[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Mass bugs filing: bogus debian/watch files

On 05/11/07 at 16:19 -0600, Raphael Geissert wrote:
> Lucas Nussbaum wrote:
> > I like the idea, and think that you should go ahead with reporting the
> > bugs.
> > 
> > However, if you are interested in that stuff, it could be a good idea to
> > contact the DEHS maintainer: there might be a way to integrate this into
> > DEHS, and avoid the HTML parsing ;)
> I've already tough about that, but first of all I would like to receive some
> kind of acception before I start mass filing bug reports :)

There isn't really any kind of "official" acceptation in Debian. By
asking on -devel@, you basically did everything you could, and if people
complain, you can always say "see, I asked on -devel@, nobody
complained." This works quite well for my monthly emails. ;)

> (and if no one
> else has something to say I'll let my scripts send all the reports tomorrow
> afternoon, GMT-6).


> I also need to find a way to detect when a bug report has been closed and
> automatically remove it from the internal database used by my scripts.
> It would of course be easier if there was some other kind of way to retrieve
> information from the BTS other than email and SOAP.

You don't really need that: working with usertags is probably enough.
For archive rebuilds, my workflow is:
 rebuild n:
   get a list of failures
   file all the bugs, and tag them with qa-ftbfs-200710XX and qa-ftbfs
   using the usertag and a script, automatically add the bug numbers
     to my list of failures
 rebuild n+1:
   get a new list of failures
   using a script, merge the old list of failures, and the new list of
     failures, so I won't file duplicate bugs
   using a script, tag all the bugs in this list. This way, I can look
     at the BTS and see if all bugs that should be open are still open
     (no bugs wrongly closed).
   file all the new bugs, and tag them with qa-ftbfs-200711YY and
   using a the usertag and a script, automatically add the bug numbers
     to my list of failures

That workflow allows to make sure that:
 * all failures are reported
 * if bugs are wrongly closed, I can notice it and reopen them

The only thing remaining is to deal with failures that were fixed
without the bug being closed (for example, when the failure is caused by
another package). To do that, from time to time, I look at all the
packages with a qa-ftbfs-tagged bug, and compare that with my list of
known failures. I don't think I have a script for that: I simply write
something ad-hoc using the SOAP interface ;)

You can use my scripts for all those tasks: everything is in the
collab-qa svn repository on alioth, but isn't very well docummented.

  archive-rebuilds/merge-results.rb: merge two results files
  archive-rebuilds/mass-tag.rb: extracts bug numbers from a file and
    prepare a mail to request@bugs.d.o to usertag them.
  tools/bugs2packagelist/fetchbugs.rb: replace NNN with bug numbers
| Lucas Nussbaum
| lucas@lucas-nussbaum.net   http://www.lucas-nussbaum.net/ |
| jabber: lucas@nussbaum.fr             GPG: 1024D/023B3F4F |

Reply to: