Re: Yow: Maciej.

[Jim Blandy <jimb@red-bean.com>]

I think when I said "I want to get rid of the warnings," everyone
assumed I was talking about really hairy stuff.  But it's all really
simple things, just ninety-five files worth of dumb fixes.  It was a
problem of volume, not difficulty.

It looks like we're pretty much taken care of on Red Hat.  Thanks to
everyone who submitted a patch, however small.  Eventually we'll bring
Guile up to the EGCS Gold Standard. :)


> guile-core compiles almost warning-less using gcc 2.8.1 on my Solaris
> 2.5.1 box. Apart from the longjmp ones in eval.c, I get:
> 
> inet_aton.c:72: warning: no previous prototype for `inet_aton'
> inet_aton.c: In function `inet_aton':
> inet_aton.c:95: warning: subscript has type `char'
> inet_aton.c:100: warning: subscript has type `char'
> inet_aton.c:102: warning: subscript has type `char'
> inet_aton.c:124: warning: subscript has type `char' 
> 
> Here's a (minimal :-) patch for the first one. Do we want to worry
> about the other four (they're only due to Sun's implementation of
> isascii() using arrays)?

Yes --- the idea is, when Jane Random Programmer is hacking on Guile,
we want her to be able to assume that any warning is something she
should attend to.  If we have categories of benign things, I think 
people won't be strict, and things will get overgrown again.

I'd like to fix the ones in eval.c, too, but that seemed kind of
hairy, last I looked.

Thanks for the patch!  I've taken care of the other inet_aton
warnings.