This is the mail archive of the
cygwin-apps@cygwin.com
mailing list for the Cygwin project.
Re: Move zlib up one level?
----- Original Message -----
From: "Christopher Faylor" <cgf@redhat.com>
> >There are some things I believe we should be able to do with
> >/etc/setup...
> >1) Detect cross-package conflicts. Say foo and bar both contain
> >/usr/bin/ld.exe.
> >2) The lst files are currently gz' files, I think it would be good to
> >change to using bz2 in the future.
> >
> >Doing the 1st one really requires a more database orientated
approach --
> >long term of course.
>
> Right. But in the meantime we have a huge user base who can't easily
> figure out what packages they have installed.
Perhaps a new view in setup - categories/full/partial/installed ?
> >> Also, since I wrote most (all?) of what I'm pulling out, I felt
pretty
> >> comfortable with it...
> >
> >Oh, sure :}. I wasn't meaning you weren't capable, just that having
the
> >same code in two places is bad because of the capability for skew.
>
> Absolutely. I really hate this but I was doing it the wrong way for
1.3.5.
> I am optimistically thinking that there won't be a new cygwin release
for
> a while after that and I wanted to get something that works into
1.3.5.
Makes sense. Hopefully the daemon will go in immedaitely post 1.3.5, and
that will need a little ironing out I'm sure.
> I also didn't want to be pulling things apart while you are actively
working
> on this but maybe this is an important enough goal that it is worth
doing
> ASAP.
As long as you're happy to rework cygcheck twice :].
> Btw, I considered just suggesting that this functionality should go
into
> setup.exe. It would be sort of nice to be able to say:
>
> setup --dump
>
> or something to get the current list of installed packages.
>
> However, there are two problems with that, IMO:
>
> 1) setup.exe is currently a windows app so it can't write to the
console.
That's a small problem :}.
> 2) Cygcheck should really be the one stop place for all debugging
output
> so, at the very least, cygcheck would have to call a 'dump capable'
> setup.exe to get its output but, that wouldn't be useful for
debugging
> cases where setup.exe wasn't available.
I think cygcheck should have the functionality built in statically, just
the source shouldn't be split.
Rob