Re: <stdint.h> on IRIX (was Re: [Cvs-test-results] CVS trunk testing results (SGI IRIX))

[Derek R. Price]

Digging a bit deeper, it looks like ccvs/m4/stdint.m4 should be creating
lib/stdint.h from lib/stdint_.h when it can't compile a simple program
which just includes <stdint.h>.  Since that compile should fail on IRIX
in c89 mode, the substitute file should be used on IRIX instead of the
system one.  Is there something in the way I included inttypes.h that
precluded lib/stdint.h from being found?

(Probably, there should be an AC or GNULIB test which finds and
subsequently uses c99 when available on IRIX, but this is looking more
and more like an unrelated issue since some form of <stdint.h> worked in
CVS on IRIX previously.)

Regards,

Derek


Derek R. Price wrote:
> Larry,
>
> Any luck figuring out why the unconditional #include of <stdint.h> via
> lib/md5.h has been working for 6 months and the #include of the same
> file from lib/inttypes.h fails?
>
> Thanks,
>
> Derek
>
>
> Derek R. Price wrote:
>   
>> Larry Jones wrote:
>>   
>>     
>>> Yep, "-c99" (or you can use the c99 command instead of cc).  It seems to
>>> me that autoconf should either add that option or detect that the header
>>> is unusable.  
>>>     
>>>       
>> Hrm.  This is interesting too.  Why does the unconditional #include
>> <stdint.h> from client.c and update.c (both via #include "md5.h") work
>> and the #include <stdint.h> from lib/inttypes.h fail with the c99
>> error?  We have been doing the include via lib/md5.h for six months or so.
>>
>> Regards,
>>
>> Derek
>>
>>   
>>     
>
>
>   


-- 
Derek R. Price
CVS Solutions Architect
Get CVS support at Ximbiot <http://ximbiot.com>!
v: +1 248.835.1260
f: +1 248.835.1263
<mailto:address@hidden>