Re: co-mentor for a GSoC proposal wanted: debbugs web submission
On Fri, Mar 16, 2007 at 01:01:06PM -0700, Don Armstrong <don@debian.org> wrote:
> On Fri, 16 Mar 2007, Josselin Mouette wrote:
> > Le jeudi 15 mars 2007 à 20:34 -0400, Roberto C. Sanchez a écrit :
> > > What will a web interface provide that cannot be found/done by browsing
> > > http://bugs.debian.org/src:<pkgname> ?
> >
> > A way for us to manage our 1600 bugs?
>
> You can do this by using usertags, and then querying for things like:
>
> bugs.debian.org/cgi-bin/pkgreport.cgi?pkg=foopkg&tag=accessiblity; or
> similar.
>
> I personally haven't run into a case where it'd be useful to just show
> an overview page of the separation of bugs into categories, but that's
> just because I personally haven't dealt with packages with more than
> around 300 bugs.
>
> If you want features like this, the right thing to do is file wishlist
> bugs against debbugs or bugs.debian.org instead of complaining on
> -devel or worse, silently suffering.
The problem with filing a bug is that it is one person talking of his
own problem, while I think it would be much better if we all talked
about it.
Being one of those developpers who have to deal with a huge backlog of
bugs, I can tell you debbugs doesn't help up. At all.
And usertags are far from being a solution. Because of what they are:
tags.
Sure, they can help in some ways, but that really not enough.
And normal tags have the same problem too: they are tags.
The problem the BTS has with tags, is that they have no meaning, and
that doesn't help when you want to filter, in or out, bugs, to make
triaging more efficient. Bugzilla is much better on this matter.
I already hear someone saying "oh but tags do have a meaning, read
http://www.debian.org/Bugs/Developer";
Well, okay, let's take a look at this page, and to the definition of the
most interesting ones to deal with bug backlog and unresponsive
reporters:
- moreinfo
This bug can't be addressed until more information is provided by
the submitter. (...)
- unreproducible
This bug can't be reproduced on the maintainer's system. Assistance
from third parties is needed in diagnosing the cause of the problem.
- confirmed
The maintainer has looked at, understands, and basically agrees with
the bug, but has yet to fix it. (...)
A combination of moreinfo and unreproducible is understandable. Any
other combination is meaningless and fails in every possible way to make
a bug identificable from the summary page, which you really would like
to be the case when you have to deal with hundreds of bugs.
Yet, any of these combinations are possible, because tags are tags, and
have no damn meaning. Setting a tag doesn't unset another one when these
tags'meaning are antagonists.
And yes, it happens very easily, look at #310882, #359965, and #342305.
And there are dozens of these kind of tagging.
I'll stop here for today, but this is far from being the sole PITA with
using debbugs.
Now, the thing is, someone could file a bug about this, but until we all
decide what would be best for us all to deal with our bugs, what painful
experiences we have with debbugs, nothing interesting is going to happen.
There are a lot of possibilities to improve the situation, but all might
not be suitable for all use cases.
This is why it is better to have a thread on -devel now, than a useless
bug opened on bugs.debian.org.
Mike
Reply to: