Bug#551521: [UDD] please expose a list of RC-buggy and/or ANY-buggy packages

[George Danchev]

Package: qa.debian.org

Hi,

It would be nice to get impression of the current (topmost) bug numbers, for
instance:

packages with more than 1 open RC-bug
packages with more than 10/100 open bugs (any kind of)

and eventually reports about packages with bugs tagged as 'request for help',
'more info' and 'wontfix'.

thanks
--
pub 4096R/0E4BD0AB <people.fccf.net/danchev/key pgp.mit.edu>



--
To UNSUBSCRIBE, email to debian-bugs-dist-REQUEST@...
with a subject of "unsubscribe". Trouble? Contact listmaster@...

[Lucas Nussbaum]

On 18/10/09 at 22:57 +0200, George Danchev wrote:

> Package: qa.debian.org
>
> Hi,
>
> It would be nice to get impression of the current (topmost) bug numbers, for
> instance:
>
> packages with more than 1 open RC-bug
> packages with more than 10/100 open bugs (any kind of)
>
> and eventually reports about packages with bugs tagged as 'request for help',
> 'more info' and 'wontfix'.

I agree. Maybe you could work on writing the SQL queries for those
lists, and such lists could be provided as a CGI on
http://udd.debian.org/?
--
| Lucas Nussbaum
| lucas@...   http://www.lucas-nussbaum.net/ |
| jabber: lucas@...             GPG: 1024D/023B3F4F |



--
To UNSUBSCRIBE, email to debian-bugs-dist-REQUEST@...
with a subject of "unsubscribe". Trouble? Contact listmaster@...

[George Danchev]

> On 18/10/09 at 22:57 +0200, George Danchev wrote:
> > Package: qa.debian.org
> >
> > Hi,
> >
> > It would be nice to get impression of the current (topmost) bug numbers,
> > for instance:
> >
> > packages with more than 1 open RC-bug
> > packages with more than 10/100 open bugs (any kind of)
> >
> > and eventually reports about packages with bugs tagged as 'request for
> > help', 'more info' and 'wontfix'.
>
> I agree. Maybe you could work on writing the SQL queries for those
> lists, and such lists could be provided as a CGI on
> http://udd.debian.org/?

Yeah, providing thes next to existing scripts is what I meant, and I'm willing
to invest some time after it. Should I download udd.sql.gz, install plperl
(createlang, etc... which might take some time) at my site or I can login to
samosa and use local database there to try some queries?

I also noticed that there are some nice VIEWs at [1], which could be used.

Extracting interesting bug-profiles might need further and broader discussion,
but on my second thoughts I believe that these concerning testing are the
hottest ones, so we are best to start with them first. My revised lists are:

packages in testing with more than 0/5 open RC-bug
packages in testing with more than 10/100 open bugs (any kind of)

packages in testing with bugs tagged as 'request for help', 'more info' and
'wontfix'.

[1] http://udd.debian.org/schema/udd.html
 
--
pub 4096R/0E4BD0AB <people.fccf.net/danchev/key pgp.mit.edu>



--
To UNSUBSCRIBE, email to debian-bugs-dist-REQUEST@...
with a subject of "unsubscribe". Trouble? Contact listmaster@...

[Lucas Nussbaum]

On 20/10/09 at 22:15 +0200, George Danchev wrote:

> > On 18/10/09 at 22:57 +0200, George Danchev wrote:
> > > Package: qa.debian.org
> > >
> > > Hi,
> > >
> > > It would be nice to get impression of the current (topmost) bug numbers,
> > > for instance:
> > >
> > > packages with more than 1 open RC-bug
> > > packages with more than 10/100 open bugs (any kind of)
> > >
> > > and eventually reports about packages with bugs tagged as 'request for
> > > help', 'more info' and 'wontfix'.
> >
> > I agree. Maybe you could work on writing the SQL queries for those
> > lists, and such lists could be provided as a CGI on
> > http://udd.debian.org/?
>
> Yeah, providing thes next to existing scripts is what I meant, and I'm willing
> to invest some time after it. Should I download udd.sql.gz, install plperl
> (createlang, etc... which might take some time) at my site or I can login to
> samosa and use local database there to try some queries?
>
> I also noticed that there are some nice VIEWs at [1], which could be used.
>
> Extracting interesting bug-profiles might need further and broader discussion,
> but on my second thoughts I believe that these concerning testing are the
> hottest ones, so we are best to start with them first. My revised lists are:
>
> packages in testing with more than 0/5 open RC-bug
> packages in testing with more than 10/100 open bugs (any kind of)
>
> packages in testing with bugs tagged as 'request for help', 'more info' and
> 'wontfix'.

The best would be to have some use cases for this data in mind: it's
easy to generate data, harder to generate useful data :-)

See http://wiki.debian.org/UltimateDebianDatabase for information about
how to connect to UDD.
--
| Lucas Nussbaum
| lucas@...   http://www.lucas-nussbaum.net/ |
| jabber: lucas@...             GPG: 1024D/023B3F4F |



--
To UNSUBSCRIBE, email to debian-bugs-dist-REQUEST@...
with a subject of "unsubscribe". Trouble? Contact listmaster@...

[Lucas Nussbaum]

On 20/10/09 at 23:50 +0200, George Danchev wrote:

> > On 20/10/09 at 22:15 +0200, George Danchev wrote:
> > > packages in testing with more than 0/5 open RC-bug
> > > packages in testing with more than 10/100 open bugs (any kind of)
> > >
> > > packages in testing with bugs tagged as 'request for help', 'more info'
> > > and 'wontfix'.
> >
> > The best would be to have some use cases for this data in mind:
>
> Well, the test cases are as following:
>
> * identify the group of packages holding up our release (RC>0)
>
> * identify the group of packages which could be *eventually* removed, if they
> are leaf packages with RC>5; if they are not leaf packages and a fair amount
> of packages depend on them, then we are in real trouble, hence such a trend is
> best to be prevented in advance if at all possible.
>
> The 'request for help' and 'more info' group could be interesting to identify
> since a broader amount of users (not only reporter and maintainer) could
> eventually supply the needed data or hopefully a solution if they knew that in
> the first place.

Well, my point is that identifying such packages is basically useless if
no action is taken. So, at some point, it's a good to think of ways this
data could be used as a basis for QA work.

You might be interested in taking a look at bapase
<http://udd.debian.org/cgi-bin/bapase.cgi>. It combines different kind
of data to identify "interesting" packages, and is coupled with a
process to orphan or remove those packages. (work around bapase hasn't
been very active recently, though)
--
| Lucas Nussbaum
| lucas@...   http://www.lucas-nussbaum.net/ |
| jabber: lucas@...             GPG: 1024D/023B3F4F |



--
To UNSUBSCRIBE, email to debian-qa-REQUEST@...
with a subject of "unsubscribe". Trouble? Contact listmaster@...

[George Danchev]

> On 20/10/09 at 22:15 +0200, George Danchev wrote:
> > packages in testing with more than 0/5 open RC-bug
> > packages in testing with more than 10/100 open bugs (any kind of)
> >
> > packages in testing with bugs tagged as 'request for help', 'more info'
> > and 'wontfix'.
>
> The best would be to have some use cases for this data in mind:

Well, the test cases are as following:

* identify the group of packages holding up our release (RC>0)

* identify the group of packages which could be *eventually* removed, if they
are leaf packages with RC>5; if they are not leaf packages and a fair amount
of packages depend on them, then we are in real trouble, hence such a trend is
best to be prevented in advance if at all possible.

The 'request for help' and 'more info' group could be interesting to identify
since a broader amount of users (not only reporter and maintainer) could
eventually supply the needed data or hopefully a solution if they knew that in
the first place.

> it's easy to generate data, harder to generate useful data :-)

I don't believe that ;-)
 
> See http://wiki.debian.org/UltimateDebianDatabase for information about
> how to connect to UDD.

Thanks.
 
--
pub 4096R/0E4BD0AB <people.fccf.net/danchev/key pgp.mit.edu>



--
To UNSUBSCRIBE, email to debian-bugs-dist-REQUEST@...
with a subject of "unsubscribe". Trouble? Contact listmaster@...

[Rene Engelhard]

Hi,

On Sun, Oct 18, 2009 at 10:57:26PM +0200, George Danchev wrote:
> packages with more than 10/100 open bugs (any kind of)

That is a nonsensical measure. Big packages have many bugs. Many of them
are upstream bugs. It doesn't make sense to use a hardcoded value like this.

> and eventually reports about packages with bugs tagged as 'request for help',
> 'more info' and 'wontfix'.

True for the last ones. I disagree about wontfix, that will just prompt
people to close them if they get a useless message about that.

Grüße/Regards,

Rene
--
 .''`.  René Engelhard -- Debian GNU/Linux Developer
 : :' : http://www.debian.org | http://people.debian.org/~rene/
 `. `'  rene@... | GnuPG-Key ID: D03E3E70
   `-   Fingerprint: E12D EA46 7506 70CF A960 801D 0AA0 4571 D03E 3E70



--
To UNSUBSCRIBE, email to debian-bugs-dist-REQUEST@...
with a subject of "unsubscribe". Trouble? Contact listmaster@...

[George Danchev]

> Hi,
>
> On Sun, Oct 18, 2009 at 10:57:26PM +0200, George Danchev wrote:
> > packages with more than 10/100 open bugs (any kind of)
>
> That is a nonsensical measure. Big packages have many bugs.

Debian archive contains more small packages than large ones, so it makes sense
to catch these with many bugs opened, doesn't it?

> Many of them are upstream bugs.

What difference does it make, they are still bugs.

> It doesn't make sense to use a hardcoded value like  this.

That could be any BigInt or user-defined of course.
 
> > and eventually reports about packages with bugs tagged as 'request for
> > help', 'more info' and 'wontfix'.
>
> True for the last ones. I disagree about wontfix, that will just prompt
> people to close them if they get a useless message about that.

wontfix could be abused, so it might be interesting to know how popular it is.

--
pub 4096R/0E4BD0AB <people.fccf.net/danchev/key pgp.mit.edu>



--
To UNSUBSCRIBE, email to debian-bugs-dist-REQUEST@...
with a subject of "unsubscribe". Trouble? Contact listmaster@...