On Feb 4 22:09, Christopher Faylor wrote:
> On Sun, Feb 04, 2007 at 11:29:59PM -0000, Dave Korn wrote:
> >There have been no serious objections, so I think we should go ahead.
> >Perhaps we should replace the gcc, g++, g77 etc. drivers with shell
> >scripts that look for -mno-cygwin on the command line and redirect to
> >i686-pc-mingw-{gcc,g++,etc} or to i686-pc-cygwin-{gcc,g++,etc}, just to
> >make life easier for the backwardly-compatible, but I don't see any
> >reason not to go ahead and remove the option from the driver.
>
> I can remove the driver from the upstream trunk sources but, as I've
> repeatedly said, I don't want to make -mno-cygwin available by default.
> If we still, by default, have the option then we still have problems
> with people not getting it.
I'm somewhat concerned about this step, though. How is the configure in
winsup supposed to work, if you suddenly need two compilers to build
Cygwin? The top-level configury only takes one target, not two. If you
only have a linux-x-cygwin cross compiler, your build suddenly fails if
you don't have a linux-x-mingw compiler as well.
So, to build Cygwin from scratch, you first have to build two
compilers and two binutils. This sounds not feasible to me.
Or, the configury in winsup must be changed so that you can build
everything with the i686-pc-cygwin compiler. This should be possible,
shouldn't it? It's just a question of using the right start files and
import libs, isn't it?
I would like to have this cleared before we do this crucial step.
Corinna
--
Corinna Vinschen Please, send mails regarding Cygwin to
Cygwin Project Co-Leader cygwin AT cygwin DOT com
Red Hat
--
Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple
Problem reports: http://cygwin.com/problems.html
Documentation: http://cygwin.com/docs.html
FAQ: http://cygwin.com/faq/
