[libgfortran,patch] Use __mingw_snprintf instead of snprintf when possible

[Tobias Burnus]

:REVIEWMAIL:

Danny Smith wrote:
>> Built and tested (but not regtested, as it requires a very weird setup
>> on mingw) on i386-pc-mingw32, OK to commit?
>>     
> How 'weird'?  Regtested on native build of mingw32 (using cygwin's
> dejagnu support) with no new regressions.
> I have no objections.
>   
Looks almost trivial and as the mingw maintainer has no objections:
OK.

Thanks for the fix!

Tobias


PS: Anyone interested in doing some reviews? Outstanding are:

USE operator patch by me:
http://gcc.gnu.org/ml/gcc-patches/2007-08/msg01081.html

-frecursive/OpenMP patch by Asher and me:
http://gcc.gnu.org/ml/gcc-patches/2007-08/msg01253.html

COMMON checking patch by me:
http://gcc.gnu.org/ml/gcc-patches/2007-08/msg01299.html

GAMMA/LGAMMA patch by me:
http://gcc.gnu.org/ml/gcc-patches/2007-08/msg01386.html

And, awaiting review for a long time, Thomas' byte-wise access to masks
http://gcc.gnu.org/ml/gcc-patches/2007-08/msg01468.html