Bug#695508: udd: bugs.cgi - bugs filed against multiple packages is not associated with a source
On 10/12/12 at 00:20 +0100, Ivo De Decker wrote:
> Hi Niels,
>
> On Sun, Dec 09, 2012 at 01:22:39PM +0100, Niels Thykier wrote:
> > The bug #694368 is filed against libfuzzy2, libfuzzy-dev (src:
> > ssdeep). At least in the view [1], this means that the "package"
> > column gets the PTS link wrong. It links to p.qa.d.o/libfuzzy2 and
> > p.qa.d.o/libfuzzy-dev (respectively), but it should have linked to
> > p.qa.d.o/ssdeep in both cases. This is not a huge problem as the
> > PTS fixes this with a redirect, but ...
> >
> > I suspect also causes bugs.cgi not discovering the unblock hint for
> > ssdeep. On a related note; as ssdeep is unblocked the package will
> > problably migrate in a week (causing the example disappear from the
> > view). Let me know if I should find more information before then.
>
> This bug isn't caused by the cgi, but by the import script. In
> bugs_gatherer.pl (l 218):
>
> my $srcpkg;
> if ($bug{package} =~ /^src:(.*)/)
> {
> $srcpkg = $1;
> } else {
> $srcpkg = exists($pkgsrc{$bug{package}}) ? $pkgsrc{$bug{package}} : $bug{package};
> }
>
> For the bug mentioned, the package is "libfuzzy2,libfuzzy-dev". There is no
> corresponding source package, so in the database the source package also is
> "libfuzzy2,libfuzzy-dev". The package should be split in parts and those parts
> should be converted to source packages.
>
> I could try to create a patch, but I don't know where to get the data to do
> the test, so I'll leave that to someone else.
Hi,
If you could write a basic patch, I could try to test it myself.
Alternatively, you could get the BTS data using rsync:
$ rsync bugs-mirror.debian.org::
bts-spool-db [bugs-mirror.debian.org] active bug spool
bts-spool-archive [bugs-mirror.debian.org] archived bug spool
bts-spool-index [bugs-mirror.debian.org] bug index files
bts-versions [bugs-mirror.debian.org] bts package version information
But I'm not sure it's worth the trouble.
> If the import is fixed, the cgi probably still won't do the right thing if the
> bug is filed against different source packages (or binary packages in
> different source packages). I don't know if there are examples like that.
There are, unfortunately.
Lucas
Reply to: