Re: BTS really broken
John Goerzen wrote:
> For well over two years, the Debian BTS has had the wrong maintainer
> for prc-tools. I reported this bug as #60935 but said bug has been
> duly ignored completely. I am getting highly annoyed with this
> situation of brokenness and non-accountability in the BTS
> maintenance. Can someone please investigate the situation and fix the
> bugs in the bug system?!
Hm.. The bug report says Stephen Zaner <gibreel@debian.org> is the
maintainer.
Master is using stable right now, and has:
joeyh@master:~>dpkg --print-avail prc-tools
Package: prc-tools
Priority: extra
Section: otherosfs
Installed-Size: 9235
Maintainer: John Goerzen <jgoerzen@complete.org>
My work box is running frozen right now, and has:
joey@gumdrop:~>dpkg -p prc-tools
Package: prc-tools
Priority: extra
Section: otherosfs
Installed-Size: 8774
Maintainer: Stephen Zander <gibreel@debian.org>
My home box is running unstable right now, and has:
joey@kite:~>dpkg -p prc-tools
Package `prc-tools' is not available.
Indeed, a grep of a mirror finds prc-tools only has a .orig.tar.gz in
woody, no debs, no diff, no dsc.
/etc/debbugs/Maintainers on master has:
prc-tools John Goerzen <jgoerzen@complete.org>
/org/ftp.debian.org/scripts/masterfiles/mkmaintainers{,.pl} on auric seems
to be what makes this file. The perl script essentially loads up some files
like pseudo-packages.maintainers, then iterates over all of stable, frozen,
unstable, and experimental (these are called "suites" for some reason),
parses all .dsc files files for maintainer info. Each time it gets a
maintainer, it calls this function:
sub setmaint () {
my($p, $newm, $newa, $news, $newv) = @_;
if (! $maint{$p} ||
($news == $suite{$p} || $news != $DI::dist{'experimental'}{'suite'}) &&
($news > $suite{$p} || newer($newv, $version{$p})) &&
(($arch{$p} ne 'any' && $arch{$p} ne 'all') || $newa eq 'any' || $newa eq
'all')) {
$maint{$p} = $newm;
$arch{$p} = $newa;
$suite{$p} = $news;
$version{$p} = $newv;
}
}
Ugh. That if statement really sucks. I think it must be preventing the
maintainer of a package from being seen if a maintainer was already found
earlier, but only in some cases. I haven't entirely puzzled through it yet.
Sorry, I have to get some rest -- hope this helps some.
--
see shy jo
Reply to: