Re: Debian's problems
On Sun, 12 Sep 1999, Raul Miller wrote:
> But maybe it would be better to accept unstable package uploads from
> developers only after the developer has no more important bugs on packages
> in frozen...
This should be a good idea in theory, but I think in practice people
will get frustrated, because it is impossible to say that *all* bugs
of severity `important' in the BTS fall into this simple category.
What if the bugs are fixed, but being tested on some remote machine,
and won't be done testing for several hours/days? What if the bug
can only be fixed with an upstream patch which hasn't arrived yet?
What if the maintainer *has* fixed the bug, but before he uploads,
he's waiting for a second opinion. What if the maintainer wants people
to fetch a new unstable version which has *probably* fixed the bug,
and it really seems to have, apply the patch to the frozen version?
And so on.
There are many situations where it would be silly, I think, to prevent
the maintainer getting on with work in unstable because the BTS contains
an entry under `important'. If a maintainer has two courses of action:
(i) develop funky stuff in unstable, (ii) fix bugs in frozen, and
you block off one path (the first), it does /not/ necessarily, or
even likely follow that the maintainer will decide to follow path (ii)
instead. More likely the maintainer will go away, and contribute
nothing.
--
Chris <chris@fluff.org> ( http://www.fluff.org/chris )
Reply to: