Bug#69311: PROPOSAL] Finishing the /usr/doc -> /usr/share/doc transition.
Chris Waters wrote:
> The tech committee's decision makes a lot of sense given their premise
> that Potato was about to be released, and we wouldn't have time to
> change all the packages. But that premise proved wrong, we did have
> time to change the vast majority (over 80% by JH's count) of the
> packages. We are, I suspect, just about where the ctte expected we
> would be when Woody was about to freeze. (Both in degree of
> preparedness and in actual calendric date.)
> Things have *not* gone as planned so far. So, saying "stick with the
> plan, stick with the plan" seems a bit myopic.
That depends on how you interpret the plan, doesn't it?
If you interpret the plan as "we will have FHS compliance within 1.5
years, and here's how...", then yes, we need to jump ahead now.
If you interpret the plan as "we will have FHS compliance eventually,
but without ever introducing problems in installing packages from one
distribution into a chronologically adjacent distribution, and here's
how...", then we stick with the plan, or everything we have done so far
is useless.
I interpret the plan as being mostly motivated by the latter. Phrases like
* We should not break backwards compatibility during the
transition period. This is a quality of implementation issue
and by all the talk of "potato+1", "woody+1", and so on, and the
relative scarcity of actual dates in the decision point to that being
the case. Not to mention the whole symlink thing.
If we started removing /usr/doc links right now, what would happen?
Apache's http://localhost/doc would start having holes in it for
one. Yes, it still uses /usr/doc. I think dhelp would break similarly.
Our users would become confused -- they have not yet been educated much
about /usr/share/doc. (Such education was a big part of the decision,
btw.) Woody was supposed to be the release where we dealt with these
things, and we have yet to deal with them.
--
see shy jo
Reply to: