Re: Newbie question
On Wed, Feb 12, 2003 at 07:18:40PM -0500, Earnie Boyd wrote:
> Mark Paulus wrote:
> >I guess that's a question for the group. Does cygwin present
> >enough of an extension to Microsoft windows that is could be
> >considered and independant "flavour" of a kernel, and thus
> >require another kernel-type indicator?
> Identifying by C runtime I think would be most beneficial, -cygwin or
> -msvcrt.
First, I must say that i'm not really aware of windows things, and have
never used mingw32 or cygwin. However i'm facing the same sort of
problem since i'm slowly trying to get debian packages compiled for DOS
with DJGPP...
I don't really get what you mean by "C runtime" (libc ?). However i
don't think it is appropriate (at least it doesn't follow what already
exist.) Think of hurd-i386 or (free|net)bsd-i386.
One point is that everything that could coexist on a system should be
the same debian architecture. Imagine a system with a windows kernel
(windows, but maybe wine? reactos?). One may be able to install binaries
compiled with mingw32 as well as cygwin, which could co-exist.
Another point is that the libc's used by these things are doing much to
give a UNIX-like API to windows. This will make programs built with them
badly coexist with "native" windows programs (i'm thinking of pathnames
and such things...). Maybe it's bad enough not to use w32-i386, maybe
not...
So win32 is not enough, and cygwin is too much ;) For my case, dos is
not enough, djgpp is too much. Can someone think of a word that would
reflect the situation ?
I read somewhere something from Marcus Brinkmann (a guy that is involved
in the Hurd's developement, thus in debian hurd-i386), about debian
ports to other OSes and more, which proposed more elegant solutions than
debian arch strings for such cases. I'll find the URL and post it here,
it's very interesting.
--
Jeremie Koenig <sprite@sprite.fr.eu.org>
Reply to: