y ask keyring-maint to replace your key with a
new one. :) Not to mention that it weakens the web of trust for no very
good reason.
Cheers,
--
Colin Watson [EMAIL PROTECTED]
On Thu, Jan 02, 2003 at 12:08:44PM +0100, Marcel Kolaja wrote:
> On Thu, Jan 02, 2003 at 05:01:46AM +0000, Colin Watson wrote:
> > On Thu, Jan 02, 2003 at 01:01:26AM +0100, Marcel Kolaja wrote:
> > > Note: Please Cc me, if you reply. I am not subscribed to the list. Thank
eefonts | (ttf-larabie-straight,
ttf-larabie-deco)' and you get:
Depends: ttf-freefonts | ttf-larabie-straight, ttf-freefonts |
ttf-larabie-deco
Cheers,
--
Colin Watson [EMAIL PROTECTED]
.
In general -mentors is not a good place to discuss bugs, unless they're
in your own package and you need help fixing them. The bug tracking
system is usually better than any list for this.
Cheers,
--
Colin Watson [EMAIL PROTECTED]
are probably bugs, although it would require an epoch to
fix them. It seems like at best luck that they work with the current
packaging tools, which aren't obliged to implement anything more than
what the version number specification describes.
Cheers,
--
Colin Watson [EMAIL PROTECTED]
nderstandably
> busy maintainer, he has been helpful enough already.
Oh well. :)
--
Colin Watson [EMAIL PROTECTED]
keys.debian.org -- isn't
> it enough?
keyring.debian.org is not a public keyserver, in the sense that unless
your upload refers to a Debian developer's key it will be ignored.
Cheers,
--
Colin Watson [EMAIL PROTECTED]
l
need to be familiar with how symbol versioning works, though, and
frankly I suspect it would be a lot of effort for little gain. People
with those skills would be better occupied fixing the bugs in glibc
instead.
Cheers,
--
Colin Watson [EMAIL PROTECTED]
ore fragile, and should be avoided unless absolutely
necessary; you shouldn't use them just to "promote" a package from
non-free to contrib.
Cheers,
--
Colin Watson [EMAIL PROTECTED]
. Colin, is that possible?
Not for tags in general, but lists for 'help', 'security', and
'unreproducible' are linked off http://qa.debian.org/ (the third set of
links under "Work needed").
Cheers,
--
Colin Watson [EMAIL PROTECTED]
n.org were wired into the PTS, though.)
--
Colin Watson [EMAIL PROTECTED]
e a tighter versioned dependency on perl.
--
Colin Watson [EMAIL PROTECTED]
on to address so they will ask ftpmaster to
> drop it?
That would do no harm, but really the QA group now gets to decide, so
you could just reference this discussion in a bug against
ftp.debian.org. I agree that libiniconf-perl should be removed.
Cheers,
--
Colin Watson [EMAIL PROTECTED]
t (currently) attempt to calculate general module
dependencies, only core dependencies.
Cheers,
--
Colin Watson [EMAIL PROTECTED]
in debian/control. The binary-indep target in
debian/rules can then just copy everything into the right place under
debian/$package.
Cheers,
--
Colin Watson [EMAIL PROTECTED]
#x27;ll be moot, so thanks again.
>
> Some packages send reminder mail to the root during install.
That's usually due to debconf. Keep debconf notes to a minimum, please.
--
Colin Watson [EMAIL PROTECTED]
On Sat, Feb 01, 2003 at 12:57:31PM +, Roger Leigh wrote:
> Colin Watson <[EMAIL PROTECTED]> writes:
> > dh_perl doesn't (currently) attempt to calculate general module
> > dependencies, only core dependencies.
>
> OK, thanks.
>
> I guess that, due to
tives for virtual packages should come first, not
second. As somebody else said, though, this is probably a false
positive.
--
Colin Watson [EMAIL PROTECTED]
ould not be the
> right thing. Therefore I will declare this to be a false positive
> since there is a real package rsh-client it should not create a
> warning.
It's a false positive, yes. Please file it as a bug against lintian
(also compare #179614).
Cheers,
--
Colin Watson [EMAIL PROTECTED]
larly if you
affect binary compatibility by mistake. Dealing with versioning of the
same package is also problematic, particularly if the upstream author
hasn't been careful.
http://www.netfort.gr.jp/~dancer/column/libpkg-guide/libpkg-guide.html
goes into many of the issues invol
ps) and you'll see italics.
(.BI is alternating bold and italics rather than bold and italics
simultaneously. I'm not sure from your post whether you knew that.)
Cheers,
--
Colin Watson [EMAIL PROTECTED]
going to be a name clash between your pages and somebody
else's. If that's not the case, extensions do no harm but aren't needed.
Cheers,
--
Colin Watson [EMAIL PROTECTED]
ble, then that would be good.
--
Colin Watson [EMAIL PROTECTED]
ld fail for some people right out of the box.
I think there was a technical committee discussion about this ...
Ah, yes. #119517.
Cheers,
--
Colin Watson [EMAIL PROTECTED]
n-free. The sort of things that disqualify a package from inclusion
into non-free are more like "may not redistribute to third parties" or
"may not distribute together with our competitors' software".
--
Colin Watson [EMAIL PROTECTED]
number, so that the program can
automatically regenerate the cache whenever it finds that its version is
different from that of the cache?
--
Colin Watson [EMAIL PROTECTED]
sh$'`
That will be wrong in the event that somebody has run 'dselect update'
but not yet upgraded.
Cheers,
--
Colin Watson [EMAIL PROTECTED]
gt;
> I thought so as well, but it looks like postinst is only called with
> the most recently configured version, not with the current version.
Why not just preprocess the postinst in debian/rules to embed its own
version number directly into it?
--
Colin Watson [EMAIL PROTECTED]
On Tue, Mar 04, 2003 at 11:52:09PM +0100, Marc Haber wrote:
> On Tue, 4 Mar 2003 16:41:48 +0000, Colin Watson <[EMAIL PROTECTED]>
> wrote:
> >Why not just preprocess the postinst in debian/rules to embed its own
> >version number directly into it?
>
> Nice idea,
it into WNPP, I feel.
Cheers,
--
Colin Watson [EMAIL PROTECTED]
in /var/log, or
> /var/log/packagename? it's a single logfile, but it gets logrotated,
> so i can see people getting grumpy about it in /var/log/sugarplum.*
> after a couple of months...
You could always tell logrotate to keep only so many rotations. Since
logrotate config
evelopers disagree with joe's proposal?
I think that WNPP is an awkward place to handle sponsorship, and that a
home in the new-maintainer database would be much cleaner. See my other
post.
--
Colin Watson [EMAIL PROTECTED]
no stake in this. wnpp is just another
pseudo-package as far as we're concerned. I think the implementation of
any changes there is up to debian-www.
Cheers,
--
Colin Watson [EMAIL PROTECTED]
eople should worry
about this lintian error.
--
Colin Watson [EMAIL PROTECTED]
who would need to be asked to modify the wnpp emails to include
> the RFS category? the list-admins?
Marcelo Magallon maintains that (so probably should be asked about the
web implementation too).
--
Colin Watson [EMAIL PROTECTED]
't realize how often it's needed for
porting to even some relatively ordinary architectures. If all it takes
to port something to mips is to update config.guess and config.sub, I'll
do it; obviously I'd let upstream know, but their timescales are
undoubtedly different and I don't
have made, particularly if many unrelated changes are needed.
Given the clumsiness of existing tools to handle multiple patch sets,
this is usually only worth it for very large packages, IME.
--
Colin Watson [EMAIL PROTECTED]
ular it turns out not to be as simple as just grabbing up-to-date
Packages files every day.
This is probably the highest-priority issue facing debbugs right now,
although also probably the most difficult.
See the archives of debian-debbugs for (a little) more discussion on
this subject.
Cheers,
--
Colin Watson [EMAIL PROTECTED]
number > 1, you know. You just have to be careful that your
> > upload includes the .orig.tar.gz (debuild -sa, in other words).
>
> Well, now I'm confused, because this is what my sponsor told me.
Please educate your sponsor. :-)
--
Colin Watson [EMAIL PROTECTED]
and is a total absence of
solid facts! I demand that I may or may not be Vroomfondel!"
-- "The Hitch Hiker's Guide to the Galaxy", Douglas Adams
--
Colin Watson [EMAIL PROTECTED]
t; But... What is "sh"? I guess it isn't talking about /bin/sh (although
> that is used inside the package) And if is a platform, then what sort
> of machinery is that?
SuperH, the Hitachi chip in the Dreamcast console. Ignore it for now;
they have no packages other
icy. Can someone
> help me ?
Well, those two should be a pretty good start. What are you having
problems with?
--
Colin Watson [EMAIL PROTECTED]
ably if you're building on a semi-hacked-up woody
system you aren't building packages for upload anyway, so you don't
necessarily have to worry too much about complying with the letter of
policy.
Cheers,
--
Colin Watson [EMAIL PROTECTED]
^^^
[EMAIL PROTECTED] is an alias for that. Those with accounts can see
/etc/aliases on master.
--
Colin Watson [EMAIL PROTECTED]
D]
> [EMAIL PROTECTED]
> [EMAIL PROTECTED]
Look at the headers of the mails you get to determine the address. Mine
contain
"[EMAIL PROTECTED]",
decoding to [EMAIL PROTECTED]
> For the love of God. Someone, please help.
You need to ask [EMAIL PROTEC
's a feature. It's a good idea saving your answers
> when a package is reinstalled more times.
It's definitely not a feature if you've purged in between. That's always
supposed to delete configuration.
I thought that db_purge did that job, though?
--
Colin Watson [EMAIL PROTECTED]
u've
closed all the bugs you meant to before uploading, look in the Closes:
line of the .changes file generated by the build.
Cheers,
--
Colin Watson [EMAIL PROTECTED]
ed in all those distributions. Otherwise it
is left up to the maintainer's discretion to close it appropriately.
Common sense ought to apply.
Cheers,
--
Colin Watson [EMAIL PROTECTED]
On Thu, Apr 03, 2003 at 06:52:47PM -0600, Drew Scott Daniels wrote:
> On Fri, 4 Apr 2003 00:39:34 +0100, Colin Watson wrote:
> > On Thu, Apr 03, 2003 at 01:19:54PM -0600, Drew Scott Daniels wrote:
> > > Where might I get a copy of the bug database? Is there any way to search
>
On Fri, Apr 04, 2003 at 06:34:37PM +1000, Matthew Palmer wrote:
> On Fri, 4 Apr 2003, Colin Watson wrote:
> > Bits and pieces of state. There are no hidden bugs, if that's what you
> > mean.
>
> Oooh, the irony. I've just this minute been poking around in the a
pkg-source
build the source package in source+diff format; that way the
Debian-specific changes are easier to review.
Regards,
--
Colin Watson [EMAIL PROTECTED]
[Please don't send me private copies of replies to list mail - thanks.]
On Fri, Apr 04, 2003 at 05:09:22AM -0500, B. Douglas Hilton wrote:
> Colin Watson wrote:
> >On Fri, Apr 04, 2003 at 03:08:22AM -0500, B. Douglas Hilton wrote:
> >>Still, I am back on my "must jo
ere used to be a 'dpkg-signpackage'
> command, but I can't find it anymore
debsign, maybe?
Cheers,
--
Colin Watson [EMAIL PROTECTED]
On Fri, Apr 04, 2003 at 12:18:04PM -0600, Drew Scott Daniels wrote:
> On Fri, 4 Apr 2003, Colin Watson wrote:
> > On Thu, Apr 03, 2003 at 06:52:47PM -0600, Drew Scott Daniels wrote:
> > > Would anything else be there?
> >
> > Bits and pieces of state. There are no h
t; file(s) as the argument.
>
> [EMAIL PROTECTED]:~$ cat bin/dsign
> #!/bin/sh
> debsign [EMAIL PROTECTED] $*
You could get rid of this script if you liked, by putting
[EMAIL PROTECTED] in ~/.devscripts.
Cheers,
--
Colin Watson [EMAIL PROTECTED]
to have a Debian-specific diff against your own upstream release for
the assistance of other Debian developers and users, not to have a diff
against the *previous* upstream release.
Cheers,
--
Colin Watson [EMAIL PROTECTED]
pstream
> author (and nnn-forwarded) should work much better. - The discussion
> is archived in the BTS and I get the messages anyway, because I'm
> either the maintainer or at least subscribed to the package via PTS.
That works too. Take your pick. The point is that to encourage t
On Thu, Apr 10, 2003 at 06:54:37PM +0200, Andreas Metzler wrote:
> On Thu, Apr 10, 2003 at 05:35:04PM +0100, Colin Watson wrote:
> > If a bug is already marked as forwarded, then cc'ing -forwarded will log
> > the message in the bug report but not change the forwarded-to a
ge in the Debian archive always share files, even if
only /usr/share/doc//copyright. (If you need cc'ed responses
then you need to say so or you usually won't get them.)
dpkg doesn't support this, anyway.
Cheers,
--
Colin Watson [EMAIL PROTECTED]
variable in ~/.devscripts to whatever set of options you usually pass.
Mine ends with "-ICVS -I.svn -uc -us".
Cheers,
--
Colin Watson [EMAIL PROTECTED]
I hadn't had time to fix and then sent me a quiet mail asking me
if I'd advocate their NM application, I'd almost certainly say yes. I'm
much less likely to say yes when people pop up out of the blue.
--
Colin Watson [EMAIL PROTECTED]
, deployed, widely-understood system; you don't go
changing that under people for the sake of some web interface.
Cheers,
--
Colin Watson [EMAIL PROTECTED]
ore importantly) _in_which_packages_ would make the most
>difference for the whole project?
Have you tried http://bugs.qa.debian.org/?
--
Colin Watson [EMAIL PROTECTED]
On Sun, Apr 13, 2003 at 01:02:05AM +0100, Andrew Suffield wrote:
> On Sat, Apr 12, 2003 at 03:44:54PM +0100, Colin Watson wrote:
> > Have you tried http://bugs.qa.debian.org/?
>
> http://bugs.qa.debian.org/cgi-bin/standard.cgi says that bison has 3
> serious bugs.
>
> P
On Sun, Apr 13, 2003 at 12:21:49PM +1000, Martin Michlmayr wrote:
> * Colin Watson <[EMAIL PROTECTED]> [2003-04-13 02:25]:
> > The fact that pkgreport.cgi doesn't display them is a bug in
> > debbugs, not
>
> Well, then we can still blame you. :-P
You'll
On Sun, Apr 13, 2003 at 03:19:03PM +0100, Andrew Suffield wrote:
> On Sun, Apr 13, 2003 at 02:25:43AM +0100, Colin Watson wrote:
> > (Hint for the unaware: #166130, #166897, #167054. The fact that
> > pkgreport.cgi doesn't display them is a bug in debbugs, not
> > bu
mal procedure (when we
have a proper maintainer for the package) is indeed to send changes
upstream wherever possible.
Regards,
--
Colin Watson [EMAIL PROTECTED]
s around
dpkg-buildpackage like debuild (but then you're still using
dpkg-buildpackage, of course). Don't build source archives by hand; it's
error-prone and a waste of your time.
Cheers,
--
Colin Watson [EMAIL PROTECTED]
ractively from source.
If it's just for private use, then there's no HOWTO for it because it's
easier than the normal case. :) Just have a "source" package full of
binaries and make debian/rules copy those into the temporary build
directory.
Cheers,
--
Colin Watson [EMAIL PROTECTED]
27;m surprised the list doesn't set the reply-to field.
There are good reasons for this that have been gone over uncountable
times. It's best to use a mailer that has a reply-to-list function.
Cheers,
--
Colin Watson [EMAIL PROTECTED]
time.
Matthias (Klose), is this right? If so somebody should tell aj.
--
Colin Watson [EMAIL PROTECTED]
ven her instructions on how to find the PID of the one
> using port 80 and kill it, but how in the world would sshd have taken
> over port 80? That seems like a pretty huge mistake in config...
Wow, I'm impressed. As far as I know (as co-maintainer) there's nothing
in the ssh packa
/share/doc in order to work.
--
Colin Watson [EMAIL PROTECTED]
hat. Oh well.
Magic options to dpkg-gencontrol (try -v). I'd advise against it unless
it's really painful to just add an epoch to the whole lot, though.
--
Colin Watson [EMAIL PROTECTED]
ing to hack around with them.
> I say simple, but then, I didn't write the decompiler. :)
Note that, IIRC, the Java .class format doesn't store the names of local
variables in methods, so you lose a good deal of intelligibility during
the decompilation process.
Cheers,
--
Colin Watson [EMAIL PROTECTED]
On Tue, Apr 22, 2003 at 08:55:49PM +0100, Pete Ryland wrote:
> On Tue, Apr 22, 2003 at 06:36:13PM +0100, Colin Watson wrote:
> > > I say simple, but then, I didn't write the decompiler. :)
(Actually I didn't write that ...)
> > Note that, IIRC, the Java .class forma
ugs in Debian, so "this
piece of software isn't packaged" is a valid wishlist bug, or "this
package is without a maintainer" is a valid normal bug. But "this
software cannot be packaged" is not our bug, it's just a statement of
fac
ew-maintainer report a few months later.
Also, there are a lot of bugs of the "won't build from source" variety
and similar, and it doesn't usually take an enormous amount of patching
to fix them. Helping with these is very valuable, and the sort of people
who spend their time going through these will very likely notice.
Cheers,
--
Colin Watson [EMAIL PROTECTED]
arious tools for doing so.
Ugh! Don't do that! Get the source package and modify it.
> I ran lintian (which produced a couple of mild complaints), but I
> don't have "linda" and can't find it.
'apt-get install linda' on testing or unstable.
Cheers,
--
n policy also requires that upstream changelogs be
/usr/share/doc//changelog.gz, rather than ChangeLog.gz (or at
least be a symlink to the latter).
Cheers,
--
Colin Watson [EMAIL PROTECTED]
On Thu, May 01, 2003 at 12:03:04PM -0800, Fielder George Dowding wrote:
> On Thu, 1 May 2003 11:46:39 +0100
> Colin Watson <[EMAIL PROTECTED]> wrote:
> > On Wed, Apr 30, 2003 at 12:39:00PM -0800, Fielder George Dowding
> > wrote:
> > > I am attempting to become the
On Fri, May 02, 2003 at 10:48:35PM -0700, Brian Nelson wrote:
> Colin Watson <[EMAIL PROTECTED]> writes:
> > You can use the pbuilder package to set yourself up an unstable chroot,
> > or, if you want to roll your own, see:
> >
> >
> > http://www.debian.or
On Tue, May 06, 2003 at 01:48:37AM -0700, Osamu Aoki wrote:
> On Mon, May 05, 2003 at 06:04:59PM +0100, Colin Watson wrote:
> > On Fri, May 02, 2003 at 10:48:35PM -0700, Brian Nelson wrote:
> > > I don't think woody's debootstrap will work directly for sid chroots,
about the situation
to verify that the override's correct beforehand).
Either way, you should be telling people to use dpkg-statoverride if
they want to change the permissions.
Also, 4750 is too tight; there's no reason not to make the executables
readable to unprivileged users, i.e
it and do that the first time.
> However, the upstream source does runtime validation of permissions -- It
> will refuse to work with anything but 4750.
I'd be inclined to patch the upstream source such that it will accept
world-readability.
Cheers,
--
Colin Watson [EMAIL PROTECTED]
ected; anyone can
get a valid password just by asking, but the expectation is that only
developers should be downloading the packages, and this discourages
people from shoving it into apt lines.
--
Colin Watson [EMAIL PROTECTED]
On Tue, May 13, 2003 at 08:30:03AM +0100, Mark Brown wrote:
> On Tue, May 13, 2003 at 01:29:00AM +0100, Colin Watson wrote:
> > On Mon, May 12, 2003 at 05:41:40PM -0600, Jack Moffitt wrote:
> > > Perhaps an easy thing to do would just be to show whether or not a
> > >
bly
have been "yes, but entirely at your own legal risk", I think.)
Cheers,
--
Colin Watson [EMAIL PROTECTED]
r name"' instead (or, as I do, just use 'debuild -uc
-us' and run debsign afterwards when I've checked that the build looks
sane).
--
Colin Watson [EMAIL PROTECTED]
sted yet? Surely the order should be build, test, sign.
Accordingly, I have DEBUILD_DPKG_BUILDPACKAGE_OPTS="-uc -us", together
with a bunch of other options that don't matter here, in ~/.devscripts.
I've never yet found that I want the other behaviour.
--
Colin Watson [EMAIL PROTECTED]
#x27;s in the
.orig.tar.gz just gets you a warning, not a fatal error, so you could do
that. Alternatively, you could move the original version aside in the
build target and move it back in clean.
Cheers,
--
Colin Watson [EMAIL PROTECTED]
s are available at
> http://www.isotton.com/debian/sitemap/
>
> (Version 2.3-3).
>
> Could somebody please upload this update?
Done. Sorry for the delay.
--
Colin Watson [EMAIL PROTECTED]
this method should be enough.
This is vague memory, but there was a conversation on IRC a few weeks
back in which it was observed that using 'dpkg -l' from a postinst (i.e.
from within dpkg) was potentially risky. I'd stick to looking for files
myself.
Cheers,
--
Colin Watson
hangelog entries from the NMU
fixes as well, though.
--
Colin Watson [EMAIL PROTECTED]
ebian.org/debian-perl-0301/msg0.html
I suggest adding a lintian override. The modules in question are closely
tied to their XS implementations; if they ever got out of sync (i.e. if
somebody was seriously sharing /usr/share between machines, rather than
it just being a theoretical exerc
; libsigcx-gtk-0.6-1 libsigcx-gtk-0.6-dev | grep testing
> | [EMAIL PROTECTED]:~%
> `
>
> And furthermore: Why do they have no testing canditate?
It looks like Bj?rn's script is buggy. It should be using source package
names there rather than binary package names.
--
Colin Watson [EMAIL PROTECTED]
On Tue, Jun 03, 2003 at 02:27:19PM +0100, Colin Watson wrote:
> The problem is that libsigcx's soname has changed since the last version
> in testing. This usually means that it needs (a) everything that depends
> on it to be updated and (b) a manual kick by aj. It's possible
gt; chowning if the account is not present? That way, I would be able to
> give back patches to upstream.
That'll make the package build differently depending on whether it's
already installed or not. I'm inclined to think that this is a bad idea.
--
Colin Watson [EMAIL PROTECTED]
g in postinst anyway, can't you just change the
> package to compile with say, nobody as the user?
The nobody user should never own any files. root would do, though.
--
Colin Watson [EMAIL PROTECTED]
t in testing and still be a
valid candidate, as they could all be upgraded in a single testing run -
they might even form a loop.
> Does this require action by me?
No, don't worry about it; it's just an extra note on why
directory-administrator isn't a valid candidate. You'd already worked
that out for yourself.
Cheers,
--
Colin Watson [EMAIL PROTECTED]
901 - 1000 of 1012 matches
Mail list logo
