Re: [PATCH] eruby on tiger

On Tue, Jul 19, 2005 at 07:12:14PM +0200, Jb Evain wrote:
> Bonjour !
> 
> >   if (isspace ( *(unsigned char *)s ) ) { ... }
> 
> Yeah, or 
> 
> if (isspace ((unsigned char) *s)) ...
> 
> or ... :)

I'm personally less happy about casting an int to an unsigned char and then
implicitly back to an int again, as I don't trust all compilers to get it
right. If you want to do it that way, then I'd say it's safer to do

  if (isspace (*s & 0xff)) ...

But if you cast the pointer to an unsigned char *, then it's not an issue.

> Attached is a patch that respects the isspace desire to handle
> unsigned char

isspace takes an _int_. If isspace were declared to take an unsigned char,
then I think (in ANSI C at least) the problem wouldn't happen. But it's
declared to take an int, because historically that's what it always did.

Blame K&R for allowing compilers to choose whether chars were signed or
unsigned :-(

Regards,

Brian.