Re: Glibc for win32
On Tue, Sep 30, 2003 at 11:45:37AM +0200, Jakob Eriksson wrote:
>On Mon, Sep 29, 2003 at 08:38:55PM -0400, Christopher Faylor wrote:
>> On Mon, Sep 29, 2003 at 10:09:29PM +0100, Carlos Sousa wrote:
>> >On Mon, 29 Sep 2003 16:34:07 +0200 Jakob Eriksson wrote:
>> >> http://line.sourceforge.net/
>> >Sorry to butt in, I'm just a lurker here, but isn't LINE a dead project?
>> >Doesn't seem to have any activity since mid-2001...
>> >On the other hand, bringing a LINE fork under debian-win32 might be the
>> >safest thing to do, in that scenario.
>> AFAIK, it's a dead project and it uses cygwin at its core. Wasn't that
>> what you were trying to avoid?
>I have no problem using cygwin derived code per se.
>But LINE uses cygwin in an original way.
>My point is: with LINE, Debian-i386 can be used as is. No porting
>required. (LINE already runs actual Linux programs.)
>With traditional cygwin, all Debian has to be recompiled and probably
>Glibc needs to be ported to ease that recompile.
There is 0% chance that a glibc-for-windows port would be accepted back
into the main glibc base so, you'd be committed to maintaining a glibc
fork in this scenario, too.
>It is anybodys shot which way is the easiest, but I believe LINE is.
>(Since LINE already can run complex Linux applications.)
But you're still going through two levels of emulation. It's bound to
be pretty slow.
>Of course, none of this matters until somebody digs in, one approach
>or another. I am all talk and nothing else at this stage. If you decide
>to go for either approach, more power to you.
I'm just a bystander here. I already maintain cygwin. I'm not volunteering
anything except semi-expert observations -- and those are worth as much as
you paid for them, of course.