr build

[rguenth at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=54701

--- Comment #3 from Richard Guenther <rguenth at gcc dot gnu.org> 2012-09-25 11:47:37 UTC ---
(In reply to comment #2)
> (In reply to comment #1)
> > Fixing it would probably have been as fast as reporting this bug.  Confirmed.
>
> So are bugs invalid that are easy to fix for others?

Obviously not, see my "Confirmed" note.

> If I fix this one the next error comes up:
> 
> ../../../gcc.gnu.org/trunk/gcc/config/avr/avr-log.c: In function 'unsigned int
> avr_double_int_pop_digit(double_int*, unsigned int)':
> ../../../gcc.gnu.org/trunk/gcc/config/avr/avr-log.c:151: error:
> 'div_and_round_double' was not declared in this scope
> make[2]: *** [avr-log.o] Error 1
> make[2]: Leaving directory `/local/gnu/build/gcc-trunk-avr/gcc'
> make[1]: *** [all-gcc] Error 2
> make[1]: Leaving directory `/local/gnu/build/gcc-trunk-avr'
> make: *** [all] Error 2
> 
> 
> I never understood why double-int.c:double_int_split_digit is static, but now
> div_and_round_double is also static :-(

Functions are static when they are not used outside a TU or they are not
considered the correct interface to a problem.  div_and_round_double is now
just called divmod.