Mail Archives: cygwin/2001/03/29/19:44:44
On Wed, Mar 28, 2001 at 06:30:42PM -0800, Paul Garceau wrote:
>Hi folks,
>
> Forgive me, couldn't let this pass...
>
>On 27 Mar 2001, at 23:49, the Illustrious Kenton Varda wrote:
>
>> I think that's what I was originally looking for. :) Why isn't this
>> part of cygwin's mingw package already?
>
> *ahem*
>
> Mingw _does not belong to Cygwin_, Mingw is a separate and optional
>set of development tools that just happens to run alongside the Cygwin
>Development Tools and that _may be_ installed with the Cygwin
>Development Tools...it is _not_ a Cygwin development tool, it is a
>Cygwin "Option" in the same way that the "-mno-cygwin" switch is an
>"optional" switch.
> It has only achieved this after a whole bunch of volunteers spent a
>great deal of time and energy (man-years) attempting to integrate the
>two to work with a simple switch called "-mno-cygwin". It has _not
>been_ a "trivial task".
I can attest to that. gcc is really not designed to accomodate the
kind of magic that -mno-cygwin is attempting. I try to push the
envelope with every release of gcc but I think I've reached the point
where I now have to branch out into binutils to achieve perfect
separation. The current testing version of gcc actually has some
hacks that I probably would never send back to gcc.gnu.org.
And, I really am trying to be sensitive to the fact that mingw is not
"my" project. I don't want to cloud the issue by releasing too much of
mingw with Cygwin. I'm sure that the mingw maintainers have enough to
do with their own distribution and don't want to be bothered spending
excessive amounts of time trying to integrate things with Cygwin.
Sometimes I think that the best way to handle all of this is to have a
completely different compiler which defaults to a totally separate
/usr/include and /usr/lib area. i686-pc-mingw-gcc. Hmm. It has a nice
ring to it.
cgf
--
Want to unsubscribe from this list?
Check out: http://cygwin.com/ml/#unsubscribe-simple
- Raw text -