Re: [Result] Moving to the FHS: ...
Hi,
>>"Santiago" == Santiago Vila <sanvila@unex.es> writes:
Santiago> On 5 Sep 1999, Manoj Srivastava wrote:
can get rid of symlinks later.
Santiago> Mmm, let's see if I understood this "deprecated vs illegal" thing.
Santiago> Suppose I decide, for the packages I have not converted
Santiago> yet, not to move to /usr/share/doc until the "last
Santiago> minute". Will I get bug reports? Will my packages be
Santiago> NMUed?
Possibley, depending on how things are done. Not immediately,
since the old way of foing things is not illegal immediately.
Santiago> I imagine the following scenario:
Santiago> slink: everybody uses /usr/doc.
Santiago> potato: mix /usr/share/doc and /usr/doc but people using
Santiago> /usr/share/doc should use symlinks.
*must*, or be in violation of policy.
Santiago> potato+1: we deprecate symlinks but still allow them, make /usr/doc
Santiago> illegal, and start filing bugs against packages
Santiago> still using /usr/doc. [ This is of course not
Santiago> decided yet, it's just an hypothesis ].
What about this: (note that I am nt tying these states to a
release cycle; they may happen more than one states/cycle, or
more than one cycle/state)
.
Next stage: : /usr/doc is illegal, and we start filing bugs against
packages still using /usr/doc. Symlinks are still
required.
Next stage: Symlinks are now deprecated, but still legal
Next stage: we make symlinks illegal and start filing bugs
against packages using them.
Next stage: We create a pacakge/basefile postinst that actively
removes old symlinks
Santiago> [ BTW: How do these hypothesis sound as a proposal? ]
I am not sure. I would rather introduce more states, though
possibly keeping the same timetable as you did.
Santiago> Since I don't think maintaining Debian packages should be
Santiago> *gratuitously* painful, what kind of technical problems
Santiago> would my packages cause to the average Debian user if I
Santiago> decide to move from /usr/doc to /usr/share/doc in one shot
Santiago> and without symlinks during the unstable stage of potato+1?
If we go your route, possibly none. If we go a more gradual
route as I show above, your packages would be buggy in between one of
the states (which may be a reasonable argument for your strategy).
Santiago> I guess if we are going to be "permissive" about packages
Santiago> still using /usr/doc in potato, we should probably be
Santiago> permissive as well about packages not using symlinks when
Santiago> they are not really needed.
Depends on whether there is techniocal grounds for using the
smoother transition scheme detailed above.
manoj
--
There is no choice before us. Either we must Succeed in providing the
rational coordination of impulses and guts, or for centuries
civilization will sink into a mere welter of minor excitements. We
must provide a Great Age or see the collapse of the upward striving
of the human race. Alfred North Whitehead
Manoj Srivastava <srivasta@debian.org> <http://www.debian.org/%7Esrivasta/>
Key C7261095 fingerprint = CB D9 F4 12 68 07 E4 05 CC 2D 27 12 1D F5 E8 6E
Reply to: