Re: woody is getting worse...
Branden Robinson wrote:
Don't worry, I do indeed regret that this bug made it into testing, if
for no other reason than that all the personal attacks are exasperating.
I have not seen any personal attacks against you, but of course I
can only read debian-devel. In any case, if this is the only
reason you regret the propagation of this bug, then that is
unfortunate. I don't believe that's true, but your wording here
implies it.
For bugs to be merged, they have to have the same severity. The
original bug was filed as "important", which isn't RC. I downgraded it
to normal, which people may or may not agree with, but neither in its
When subsequent bugs were filed, I merged them with the earliest report
so that they would all be closed automatically when 4.1.0-8 was
dinstalled. After all, it's good not to have erroneous bugs cluttering
the BTS, right?
Well, maybe not. Maybe people would rather bug severities were never
downgraded, and bugs never merged. That way every bug submitter is
I would say rather that bugs should be merged to the highest severity
of the group, not the lowest (or the first, or whatever).
empowered to exercise a single-handed veto on the progress of any given
package into woody.
If someone is clearly trying to make trouble, those bugs can easily
be closed. Ok, maybe someone could theoretically launch a DoS against
a package, but that possibility exists now, regardless of merging
procedures.
Alternatively, such matters can be left to the maintainer's discretion,
and when his opinion is disagreed with, people can upgrade the bug in
conjunction with a persuasive argument. In this case, nobody bothered.
Well, in a way they did by filing equivalent bugs with higher
severity. I agree this is non-optimal, but there it is. As was
mentioned earlier in this thread, searching bugs is quite painful.
in such a state. Debian is a collaborative, volunteer project, and if
we care only about the little fiefdom that is our own package running on
our own architecture, we're depriving ourselves of some excellent
synergistic effects.
Agreed. But see my point above about duplicate bugs with higher
severity.
I left the bug at normal severitity because I:
1) thought that users of testing and unstable, by and large, read
debian-devel;
A reasonable assumption, but not true in all cases. There are many
days when I just don't have time. That said, I _always_ check
d-d before a dist-upgrade that affects critical components, and
fortunately I caught this bug before it caused me any problems.
Even so, there is some latency between noticing a problem and
finding a cause. In that time many users can be affected and
there is little to no information available.
2) thought that users of testing and unstable, by and large, bothered to
check for the existence of previously filed bugs on a subject;
Also reasonable, though the bug-searching interface is non-optimal.
3) thought that users of testing and unstable, by and large, wouldn't
get the screaming willies at the notion of editing 2 characters out of a
two-line conffile;
That's not the issue. The issue is that RC-bugs should be
investigated. Obviously you did that. I understand your
explanation about downgrading them to merge, I just feel
that's the wrong approach. The original bug should have
been _upgraded_ to merge.
4) thought that users of testing and unstable, by and large, would take
the initiative to challenge my assessment of the bug's severity before
the package went into woody, if they sincerely disagreed with that
assessment.
I believe they did by filing duplicate bugs. Again, non-ideal,
but you can't claim _nothing_ was done.
experience will inform my future decisions. I did not appreciate the
degree to which testing had supplanted potato as an end-users'
distribution, in which almost complete faith is placed in the developers
to hammer all the signficiant bugs out of a package release in less than
10 days.
Unfortunately, it's really the only option users have if they
want any sort of up-to-date system. This is bad for many, many
reasons. I won't go on about long release cycles because this
is a volunteer effort. But to expect every non-developer to stay
at stable is simply unreasonable.
-David
--
"Some little people have music in them, but Fats, he was all music,
and you know how big he was." -- James P. Johnson
Reply to: