On Thu, 23 Sep 2004, Harshal Pradhan wrote:

> So, what is right solution in this particular case? And what kind of patch
> would be accepted upstream? I think we are not really concerned with locales
> etc. for this particular case and are only really looking for the classical
> ASCII whitespaces, right?
>
> I can shut it up by casting like so
>
> isspace((int)(unsigned char)what[len-1])
>
> or doing something like
>
> (what[len-1] > 0) && isspace((int)what[len-1])
>
> Or something else ..... Thoughts?

I think we should modify the 'char *' variables to become 'unsigned char *'
variables as then typecasting it to 'int' like we do won't produce any
negative values, which I assume is the actual problem you face?

-- 
      Daniel Stenberg -- http://curl.haxx.se -- http://daniel.haxx.se
       Dedicated custom curl help for hire: http://haxx.se/curl.html