|
Home > Archive > Debian Developers > April 2004 > Raising severity of Contains /usr/share/info/dir.gz if rebuilt on current sid"-bu
You are viewing an archived Text-only version of the thread.
To view this thread in it's original format and/or if you want to reply to
this thread please [click here]
| Author |
Raising severity of Contains /usr/share/info/dir.gz if rebuilt on current sid"-bu
|
|
| Andreas Metzler 2004-03-17, 9:42 am |
| Hello,
About 4 months ago I digged through Debian to find all packages which
would contain /usr/share/info/dir.gz if rebuilt (see #214769 for
details.) and submitted (iirc) about 30 bugs (about 20 still open)
with Severity important. Every single one of these bugs has patch.
I.e. any of these package has a RC bug if it is recompieled without
fixing this bug. Today an NMU happened (anubis) which did exactly
this. To stop this from happening again I want to raise the severity
to serious, but not without asking here, I have CCed release but want
discussion to happen on -devel.
#218471 anubis
#218478 bison
#218479 bogosort
#218483 bookview
#218941 fsh
#218946 gawk
#218964 gnuplot-doc
#218976 texinfo
#218978 jwhois
#219599 libgmp3-doc
#219607 lightning
#219609 mdk
#219612 ndtpd
#219988 ssed
#219990 stow
#219991 texi2html
#219994 which
anubis Janusz A. Urbanowicz <alex@bofh.net.pl>
bison Vincent Renardias <vincent@debian.org>
bogosort Michael Janssen <jamuraa@debian.org>
bookview Susumu OSAWA <susumuo@debian.org>
fsh Tommi Virtanen <tv@debian.org>
gawk James Troup <james@nocrew.org>
texinfo Josip Rodin <joy-packages@debian.org>
jwhois Peter Karlsson <peterk@debian.org>
lightning Marco Kuhlmann <mk@debian.org>
mdk Baruch Even <baruch@debian.org>
ndtpd Masayuki Hatta <mhatta@debian.org>
ssed Ganesan Rajagopal <rganesan@debian.org>
stow Guillaume Morin <gemorin@debian.org>
texi2html Ivo Timmermans <ivo@debian.org>
which Ian Zimmerman <itz@speakeasy.org>
http://bugs.debian.org/cgi-bin/pkgr...ic.univie.ac.at
--
NMUs aren't an insult, they're not an attack, and they're
not something to avoid or be ashamed of.
Anthony Towns in 2004-02 on debian-devel
--
To UNSUBSCRIBE, email to debian-devel-request@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
| |
| Andreas Barth 2004-03-17, 9:43 am |
| Hi,
* Andreas Metzler (ametzler@downhill.at.eu.org) [040317 11:10]:
> I.e. any of these package has a RC bug if it is recompieled without
> fixing this bug. Today an NMU happened (anubis) which did exactly
> this. To stop this from happening again I want to raise the severity
> to serious, but not without asking here, I have CCed release but want
> discussion to happen on -devel.
I'd say: just go ahead. It's definitly required that a package is
recompilable.
Cheers,
Andi
--
http://home.arcor.de/andreas-barth/
PGP 1024/89FB5CE5 DC F1 85 6D A6 45 9C 0F 3B BE F1 D0 C5 D1 D9 0C
--
To UNSUBSCRIBE, email to debian-devel-request@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
| |
| Josip Rodin 2004-03-17, 9:43 am |
| On Wed, Mar 17, 2004 at 10:45:56AM +0100, Andreas Metzler wrote:
> I.e. any of these package has a RC bug if it is recompieled without
> fixing this bug.
> #218976 texinfo
The above assertion is incorrect in that case, as evidenced in the bug log. 
--
2. That which causes joy or happiness.
--
To UNSUBSCRIBE, email to debian-devel-request@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
| |
| Andreas Metzler 2004-03-17, 9:43 am |
| On Wed, Mar 17, 2004 at 10:45:56AM +0100, Andreas Metzler wrote:
> About 4 months ago I digged through Debian to find all packages which
> would contain /usr/share/info/dir.gz if rebuilt
[...]
> Today an NMU happened (anubis) which did exactly
> this. To stop this from happening again
[...]
James Troup pointed out on IRC that this reasoning ("botched NMUs
break stuff") is not very convincing, I want to change it to: "We want
to be able to recompile sarge without introducing RC-bugs just by
rebuilding, don't we?
On a sidenote somebody proposed a lintian test for this. - This won't
work usually because the bug only pops up if /usr/sbin is in your
$PATH when you compile the package (most of us have not but the
autobuilders do). - So even if *your* home-compiled package checks out
ok with lintian and debdiff the autobuilt one will be broken.
cu andreas
--
To UNSUBSCRIBE, email to debian-devel-request@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
| |
| Andreas Metzler 2004-03-17, 9:43 am |
| retitle 218976 Contains /usr/share/info/dir.old.gz if rebuilt on current sid
thanks
On Wed, Mar 17, 2004 at 12:20:08PM +0100, Josip Rodin wrote:
> On Wed, Mar 17, 2004 at 10:45:56AM +0100, Andreas Metzler wrote:
[color=darkred]
[color=darkred]
> The above assertion is incorrect in that case, as evidenced in the
> bug log.
texinfo indeed would not contain /usr/share/info/dir.gz but it would
contain /usr/share/info/dir.old.gz which would be a RC, afaik.
cu andreas
--
To UNSUBSCRIBE, email to debian-devel-request@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
| |
| Anthony Towns 2004-03-17, 9:43 am |
| On Wed, Mar 17, 2004 at 10:45:56AM +0100, Andreas Metzler wrote:
> About 4 months ago I digged through Debian to find all packages which
> would contain /usr/share/info/dir.gz if rebuilt (see #214769 for
> details.) and submitted (iirc) about 30 bugs (about 20 still open)
> with Severity important. Every single one of these bugs has patch.
> NMUs aren't an insult, they're not an attack, and they're
> not something to avoid or be ashamed of.
I can't see any reason not to just NMU them instead. There's no need to
drop the packages as they are.
Cheers,
aj
--
Anthony Towns <aj@humbug.org.au> <http://azure.humbug.org.au/~aj/>
I don't speak for anyone save myself. GPG signed mail preferred.
Linux.conf.au 2004 -- Because we could.
http://conf.linux.org.au/ -- Jan 12-17, 2004
--
To UNSUBSCRIBE, email to debian-devel-request@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
| |
| Josip Rodin 2004-03-17, 9:43 am |
| On Wed, Mar 17, 2004 at 12:37:26PM +0100, Andreas Metzler wrote:
>
>
>
> texinfo indeed would not contain /usr/share/info/dir.gz but it would
> contain /usr/share/info/dir.old.gz which would be a RC, afaik.
Um, how? What could possibly be bothered by dir.old*?
--
2. That which causes joy or happiness.
--
To UNSUBSCRIBE, email to debian-devel-request@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
| |
| Andreas Metzler 2004-03-17, 4:39 pm |
| On Wed, Mar 17, 2004 at 03:22:47PM +0100, Josip Rodin wrote:
> On Wed, Mar 17, 2004 at 12:37:26PM +0100, Andreas Metzler wrote:
[color=darkred]
[color=darkred]
[color=darkred]
>
> Um, how? What could possibly be bothered by dir.old*?
Any other package with this bug. (Trying to overwrite ... which is
also in....)
cu andreas
--
To UNSUBSCRIBE, email to debian-devel-request@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
| |
| Andreas Metzler 2004-03-17, 4:39 pm |
| On Wed, Mar 17, 2004 at 09:09:12PM +1000, Anthony Towns wrote:
> On Wed, Mar 17, 2004 at 10:45:56AM +0100, Andreas Metzler wrote:
[color=darkred]
[color=darkred]
> I can't see any reason not to just NMU them instead. There's no need to
> drop the packages as they are.
I did not suggest removal. I've prepared NMUs for the first four items
on the list which I'll upload on monday, though.
cu andreas
--
To UNSUBSCRIBE, email to debian-devel-request@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
| |
| Andreas Metzler 2004-03-17, 4:40 pm |
| On Wed, Mar 17, 2004 at 10:45:56AM +0100, Andreas Metzler wrote:
> Hello,
> About 4 months ago I digged through Debian to find all packages which
> would contain /usr/share/info/dir.gz if rebuilt (see #214769
[...]
After reading the thread and finding another borked NMU I have decided
to raise the severity and NMU if the maintainer does not react.
> #218471 anubis
already rc.
> #218478 bison
NMU pending
> #218479 bogosort
NMU pending
> #218483 bookview
NMU pending
> #218941 fsh
NMU pending
> #218946 gawk
> #218964 gnuplot-doc
NMU pending
> #218976 texinfo
> #218978 jwhois
> #219599 libgmp3-doc
> #219607 lightning
already rc.
> #219609 mdk
> #219612 ndtpd
> #219988 ssed
> #219990 stow
> #219991 texi2html
> #219994 which
[...]
cu andreas
--
To UNSUBSCRIBE, email to debian-devel-request@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
| |
| Anthony Towns 2004-03-18, 9:41 am |
| On Wed, Mar 17, 2004 at 03:46:50PM +0100, Andreas Metzler wrote:
> On Wed, Mar 17, 2004 at 09:09:12PM +1000, Anthony Towns wrote:
> I did not suggest removal.
Yes, you did. That's the only difference between RC bugs and non-RC
bugs: packages with the former don't get released (ignoring exceptional
circumstances, which this isn't).
Cheers,
aj
--
Anthony Towns <aj@humbug.org.au> <http://azure.humbug.org.au/~aj/>
I don't speak for anyone save myself. GPG signed mail preferred.
Linux.conf.au 2004 -- Because we could.
http://conf.linux.org.au/ -- Jan 12-17, 2004
| |
| Anthony Towns 2004-03-18, 9:41 am |
| On Wed, Mar 17, 2004 at 11:59:04AM +0100, Andreas Barth wrote:
> * Andreas Metzler (ametzler@downhill.at.eu.org) [040317 11:10]:
> I'd say: just go ahead. It's definitly required that a package is
> recompilable.
Uh, no, strictly speaking it's not. We're getting closer to achieving
that, but we haven't achieved it yet. Our (re)buildability requirements
are still relatively limited.
Cheers,
aj
--
Anthony Towns <aj@humbug.org.au> <http://azure.humbug.org.au/~aj/>
I don't speak for anyone save myself. GPG signed mail preferred.
Linux.conf.au 2004 -- Because we could.
http://conf.linux.org.au/ -- Jan 12-17, 2004
| |
| Matt Kraai 2004-03-19, 9:38 am |
| On Thu, Mar 18, 2004 at 02:44:12PM +1000, Anthony Towns wrote:
> On Wed, Mar 17, 2004 at 11:59:04AM +0100, Andreas Barth wrote:
>
> Uh, no, strictly speaking it's not. We're getting closer to achieving
> that, but we haven't achieved it yet. Our (re)buildability requirements
> are still relatively limited.
Does that mean that FTBFS bugs found by a pbuilder run should not
be considered release-critical?
--
Matt Kraai kraai@ftbfs.org http://ftbfs.org/
--
To UNSUBSCRIBE, email to debian-devel-request@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
| |
| Anthony Towns 2004-03-19, 9:39 am |
| On Thu, Mar 18, 2004 at 08:08:56PM -0800, Matt Kraai wrote:
> On Thu, Mar 18, 2004 at 02:44:12PM +1000, Anthony Towns wrote:
> Does that mean that FTBFS bugs found by a pbuilder run should not
> be considered release-critical?
No; we've got a few "buildability" requirements that are required, and
a few that still aren't. The aim is to work towards having them all be
guaranteed; but we don't really have enough time left over from fixing
all the things we're currently trying to guarantee to add new things.
Expecting things to autobuild correctly first time round, expecting
things to stay buildable in unstable (although maybe not building the
exactly correct thing), and having roughly the complete source code for
everything are the basic requirements we have at the moment.
Cheers,
aj
--
Anthony Towns <aj@humbug.org.au> <http://azure.humbug.org.au/~aj/>
I don't speak for anyone save myself. GPG signed mail preferred.
Linux.conf.au 2004 -- Because we could.
http://conf.linux.org.au/ -- Jan 12-17, 2004
| |
| Wookey 2004-03-23, 10:35 am |
| +++ Adrian Bunk [04-03-18 11:46 +0100]:
> On Wed, Mar 17, 2004 at 11:59:04AM +0100, Andreas Barth wrote:
>
> I'd like to agree with you.
>
> But the first thing to achieve this would be to fix testing to honor
> build dependencies...
As the subject has come up I'd just like to add that the Embedded Debian
sub-project would _really really_ like to see the testing migration rules
changed to enforce buildability.
The fact that Debian can't build itself at any given time is a real problem
for emdebian, because that really does need to be (cross) buildable from
source at all times.
We can deal with this for the time being by maintaining our own emdebian
tree which does enforce buildability in its migration rules, but this has
it's own problems (getting out of sync).
I hope that everyone can agree that it should be a goal of Debian's to be
able to build itself, and that we can move towards making that policy. Can
someone tell me why we don't already do this - presumably there is some
reason of expediency?
[please cc: me]
Wookey
--
Aleph One Ltd, Bottisham, CAMBRIDGE, CB5 9BA, UK Tel +44 (0) 1223 811679
work: http://www.aleph1.co.uk/ play: http://www.chaos.org.uk/~wookey/
--
To UNSUBSCRIBE, email to debian-devel-request@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
| |
| Colin Watson 2004-03-23, 11:38 am |
| On Tue, Mar 23, 2004 at 01:31:03PM +0000, Wookey wrote:
> +++ Adrian Bunk [04-03-18 11:46 +0100]:
>
> As the subject has come up I'd just like to add that the Embedded Debian
> sub-project would _really really_ like to see the testing migration rules
> changed to enforce buildability.
>
> The fact that Debian can't build itself at any given time is a real problem
> for emdebian, because that really does need to be (cross) buildable from
> source at all times.
>
> We can deal with this for the time being by maintaining our own emdebian
> tree which does enforce buildability in its migration rules, but this has
> it's own problems (getting out of sync).
>
> I hope that everyone can agree that it should be a goal of Debian's to be
> able to build itself, and that we can move towards making that policy. Can
> someone tell me why we don't already do this - presumably there is some
> reason of expediency?
Not really. It's a change that (IMHO) really needs to be made at the
start of a release cycle rather than at the end, though, so that we have
the proper time to fix everything and don't introduce nasty surprises in
bits of testing migration that we need to happen; and it didn't happen
early in the sarge cycle. Early sarge+1 seems like the right time to
start enforcing this.
--
Colin Watson [cjwatson@flatline.org.uk]
--
To UNSUBSCRIBE, email to debian-devel-request@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
| |
| Anthony Towns 2004-03-23, 10:34 pm |
| On Tue, Mar 23, 2004 at 01:31:03PM +0000, Wookey wrote:
> I hope that everyone can agree that it should be a goal of Debian's to be
> able to build itself, and that we can move towards making that policy. Can
> someone tell me why we don't already do this - presumably there is some
> reason of expediency?
Basically just that it's already difficult enough to keep testing flowing
smoothly that adding more friction seems like a bad idea. Getting the maximum
time-to-fix of RC bugs down to a month or at most two really needs to happen
before we start doing other major things like this.
We're doing well enough at the moment that this is probably worth
reconsidering, but probably too much of that smoothness is due solely
to the impending release and won't last once sarge is out the door.
> [please cc: me]
(Please look into setting your Mail-Followup-To: header)
Cheers,
aj
--
Anthony Towns <aj@humbug.org.au> <http://azure.humbug.org.au/~aj/>
I don't speak for anyone save myself. GPG signed mail preferred.
Linux.conf.au 2004 -- Because we could.
http://conf.linux.org.au/ -- Jan 12-17, 2004
| |
| Andreas Metzler 2004-03-30, 6:36 am |
| Andreas Metzler <ametzler@logic.univie.ac.at> wrote:
> On Wed, Mar 17, 2004 at 10:45:56AM +0100, Andreas Metzler wrote:
> [...]
> After reading the thread and finding another borked NMU I have decided
> to raise the severity and NMU if the maintainer does not react.
[...]
Almost[1] all the occurrences of this bug that I have found[2] have
been dealt with either by a maintaner-upload or a NMU.
cu andreas
[1] There is a singele one open but the maintainer did not want me to
NMU.
[2] It is possible that other occurences of this bug are still hidden,
especially if the package uses a
tarball-in-tarball-without-standard-way-to-unpack-the-50-different-implementations
system.
--
NMUs aren't an insult, they're not an attack, and they're
not something to avoid or be ashamed of.
Anthony Towns in 2004-02 on debian-devel
--
To UNSUBSCRIBE, email to debian-devel-request@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
| |
| Anthony Towns 2004-04-02, 3:37 am |
| On Tue, Mar 30, 2004 at 12:26:24PM +0200, Andreas Metzler wrote:
> Almost[1] all the occurrences of this bug that I have found[2] have
> been dealt with either by a maintaner-upload or a NMU.
That sure makes raising the severity pretty plausible now.
Cheers,
aj
--
Anthony Towns <aj@humbug.org.au> <http://azure.humbug.org.au/~aj/>
I don't speak for anyone save myself. GPG signed mail preferred.
Linux.conf.au 2004 -- Because we could.
http://conf.linux.org.au/ -- Jan 12-17, 2004
|
|
|
|
|