From: Jeff <jeff@pc-4374.bc.rogers.wave.ca>
  To  : ggi-develop@eskimo.com
  Date: 31 Jul 1998 10:09:43 -0700

Re: CVS and source

Andrew Apted <ajapted@netspace.net.au> writes:

> Hartmut writes:
> 
> >  Jeff,
> >  You'r probably not the first one who stumbled across this ...
> >  It looks like we have to be far more careful with the libggi documentation
> >  changelog, where all of this should be documented.
> >  
> >  AFAIK: libggi.h -> ggi.h, you guessed correctly.
> >  ggiSetPaletteVec -> ggiSetPalette, with identical parameters and functionality
> >  ggiKeyPoll: no idea. Andrew, did you change that?
> 
> Yeah, all the programs in the repository use ggiKbhit() (identical
> functionality btw), and it seemed silly to have two names for the one
> function in the libggi API.
> 
> It wouldn't be hard to put back (add a #define to ...), but seeing that
> we're in "breaking things" mode, shouldn't we clean up little things
> like this ?


yes, definitely, clean up etc, but a nice changelog of all changes to
the CVS tree would be cool so that when something breaks, we have a
nice easy way of looking it up.....

btw as a side note to last night additions:

__i386 doesn't seem to define itself anywhere on my machine btw
(types.h looks for it to define endianess, line 183), I think what you want is
__i386__ (at least in 2.1.112)

Note: do you want bugs that I notice like this posted to this mailing
list, should I fix them myself, or.......what is the prefered method
of me communicating these problems? :)  also where can one find a list
of things to do in the GGI development in case one is interested in
casually helping out on occasion :)


jeff