Re: [Mingw-w64-public] warning when following autoconf manual

[JonY <jo...@us...>]

On 11/2/2011 07:07, Vincent Torri wrote:
> On Tue, Nov 1, 2011 at 11:43 PM, JonY <jo...@us...> wrote:
> 
>> On 11/1/2011 22:58, Vincent Torri wrote:
>>> On Tue, Nov 1, 2011 at 10:45 AM, JonY <jo...@us...>
>> wrote:
>>>
>>>> On 11/1/2011 14:43, Ozkan Sezer wrote:
>>>>> On Tue, Nov 1, 2011 at 1:50 AM, Vincent Torri <vin...@gm...
>>>
>>>> wrote:
>>>>>> hey
>>>>>>
>>>>>> When using alloca in an autotooled project, the autoconf manual says
>>>> this:
>>>>>>
>>>>>>
>>>>
>> http://www.gnu.org/software/autoconf/manual/autoconf.html#Particular-Functions
>>>>>>
>>>>>> but adding that piece of code leads to that warning:
>>>>>>
>>>>>> eina_prefix.c:34:0: warning: "alloca" redefined [enabled by default]
>>>>>>
>>>>
>> /home/torri/local/opt/mingw-w64-x86_32/bin/../lib/gcc/i686-w64-mingw32/4.7.0/../../../../i686-w64-mingw32/include/malloc.h:183:0:
>>>>>> note: this is the location of the previous definition
>>>>>>
>>>>>
>>>>> #elif defined _MSC_VER  is the key, I guess. Better check would be
>>>>> #elif defined _WIN32
>>>>>
>>>>
>>>> You can also check if alloca.h exists and include that instead, at least
>>>> that's what binutils does.
>>>>
>>>
>>> it's already done by
>>>
>>> #ifdef HAVE_ALLOCA_H
>>> # include <alloca.h>
>>>
>>>  Vincent Torri
>>
>> Well, remove the one in eina_prefix.c, it has no business declaring it
>> there. Alternatively, guard the parts that require alloca with
>> HAVE_ALLOCA_H.
>>
> 
> If you think that it's correct, then it's something you have to tell to
> autoconf developpers
> 
> Vincent Torri

I don't see how this is autoconf's fault at all.

Isn't eina_prefix.c part of enlightenment? Is HAVE_ALLOCA_H defined? Is
autoconf failing to detect alloca? Do you have the config.log entries
about searching for alloca.h?