Re: More compilation fixes for Camel

On Wed, 2006-10-18 at 13:59 +0300, Tor Lillqvist wrote:
> Instead of this cast-o-rama, have you investigated whether it would be
> simpler and more correct to change the prototypes of the affected
> functions instead (changing signedness or constness of char * parameters
> as appropriate), and/or changing signedness/constness of the affected
> char* variables? 
> 
> Maybe then the need for casts all over the place to avoid warnings would
> diminish drastically?

Indeed. I have been thinking about correcting the prototypes too. But I
figured that most of such corrections would imply changing the API (at
least a little bit).

I also noticed that most costness and signedness of the functions are
indeed correct. It were mostly temporary variables, mostly called "p"
and "intPtr" which didn't have the same constness and/or signedness. And
often the temporary variable is indeed used by a mixed set of functions
that do and don't change the pointer and that do and don't expect the
signedness.

Another example where the strlen, strcpy, strcmp functions being used
with temporary variables that have correct signed and constness.

Very few assignments where wrong. So the return values of the functions
where probably used to decide about the constness and signedness of the
temporary variables.

-- 
Philip Van Hoof, software developer
home: me at pvanhoof dot be
gnome: pvanhoof at gnome dot org
work: vanhoof at x-tend dot be
blog: http://pvanhoof.be/blog