Re: How to create multi-source tarball with different submodules for scipy
On Mon, Jan 16, 2023 at 05:05:39PM +0100, Andreas Tille wrote: > Hi, > > I tried to create a multi-source tarball for scipy in its experimental > branch[1]. Upstream includes a set of git submodules in its build > process. I intended to merge all these submodules in a single > scipy_1.10.0.orig-submodules.tar.gz. This tarball is created with a > script[2] which makes sure that the exact directory structure as it is > used by upstream is conserved. This directory layout is needed in the > build process. Unfortunately `dpkg-source -x` extracts the content of > the submodules tarball into a subdirectory submodules/. > > Is there any trick to unpack this tarball right into the root? > Otherwise I need to do some symlinks workaround in d/rules to provide > all files where these are needed. Not that I know of; this is the design of the multi-source tarball setup: each component tarball is extracted into a directory with the name of the component. Best wishes, Julian
Re: [j...@debian.org: Your Debian package(s) will not migrate to testing]
On Thu, Sep 05, 2019 at 10:10:00AM +0200, Andre Noll wrote: > Hi Adam, > > looks like tfortune needs a new source-only upload to migrate to > testing. Could you please do this for me? Here's the commands to > build the source package: > > git archive --prefix tfortune-1.0.0/ origin/master > > ../tfortune_1.0.0.orig.tar.xz > dpkg-buildpackage -S > > There have been no changes since June, so your existing repo should > still be uptodate. > > Thanks > Andre Hi Adam and Andre, You will need to upload a new version, though; the system will not accept a package with the same version number. Best wishes, Julian > - Forwarded message from Julian Gilbey - > > Subject: Your Debian package(s) will not migrate to testing > From: Julian Gilbey > To: Andre Noll > Message-Id: > Date: Wed, 04 Sep 2019 22:16:56 +0100 > > > Dear Andre Noll, > > This is a courtesy message about your package(s) which is/are > stuck in sid, and will not migrate to testing, in case you > are unaware; they are listed below. > > The release managers announced on 7th July 2019 in their email > to debian-devel-announce and debian-release > (Message-ID: <20190707014700.gf15...@powdarrmonkey.net>) > that only source-only uploads would transition to testing. > > Here is the relevant part of what they wrote: > > ~ > No binary maintainer uploads for bullseye > = > > The release of buster also means the bullseye release cycle is about to begin. > >From now on, we will no longer allow binaries uploaded by maintainers to > migrate to testing. This means that you will need to do source-only uploads if > you want them to reach bullseye. > > Q: I already did a binary upload, do I need to do a new (source-only) > upload? > A: Yes (preferably with other changes, not just a version bump). > > Q: I needed to do a binary upload because my upload went to the NEW queue, > do I need to do a new (source-only) upload for it to reach bullseye? > A: Yes. We also suggest going through NEW in experimental instead of > unstable > where possible, to avoid disruption in unstable. > > Q: Does this also apply to contrib and non-free? > A: No. Not all packages in contrib and non-free can be built on the buildds, > so maintainer uploads will still be allowed to migrate for packages > outside main. > ~ > > To perform a source-only build and upload, run dpkg-buildpackage -S > and then upload the relevant files in the normal way (for example, > using dupload or dput). > > Your package(s) involved is/are as follows: > > Source package: tfortune > Version: 1.0.0-2 > One relevant line of the excuse file (including the uploader): > Not built on buildd: arch amd64 binaries uploaded by kilobyte > Transition verdict: REJECTED_PERMANENTLY > > > > I hope this is of help to you! > > Best wishes, > >Julian > > - End forwarded message -
Re: debian/watch and different action than `uupdate'
On Mon, Aug 27, 2007 at 12:30:41AM +0200, Daniel Leidert wrote: 2) Is it possible to use a different action than uupdate, e.g. `fakeroot debian/rules get-orig-source'? Nice idea. Unfortunately uscan --force-download does not call the script. Instead it tries to get the archive I'm testing for - but that's not, what I want - the get-orig-source gets the necessary archives, merges them and creates the source tarball. Is the problem, that the above idea fails, related to the fact, that I'm using version 3 of the debian/watch format? It may also call the script; at present, uscan assumes that the user wishes to download the new archive *before* running the update action. I wonder whether the best way around this one would be to have a new option nodownload or something like that? Julian -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: debian/watch and different action than `uupdate'
On Sun, Aug 26, 2007 at 07:10:17PM +0200, Daniel Leidert wrote: 2) Is it possible to use a different action than uupdate, e.g. `fakeroot debian/rules get-orig-source'? And this too. Seems to be impossible atm. Would be a nice-to-have for me. $ cat debian/get-source #! /bin/sh fakeroot debian/rules get-orig-source $ cat debian/watch http://.../.../pkg-(.*).tar.gz debian debian/get-source Julian -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: debian/watch file and berlios
On Tue, Nov 21, 2006 at 12:43:33PM -0500, Justin Pryzby wrote: On Mon, Nov 20, 2006 at 11:21:45AM +0100, Andreas Bresser wrote: Hello, The upstream-author of my package spe uploaded a new release on berlios.de (on sourceforge is still the old version), so I changed the debian/watch-file from sf.net to berlios: version=3 opts=downloadurlmangle=s/prdownload/download/ \ http://developer.berlios.de/project/showfiles.php?group_id=4161 \ http://prdownload.berlios.de/python/SPE-(.*)-wx.*\.tar\.gz but this does not work: ~/debian/spe2/spe-0.8.2a+repack$ uscan spe: Newer version (0.8.3.c) available on remote site: http://download.berlios.de/python/SPE-0.8.3.c-wx2.6.1.0.tar.gz (local version is 0.8.2a+repack) uscan warning: In directory ., downloading http://download.berlios.de/python/SPE-0.8.3.c-wx2.6.1.0.tar.gz failed: 403 Forbidden when I try to download this file directly with wget it works: Berlios apparently rejects based on User-Agent. wget -q --header 'User-Agent: libwww-perl/5.805' 'http://developer.berlios.de/project/showfiles.php?group_id=4161' /dev/null ; echo $? 1 wget -q 'http://developer.berlios.de/project/showfiles.php?group_id=4161' /dev/null ; echo $? 0 This was fixed in devscripts 2.9.24, and was bug#397354. Do you still observe this erroneous behaviuur with the latest devscripts? I also note the following strange headers sent by uscan, observed with ethereal^Wwireshark: Connection: TE, close TE: deflate,gzip;q=0.3 Not a clue, sorry. Problems aside, I think you will want something like: version=3 http://developer.berlios.de/project/showfiles.php?group_id=4161 SPE-(.*)-wx.*\.tar\.gz Julian -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Bug#194678: devscripts: making -uc -us the default signing options
On Sun, May 25, 2003 at 12:07:19PM -0600, Bob Proulx wrote: Package: devscripts Version: 2.7.0 Severity: wishlist File: /usr/bin/debuild Recently on debian-mentors a discussion[1] took place about the default options to 'debuild'. There at least one developer stated[2] that they always override the default signing with debuild -uc -us in their personal override files and then debsign later after testing. I also always override, but on the command line, and proposed[3] changing the default options. It was suggested[4] that I file a wishlist bug. Here it is. Would it be possible to change the default options to debuild so that the options -uc -us are the default? And change the documentation to suggest using debsign after debuild so that the developer is directed along the path of build, test, sign? This would make the simple command 'debuild' the simple case and the default for most developers. I believe it simplifies the process. Hmm. Interesting idea. My problems with this are: (1) It will change the default behaviour of the program, and I'm not entirely happy with doing this. Maybe ask on -devel and see whether there's a broad agreement on this point? (2) It will make it's default behaviour different from that of dpkg-buildpackage, which it is essentially emulating with a few bells and whistles. (3) It is easy to modify /etc/devscripts.conf or ~/.devscripts to do this. On the positive side, it might encourage people to test their packages before releasing them, or it might just encourage them to add --sign (or whatever the signing command-line option would become) to their config file, gaining very little. I'm not sure which way to go on this one. Julian -- =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- Julian Gilbey, website: http://www.polya.uklinux.net/ Debian GNU/Linux Developer, see: http://people.debian.org/~jdg/ Visit http://www.thehungersite.com/ to help feed the hungry
Re: latex build problem
On Wed, Dec 25, 2002 at 05:44:16PM -0800, Osamu Aoki wrote: I encounter errors in latex when building PS file (or PDF file). I do not understand (\end occurred inside a group at level 31) What should I look for for this error? I tried -v option in debiandoc2latexpsand essentially same output with pool size. Probably problems with the source file. I have big enough pool size and error messages are as follows . Appendice A. [177] [178] Underfull \hbox (badness 4899) in paragraph at lines 9373--9375 []\T1/ppl/m/n/10.95 I fi-les SGML ori-gi-na-li usa-ti per crea-re il do-cu-men- to so-no di-spo-ni-bi-li in CVS pres- Not an error, just a warning. Underfull \hbox (badness 5359) in paragraph at lines 9373--9375 \T1/ppl/m/n/10.95 so: \T1/pcr/m/n/10.95 :pserver:[EMAIL PROTECTED]/cvsr oot/qref \T1/ppl/m/n/10.95 op-pu-re [][]$\T1/pcr/m/n/10.95 http : / / qref . [179] Underfull \hbox (badness 1122) in paragraph at lines 9384--9386 []\T1/ppl/m/n/10.95 Le uti-li-t^^e0 per la con-ver-sio-ne da De-bian-Doc SGML s o-no di-spo-ni-bi-li nel pac-chet-to De-bian [180] (./reference.it.aux) Ditto. LaTeX Warning: Label(s) may have changed. Rerun to get cross-references right. Ditto. ) (\end occurred inside a group at level 31) That's a bug somewhere in the TeX-generating code, most likely. Perhaps you can place the SGML file somewhere where it can be downloaded for checking. Also, please let us know which version of the SGML tools packages you are using. Julian -- =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- Julian Gilbey, website: http://www.polya.uklinux.net/ Debian GNU/Linux Developer, see: http://people.debian.org/~jdg/ Visit http://www.thehungersite.com/ to help feed the hungry
Re: pure python and postinst
On Tue, Nov 19, 2002 at 08:17:28PM -0500, Steve M. Robbins wrote: One thing I noticed in the packages (that isn't covered in the policy) is the practice of deleting the .pyc files in debian/rules, and then running compileall.py in the postinst. Would someone comment on this? Is it a requirement -- is it a bug to include the compiled python files in the .deb? Are compiled python files architecture-independent? Julian -- =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- Julian Gilbey, website: http://www.polya.uklinux.net/ Debian GNU/Linux Developer, see: http://people.debian.org/~jdg/ Visit http://www.thehungersite.com/ to help feed the hungry -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: pure python and postinst
On Tue, Nov 19, 2002 at 08:17:28PM -0500, Steve M. Robbins wrote: One thing I noticed in the packages (that isn't covered in the policy) is the practice of deleting the .pyc files in debian/rules, and then running compileall.py in the postinst. Would someone comment on this? Is it a requirement -- is it a bug to include the compiled python files in the .deb? Are compiled python files architecture-independent? Julian -- =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- Julian Gilbey, website: http://www.polya.uklinux.net/ Debian GNU/Linux Developer, see: http://people.debian.org/~jdg/ Visit http://www.thehungersite.com/ to help feed the hungry
Re: autobuilding arch independent packages ...
On Mon, Sep 23, 2002 at 11:38:02AM +0200, Remi VANICAT wrote: Another possibility is to go to one or another debian.org developer machine an to make the build there. You may even be able to use screen to not been forced to be connected for the whole build. You should be able to build offline; no interaction should be needed (or your build is broken). Of course this mean that you need to have your gnupg secret key there, and this may be unwanted. See debsign in the devscripts package ;-) Julian -- =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- Julian Gilbey, Dept of Maths, Queen Mary, Univ. of London website: http://www.maths.qmul.ac.uk/~jdg/ Debian GNU/Linux Developer, see: http://people.debian.org/~jdg/ Visit http://www.thehungersite.com/ to help feed the hungry -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: autobuilding arch independent packages ...
On Mon, Sep 23, 2002 at 11:38:02AM +0200, Remi VANICAT wrote: Another possibility is to go to one or another debian.org developer machine an to make the build there. You may even be able to use screen to not been forced to be connected for the whole build. You should be able to build offline; no interaction should be needed (or your build is broken). Of course this mean that you need to have your gnupg secret key there, and this may be unwanted. See debsign in the devscripts package ;-) Julian -- =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- Julian Gilbey, Dept of Maths, Queen Mary, Univ. of London website: http://www.maths.qmul.ac.uk/~jdg/ Debian GNU/Linux Developer, see: http://people.debian.org/~jdg/ Visit http://www.thehungersite.com/ to help feed the hungry
Re: autobuilding arch independent packages ...
On Mon, Sep 23, 2002 at 11:25:59AM +0100, Colin Watson wrote: On Mon, Sep 23, 2002 at 12:06:29PM +0200, Dagfinn Ilmari Manns?ker wrote: Colin Watson [EMAIL PROTECTED] writes: On Mon, Sep 23, 2002 at 11:38:02AM +0200, Remi VANICAT wrote: Of course this mean that you need to have your gnupg secret key there, and this may be unwanted. No, build it unsigned on the remote machine, then copy the .dsc and .changes back for signing on a local trusted machine. Or just use debrsign to do it automatically. Only problem with debrsign is that your passphrase needs to go across the network and back (albeit over ssh if you're sane), because it operates by copying to a remote machine and signing remotely rather than copying from a remote machine and signing locally. It'd be nice if it could operate the other way round. debsign can do this. I should add a comment to the debrsign manpage Now done in my local tree. Julian -- =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- Julian Gilbey, Dept of Maths, Queen Mary, Univ. of London website: http://www.maths.qmul.ac.uk/~jdg/ Debian GNU/Linux Developer, see: http://people.debian.org/~jdg/ Visit http://www.thehungersite.com/ to help feed the hungry
Re: What to do on accidential upload
On Sun, Aug 18, 2002 at 08:19:35PM +0200, Andreas Rottmann wrote: Hi! I uploaded a package (quick-lounge-applet) to unstable that sould have gone to experimental (mostly due to a dependency on libpanel-applet2). /me slaps himself fiercly How can I try to clean this mess up? It's a new package, right? So I'd just email a bug report against ftp.debian.org asking for this package not to be installed in unstable, and upload it again to experimental. Julian -- =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- Julian Gilbey, Dept of Maths, Queen Mary, Univ. of London website: http://www.maths.qmul.ac.uk/~jdg/ Debian GNU/Linux Developer, see: http://people.debian.org/~jdg/ Visit http://www.thehungersite.com/ to help feed the hungry -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: What to do on accidential upload
On Sun, Aug 18, 2002 at 08:19:35PM +0200, Andreas Rottmann wrote: Hi! I uploaded a package (quick-lounge-applet) to unstable that sould have gone to experimental (mostly due to a dependency on libpanel-applet2). /me slaps himself fiercly How can I try to clean this mess up? It's a new package, right? So I'd just email a bug report against ftp.debian.org asking for this package not to be installed in unstable, and upload it again to experimental. Julian -- =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- Julian Gilbey, Dept of Maths, Queen Mary, Univ. of London website: http://www.maths.qmul.ac.uk/~jdg/ Debian GNU/Linux Developer, see: http://people.debian.org/~jdg/ Visit http://www.thehungersite.com/ to help feed the hungry
Re: problem with gnupg 1.0.7 and debsign
On Wed, Aug 14, 2002 at 11:44:30PM +0200, C?dric Delfosse wrote: Hello to the list. I have this problem since my complete upgrade to testing. When I tried to build my package, the process fails at debsign: dh_builddeb dpkg-deb: building package `darkstat' in `../darkstat_2.1-1_i386.deb'. signfile darkstat_2.1-1.dsc gpg: skipped `C?dric Delfosse [EMAIL PROTECTED]': secret key not available [...] But if I use my fingerprint: You need a passphrase to unlock the secret key for user: C?dric Delfosse [EMAIL PROTECTED] Note that [EMAIL PROTECTED] != [EMAIL PROTECTED] You need to add your @debian.org address to your key (and then upload it to keyring.debian.org) or use your @laposte.net address in your packages. Julian -- =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- Julian Gilbey, Dept of Maths, Queen Mary, Univ. of London website: http://www.maths.qmul.ac.uk/~jdg/ Debian GNU/Linux Developer, see: http://people.debian.org/~jdg/ Visit http://www.thehungersite.com/ to help feed the hungry -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: problem with gnupg 1.0.7 and debsign
On Wed, Aug 14, 2002 at 11:44:30PM +0200, C?dric Delfosse wrote: Hello to the list. I have this problem since my complete upgrade to testing. When I tried to build my package, the process fails at debsign: dh_builddeb dpkg-deb: building package `darkstat' in `../darkstat_2.1-1_i386.deb'. signfile darkstat_2.1-1.dsc gpg: skipped `C?dric Delfosse [EMAIL PROTECTED]': secret key not available [...] But if I use my fingerprint: You need a passphrase to unlock the secret key for user: C?dric Delfosse [EMAIL PROTECTED] Note that [EMAIL PROTECTED] != [EMAIL PROTECTED] You need to add your @debian.org address to your key (and then upload it to keyring.debian.org) or use your @laposte.net address in your packages. Julian -- =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- Julian Gilbey, Dept of Maths, Queen Mary, Univ. of London website: http://www.maths.qmul.ac.uk/~jdg/ Debian GNU/Linux Developer, see: http://people.debian.org/~jdg/ Visit http://www.thehungersite.com/ to help feed the hungry
Re: debuild, dupload and then?
On Tue, Aug 06, 2002 at 07:14:18AM +0900, Oohara Yuuma wrote: On Sun, 4 Aug 2002 21:08:18 +0200, Holger Kubiak [EMAIL PROTECTED] wrote: Is there a tool that generates the /debian/dists/-tree? debarchiver (or the unpackaged katie) Something like dpkg-scanpackages maybe? Julian -- =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- Julian Gilbey, Dept of Maths, Queen Mary, Univ. of London website: http://www.maths.qmul.ac.uk/~jdg/ Debian GNU/Linux Developer, see: http://people.debian.org/~jdg/ Visit http://www.thehungersite.com/ to help feed the hungry
Re: Bug#149486: FTBFS: bad build depends
On Sun, Jun 09, 2002 at 04:13:45PM -0700, David Caldwell wrote: Hello, I just received a bug report: Package: simpleproxy Version: 3.1-1 Severity: important This package fails to build from source on the ia64 autobuilder with the error: make: dh_testdir: Command not found It would appear that you need to add a build dependency on debhelper. A complete build log is available at buildd.debian.org. I already have this in the control file: Build-Depends: debhelper ( 3.0.0) No you don't. Julian -- =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- Julian Gilbey, Dept of Maths, Queen Mary, Univ. of London website: http://www.maths.qmul.ac.uk/~jdg/ Debian GNU/Linux Developer, see: http://people.debian.org/~jdg/ Visit http://www.thehungersite.com/ to help feed the hungry -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Bug#149486: FTBFS: bad build depends
On Sun, Jun 09, 2002 at 04:13:45PM -0700, David Caldwell wrote: Hello, I just received a bug report: Package: simpleproxy Version: 3.1-1 Severity: important This package fails to build from source on the ia64 autobuilder with the error: make: dh_testdir: Command not found It would appear that you need to add a build dependency on debhelper. A complete build log is available at buildd.debian.org. I already have this in the control file: Build-Depends: debhelper ( 3.0.0) No you don't. Julian -- =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- Julian Gilbey, Dept of Maths, Queen Mary, Univ. of London website: http://www.maths.qmul.ac.uk/~jdg/ Debian GNU/Linux Developer, see: http://people.debian.org/~jdg/ Visit http://www.thehungersite.com/ to help feed the hungry -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Package that can work with something in non-free - to build with it or not?
On Thu, May 02, 2002 at 08:09:44PM +0200, Hugo van der Merwe wrote: I am currently working on a package that can optionally support scilab. For the time being I'm not build-depending on scilab, and not building against it either, as scilab is in non-free. Can I be sure that the buildd's won't build with scilab? If scilab is installed, the configure script will use it. So I should probably look into some --disable-scilab type configure parameter? Build-Conflicts, perhaps? Furthermore, how is something like this usually handled, if the user wants scilab support, he can rebuild the package himself ? Tell the user to remove the Build-Conflicts line and rebuild. Julian -- =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- Julian Gilbey, Dept of Maths, Queen Mary, Univ. of London website: http://www.maths.qmul.ac.uk/~jdg/ Debian GNU/Linux Developer, see: http://people.debian.org/~jdg/ Visit http://www.thehungersite.com/ to help feed the hungry -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: pre-depending
On Thu, Apr 04, 2002 at 03:55:29PM -0600, Gunnar Wolf wrote: Hi, I need debconf and wwwconfig-common for the package I am building. They need to be correctly installed by the time I run the postinst - I think that they should be mentioned as pre-depends... Now, as I read, before including somehting as pre-depends, I should check this with debian-devel. Only needs to be a pre-depends if you need them correctly installed and configured before you run your PREinst. Julian -- =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- Julian Gilbey, Dept of Maths, Queen Mary, Univ. of London website: http://www.maths.qmul.ac.uk/~jdg/ Debian GNU/Linux Developer, see: http://people.debian.org/~jdg/ Visit http://www.thehungersite.com/ to help feed the hungry -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: lintian: executable-not-elf-or-script
On Tue, Mar 19, 2002 at 06:28:23AM +0100, J?rgen H?gg wrote: I'm writing a program that (for my conveniance) is split up in several perl modules. These are put in my programs private directory, /usr/lib/ham, because they might some day interfere with other 'official' perl modules. But lintian does not like this: W: ham-server: executable-not-elf-or-script ./usr/lib/ham/Ham/Sched.pm [...] But the modules are not executable, it's -rw-r--r--. What version of lintian are you using? Have you checked the output of dpkg-deb -c on your .deb file to ensure that the mode really is what you think it is? Julian -- =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- Julian Gilbey, Dept of Maths, Debian GNU/Linux Developer Queen Mary, Univ. of London see http://people.debian.org/~jdg/ http://www.maths.qmul.ac.uk/~jdg/ or http://www.debian.org/ Visit http://www.thehungersite.com/ to help feed the hungry -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: lintian: executable-not-elf-or-script
On Tue, Mar 19, 2002 at 06:28:23AM +0100, J?rgen H?gg wrote: I'm writing a program that (for my conveniance) is split up in several perl modules. These are put in my programs private directory, /usr/lib/ham, because they might some day interfere with other 'official' perl modules. But lintian does not like this: W: ham-server: executable-not-elf-or-script ./usr/lib/ham/Ham/Sched.pm [...] But the modules are not executable, it's -rw-r--r--. What version of lintian are you using? Have you checked the output of dpkg-deb -c on your .deb file to ensure that the mode really is what you think it is? Julian -- =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- Julian Gilbey, Dept of Maths, Debian GNU/Linux Developer Queen Mary, Univ. of London see http://people.debian.org/~jdg/ http://www.maths.qmul.ac.uk/~jdg/ or http://www.debian.org/ Visit http://www.thehungersite.com/ to help feed the hungry
Re: watch file syntax
On Thu, Mar 14, 2002 at 12:00:31PM -0800, Ian Duggan wrote: And go for version 2.6.91 (which has just been uploaded) if you have any FTP sites; version 2.6.90 was broken for them. See uscan(1) in the new package for a description of the new watch file format. The new devscripts also features a funky new config file. It was probably my longest ever changelog ;-) I looked on the mirrors and the latest I can find is 2.6.90. Where would I look to find 2.6.91? I looked in the /pool/d/devscripts/ dir on mirrors.kernel.org/debian and http.us.debian.org. Where do new uploads go to, and how often do the mirrors update? If 2.6.90 was there when you looked, by the same time tomorrow you should see 2.6.91. Julian -- =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- Julian Gilbey, Dept of Maths, Debian GNU/Linux Developer Queen Mary, Univ. of London see http://people.debian.org/~jdg/ http://www.maths.qmul.ac.uk/~jdg/ or http://www.debian.org/ Visit http://www.thehungersite.com/ to help feed the hungry -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: watch file syntax
On Thu, Mar 14, 2002 at 12:00:31PM -0800, Ian Duggan wrote: And go for version 2.6.91 (which has just been uploaded) if you have any FTP sites; version 2.6.90 was broken for them. See uscan(1) in the new package for a description of the new watch file format. The new devscripts also features a funky new config file. It was probably my longest ever changelog ;-) I looked on the mirrors and the latest I can find is 2.6.90. Where would I look to find 2.6.91? I looked in the /pool/d/devscripts/ dir on mirrors.kernel.org/debian and http.us.debian.org. Where do new uploads go to, and how often do the mirrors update? If 2.6.90 was there when you looked, by the same time tomorrow you should see 2.6.91. Julian -- =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- Julian Gilbey, Dept of Maths, Debian GNU/Linux Developer Queen Mary, Univ. of London see http://people.debian.org/~jdg/ http://www.maths.qmul.ac.uk/~jdg/ or http://www.debian.org/ Visit http://www.thehungersite.com/ to help feed the hungry
Re: watch file syntax
On Thu, Mar 14, 2002 at 12:50:33AM -0500, Joey Hess wrote: Ian Duggan wrote: When specifying the regexp to be used in a watch file, are backslashes supposed to be escaped? The New Maintainer's Guide says one thing, but the man page for uscan says another. Which is preferred/correct? Install the uscan from unstable, and do not escape backslashes. The new devscripts features a rewritten uscan, that also has a new, more powerful and easier to use watch file format. And it doesn't have the escaping problems the old uscan did. And go for version 2.6.91 (which has just been uploaded) if you have any FTP sites; version 2.6.90 was broken for them. See uscan(1) in the new package for a description of the new watch file format. The new devscripts also features a funky new config file. It was probably my longest ever changelog ;-) Julian -- =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- Julian Gilbey, Dept of Maths, Debian GNU/Linux Developer Queen Mary, Univ. of London see http://people.debian.org/~jdg/ http://www.maths.qmul.ac.uk/~jdg/ or http://www.debian.org/ Visit http://www.thehungersite.com/ to help feed the hungry
Re: dh_movefiles
On Wed, Jan 16, 2002 at 11:19:27PM -0500, christophe barb? wrote: I'm trying to use dh_movefiles to fill a package-dev directory but something change and I don't understand what is expected from me. # dh_movefiles dh_movefiles: I was asked to move files from debian/tmp to debian/tmp. Perhaps you should set DH_COMPAT=2? I've tried several thing like 'dh_movefiles package-dev' but nothing help. I've the two files package-dev.(files|dirs) in the debian directory. dh_movefiles acts on *all* of your packages, moving files from debian/tmp to debian/$pkg for each package, according to the list of files in $pkg.files. If you are using DH_COMPAT=2, then your primary package directory will be debian/$pkg; with DH_COMPAT=1, it will be debian/tmp. So what happens in your case is you are trying to move files from debian/tmp to debian/tmp when you handle the primary package. Easiest solution: don't let dh_movefiles act on your primary package, so use dh_movefiles -Xpackage. Julian -- =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- Julian Gilbey, Dept of Maths, Debian GNU/Linux Developer Queen Mary, Univ. of London see http://people.debian.org/~jdg/ http://www.maths.qmul.ac.uk/~jdg/ or http://www.debian.org/ Visit http://www.thehungersite.com/ to help feed the hungry -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: dh_movefiles
On Wed, Jan 16, 2002 at 11:19:27PM -0500, christophe barb? wrote: I'm trying to use dh_movefiles to fill a package-dev directory but something change and I don't understand what is expected from me. # dh_movefiles dh_movefiles: I was asked to move files from debian/tmp to debian/tmp. Perhaps you should set DH_COMPAT=2? I've tried several thing like 'dh_movefiles package-dev' but nothing help. I've the two files package-dev.(files|dirs) in the debian directory. dh_movefiles acts on *all* of your packages, moving files from debian/tmp to debian/$pkg for each package, according to the list of files in $pkg.files. If you are using DH_COMPAT=2, then your primary package directory will be debian/$pkg; with DH_COMPAT=1, it will be debian/tmp. So what happens in your case is you are trying to move files from debian/tmp to debian/tmp when you handle the primary package. Easiest solution: don't let dh_movefiles act on your primary package, so use dh_movefiles -Xpackage. Julian -- =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- Julian Gilbey, Dept of Maths, Debian GNU/Linux Developer Queen Mary, Univ. of London see http://people.debian.org/~jdg/ http://www.maths.qmul.ac.uk/~jdg/ or http://www.debian.org/ Visit http://www.thehungersite.com/ to help feed the hungry
Re: How to change manual's section ?
On Thu, Jan 03, 2002 at 10:29:36PM +0100, Gergely Nagy wrote:
The best thing would be to convince upstream to Do The Right
Thing(tm). Meanwhile, you can add stuff like this to your install target
in debian/rules:
DMANDIR=$(CURDIR)/debian/foobar/usr/share/man
install: build-stamp
${MAKE} install DESTDIR=$(CURDIR)/debian/foobar
mv ${DMANDIR}/man1/thisbelongsto5.1 \
${DMANDIR}/man5/thisbelongsto5.5
...
Not the nicest, but works. If you have lots of manual pages to move,
you might want to store them in a variable, and do a for loop or
something.
Note that you should also change the .TH line at the start of the man
pages to refer to the new section.
Julian
--
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
Julian Gilbey, Dept of Maths, Debian GNU/Linux Developer
Queen Mary, Univ. of London see http://people.debian.org/~jdg/
http://www.maths.qmul.ac.uk/~jdg/ or http://www.debian.org/
Visit http://www.thehungersite.com/ to help feed the hungry
Also: http://www.helpthehungry.org/
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Package naming vs. versioning
On Wed, Nov 28, 2001 at 11:50:11AM -0800, Grant Bowman wrote: Hello, I have seen packages change their package names to include a version number. Reading the policy, there is little guidance on this subject. It would seem that's what Epochs are designed for. However I am aware there could be reasons for wanting to change the package name. One reason might be wanting to run multiple versions concurrently like the libdb or kernel-image packages. Autoconf only renames the older version which seems proper. Another reason might be that a previous packager is AWOL. (1) libdb: Shared library packages have the major version number in their name; see policy 11.3. (2) Epochs are designed for when a mistake is made or the version numbering scheme changes. Remember that epochs are often not shown with the version number, so use them only when essential. (3) Kernel example: yes, you are right. (4) autoconf2.13/autoconf and other examples, such as fvwm1/fvwm: this is only done in cases where the newer package differs so significantly from the older package that the maintainer believes that there is a real need for the old package to continue to exist. In this case, the newer package usually supersedes the old one, but the old one continues to be available if specifically wanted. (5) apache (1.3)/apache2, in the past fvwm (1.x)/fvwm2. Here are examples where the newer package is still in alpha or beta state and not yet ready to be released into the wild. In this case, we don't want people to be automatically upgraded, and so we name the unstable package something different, so people only get the newer package if they specifically request it. At a later stage, when the newer version is more stable, they may rename the newer one to the original name and provide a dummy transition package for brave people. The trouble with changing the package name is lack of clarity. The same software is under different names, perhaps in different distributions like woody vs. potato. It just seems wrong to me in a general case for a standard package to use a version in the package name. What is the right way to handle this? See above for my thoughts. Julian -- =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- Julian Gilbey, Dept of Maths, Debian GNU/Linux Developer Queen Mary, Univ. of London see http://people.debian.org/~jdg/ http://www.maths.qmul.ac.uk/~jdg/ or http://www.debian.org/ Visit http://www.thehungersite.com/ to help feed the hungry Also: http://www.helpthehungry.org/ -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Package naming vs. versioning
On Wed, Nov 28, 2001 at 11:50:11AM -0800, Grant Bowman wrote: Hello, I have seen packages change their package names to include a version number. Reading the policy, there is little guidance on this subject. It would seem that's what Epochs are designed for. However I am aware there could be reasons for wanting to change the package name. One reason might be wanting to run multiple versions concurrently like the libdb or kernel-image packages. Autoconf only renames the older version which seems proper. Another reason might be that a previous packager is AWOL. (1) libdb: Shared library packages have the major version number in their name; see policy 11.3. (2) Epochs are designed for when a mistake is made or the version numbering scheme changes. Remember that epochs are often not shown with the version number, so use them only when essential. (3) Kernel example: yes, you are right. (4) autoconf2.13/autoconf and other examples, such as fvwm1/fvwm: this is only done in cases where the newer package differs so significantly from the older package that the maintainer believes that there is a real need for the old package to continue to exist. In this case, the newer package usually supersedes the old one, but the old one continues to be available if specifically wanted. (5) apache (1.3)/apache2, in the past fvwm (1.x)/fvwm2. Here are examples where the newer package is still in alpha or beta state and not yet ready to be released into the wild. In this case, we don't want people to be automatically upgraded, and so we name the unstable package something different, so people only get the newer package if they specifically request it. At a later stage, when the newer version is more stable, they may rename the newer one to the original name and provide a dummy transition package for brave people. The trouble with changing the package name is lack of clarity. The same software is under different names, perhaps in different distributions like woody vs. potato. It just seems wrong to me in a general case for a standard package to use a version in the package name. What is the right way to handle this? See above for my thoughts. Julian -- =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- Julian Gilbey, Dept of Maths, Debian GNU/Linux Developer Queen Mary, Univ. of London see http://people.debian.org/~jdg/ http://www.maths.qmul.ac.uk/~jdg/ or http://www.debian.org/ Visit http://www.thehungersite.com/ to help feed the hungry Also: http://www.helpthehungry.org/
Re: Semi-contrib packages
On Sat, Nov 24, 2001 at 09:49:29PM -0500, Adam C Powell IV wrote: James Troup wrote: Adam C Powell IV [EMAIL PROTECTED] writes: Can I Build-Depends: ccc [alpha], cfal [alpha] and still have the source package in main? No, that would violate policy (2.1.2). Right, thanks for pointing this out (I need to RTFP :-). So the source would become contrib, because some of the binaries would need contrib software to build on one platform. But then, can I put the gcc/g77-built binaries in main? Hmm, doubtful. No. I guess the proper way to do this would be to assemble a separate contrib petsc-dec source package. But then, the additional petsc_2.1.0.orig.tar.gz would take another ~7 megs on the mirrors! Correct. So maybe this wasn't such a good idea after all. I'll just leave it as-is, with instructions for building the faster version with ccc/cfal. Sounds good. Julian -- =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- Julian Gilbey, Dept of Maths, Debian GNU/Linux Developer Queen Mary, Univ. of London see http://people.debian.org/~jdg/ http://www.maths.qmul.ac.uk/~jdg/ or http://www.debian.org/ Visit http://www.thehungersite.com/ to help feed the hungry Also: http://www.helpthehungry.org/ -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Semi-contrib packages
On Sat, Nov 24, 2001 at 09:49:29PM -0500, Adam C Powell IV wrote: James Troup wrote: Adam C Powell IV [EMAIL PROTECTED] writes: Can I Build-Depends: ccc [alpha], cfal [alpha] and still have the source package in main? No, that would violate policy (2.1.2). Right, thanks for pointing this out (I need to RTFP :-). So the source would become contrib, because some of the binaries would need contrib software to build on one platform. But then, can I put the gcc/g77-built binaries in main? Hmm, doubtful. No. I guess the proper way to do this would be to assemble a separate contrib petsc-dec source package. But then, the additional petsc_2.1.0.orig.tar.gz would take another ~7 megs on the mirrors! Correct. So maybe this wasn't such a good idea after all. I'll just leave it as-is, with instructions for building the faster version with ccc/cfal. Sounds good. Julian -- =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- Julian Gilbey, Dept of Maths, Debian GNU/Linux Developer Queen Mary, Univ. of London see http://people.debian.org/~jdg/ http://www.maths.qmul.ac.uk/~jdg/ or http://www.debian.org/ Visit http://www.thehungersite.com/ to help feed the hungry Also: http://www.helpthehungry.org/
Re: PDF versions of policy documents?
On Fri, Nov 23, 2001 at 10:27:47AM -0800, Sean 'Shaleh' Perry wrote: On 23-Nov-2001 Alexander Karelas wrote: Are there PDF versions (or any version other than HTML for that matter) of the Debian policy manuals somewhere? I would like to print them. policy.sgml.gz is the original source of the policy and it is in docbook's dtd. You can use one of the dockbook converters and output a ps or something else to print. s/docbook/debiandoc/ You could also print the postscript. Or run debiandoc2latexpdf on the sgml file found in the package itself. Julian -- =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- Julian Gilbey, Dept of Maths, Debian GNU/Linux Developer Queen Mary, Univ. of London see http://people.debian.org/~jdg/ http://www.maths.qmul.ac.uk/~jdg/ or http://www.debian.org/ Visit http://www.thehungersite.com/ to help feed the hungry Also: http://www.helpthehungry.org/ -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: PDF versions of policy documents?
On Fri, Nov 23, 2001 at 10:27:47AM -0800, Sean 'Shaleh' Perry wrote: On 23-Nov-2001 Alexander Karelas wrote: Are there PDF versions (or any version other than HTML for that matter) of the Debian policy manuals somewhere? I would like to print them. policy.sgml.gz is the original source of the policy and it is in docbook's dtd. You can use one of the dockbook converters and output a ps or something else to print. s/docbook/debiandoc/ You could also print the postscript. Or run debiandoc2latexpdf on the sgml file found in the package itself. Julian -- =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- Julian Gilbey, Dept of Maths, Debian GNU/Linux Developer Queen Mary, Univ. of London see http://people.debian.org/~jdg/ http://www.maths.qmul.ac.uk/~jdg/ or http://www.debian.org/ Visit http://www.thehungersite.com/ to help feed the hungry Also: http://www.helpthehungry.org/
Re: Bugs not closed by installer
On Tue, Nov 20, 2001 at 03:16:39PM +0100, Martin Michlmayr wrote: * Andreas Rottmann [EMAIL PROTECTED] [20011120 15:07]: In the 'package installed' mail the installer tells me: However the relevant part of the changelog reads: Take a look at the changes file: Closes: 87298 87309 97112 98114 104693 105938 Changes: gql (0.2.0-1) unstable; urgency=low . * Now using uCXX (closes: #98114) * Integrated gql-drivers and GDBI. * Integrated NMU fixes (closes: #87298, #87309) * Changed section to non-US (closes: #97112) * Fixed Build-Depends (closes: #104693, #105938) Files: Note that this changelog entry is significantly different from the one posted to the list; it's not even the same as that of revision 0.2.0-1 there. Andreas, you probably want to investigate this. Julian -- =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- Julian Gilbey, Dept of Maths, Debian GNU/Linux Developer Queen Mary, Univ. of London see http://people.debian.org/~jdg/ http://www.maths.qmul.ac.uk/~jdg/ or http://www.debian.org/ Visit http://www.thehungersite.com/ to help feed the hungry Also: http://www.helpthehungry.org/ -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Bugs not closed by installer
On Tue, Nov 20, 2001 at 03:16:39PM +0100, Martin Michlmayr wrote: * Andreas Rottmann [EMAIL PROTECTED] [20011120 15:07]: In the 'package installed' mail the installer tells me: However the relevant part of the changelog reads: Take a look at the changes file: Closes: 87298 87309 97112 98114 104693 105938 Changes: gql (0.2.0-1) unstable; urgency=low . * Now using uCXX (closes: #98114) * Integrated gql-drivers and GDBI. * Integrated NMU fixes (closes: #87298, #87309) * Changed section to non-US (closes: #97112) * Fixed Build-Depends (closes: #104693, #105938) Files: Note that this changelog entry is significantly different from the one posted to the list; it's not even the same as that of revision 0.2.0-1 there. Andreas, you probably want to investigate this. Julian -- =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- Julian Gilbey, Dept of Maths, Debian GNU/Linux Developer Queen Mary, Univ. of London see http://people.debian.org/~jdg/ http://www.maths.qmul.ac.uk/~jdg/ or http://www.debian.org/ Visit http://www.thehungersite.com/ to help feed the hungry Also: http://www.helpthehungry.org/
Re: dpkg-statoverride question
On Wed, Nov 14, 2001 at 09:35:48AM -0200, Pedro Zorzenon Neto wrote: Hi Mentors, My package (avrprog) uses dpkg-statoverride in the postinst script. Should I put a pre-depends on dpkg? Please be certain that you are using it correctly; you should be vary careful about doing anything more than dpkg-statoverride --list in a maintainer script. Julian -- =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- Julian Gilbey, Dept of Maths, Debian GNU/Linux Developer Queen Mary, Univ. of London see http://people.debian.org/~jdg/ http://www.maths.qmw.ac.uk/~jdg/ or http://www.debian.org/ Visit http://www.thehungersite.com/ to help feed the hungry Also: http://www.helpthehungry.org/
Re: closes: in changelog and closing bug reports...
On Tue, Oct 30, 2001 at 10:55:02PM +0200, Hugo van der Merwe wrote: I have the following in my changelog: * New maintainer (closes: #98251) * Install upstream man page (closes: #93474) * Removed Suggests: wmaker (closes: #82998) however the bugs weren't closed automatically. Might it be because of: * Not uploaded, found new upstream release (I made another new version, and only uploaded then. So this was never the topmost paragraph. Might that be the problem?) Yes. Have a look at the -v option to dpkg-buildpackage. I guess I should just close them by hand then, but am wondering why this happens, is Closes: case sensitive? No. Julian -- =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- Julian Gilbey, Dept of Maths, Queen Mary, Univ. of London Debian GNU/Linux Developer, see http://people.debian.org/~jdg NEW: Visit http://www.helpthehungry.org/ to do just that ALSO http://www.thehungersite.com/ is back! -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: closes: in changelog and closing bug reports...
On Tue, Oct 30, 2001 at 10:55:02PM +0200, Hugo van der Merwe wrote: I have the following in my changelog: * New maintainer (closes: #98251) * Install upstream man page (closes: #93474) * Removed Suggests: wmaker (closes: #82998) however the bugs weren't closed automatically. Might it be because of: * Not uploaded, found new upstream release (I made another new version, and only uploaded then. So this was never the topmost paragraph. Might that be the problem?) Yes. Have a look at the -v option to dpkg-buildpackage. I guess I should just close them by hand then, but am wondering why this happens, is Closes: case sensitive? No. Julian -- =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- Julian Gilbey, Dept of Maths, Queen Mary, Univ. of London Debian GNU/Linux Developer, see http://people.debian.org/~jdg NEW: Visit http://www.helpthehungry.org/ to do just that ALSO http://www.thehungersite.com/ is back!
Re: [BTS] Bug number $bug not found.
On Fri, Oct 26, 2001 at 06:23:53PM +0200, Amaya wrote: I am trying to reopen a bug, tag it and merge it with a more recent one. This is the answer I get from the BTS: Debian Bug Tracking System said: Processing commands for [EMAIL PROTECTED]: reopen 109629 Bug number 109629 not found. tags 109629 upstream Bug number 109629 not found. merge 117116 109629 Bug number 109629 not found. I guess it's older than 28 days, but... can't I access it anymore? Thank you for any ideas. Send a note to 117116 referring to the old bug page if it's the same bug. If it's different and not yet fixed, then resubmit it. Julian -- =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- Julian Gilbey, Dept of Maths, Queen Mary, Univ. of London Debian GNU/Linux Developer, see http://people.debian.org/~jdg NEW: Visit http://www.helpthehungry.org/ to do just that ALSO http://www.thehungersite.com/ is back!
Re: [BTS] Bug number $bug not found.
On Fri, Oct 26, 2001 at 06:23:53PM +0200, Amaya wrote: I am trying to reopen a bug, tag it and merge it with a more recent one. This is the answer I get from the BTS: Debian Bug Tracking System said: Processing commands for [EMAIL PROTECTED]: reopen 109629 Bug number 109629 not found. tags 109629 upstream Bug number 109629 not found. merge 117116 109629 Bug number 109629 not found. I guess it's older than 28 days, but... can't I access it anymore? Thank you for any ideas. Send a note to 117116 referring to the old bug page if it's the same bug. If it's different and not yet fixed, then resubmit it. Julian -- =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- Julian Gilbey, Dept of Maths, Queen Mary, Univ. of London Debian GNU/Linux Developer, see http://people.debian.org/~jdg NEW: Visit http://www.helpthehungry.org/ to do just that ALSO http://www.thehungersite.com/ is back! -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: to conffile or not to conffile
On Tue, Oct 23, 2001 at 05:30:03PM -0200, Henrique de Moraes Holschuh wrote: worry that some package will come along and wipe out their changes simply because it wasn't a conffile. A package that does anything that braindamaged, must not be allowed outside unstable (and its maintainer needs a carefully applied cluebat for maximum skull damage). s/unstable/experimental/ Julian -- =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- Julian Gilbey, Dept of Maths, Queen Mary, Univ. of London Debian GNU/Linux Developer, see http://people.debian.org/~jdg NEW: Visit http://www.helpthehungry.org/ to do just that ALSO http://www.thehungersite.com/ is back! -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: to conffile or not to conffile
On Tue, Oct 23, 2001 at 05:30:03PM -0200, Henrique de Moraes Holschuh wrote: worry that some package will come along and wipe out their changes simply because it wasn't a conffile. A package that does anything that braindamaged, must not be allowed outside unstable (and its maintainer needs a carefully applied cluebat for maximum skull damage). s/unstable/experimental/ Julian -- =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- Julian Gilbey, Dept of Maths, Queen Mary, Univ. of London Debian GNU/Linux Developer, see http://people.debian.org/~jdg NEW: Visit http://www.helpthehungry.org/ to do just that ALSO http://www.thehungersite.com/ is back!
Re: to conffile or not to conffile
On Mon, Oct 22, 2001 at 01:36:53PM -0600, Matt Armstrong wrote: Julian Gilbey [EMAIL PROTECTED] writes: On Fri, Oct 19, 2001 at 02:10:02PM -0500, Warren Turkal wrote: Why not just create a symlink to a device node /dev/flipit and just link it to the right file. Warren Why/how would that help? Julian Then /etc/flipit.conf could just always have /dev/flipit, and the package's maintainer scripts would control where /dev/flipit pointed to. Yuck, yuck, yuck, yuck. Let /etc/flipit.conf just be the symlink or contents directly and not be a conffile. Where's the big problem with that? Incidentally, this doesn't solve the problems, it only hides them: what if the sysadmin wants to change something in the config? Then he has to adjust -- what? /etc/flipit.conf? No, that only contains a #include /dev/flipit command. /dev/flipit? No, that's only a symlink. The target of /dev/flipit? Well, won't that be reset on the next upgrade? This seems a ridiculously complex way of *not* solving the problem. Julian -- =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- Julian Gilbey, Dept of Maths, Queen Mary, Univ. of London Debian GNU/Linux Developer, see http://people.debian.org/~jdg NEW: Visit http://www.helpthehungry.org/ to do just that ALSO http://www.thehungersite.com/ is back! -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: to conffile or not to conffile
On Mon, Oct 22, 2001 at 01:36:53PM -0600, Matt Armstrong wrote: Julian Gilbey [EMAIL PROTECTED] writes: On Fri, Oct 19, 2001 at 02:10:02PM -0500, Warren Turkal wrote: Why not just create a symlink to a device node /dev/flipit and just link it to the right file. Warren Why/how would that help? Julian Then /etc/flipit.conf could just always have /dev/flipit, and the package's maintainer scripts would control where /dev/flipit pointed to. Yuck, yuck, yuck, yuck. Let /etc/flipit.conf just be the symlink or contents directly and not be a conffile. Where's the big problem with that? Incidentally, this doesn't solve the problems, it only hides them: what if the sysadmin wants to change something in the config? Then he has to adjust -- what? /etc/flipit.conf? No, that only contains a #include /dev/flipit command. /dev/flipit? No, that's only a symlink. The target of /dev/flipit? Well, won't that be reset on the next upgrade? This seems a ridiculously complex way of *not* solving the problem. Julian -- =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- Julian Gilbey, Dept of Maths, Queen Mary, Univ. of London Debian GNU/Linux Developer, see http://people.debian.org/~jdg NEW: Visit http://www.helpthehungry.org/ to do just that ALSO http://www.thehungersite.com/ is back!
Re: to conffile or not to conffile
On Fri, Oct 19, 2001 at 02:10:02PM -0500, Warren Turkal wrote: Why not just create a symlink to a device node /dev/flipit and just link it to the right file. Warren Why/how would that help? Julian On Tue, Oct 16, 2001 at 10:46:52PM -0600, Matt Armstrong wrote: I'm packaging my own program (http://www.lickey.com/flipit/) and have a question about configuration files. Section 11.7.3 of the Debian Policy Manual states that the configuration files in /etc must be left alone by the maintainer scripts if it is marked a conffile: The easy way to achieve this behavior is to make the configuration file a conffile. This is appropriate only if it is possible to distribute a default version that will work for most installations, although some system administrators may choose to modify it. This implies that the default version will be part of the package distribution, and must not be modified by the maintainer scripts during installation (or at any other time). So I'm left in a situation where: - Policy forbids me from editing the /etc/flipit.conf if I mark it a conffile. - The program won't work if /etc/flipit.conf isn't edited. - I really do want to make it a conffile, so the user is notified of future options. - I really do want to ask the user what serial port to use when the thing is installed. How do I best resolve this paradox? -- =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- Julian Gilbey, Dept of Maths, Queen Mary, Univ. of London Debian GNU/Linux Developer, see http://people.debian.org/~jdg NEW: Visit http://www.helpthehungry.org/ to do just that ALSO http://www.thehungersite.com/ is back! -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: to conffile or not to conffile
On Fri, Oct 19, 2001 at 02:10:02PM -0500, Warren Turkal wrote: Why not just create a symlink to a device node /dev/flipit and just link it to the right file. Warren Why/how would that help? Julian On Tue, Oct 16, 2001 at 10:46:52PM -0600, Matt Armstrong wrote: I'm packaging my own program (http://www.lickey.com/flipit/) and have a question about configuration files. Section 11.7.3 of the Debian Policy Manual states that the configuration files in /etc must be left alone by the maintainer scripts if it is marked a conffile: The easy way to achieve this behavior is to make the configuration file a conffile. This is appropriate only if it is possible to distribute a default version that will work for most installations, although some system administrators may choose to modify it. This implies that the default version will be part of the package distribution, and must not be modified by the maintainer scripts during installation (or at any other time). So I'm left in a situation where: - Policy forbids me from editing the /etc/flipit.conf if I mark it a conffile. - The program won't work if /etc/flipit.conf isn't edited. - I really do want to make it a conffile, so the user is notified of future options. - I really do want to ask the user what serial port to use when the thing is installed. How do I best resolve this paradox? -- =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- Julian Gilbey, Dept of Maths, Queen Mary, Univ. of London Debian GNU/Linux Developer, see http://people.debian.org/~jdg NEW: Visit http://www.helpthehungry.org/ to do just that ALSO http://www.thehungersite.com/ is back!
Re: to conffile or not to conffile
On Tue, Oct 16, 2001 at 10:46:52PM -0600, Matt Armstrong wrote: I'm packaging my own program (http://www.lickey.com/flipit/) and have a question about configuration files. So I'm left in a situation where: - Policy forbids me from editing the /etc/flipit.conf if I mark it a conffile. Correct. - The program won't work if /etc/flipit.conf isn't edited. Who needs to edit it, the sysadmin or the maintainer scripts? - I really do want to make it a conffile, so the user is notified of future options. There are other ways to do this -- see below. - I really do want to ask the user what serial port to use when the thing is installed. Use debconf to get this info. Appropriate solution: The first time this package is installed, use debconf to get the info, then write a config file to /etc. Maybe add a couple of comment lines: # Please do not modify the next line # config_version=0.3 On upgrades, check the config_version line, and if you are now introducing a later version, either use debconf to get the desired newer config settings or just append the default values to the existing config file with appropriate comments, taking care to update the config_version line and not to affect any of the sysadmin's settings. Remember that if every package told the sysadmin of all the new bells and whistles, upgrades would be dreadful; you need only ask them for new information really needed via debconf. Julian -- =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- Julian Gilbey, Dept of Maths, Queen Mary, Univ. of London Debian GNU/Linux Developer, see http://people.debian.org/~jdg NEW: Visit http://www.helpthehungry.org/ to do just that ALSO http://www.thehungersite.com/ is back!
Re: hosed build-arch and build-indep
On Sat, Oct 13, 2001 at 10:55:00AM +0200, Michael Banck wrote: 1. Shouldn't lintian complain about this when Architecture is set to 'any' and binary-arch is empty but binary-indep is crowded? File a wishlist bug against lintian. 2. Yesterday evening I uploaded another (new) python program. Same problems, this time because the package got a C-library added since I first packaged it and I didn't think of this :( This package is now in incoming waiting for the ftp-admin to add it to the override file. Should I - wait until it is in the archive and upload a fixed version then? - delete the files from incoming and upload Debian Revision -1 again? - upload the new -2 revision now? General principle: never overwrite a file already in the incoming queue. Delete the files from incoming and upload a version -2. (You should use the -sa option to dpkg-buildpackage.) Julian -- =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- Julian Gilbey, Dept of Maths, Queen Mary, Univ. of London Debian GNU/Linux Developer, see http://people.debian.org/~jdg NEW: Visit http://www.helpthehungry.org/ to do just that ALSO http://www.thehungersite.com/ is back! -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: hosed build-arch and build-indep
On Sat, Oct 13, 2001 at 10:55:00AM +0200, Michael Banck wrote: 1. Shouldn't lintian complain about this when Architecture is set to 'any' and binary-arch is empty but binary-indep is crowded? File a wishlist bug against lintian. 2. Yesterday evening I uploaded another (new) python program. Same problems, this time because the package got a C-library added since I first packaged it and I didn't think of this :( This package is now in incoming waiting for the ftp-admin to add it to the override file. Should I - wait until it is in the archive and upload a fixed version then? - delete the files from incoming and upload Debian Revision -1 again? - upload the new -2 revision now? General principle: never overwrite a file already in the incoming queue. Delete the files from incoming and upload a version -2. (You should use the -sa option to dpkg-buildpackage.) Julian -- =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- Julian Gilbey, Dept of Maths, Queen Mary, Univ. of London Debian GNU/Linux Developer, see http://people.debian.org/~jdg NEW: Visit http://www.helpthehungry.org/ to do just that ALSO http://www.thehungersite.com/ is back!
Re: dpkg-deb vs. perl
On Fri, Oct 05, 2001 at 03:46:51PM +0200, P?sztor Gy?rgy wrote: Hi, How can I get info, or manipulate .deb packages from perl, like the dpkg-deb does in command line? Using system(dpkg-deb [options]) is a good way. Julian -- =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- Julian Gilbey, Dept of Maths, Queen Mary, Univ. of London Debian GNU/Linux Developer, see http://people.debian.org/~jdg NEW: Visit http://www.helpthehungry.org/ to do just that ALSO http://www.thehungersite.com/ is back!
Re: GPG Key Signing in UK
On Fri, Sep 21, 2001 at 08:02:30PM +0100, Scott James Remnant wrote: I live in Birmingham, UK and work in London, UK, so arranging a meet in/around these two cities is usually possible at any time. You may want to join the debian-uk list: see http://www.chiark.greenend.org.uk/mailman/listinfo/debian-uk Julian -- =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- Julian Gilbey, Dept of Maths, Queen Mary, Univ. of London Debian GNU/Linux Developer, see http://people.debian.org/~jdg NEW: Visit http://www.helpthehungry.org/ to do just that -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: GPG Key Signing in UK
On Fri, Sep 21, 2001 at 08:02:30PM +0100, Scott James Remnant wrote: I live in Birmingham, UK and work in London, UK, so arranging a meet in/around these two cities is usually possible at any time. You may want to join the debian-uk list: see http://www.chiark.greenend.org.uk/mailman/listinfo/debian-uk Julian -- =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- Julian Gilbey, Dept of Maths, Queen Mary, Univ. of London Debian GNU/Linux Developer, see http://people.debian.org/~jdg NEW: Visit http://www.helpthehungry.org/ to do just that
Re: trouble with HTML changelogs and SGML catalogs
On Wed, Sep 19, 2001 at 04:23:47PM -0400, Colin Walters wrote: I have two problems with it. First, lintian complains: E: xml-resume-library: changelog-file-not-compressed changelog 13585 You can add a lintian override; also see the other responses. Julian -- =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- Julian Gilbey, Dept of Maths, Queen Mary, Univ. of London Debian GNU/Linux Developer, see http://people.debian.org/~jdg NEW: Visit http://www.helpthehungry.org/ to do just that
Re: Native packages
On Thu, Sep 06, 2001 at 04:00:25PM +0200, peter karlsson wrote: Colin Watson: If you can't build the Debian package as part of the process of generating the tarball from CVS, I suppose you could use -b and hack the .changes by hand to include the source, although that's rather ugly. I tried hacking the changes file by hand, but my upload got rejected three times so I gave up... :-/ For what reason was it rejected? If you upload source files, you'll also need a .dsc file, and to resign everything using debsign. Julian -- =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- Julian Gilbey, Dept of Maths, Queen Mary, Univ. of London Debian GNU/Linux Developer, see http://people.debian.org/~jdg NEW: Visit http://www.helpthehungry.org/ to do just that -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Native packages
On Thu, Sep 06, 2001 at 04:00:25PM +0200, peter karlsson wrote: Colin Watson: If you can't build the Debian package as part of the process of generating the tarball from CVS, I suppose you could use -b and hack the .changes by hand to include the source, although that's rather ugly. I tried hacking the changes file by hand, but my upload got rejected three times so I gave up... :-/ For what reason was it rejected? If you upload source files, you'll also need a .dsc file, and to resign everything using debsign. Julian -- =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- Julian Gilbey, Dept of Maths, Queen Mary, Univ. of London Debian GNU/Linux Developer, see http://people.debian.org/~jdg NEW: Visit http://www.helpthehungry.org/ to do just that
Re: Native packages
On Wed, Sep 05, 2001 at 10:49:51PM +0200, peter karlsson wrote: Hi! How do I get dpkg-buildpackage not to re-build the source tarball when building a native package? No matter what I do, it rebuilds it, which prevents me from keeping the tarball I created from my CVS tree, which also is what I distribute elsewhere. You can't, AFAIK. If you distribute it elsewhere (presumably beyond Debian), you may want to consider treating it as a non-native package instead. Julian -- =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- Julian Gilbey, Dept of Maths, Queen Mary, Univ. of London Debian GNU/Linux Developer, see http://people.debian.org/~jdg NEW: Visit http://www.helpthehungry.org/ to do just that
Re: dpkg can't purge init script
On Thu, Aug 23, 2001 at 11:35:56AM +0200, Eric Van Buggenhaut wrote: *) echo prerm called with unknown argument \`$1' 2 exit 0 This should probably be exit 1. Well, my script is the exact copy of the skeleton found in /usr/share/debhelper/dh_make/debian/prerm.ex where it is exit 0 Hmm. I still think I'm probably right; an error occurred. Again, /etc/init.d/superviser-server is a copy of the skeleton found in /usr/share/debhelper/debian/init.d.ex : [...] Problem seems to be that prerm remove fails : Because : curitiba_POTATO:/# start-stop-daemon --stop --quiet --pidfile /var/run/superviser-server.pid --exec /usr/sbin/superviser-server.pl ; echo $? 1 curitiba_POTATO:/# If you have any idea ... As somebody pointed out in this thread, the call to start-stop-daemon should use the --oknodo option. Julian -- =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- Julian Gilbey, Dept of Maths, Queen Mary, Univ. of London Debian GNU/Linux Developer, see http://people.debian.org/~jdg -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: dpkg can't purge init script
On Fri, Aug 24, 2001 at 12:35:11AM +0200, Eric Van Buggenhaut wrote: That's the only thing it could be. Your init.d script is failing and returning 1. It is bugged, fix it. Hint: --oknodo parameter to start-stop-daemon (look it up). OK, --oknodo did the trick. Thanks. I don't see the point why this option isn't shipped in the default file. Bug filed against dh-make package. Julian -- =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- Julian Gilbey, Dept of Maths, Queen Mary, Univ. of London Debian GNU/Linux Developer, see http://people.debian.org/~jdg -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: dpkg can't purge init script
On Thu, Aug 23, 2001 at 11:35:56AM +0200, Eric Van Buggenhaut wrote: *) echo prerm called with unknown argument \`$1' 2 exit 0 This should probably be exit 1. Well, my script is the exact copy of the skeleton found in /usr/share/debhelper/dh_make/debian/prerm.ex where it is exit 0 Hmm. I still think I'm probably right; an error occurred. Again, /etc/init.d/superviser-server is a copy of the skeleton found in /usr/share/debhelper/debian/init.d.ex : [...] Problem seems to be that prerm remove fails : Because : curitiba_POTATO:/# start-stop-daemon --stop --quiet --pidfile /var/run/superviser-server.pid --exec /usr/sbin/superviser-server.pl ; echo $? 1 curitiba_POTATO:/# If you have any idea ... As somebody pointed out in this thread, the call to start-stop-daemon should use the --oknodo option. Julian -- =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- Julian Gilbey, Dept of Maths, Queen Mary, Univ. of London Debian GNU/Linux Developer, see http://people.debian.org/~jdg
Re: dpkg can't purge init script
On Fri, Aug 24, 2001 at 12:35:11AM +0200, Eric Van Buggenhaut wrote: That's the only thing it could be. Your init.d script is failing and returning 1. It is bugged, fix it. Hint: --oknodo parameter to start-stop-daemon (look it up). OK, --oknodo did the trick. Thanks. I don't see the point why this option isn't shipped in the default file. Bug filed against dh-make package. Julian -- =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- Julian Gilbey, Dept of Maths, Queen Mary, Univ. of London Debian GNU/Linux Developer, see http://people.debian.org/~jdg
Re: dpkg can't purge init script
On Wed, Aug 22, 2001 at 10:28:42PM +0100, Andrew Suffield wrote: Yes actually, it should be guarded with a -e: if [ -e /etc/init.d/superviser-ircd ]; then /etc/init.d/superviser-ircd stop fi Should better be a -x, not -e. But that's a small point. Julian -- =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- Julian Gilbey, Dept of Maths, Queen Mary, Univ. of London Debian GNU/Linux Developer, see http://people.debian.org/~jdg -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: dpkg can't purge init script
On Wed, Aug 22, 2001 at 10:28:42PM +0100, Andrew Suffield wrote: Yes actually, it should be guarded with a -e: if [ -e /etc/init.d/superviser-ircd ]; then /etc/init.d/superviser-ircd stop fi Should better be a -x, not -e. But that's a small point. Julian -- =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- Julian Gilbey, Dept of Maths, Queen Mary, Univ. of London Debian GNU/Linux Developer, see http://people.debian.org/~jdg
Re: proud maintainer that do no want NUM :-)
On Mon, Jul 09, 2001 at 05:32:31PM +0200, Samuel Tardieu wrote: On 9/07, Stefano Zacchiroli wrote: | I know that is a really silly question, but I really dislike fixed in | NMU on my own bug pages :-))) Send a mail containing close ... in the body (by using the right bug number) to [EMAIL PROTECTED] Don't do that; email [EMAIL PROTECTED] with a copy of the changelog entry instead, otherwise the submitter will get this uninformative message that the bug has been closed with no rationale. Julian -- =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- Julian Gilbey, Dept of Maths, Queen Mary, Univ. of London Debian GNU/Linux Developer, see http://people.debian.org/~jdg Donate free food to the world's hungry: see http://www.thehungersite.com/ -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: proud maintainer that do no want NUM :-)
On Mon, Jul 09, 2001 at 05:32:31PM +0200, Samuel Tardieu wrote: On 9/07, Stefano Zacchiroli wrote: | I know that is a really silly question, but I really dislike fixed in | NMU on my own bug pages :-))) Send a mail containing close ... in the body (by using the right bug number) to [EMAIL PROTECTED] Don't do that; email [EMAIL PROTECTED] with a copy of the changelog entry instead, otherwise the submitter will get this uninformative message that the bug has been closed with no rationale. Julian -- =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- Julian Gilbey, Dept of Maths, Queen Mary, Univ. of London Debian GNU/Linux Developer, see http://people.debian.org/~jdg Donate free food to the world's hungry: see http://www.thehungersite.com/
Re: 1st try at packaging perl program
On Fri, Jun 29, 2001 at 04:35:11PM +0100, Etienne Grossmann wrote: From: Paolo Molaro [EMAIL PROTECTED] # Would anyone be willing to send me the source tree of a debian # package that wraps a perl program/module (or send me the url for one)? # # apt-get source lib*-perl should do it. Silly question : where do I get source packages? I did what I think correct from the sources.list, apt-get manpages (put deb-src lines in sources.list, just like deb lines), but still get : anonimo:/home/etienne/prog/perl/bloksi/tmp# apt-get source mailtools Reading Package Lists... Done Building Dependency Tree... Done E: Could not open file /var/state/apt/lists/sunsite.org.uk_Mirrors_ftp.debian.org_pub_debian_dists_potato_contrib_source_Sources - open (2 No such file or directory) Have you done an apt-get update or dselect update? Julian -- =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- Julian Gilbey, Dept of Maths, Queen Mary, Univ. of London Debian GNU/Linux Developer, see http://people.debian.org/~jdg Donate free food to the world's hungry: see http://www.thehungersite.com/
Re: binary-all packaging
On Wed, Jun 27, 2001 at 03:13:31AM +0200, Robert Millan wrote: dpkg-genchanges: failure: cannot read files list file: No such file or directory What am i doing wrong? I guess i have a problem in my rules file. just in case i post it here too: Your rules file is completely ill. Scrap it, copy /usr/share/doc/debhelper/examples/rules.indep to debian/rules and modify it as necessary to build your package. Julian -- =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- Julian Gilbey, Dept of Maths, Queen Mary, Univ. of London Debian GNU/Linux Developer, see http://people.debian.org/~jdg Donate free food to the world's hungry: see http://www.thehungersite.com/ -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: binary-all packaging
On Wed, Jun 27, 2001 at 03:13:31AM +0200, Robert Millan wrote: dpkg-genchanges: failure: cannot read files list file: No such file or directory What am i doing wrong? I guess i have a problem in my rules file. just in case i post it here too: Your rules file is completely ill. Scrap it, copy /usr/share/doc/debhelper/examples/rules.indep to debian/rules and modify it as necessary to build your package. Julian -- =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- Julian Gilbey, Dept of Maths, Queen Mary, Univ. of London Debian GNU/Linux Developer, see http://people.debian.org/~jdg Donate free food to the world's hungry: see http://www.thehungersite.com/
Re: lintian: empty (transition) packages shouldn't need a copyright file
On Tue, Jun 26, 2001 at 09:00:18AM -0500, Manoj Srivastava wrote: Even an otherwise empty package contains metadata files (or else what good does it do?), and these metadata files need a copyright file giving the license under which they are distributable, So I don't think there is anything wrong with Lintian complaining about the lack of a copyright notice -- there should always be one, if only for the metadata. If this is an issue, we have a problem, as (almost?) no non-native packages have copyright statements on their metadata. Julian -- =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- Julian Gilbey, Dept of Maths, Queen Mary, Univ. of London Debian GNU/Linux Developer, see http://people.debian.org/~jdg Donate free food to the world's hungry: see http://www.thehungersite.com/ -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: lintian: empty (transition) packages shouldn't need a copyright file
On Tue, Jun 26, 2001 at 09:00:18AM -0500, Manoj Srivastava wrote: Even an otherwise empty package contains metadata files (or else what good does it do?), and these metadata files need a copyright file giving the license under which they are distributable, So I don't think there is anything wrong with Lintian complaining about the lack of a copyright notice -- there should always be one, if only for the metadata. If this is an issue, we have a problem, as (almost?) no non-native packages have copyright statements on their metadata. Julian -- =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- Julian Gilbey, Dept of Maths, Queen Mary, Univ. of London Debian GNU/Linux Developer, see http://people.debian.org/~jdg Donate free food to the world's hungry: see http://www.thehungersite.com/
Re: Bug#101325: lintian: empty (transition) packages shouldn't n
On Thu, Jun 21, 2001 at 10:53:07AM +0300, Richard Braakman wrote: On Wed, Jun 20, 2001 at 05:10:36PM +0100, Julian Gilbey wrote: That's what the next PACKAGE; achieves. Alternatively, in such a case, set a variable $empty_package = 1 and then test the value of this variable around the checks block (lines 1127--1221). There may be some other things which should be blocked out, but this does not seem intrinsically difficult. It's slightly more difficult than that, because you do want to keep the warnings about the control file and other metadata. Oh, hadn't thought of that. You are right. Julian -- =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- Julian Gilbey, Dept of Maths, Queen Mary, Univ. of London Debian GNU/Linux Developer, see http://people.debian.org/~jdg Donate free food to the world's hungry: see http://www.thehungersite.com/ -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Bug#101325: lintian: empty (transition) packages shouldn't n
On Thu, Jun 21, 2001 at 10:53:07AM +0300, Richard Braakman wrote: On Wed, Jun 20, 2001 at 05:10:36PM +0100, Julian Gilbey wrote: That's what the next PACKAGE; achieves. Alternatively, in such a case, set a variable $empty_package = 1 and then test the value of this variable around the checks block (lines 1127--1221). There may be some other things which should be blocked out, but this does not seem intrinsically difficult. It's slightly more difficult than that, because you do want to keep the warnings about the control file and other metadata. Oh, hadn't thought of that. You are right. Julian -- =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- Julian Gilbey, Dept of Maths, Queen Mary, Univ. of London Debian GNU/Linux Developer, see http://people.debian.org/~jdg Donate free food to the world's hungry: see http://www.thehungersite.com/
Re: Bug#101325: lintian: empty (transition) packages shouldn't n
On Tue, Jun 19, 2001 at 08:16:59PM -0700, Sean 'Shaleh' Perry wrote:
On 19-Jun-2001 Julian Gilbey wrote:
On Mon, Jun 18, 2001 at 06:24:52PM -0700, Sean 'Shaleh' Perry wrote:
And how am I to tell a package that is empty on purpose from one created by
accident? What constitutes empty?
Maybe have a lintian warning about an empty package, but no error
messages.
while not opposed to this, it is definately non-trivial to implement, sorry all.
Somewhere around line 1100 of /usr/bin/lintian (not sure exactly
where), you could essentially have the following code:
{
local $/;
open FILELIST, $base/index;
my $filelist = FILELIST;
if ($filelist eq ./\n) {
print W: no files in binary package $pkg\n;
next PACKAGE;
}
}
Julian
--
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
Julian Gilbey, Dept of Maths, Queen Mary, Univ. of London
Debian GNU/Linux Developer, see http://people.debian.org/~jdg
Donate free food to the world's hungry: see http://www.thehungersite.com/
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Bug#101325: lintian: empty (transition) packages shouldn't n
On Wed, Jun 20, 2001 at 08:17:03AM -0700, Sean 'Shaleh' Perry wrote:
Somewhere around line 1100 of /usr/bin/lintian (not sure exactly
where), you could essentially have the following code:
{
local $/;
open FILELIST, $base/index;
my $filelist = FILELIST;
if ($filelist eq ./\n) {
print W: no files in binary package $pkg\n;
next PACKAGE;
}
}
correct. The hard part is not outputting any other errors because of it.
Avoiding the no copyright, no docs, etc would not be as easy.
That's what the next PACKAGE; achieves.
Alternatively, in such a case, set a variable $empty_package = 1 and
then test the value of this variable around the checks block (lines
1127--1221). There may be some other things which should be blocked
out, but this does not seem intrinsically difficult.
Julian
--
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
Julian Gilbey, Dept of Maths, Queen Mary, Univ. of London
Debian GNU/Linux Developer, see http://people.debian.org/~jdg
Donate free food to the world's hungry: see http://www.thehungersite.com/
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Bug#101325: lintian: empty (transition) packages shouldn't n
On Tue, Jun 19, 2001 at 08:16:59PM -0700, Sean 'Shaleh' Perry wrote:
On 19-Jun-2001 Julian Gilbey wrote:
On Mon, Jun 18, 2001 at 06:24:52PM -0700, Sean 'Shaleh' Perry wrote:
And how am I to tell a package that is empty on purpose from one created by
accident? What constitutes empty?
Maybe have a lintian warning about an empty package, but no error
messages.
while not opposed to this, it is definately non-trivial to implement, sorry
all.
Somewhere around line 1100 of /usr/bin/lintian (not sure exactly
where), you could essentially have the following code:
{
local $/;
open FILELIST, $base/index;
my $filelist = FILELIST;
if ($filelist eq ./\n) {
print W: no files in binary package $pkg\n;
next PACKAGE;
}
}
Julian
--
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
Julian Gilbey, Dept of Maths, Queen Mary, Univ. of London
Debian GNU/Linux Developer, see http://people.debian.org/~jdg
Donate free food to the world's hungry: see http://www.thehungersite.com/
Re: Bug#101325: lintian: empty (transition) packages shouldn't n
On Wed, Jun 20, 2001 at 08:17:03AM -0700, Sean 'Shaleh' Perry wrote:
Somewhere around line 1100 of /usr/bin/lintian (not sure exactly
where), you could essentially have the following code:
{
local $/;
open FILELIST, $base/index;
my $filelist = FILELIST;
if ($filelist eq ./\n) {
print W: no files in binary package $pkg\n;
next PACKAGE;
}
}
correct. The hard part is not outputting any other errors because of it.
Avoiding the no copyright, no docs, etc would not be as easy.
That's what the next PACKAGE; achieves.
Alternatively, in such a case, set a variable $empty_package = 1 and
then test the value of this variable around the checks block (lines
1127--1221). There may be some other things which should be blocked
out, but this does not seem intrinsically difficult.
Julian
--
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
Julian Gilbey, Dept of Maths, Queen Mary, Univ. of London
Debian GNU/Linux Developer, see http://people.debian.org/~jdg
Donate free food to the world's hungry: see http://www.thehungersite.com/
Re: Bug#101325: lintian: empty (transition) packages shouldn't n
On Mon, Jun 18, 2001 at 06:24:52PM -0700, Sean 'Shaleh' Perry wrote: And how am I to tell a package that is empty on purpose from one created by accident? What constitutes empty? Maybe have a lintian warning about an empty package, but no error messages. Julian -- =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- Julian Gilbey, Dept of Maths, Queen Mary, Univ. of London Debian GNU/Linux Developer, see http://people.debian.org/~jdg Donate free food to the world's hungry: see http://www.thehungersite.com/ -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: control file question
On Tue, Jun 19, 2001 at 01:39:58PM +0200, Abraham vd Merwe wrote: Hi Julian! The problem I have is that I have a library with same host architecture but different cross compilation targets, so I have things like: Build-Depends supports an arch specification, like: Build-Depends: libc-arm-version (=2.1) [arm] See policy section 7.1 for details. I know about that, but that doesn't help since it won't be compiled on that host architecture. It's for cross compiling, so for something like dietlibc-arm-dev you'll get Build-Depends that differ from dietlibc-dev on the _same_ host architecture. But surely you can determine the packages needeed when dpkg-buildpackage is run on the various different architectures? I'm not sure I understand what the problem is. Julian -- =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- Julian Gilbey, Dept of Maths, Queen Mary, Univ. of London Debian GNU/Linux Developer, see http://people.debian.org/~jdg Donate free food to the world's hungry: see http://www.thehungersite.com/ -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: removing old conffiles on upgrade
On Tue, Jun 19, 2001 at 05:54:25PM +0100, Muhammad Hussain Yusuf wrote: Hi, I've just changed the directories of the conffile, binaries etc. to conform with Policy 3.5.5.0. The old conffile dir was /etc/X11/filerunner and the new one is /etc/filerunner and the problem is that when installing new package, and an older version is installed, dpkg gives an error message, and the old conffile directory, not being empty, does not get removed. You could you copy the old version to the new in the preinst and remove the old version in the postinst. Yes, dpkg will warn that the old directory is not empty, but don't worry about it. But be careful to check the value of $1 and only to do them when the value is sensible (eg not when it's abort-*). Julian -- =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- Julian Gilbey, Dept of Maths, Queen Mary, Univ. of London Debian GNU/Linux Developer, see http://people.debian.org/~jdg Donate free food to the world's hungry: see http://www.thehungersite.com/ -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Bug#101325: lintian: empty (transition) packages shouldn't n
On Mon, Jun 18, 2001 at 06:24:52PM -0700, Sean 'Shaleh' Perry wrote: And how am I to tell a package that is empty on purpose from one created by accident? What constitutes empty? Maybe have a lintian warning about an empty package, but no error messages. Julian -- =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- Julian Gilbey, Dept of Maths, Queen Mary, Univ. of London Debian GNU/Linux Developer, see http://people.debian.org/~jdg Donate free food to the world's hungry: see http://www.thehungersite.com/
Re: control file question
On Tue, Jun 19, 2001 at 11:50:52AM +0200, Abraham vd Merwe wrote: Hi! Can I put a different Build-Depends for each binary package in a control file or can there only be a Build-Depends for the source package? You can only have one Build-Depends in the control file for the source package. The problem I have is that I have a library with same host architecture but different cross compilation targets, so I have things like: Build-Depends supports an arch specification, like: Build-Depends: libc-arm-version (=2.1) [arm] See policy section 7.1 for details. Julian -- =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- Julian Gilbey, Dept of Maths, Queen Mary, Univ. of London Debian GNU/Linux Developer, see http://people.debian.org/~jdg Donate free food to the world's hungry: see http://www.thehungersite.com/
Re: control file question
On Tue, Jun 19, 2001 at 01:39:58PM +0200, Abraham vd Merwe wrote: Hi Julian! The problem I have is that I have a library with same host architecture but different cross compilation targets, so I have things like: Build-Depends supports an arch specification, like: Build-Depends: libc-arm-version (=2.1) [arm] See policy section 7.1 for details. I know about that, but that doesn't help since it won't be compiled on that host architecture. It's for cross compiling, so for something like dietlibc-arm-dev you'll get Build-Depends that differ from dietlibc-dev on the _same_ host architecture. But surely you can determine the packages needeed when dpkg-buildpackage is run on the various different architectures? I'm not sure I understand what the problem is. Julian -- =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- Julian Gilbey, Dept of Maths, Queen Mary, Univ. of London Debian GNU/Linux Developer, see http://people.debian.org/~jdg Donate free food to the world's hungry: see http://www.thehungersite.com/
Re: removing old conffiles on upgrade
On Tue, Jun 19, 2001 at 05:54:25PM +0100, Muhammad Hussain Yusuf wrote: Hi, I've just changed the directories of the conffile, binaries etc. to conform with Policy 3.5.5.0. The old conffile dir was /etc/X11/filerunner and the new one is /etc/filerunner and the problem is that when installing new package, and an older version is installed, dpkg gives an error message, and the old conffile directory, not being empty, does not get removed. You could you copy the old version to the new in the preinst and remove the old version in the postinst. Yes, dpkg will warn that the old directory is not empty, but don't worry about it. But be careful to check the value of $1 and only to do them when the value is sensible (eg not when it's abort-*). Julian -- =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- Julian Gilbey, Dept of Maths, Queen Mary, Univ. of London Debian GNU/Linux Developer, see http://people.debian.org/~jdg Donate free food to the world's hungry: see http://www.thehungersite.com/
Re: Replaces: and pre/post rm
On Fri, Jun 15, 2001 at 10:13:53AM -0400, Richard A Nelson wrote: Ah, very good point. Why not simply add in the purge section a block around it which says: if ! [ -f some file in the .deb which characterises the sendmail(-tls) packages ] then delete a lot of files fi Because I have to allow for the user switching from non-tls to tls *OR* from tls to non-tls. The *only* differences in the contents of the two packages are: Exactly, so the postrm which is identical in both packages will say: if [ $1' = purge ]; then if ! [ -f /usr/lib/sm.bin/sendmail ]; then # We're *really* purging; neither sendmail nor sendmail-tls # is currently installed delete a lot of files fi fi Then if one of them is installed, purging the other will not actually do anything dangerous, as /usr/lib/sm.bin/sendmail will be present on the system (it's in both packages). Julian -- =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- Julian Gilbey, Dept of Maths, Queen Mary, Univ. of London Debian GNU/Linux Developer, see http://people.debian.org/~jdg Donate free food to the world's hungry: see http://www.thehungersite.com/ -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: one package, many libraries. dependency break!
On Fri, Jun 15, 2001 at 07:28:47PM +0200, Domenico Andreoli wrote: i'm working on a package (libpspell2) which installthree libraries (libpspell.so.2.0.2, libpspell-impl.so.3.0.1, libpspell-modules.so.1.0.1), as expected it exports a shlibs which lists all of the three libs. they depend on each other :( the problem is dpkg-shlibdeps that figures out the package(s) from which this should depend on. in fact it sees that libpspell-impl depends on libpspell and adds libpspell2 (which is the package that provides it, that is this) to the dependecy list. so i get a package that depends on itself. any considerations? should i split this package in three?!? What is the exact dpkg-shlibdeps command you are using? Have you set LD_LIBRARY_PATH to point to the location of the newly generated libraries in the installation tree, thus, for example, for a pspell package: LD_LIBRARY_PATH=debian/libpspell2/usr/lib dpkg-shlibdeps \ debian/pspell/usr/bin/* Julian -- =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- Julian Gilbey, Dept of Maths, Queen Mary, Univ. of London Debian GNU/Linux Developer, see http://people.debian.org/~jdg Donate free food to the world's hungry: see http://www.thehungersite.com/ -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Empty package
On Mon, Jun 18, 2001 at 01:07:57PM +1200, Michael Beattie wrote: On Sun, Jun 17, 2001 at 04:46:20PM -0400, Bob Hilliard wrote: In order to be allowed in the archive, will it be necessary to provide a copyright file? No. Er, YES. Er, NO. My package sgb-src contains 0 (zero) non-control files. It's data.tar.gz is empty (except for ./). Similarly many, many other transition packages. Julian -- =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- Julian Gilbey, Dept of Maths, Queen Mary, Univ. of London Debian GNU/Linux Developer, see http://people.debian.org/~jdg Donate free food to the world's hungry: see http://www.thehungersite.com/ -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Empty package
On Sun, Jun 17, 2001 at 04:46:20PM -0400, Bob Hilliard wrote: Thanks, Julian. This works like a charm, but Lintian has the following objections: E: dict-web1913: no-copyright-file W: dict-web1913: prerm-does-not-remove-usr-doc-link W: dict-web1913: postinst-does-not-set-usr-doc-link Will the Lintian error prevent it from being installed? Shouldn't do; there is no content there to need a copyright. Julian -- =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- Julian Gilbey, Dept of Maths, Queen Mary, Univ. of London Debian GNU/Linux Developer, see http://people.debian.org/~jdg Donate free food to the world's hungry: see http://www.thehungersite.com/ -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: one package, many libraries. dependency break!
On Mon, Jun 18, 2001 at 04:40:46PM +0200, Domenico Andreoli wrote: In the past I have solved this problem by invoking dpkg-shlibdeps with a LD_LIBRARY_PATH setting that makes it find the libaries in the build tree. This way you get a warning that it couldn't find what package provides them, but you don't get a bogus dependency. I haven't found a cleaner solution. is LD_LIBRARY_PATH solution alternative to shlibs files? and if a libspell is already installing on the building system? wouldn't i need to put a !libpspell2 in Build-Depends? (You mean using Build-Conflicts.) Using LD_LIBRARY_PATH correctly means that the shlibs file from the new package being built will be used, rather than from the one on the system (if any). So you don't need a Build-Conflicts; you just need to get the dh_shlibdeps call correct. (I've sent you a private email on how to do this; basically use the -l option.) Julian -- =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- Julian Gilbey, Dept of Maths, Queen Mary, Univ. of London Debian GNU/Linux Developer, see http://people.debian.org/~jdg Donate free food to the world's hungry: see http://www.thehungersite.com/ -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Replaces: and pre/post rm
On Fri, Jun 15, 2001 at 10:13:53AM -0400, Richard A Nelson wrote: Ah, very good point. Why not simply add in the purge section a block around it which says: if ! [ -f some file in the .deb which characterises the sendmail(-tls) packages ] then delete a lot of files fi Because I have to allow for the user switching from non-tls to tls *OR* from tls to non-tls. The *only* differences in the contents of the two packages are: Exactly, so the postrm which is identical in both packages will say: if [ $1' = purge ]; then if ! [ -f /usr/lib/sm.bin/sendmail ]; then # We're *really* purging; neither sendmail nor sendmail-tls # is currently installed delete a lot of files fi fi Then if one of them is installed, purging the other will not actually do anything dangerous, as /usr/lib/sm.bin/sendmail will be present on the system (it's in both packages). Julian -- =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- Julian Gilbey, Dept of Maths, Queen Mary, Univ. of London Debian GNU/Linux Developer, see http://people.debian.org/~jdg Donate free food to the world's hungry: see http://www.thehungersite.com/
Re: one package, many libraries. dependency break!
On Fri, Jun 15, 2001 at 07:28:47PM +0200, Domenico Andreoli wrote: i'm working on a package (libpspell2) which installthree libraries (libpspell.so.2.0.2, libpspell-impl.so.3.0.1, libpspell-modules.so.1.0.1), as expected it exports a shlibs which lists all of the three libs. they depend on each other :( the problem is dpkg-shlibdeps that figures out the package(s) from which this should depend on. in fact it sees that libpspell-impl depends on libpspell and adds libpspell2 (which is the package that provides it, that is this) to the dependecy list. so i get a package that depends on itself. any considerations? should i split this package in three?!? What is the exact dpkg-shlibdeps command you are using? Have you set LD_LIBRARY_PATH to point to the location of the newly generated libraries in the installation tree, thus, for example, for a pspell package: LD_LIBRARY_PATH=debian/libpspell2/usr/lib dpkg-shlibdeps \ debian/pspell/usr/bin/* Julian -- =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- Julian Gilbey, Dept of Maths, Queen Mary, Univ. of London Debian GNU/Linux Developer, see http://people.debian.org/~jdg Donate free food to the world's hungry: see http://www.thehungersite.com/
Re: Empty package
On Mon, Jun 18, 2001 at 01:07:57PM +1200, Michael Beattie wrote: On Sun, Jun 17, 2001 at 04:46:20PM -0400, Bob Hilliard wrote: In order to be allowed in the archive, will it be necessary to provide a copyright file? No. Er, YES. Er, NO. My package sgb-src contains 0 (zero) non-control files. It's data.tar.gz is empty (except for ./). Similarly many, many other transition packages. Julian -- =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- Julian Gilbey, Dept of Maths, Queen Mary, Univ. of London Debian GNU/Linux Developer, see http://people.debian.org/~jdg Donate free food to the world's hungry: see http://www.thehungersite.com/
Re: Empty package
On Sun, Jun 17, 2001 at 04:46:20PM -0400, Bob Hilliard wrote: Thanks, Julian. This works like a charm, but Lintian has the following objections: E: dict-web1913: no-copyright-file W: dict-web1913: prerm-does-not-remove-usr-doc-link W: dict-web1913: postinst-does-not-set-usr-doc-link Will the Lintian error prevent it from being installed? Shouldn't do; there is no content there to need a copyright. Julian -- =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- Julian Gilbey, Dept of Maths, Queen Mary, Univ. of London Debian GNU/Linux Developer, see http://people.debian.org/~jdg Donate free food to the world's hungry: see http://www.thehungersite.com/
Re: one package, many libraries. dependency break!
On Mon, Jun 18, 2001 at 04:40:46PM +0200, Domenico Andreoli wrote: In the past I have solved this problem by invoking dpkg-shlibdeps with a LD_LIBRARY_PATH setting that makes it find the libaries in the build tree. This way you get a warning that it couldn't find what package provides them, but you don't get a bogus dependency. I haven't found a cleaner solution. is LD_LIBRARY_PATH solution alternative to shlibs files? and if a libspell is already installing on the building system? wouldn't i need to put a !libpspell2 in Build-Depends? (You mean using Build-Conflicts.) Using LD_LIBRARY_PATH correctly means that the shlibs file from the new package being built will be used, rather than from the one on the system (if any). So you don't need a Build-Conflicts; you just need to get the dh_shlibdeps call correct. (I've sent you a private email on how to do this; basically use the -l option.) Julian -- =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- Julian Gilbey, Dept of Maths, Queen Mary, Univ. of London Debian GNU/Linux Developer, see http://people.debian.org/~jdg Donate free food to the world's hungry: see http://www.thehungersite.com/
Re: Empty package
On Thu, Jun 14, 2001 at 05:21:56PM -0400, Bob Hilliard wrote: I have ITP'd dict-gcide (Bug#100892). This package will provide, conflict and replace dict-web1913. I intend to provide an empty (dummy) dict-web1913 package that will depend on dict-gcide. The only thing functionally required in the dummy package is a control file, and possibly a README.Debian file advising the user that he may purge this package at any time. Don't even bother with the README.Debian file, just have an empty package and this message in the Description field in the control file. In order to be allowed in the archive, will it be necessary to provide a copyright file? No. If everything in the package is part of the ../debian directory what should I do about a .orig.tar.gz file? Would it be acceptable to make this version of dict-web1913 a debian-native package, and include the ../debian directory in the .tar.gz file? No, stuff the empty package into the dict-gcide source package's control file and generate the package from there. That will mean that it's easy to keep it up to date if things change, and it's logically nicer too. Julian -- =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- Julian Gilbey, Dept of Maths, Queen Mary, Univ. of London Debian GNU/Linux Developer, see http://people.debian.org/~jdg Donate free food to the world's hungry: see http://www.thehungersite.com/ -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Replaces: and pre/post rm
On Thu, Jun 14, 2001 at 04:29:30PM -0400, Richard A Nelson wrote: On Thu, 14 Jun 2001, Julian Gilbey wrote: So was the conflictor's postrm actually called with the purge option? It doesn't appear to have done. Not that I can find Good. Since both packages must have the same postrm, and they remove *alot* of files, things can get really nasty if I'm not careful: 1) Install sendmail 2) customize play with things 3) Decide to install sendmail-tls 4) hrm, sendmail is no longer needed, but still has some conffiles, lets dpkg --purge it At this point, many files in /etc/mail, /var/spool/mqueue, etc have now been erased... This what I don't want to happen to a user... Ah, very good point. Why not simply add in the purge section a block around it which says: if ! [ -f some file in the .deb which characterises the sendmail(-tls) packages ] then delete a lot of files fi You might also be interested in the call: conflictor's-prerm remove in-favour package new-version This might possibly help you. See the policy manual, section 6.4. I'm perusing that now, it might help -- even if I simply rm'd the older packages postrm file ! No, no, no. Don't ever fiddle with dpkg's internal databases like this. (There is one very, very rare case I can think of in which this is necessary, but this isn't it. And that is where the old prerm has a trojan or equivalent nasty in it.) Julian -- =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- Julian Gilbey, Dept of Maths, Queen Mary, Univ. of London Debian GNU/Linux Developer, see http://people.debian.org/~jdg Donate free food to the world's hungry: see http://www.thehungersite.com/ -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Empty package
On Thu, Jun 14, 2001 at 05:21:56PM -0400, Bob Hilliard wrote: I have ITP'd dict-gcide (Bug#100892). This package will provide, conflict and replace dict-web1913. I intend to provide an empty (dummy) dict-web1913 package that will depend on dict-gcide. The only thing functionally required in the dummy package is a control file, and possibly a README.Debian file advising the user that he may purge this package at any time. Don't even bother with the README.Debian file, just have an empty package and this message in the Description field in the control file. In order to be allowed in the archive, will it be necessary to provide a copyright file? No. If everything in the package is part of the ../debian directory what should I do about a .orig.tar.gz file? Would it be acceptable to make this version of dict-web1913 a debian-native package, and include the ../debian directory in the .tar.gz file? No, stuff the empty package into the dict-gcide source package's control file and generate the package from there. That will mean that it's easy to keep it up to date if things change, and it's logically nicer too. Julian -- =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- Julian Gilbey, Dept of Maths, Queen Mary, Univ. of London Debian GNU/Linux Developer, see http://people.debian.org/~jdg Donate free food to the world's hungry: see http://www.thehungersite.com/
Re: Replaces: and pre/post rm
On Thu, Jun 14, 2001 at 04:29:30PM -0400, Richard A Nelson wrote: On Thu, 14 Jun 2001, Julian Gilbey wrote: So was the conflictor's postrm actually called with the purge option? It doesn't appear to have done. Not that I can find Good. Since both packages must have the same postrm, and they remove *alot* of files, things can get really nasty if I'm not careful: 1) Install sendmail 2) customize play with things 3) Decide to install sendmail-tls 4) hrm, sendmail is no longer needed, but still has some conffiles, lets dpkg --purge it At this point, many files in /etc/mail, /var/spool/mqueue, etc have now been erased... This what I don't want to happen to a user... Ah, very good point. Why not simply add in the purge section a block around it which says: if ! [ -f some file in the .deb which characterises the sendmail(-tls) packages ] then delete a lot of files fi You might also be interested in the call: conflictor's-prerm remove in-favour package new-version This might possibly help you. See the policy manual, section 6.4. I'm perusing that now, it might help -- even if I simply rm'd the older packages postrm file ! No, no, no. Don't ever fiddle with dpkg's internal databases like this. (There is one very, very rare case I can think of in which this is necessary, but this isn't it. And that is where the old prerm has a trojan or equivalent nasty in it.) Julian -- =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- Julian Gilbey, Dept of Maths, Queen Mary, Univ. of London Debian GNU/Linux Developer, see http://people.debian.org/~jdg Donate free food to the world's hungry: see http://www.thehungersite.com/
Re: Replaces: and pre/post rm
On Thu, Jun 14, 2001 at 01:42:08PM -0400, Richard A Nelson wrote:
The problem comes when the user replaces the non-ssl with the ssl
version (or visa versa): the postrm script on the replaced version
is apparently called with the PURGE option.
[...]
No, must've been hallucinating again...
So was the conflictor's postrm actually called with the purge
option? It doesn't appear to have done.
I don't see (the echo statements indicating) the 'disappear', but
all that is left when switching from SSL to non-SSL are:
*) /var/lib/dpkg/info/${package}.postrm
*) /var/lib/dpkg/info/${package}.list
*) a couple of conffiles that are named differently:
-) /etc/logrotate.d/{sendmail,sendmail-tls}
-) /etc/cron.daily/{sendmail,sendmail-tls}
The few conffiles are because the corresponding dh_ commands don't allow
me the choice of renaming the script (ala dh_installinit).
That leaves me with the ${package}.postrm, which I guess will have to
contain all the intelligence to:
*) If the alternative package is installed, do nothing
*) If the alternative package is *NOT* installed, kill straggling
files.
Why? When one of the packages is finally purged, surely that will do
the job of cleaning out any mess from either of them (as you appear to
indicate that they use the same manually created files).
You might also be interested in the call:
conflictor's-prerm remove in-favour package new-version
This might possibly help you. See the policy manual, section 6.4.
I guess, given that, I could even have ${package}.postinst issue
a dpkg --purge for the alternative package...
Yuck -- no way!
Julian
--
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
Julian Gilbey, Dept of Maths, Queen Mary, Univ. of London
Debian GNU/Linux Developer, see http://people.debian.org/~jdg
Donate free food to the world's hungry: see http://www.thehungersite.com/
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Replaces: and pre/post rm
On Thu, Jun 14, 2001 at 01:42:08PM -0400, Richard A Nelson wrote:
The problem comes when the user replaces the non-ssl with the ssl
version (or visa versa): the postrm script on the replaced version
is apparently called with the PURGE option.
[...]
No, must've been hallucinating again...
So was the conflictor's postrm actually called with the purge
option? It doesn't appear to have done.
I don't see (the echo statements indicating) the 'disappear', but
all that is left when switching from SSL to non-SSL are:
*) /var/lib/dpkg/info/${package}.postrm
*) /var/lib/dpkg/info/${package}.list
*) a couple of conffiles that are named differently:
-) /etc/logrotate.d/{sendmail,sendmail-tls}
-) /etc/cron.daily/{sendmail,sendmail-tls}
The few conffiles are because the corresponding dh_ commands don't allow
me the choice of renaming the script (ala dh_installinit).
That leaves me with the ${package}.postrm, which I guess will have to
contain all the intelligence to:
*) If the alternative package is installed, do nothing
*) If the alternative package is *NOT* installed, kill straggling
files.
Why? When one of the packages is finally purged, surely that will do
the job of cleaning out any mess from either of them (as you appear to
indicate that they use the same manually created files).
You might also be interested in the call:
conflictor's-prerm remove in-favour package new-version
This might possibly help you. See the policy manual, section 6.4.
I guess, given that, I could even have ${package}.postinst issue
a dpkg --purge for the alternative package...
Yuck -- no way!
Julian
--
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
Julian Gilbey, Dept of Maths, Queen Mary, Univ. of London
Debian GNU/Linux Developer, see http://people.debian.org/~jdg
Donate free food to the world's hungry: see http://www.thehungersite.com/
