On Mon, Aug 31, 1998 at 09:26:16AM -0700, Jim Pick wrote: > This is going to be a pretty tough policy decision to make. Right > now, we've got several people who are holding "positions" for or > against different proposals. That's not going to move us any closer > to a resolution of this policy question. That's a sign that we might > have to move to a more formal process. Well, I withdraw my suggestion that we wax /usr/X11R6 (except as a tree of symlinks into /usr/bin, /usr/lib, etc.). Manoj convinced me. The FHS is pretty clear on this point, and if Debian unilaterally makes a change of this magnitude, in flagrant violation of the standard, we're no better than Microsoft. However, I think I have come up with a may of managing the X hierarchy so that if things change in the future, they will be more easily managed then they are at present. I'm stripping xbase down to just a directory, some symlinks, some documentation, and the maintainer scripts. The job of xbase's maintainer scripts will be to manage the upgrade process in all its hideousness (anyone who's seen the existing maintainer scripts for xbase can attest to that). For a preview of what's coming, see http://master.debian.org/~branden/xsf.html . Essentially, xbase will become standard (since xlib6g already is) but will be very small. This is not your father's xbase. xserver-common and xlib6g will each depend on xbase. Everything else that uses X eventually depends on either xserver-common or xlib6g. -- G. Branden Robinson | "I came, I saw, she conquered." The Purdue University | original Latin seems to have been branden@purdue.edu | garbled. http://www.ecn.purdue.edu/~branden/ | -- Robert Heinlein
Attachment:
pgpcJZl0e93wk.pgp
Description: PGP signature