----- Original Message -----
From: "Christoph Reichenbach" <[EMAIL PROTECTED]>
To: <[EMAIL PROTECTED]>
Sent: Monday, December 20, 1999 11:06 PM
Subject: Re: DOS port
> Hi Rink,
>
> > > > 2> FreeSCI is *very* GGI oriented now.
> > >
> > > This is absolutely not true.
> > I meant GLIB oriented... my fault. Kernel.c for example uses some GLIB
> > functions, like GTimeVal in line 390.
>
> GTimeVal is "struct timeval" <sys/time.h>, <unistd.h>.
> g_get_current_time() is equivalent to the BSD/SVR4 call "gettimeofday()",
> which is likely to be provided by DJGPP.
> Most glib functionality we use comes from renamed functions and data
> types; you should be able to retrieve most of it from the glib header
> files. For the rest, just ask.
> (Trying to build with an appropriately modified glib.h from the original
> glib package might resolve most problems. _Might_ resolve them).
Yes, I've found that out. I've now dumped most declaration stuff in
SCI_DOS.H.
>
> > > > Anyone has problems if I split this
> > > > all up?
> [..rant..]
> > Once again, I meant GLIB, and not GGI. GLIB doesn't exists for DOS.
>
> Splitting means having to maintain multiple code bases. I'd like to avoid
> that whenever possible by sane measures, and adding a few dozen typedefs
> and #defines IMHO counts as sane.
Yup, that is what I'm doing right now!
>
> > Okay. I'll continue working on it.
>
> Good luck and Happy Hacking!
Sure! I've got the whole thing compiling now. Linking, however, gives a
whole bunch of undefined externals, but most of them are g_realloc() and
g_free(). I'll continue to fix that!
>
> llap,
> Christoph
>
Rink
>
>
