Re: [Gap-dev-discuss] LuserNet: warning fix / code cleanup : char signedness

[Yavor Doganov]

Riccardo Mottola wrote:
> (Yavor, are you on gap-dev? perhaps you should so we can avoid
> sending a lots of mails privately, I continued getting bounces from
> German for example, I fear he did not get half of our conversation)

Yes, I am subscribed.  I wrote to German privately because at first
glance the gnunstep-nonfsf project does not have a developer/bugs
mailing list, and Pantomime/GNUMail had nothing to do with GAP...
Slightly confusing.

> Alexander's code is for me very difficult to read and maintain, so I
> proceed extremely carefully and only fix stuff when really

It's difficult for me too.

> On NetBSD I get an error that atoi() expects a const char*, not a 
> unsigned char*. How is it on linux and other systems?

Yes, `atoi' expects const char *.

> My first Idea was to apply the patch below instead of a cast,
> however that triggers another maze of warnings and errors.

> -       unsigned char *resp_buf;  /* read buffer */
> +       char *resp_buf;  /* read buffer */

I tried that first, too :-)

> Any proposals? I'd leave it as it is now... and not silence the cast
> if we are absolutely sure.

OK.  I don't have an idea how to silence the warning without a cast
(there's no implicit pointer conversion, hence the complaint).  It
should be safe if I understand the code correctly...