On Fri, Sep 24, 1999 at 08:22:38PM -0400, Adam Di Carlo wrote:
> Marcel wrote:
> > - work on bugs (we have one outstanding release critical bug
> > that has not yet been handled in our CVS tree)
>
> Yeah -- note all the 'closes: Bug#...' stuff I recently put in the
> changelog.
>
> > - boxes.c that handles all screen sizes properly (I will not commit
> > this before we have a stable upload, this is on the backburner..)
> >
> > - http base system fetching (I will not commit this before we have a
> > stable and fully working upload)
>
> Ok -- but isn't http base fetching a new feature? Is that really on
> the front burner, and if so, why?
It isn't on the frontburner, though I've started work on it after
finishing the untested drivers setup change yesterday.
> My vision is that once we get the core stuff going (root, boot,
> drivers, base, doc) and building right, we should do a test release
> source upload.
>
> Then we work on new features and more bug fixes, such as what you have
> above. Also, we need a GUI for tasks selection, a lilo rework for
> i386, many documentation corrections, and also hopefully some
> "automated installation" changes.
> > > What prevents us from uploading a hyper-nuke test source upload now?
> >
> > Working driver disks, at least for i386. I don't know how other platforms
> > handle the increased size of the drivers.
> >
> > I have changed dbootstrap to handle drivers the same way as the base
> > system is handled, i.e. split disks. I also changed drivers.sh. These
> > changes were implemented this afternoon on a workstation at my work. I
> > haven't tested them yet...
>
> Ok -- btw, are you changing stuff just for i386 ?
>
> I am quite sure all architectures will need to split boot/root and
> also split the drivers disk.
It should be portable. It's the same mechanism as the base disks use,
nothing fancy.
Marcel
Attachment:
pgpGtn3Yv_kOo.pgp
Description: PGP signature