Re: (Not so fresh) News about locales

[Frank Küster]

Denis Barbier <[EMAIL PROTECTED]> wrote:

> On Wed, Jul 19, 2006 at 01:12:32PM +0200, Frank Küster wrote:
>> Denis Barbier <[EMAIL PROTECTED]> wrote:
>> 
>> > Hi,
>> >
>> > locales.postinst has been modified to write locale variables
>> > into /etc/default/locale instead of /etc/environment; the
>> > latter file is a PAM configuration file, so modifying it
>> > was a policy violation.  Most packages have been modified
>> > accordingly, but if you have trouble with some programs,
>> > please file bugs against them.  
>> 
>> localeconfig.postinst does write /etc/environment, is this a bug?  
>> 
>> If yes, which package is supposed to handle the file, then?  It isn't
>> registered with dpkg, no other maintainer script except localeconfig's
>> handles it, and it isn't in /var/lib/ucf/hashfile, either.
>
> You Cc: Progeny folks, so it looks like you believe that localeconf
> is the culprit; maybe the contents of this file can shed some light.
> This package is unmaintained for years, and is mostly useless nowadays.
> I just filed #378941.

Thank you, indeed I thought that localeconfig was wrong, but I wasn't
sure.  I've written an e-mail to [EMAIL PROTECTED] about the progeny
packages.  

Anyway, my other question is not answered:  Which packages does handle
/etc/environment?  There must be one that creates it upon installations.

Regards, Frank

-- 
Frank Küster
Single Molecule Spectroscopy, Protein Folding @ Inst. f. Biochemie, Univ. Zürich
Debian Developer (teTeX/TeXLive)

Re: lilypond and python

[Raphael Hertzog]

Hi Matthias,

On Thu, 20 Jul 2006, Matthias Klose wrote:
> well, there's curently only one person spreading lies and fud about
> python packaging, so please don't talk about "lies" as well.

Please stop ranting against Josselin, in particular if you have nothing
precise/factual to criticize. You're not making it easy to colloborate
with you.

I expect more from you.

> I'm still testing uprades and fixing upgrade issues. experimental has a
> python-defaults pointing to 2.4, so you can prepare your package and
> upload it to experimental. "pending" doesn't imply "will be fixed in x
> days".

You sent an announce to debian-devel-announce announcing the upload of the
new python2.4 by default more than a month ago.

You're not living up to your own promise.

I tried to help you in numerous ways:
- I've done the new dh_python and handled the initial bugs
  in a timely fashion
- I coordinated the discussion between you and Josselin while you refused
  to discuss together
- I tried to help you setup a real Python team and you promised me to
  start using "pkg-python" SVN repository and you never did that.

Now, please accept the blame of Thomas, you're late and it's your fault.
You are the python maintainer and nobody is forcing you to keep that
responsibility and I explicitely asked you to share it with other people
having more time than you have but you haven't done anything to help that
process...

Would you please, accept the facts, open your eyes and start making
efforts to setup a real python maintenance team? 

I'll gladly assist you in that process. 

Cheers,
-- 
Raphaël Hertzog

Premier livre français sur Debian GNU/Linux :
http://www.ouaza.com/livre/admin-debian/


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Re: Using the SSL snakeoil certificate

[Martin Schulze]

Jaldhar H. Vyas wrote:
> In bug #376146, Martin Pitt wrote:
> 
> > In an effort to clean up the SSL certificate mess on Ubuntu servers, we
> > recently converted all our supported Server packages to make use of
> > the ssl-cert package instead of creating a package-specific
> > self-signed SSL certificate. This allows admins to easily replace the
> > certificate with a 'real' one without touching dozens of configuration
> > files, and also provides a consistent setup out of the box.
> 
> Is this is a good idea for Debian?  I think it is but it doesn't make sense 
> to switch dovecot over unless all the other ssl-cert using packages also do 
> it. Is this possible in the etch timeframe?

I believe that this is a good idea, however, I would like to propose a
slightly different approach.

At the moment, it seems that all applications use their own
certificates and maybe also create them upon installation or rather
configuration.

It may be useful to have a certificate for each service, but it may
also be useful to have one certificate for all services.  This may be
discussible but needs to be decided by the local admin anyway.  Hence,
we should try to make both ways easily implementable, especially if
the system is to be reviewed or redesigned.

Hence, I propose to stay with virtual per-service certificates, but to
link them to the common snakeoil certificate from ssl-certificates
during configuration and only if there is no other setting.

For example:

  Dovecot uses .

  This is a symbolic link to  if
  the above file or link does not exist during configuration of
  dovecot.

That way, the admin can easily replace the symlink with a real
certificate if they want per-service certificates.

If, however, they want to have one real certificate for everything,
they can replace the snakeoil certificate like Martin Pitt proposed.


I would like to see some coordination between maintainers of packages
that use or create such certificates.  It'll take a while to implement
this anyway, so if only a few packages start and others follow later,
that'd be an improvement anyway.

Regards,

Joey

-- 
Open source is important from a technical angle. -- Linus Torvalds

Please always Cc to me when replying to me on the lists.


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Re: Package Selection for Debian Live

[Mathieu JANIN]

Hi folks,
i could seem idiot, but wouldn't it be nice to have a minimalistic bootstrap
CD with everything that is needed to compose your own liveCD (perharps an
enhanced version of DFSbuild, with "cleaning/compressing" feature like
localepurge and so ), and only a minimalistic set of what is needed for
basic forensic.
With some precomposed packet-lists included to build rapidly the second
stage iso for people in a hurry.
Something that could be burnt on a business card or a pendrive ?

Matt.


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Re: Could there be a arch build test befre upload?

[Goswin von Brederlow]

Joseph Smidt <[EMAIL PROTECTED]> writes:

> I've been maintaining texmaker which has originally built fine on all
> architectures but failed to build on arm after the last upload.  Luckily
> for me one of the arm devs helped me out and uploaded a fixed version.
> He was able to test on arm and when it worked informed me.
>
>   The experience got me thinking, wouldn't it be nice if all
> maintainers could test there packages on all architectures before they
> uploaded?  Maybe a dbuildtest command before dupload which ran your
> package on the various architectures to test whether it builds.  You
> could view all the build logs and if everything checks out you know the
> package is ready.  
>
>   The arm dev suggested using experimental for such a purpose,
> which I admit is a good idea, but isn't that more than is needed if all
> you want to test is whether it builds on all architectures.  dbuildtest
> could be another sanity test on top of linda and lintian.  Then you know
> in advance if the build would work.
>
> I just wondered what you all thought about this idea.
>
>   Joseph Smidt

For this you need a place to put the source (experimental upload
queue), a controler (wanna-build) that tells what to build, a network
of systems to compile (buildd for each arch). Only thing you don't
need is the signing and uploading of debs. But you can read the buildd
log without waiting for an upload.


One problem with using experimental to test compile is that afaik the
buildds don't have access to incoming.d.o so you have to wait for the
daily dinstall pules before anything gets build. That pretty much
kills the idea.

Another problem is that when you find say an arm FTBFS then you want
to do many test build on arm till you found the bug. There is no need
to recompile every test upload on all the other archs.

So maybe experimental isn't quite the right place for this. But maybe
you could talk aba into reusing the experimental buildds with a
different interface and source repository so DDs can schedule builds
on specific archs directly. You would need a DAK to upload the source
to and then a way to tell the wanna-build to add the package only on
specific archs. Nothing a few lines of perl or php can't do.

MfG
Goswin


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Re: Could there be a arch build test befre upload?

[Goswin von Brederlow]

Gustavo Noronha Silva <[EMAIL PROTECTED]> writes:

> Em Wed, 19 Jul 2006 09:14:48 -0600
> Joseph Smidt <[EMAIL PROTECTED]> escreveu:
>>  The arm dev suggested using experimental for such a purpose,
>> which I admit is a good idea, but isn't that more than is needed if
>> all you want to test is whether it builds on all architectures.
>> dbuildtest could be another sanity test on top of linda and lintian.
>> Then you know in advance if the build would work.
>> 
>
> Too much load for the buildd network, I believe. Notice that you do
> have chroots in various port machines, and you can request that
> build-deps for your package be installed on them.
>
> See you,

And how often do they not have the right build-depends installed?

Also note that the load is the same if you start a build manual or if
buildd builds a package. So in overall cpu time you don't loose as
long as builds as started for specific archs.

MfG
Goswin


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Re: Package Selection for Debian Live

[John Goerzen]

On Thu, Jul 20, 2006 at 12:01:30PM +0200, Mathieu JANIN wrote:
>   Hi folks,
> i could seem idiot, but wouldn't it be nice to have a minimalistic bootstrap
> CD with everything that is needed to compose your own liveCD (perharps an
> enhanced version of DFSbuild, with "cleaning/compressing" feature like
> localepurge and so ), and only a minimalistic set of what is needed for
> basic forensic.

You could probably do that with dfsbuild and a little bit of scrubbing.
You probably just would want to do things like rm -r /usr/share/doc
/usr/share/man on the generated image.

-- John


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Bug#378996: ITP: vulcan -- chess variant inspired by a certain well-known science fiction TV series

[Miriam Ruiz]

Package: wnpp
Severity: wishlist
Owner: Miriam Ruiz <[EMAIL PROTECTED]>


* Package name: vulcan
  Version : 0.2
  Upstream Author : Mauro Persano <[EMAIL PROTECTED]>
* URL : http://www.fzort.org/mpr/projects/vulcan/
* License : GPL
  Description : chess variant inspired by a certain well-known science 
fiction TV series

  vulcan lets you play against the computer a chess variant inspired by a
  certain well-known science fiction TV series.

  The rules implemented by this program were created by Jens Meder, and
  are described here: http://private.addcom.de/meder/3dschach/chess3d.htm

  In a nutshell, here's the idea:

* viewing the board from above, you'll notice that some squares overlap.
* from that view, piece moves are the same as in regular chess, except
  that you may choose which of the overlapping squares the piece will land 
at.
* the movement of a sliding piece (bishop, rook or queen) is blocked by
  a piece in any of the overlapping squares; the sliding piece may
  land above or below the blocking piece.
* the small 2x2 boards (the "attack boards") can be moved to an adjacent
  corner at the same level if they're holding just one piece, or to a corner
  of a different level if they're empty.

  The 3D models used in the program were originally created by Renzo del
  Fabbro, and downloaded from the site www.3dcafe.com. As far as the game
  author knows, they're in the public domain, but that must be clarified.

  The game is going to be packaged inside the Debian Games Team.

-- System Information:
Debian Release: testing/unstable
  APT prefers testing
  APT policy: (500, 'testing')
Architecture: i386 (i686)
Shell:  /bin/sh linked to /bin/bash
Kernel: Linux 2.6.14
Locale: [EMAIL PROTECTED], [EMAIL PROTECTED] (charmap=ISO-8859-15)


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Re: Bug#378996: ITP: vulcan -- chess variant inspired by a certain well-known science fiction TV series

[gregor herrmann]

On Thu, 20 Jul 2006 09:54:16 +0200, Miriam Ruiz wrote:

Hi,

> * Package name: vulcan
>   Version : 0.2
>   Upstream Author : Mauro Persano <[EMAIL PROTECTED]>
> * URL : http://www.fzort.org/mpr/projects/vulcan/
> * License : GPL
>   Description : chess variant inspired by a certain well-known science 
> fiction TV series

sounds cool, especially to all who know the well-known series :-)

I just wanted to point out that there is a package in Debian that
sounds quite similar:

  Package: 3dchess
  Description: 3D chess for X11

I don't know 3dchess, maybe there are large differences to vulcan.


Cheers,
gregor
 
-- 
 .''`.   http://info.comodo.priv.at/ | gpg key ID: 0x00F3CFE4
 : :' :  debian: the universal operating system - http://www.debian.org/
 `. `'   member of https://www.vibe.at/ | how to reply: http://got.to/quote/
   `-NP: Andrew Lloyd Webber & Tim Rice: Heaven on Their Minds


signature.asc
Description: Digital signature

Re: Using the SSL snakeoil certificate

[Klaus Ethgen]

-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1

Am Do den 20. Jul 2006 um 11:24 schrieb Martin Schulze:
> > [one cert for all services]
> I believe that this is a good idea, however, I would like to propose a
> slightly different approach.
> 
> At the moment, it seems that all applications use their own
> certificates and maybe also create them upon installation or rather
> configuration.

I like this idea mentioned above. Isn't it easy to ask the admin in
debconf for every service if he want separate certs or all linked
together?

> Hence, I propose to stay with virtual per-service certificates, but to
> link them to the common snakeoil certificate from ssl-certificates
> during configuration and only if there is no other setting.

That would be another way.

Regards
   Klaus Ethgen
- -- 
Klaus Ethgenhttp://www.ethgen.de/
pub  2048R/D1A4EDE5 2000-02-26 Klaus Ethgen <[EMAIL PROTECTED]>
Fingerprint: D7 67 71 C4 99 A6 D4 FE  EA 40 30 57 3C 88 26 2B
-BEGIN PGP SIGNATURE-
Version: GnuPG v1.4.3 (GNU/Linux)

iQEVAwUBRL+ftp+OKpjRpO3lAQJtJgf+M4e/D970JQZDTbUg00z4PTBVD0ts7Hex
XffYVpJt+dhQzXh2ljO/7vuqH2wxWvSuxevix4QSpAeJl9RpFceFsgMerpT7rqIv
lPzo+JljGeWQb02kNnRZE8aRhNjsesEBr6RIDwRnB8+zKgxzTKOqdH0pvi3iTkUB
39fBJ1v3NcYxc7DDwRWcG1Aw9I6yJgsMGexiQs0w/OZ9yY3aE8HQuyiaPhy7UnJr
FvJyO2Ddv4AOMXxVIf2PRpcGsbKf0y0mX30mVVL3FzW+qPPC8PBeM/iAnex+oZGL
wrWsdvzk3L93KMaS2EdgwW+k/0AnwPq6XNj+0XoWZJIuljHLr9xPJw==
=a/xx
-END PGP SIGNATURE-


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Re: Using the SSL snakeoil certificate

[Martin Schulze]

(please copy debian-devel, feel free to bounce my mail there after
you've done so, for others to be able to comment as well).

Klaus Ethgen wrote:
> Am Do den 20. Jul 2006 um 11:24 schrieb Martin Schulze:
> > > [one cert for all services]
> > I believe that this is a good idea, however, I would like to propose a
> > slightly different approach.
> > 
> > At the moment, it seems that all applications use their own
> > certificates and maybe also create them upon installation or rather
> > configuration.
> 
> I like this idea mentioned above. Isn't it easy to ask the admin in
> debconf for every service if he want separate certs or all linked
> together?

No!  Please don't introduce more questions that are not required for
packages to work properly.  Debconf is not a general configuration
utility that should subsume all possible configurations.  Instead it
should only ask what is required for the package to work properly in
a default environment.

If an admin wants to use different certs, they should be able to do
so easily, but it's not acceptable to ask everybodo whether to use
shared or singular certificates, or to create a cert, or to install
an already created and certified one, or, or, or...

Regards,

Joey

-- 
Open source is important from a technical angle. -- Linus Torvalds

Please always Cc to me when replying to me on the lists.


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Bug#379010: ITP: scim-thai -- ITP: scim-thai -- Thai Input Method Engine for SCIM

[Theppitak Karoonboonyanan]

Package: wnpp
Severity: wishlist
Owner: Theppitak Karoonboonyanan <[EMAIL PROTECTED]>

-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1


* Package name: scim-thai
  Version : 0.1.0
  Upstream Author : Theppitak Karoonboonyanan <[EMAIL PROTECTED]>
* URL : http://libthai.sourceforge.net/
* License : GPL
  Programming Lang: C++
  Description : Thai Input Method Engine for SCIM

 SCIM (Smart Common Input Method) is an input method (IM) platform.
 .
 This package provides Thai Input Method Engine for SCIM.  Currently, it
 supports Ketmanee, TIS-820.2538 and Pattachote keyboards; three levels
 of strictness for input sequence check.
 .
 Homepage: http://libthai.sourceforge.net/

- -- System Information:
Debian Release: testing/unstable
  APT prefers unstable
  APT policy: (500, 'unstable')
Architecture: i386 (i686)
Shell:  /bin/sh linked to /bin/bash
Kernel: Linux 2.6.16
Locale: LANG=th_TH.UTF-8, LC_CTYPE=th_TH.UTF-8 (charmap=UTF-8)

-BEGIN PGP SIGNATURE-
Version: GnuPG v1.4.3 (GNU/Linux)

iD8DBQFEv520qgzR7tCLR/4RAslVAJ9ZxucxrI+4o7UbdnId7FHb4cJj9gCgjXls
pfUxqT7JA/CwHGXHbxsL/z8=
=Z3Qq
-END PGP SIGNATURE-


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Re: Jubiläum Angebot.

[Barbara Bloch]

 

Streichen Sie mich unverzüglich aus Ihrer Liste!

 

Re: (Not so fresh) News about locales

[Christian Perrier]

> Anyway, my other question is not answered:  Which packages does handle
> /etc/environment?  There must be one that creates it upon installations.


The installer..:-)...or the local admin.

More precisely localechooser's post-base-installer creates
*both* /etc/environment AND /etc/default/locale with LANG set to the
default locale and LANGUAGE set to a a list of alternatives if such a
thing exists in the localechooser reference file (actually for very
few languages, namely both Chinese flavours and Norwegian languages).

So /etc/environment doesn't technically belong to any package.





signature.asc
Description: Digital signature

Re: Jubiläum Angebot.

[Izak Burger]

On 7/20/06, Barbara Bloch <[EMAIL PROTECTED]> wrote:

Streichen Sie mich unverzüglich aus Ihrer Liste!


You can unsubscribe here:

http://lists.debian.org/debian-devel/

Re: Bits from the Package Tracking System

[Matt Zimmerman]

On Tue, Jul 18, 2006 at 10:19:02PM +0200, Denis Barbier wrote:
> So by default it is assumed that I should make Ubuntu's work and dig
> into these patches to see if some pieces should be applied into Debian?
> No thanks, I am getting tired of all those Debian developers who are
> more interested in improving Ubuntu than Debian, and just added the
> following rules to my .procmailrc:
>   :0
>   * ^From.*(ubuntu|canonical).com
>   /dev/null
> 
> Denis, pissed off
> PS: No, I am not joking

Raphael ensured us that this would be an opt-in notification, and we agreed
to provide the data feed for it under this assumption.  If there was a
configuration error which caused this not to be the case, please try not to
overreact, be patient and allow it to be corrected.

-- 
 - mdz


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Re: Bug#378863: ITP: libcsoap1-dev -- library in C for SOAP network protocol (SOAP is a form of RPC) development

[Rudi]

On Wed, Jul 19, 2006 at 04:26:18PM +0200, Magnus Holmgren wrote:
> On Wednesday 19 July 2006 14:06, Rudi Cilibrasi took the opportunity to write:
> > * Package name: libcsoap1-dev
> >   Version : 1.047
> You need only post one ITP bug per source package. And AFAICS there is just 
> one which builds both the library and the development files.
> BTW, isn't the version number of the latest stable release 1.0.4?

Dear Magnus,

Thanks so much for these tips.  I have closed both libcsoap1-dev and
libcsoap1 and opened up a new ITP based on your suggestion called
libcsoap.  Also, I have changed the version number to 1.0.4.  Thanks
again for your useful comments.  Cheers

Rudi Cilibrasi


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Bug#379048: ITP: libcsoap -- library in C for SOAP network protocol (SOAP is a form of RPC with objects)

[Rudi Cilibrasi]

Package: wnpp
Severity: wishlist
Owner: Rudi Cilibrasi <[EMAIL PROTECTED]>


* Package name: libcsoap
  Version : 1.0.4
  Upstream Author : Ferhat Ayaz <[EMAIL PROTECTED]>
* URL : http://csoap.sourceforge.net/
* License : GPL
  Programming Lang: C
  Description : library in C for SOAP network protocol (SOAP is a form of 
RPC with objects)

This library written by Ferhat Ayaz <[EMAIL PROTECTED]> and Michael
Rans <[EMAIL PROTECTED]> provides an easy way to call methods remotely
using the Simple Object Access Protocol (SOAP) using C.  It is popular
and used for providing many types of web services.  It is similar to
XMLRPC but includes more features.

-- System Information:
Debian Release: testing/unstable
  APT prefers testing
  APT policy: (500, 'testing'), (1, 'experimental')
Architecture: i386 (i686)
Shell:  /bin/sh linked to /bin/bash
Kernel: Linux 2.6.15-1-k7
Locale: LANG=en_US.UTF-8, LC_CTYPE=en_US.UTF-8 (charmap=UTF-8)


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Re: (Not so fresh) News about locales

[Denis Barbier]

[Progeny dropped from Cc]

On Thu, Jul 20, 2006 at 09:15:13AM +0200, Frank Küster wrote:
> Denis Barbier <[EMAIL PROTECTED]> wrote:
> 
> > On Wed, Jul 19, 2006 at 01:12:32PM +0200, Frank Küster wrote:
> >> Denis Barbier <[EMAIL PROTECTED]> wrote:
> >> 
> >> > Hi,
> >> >
> >> > locales.postinst has been modified to write locale variables
> >> > into /etc/default/locale instead of /etc/environment; the
> >> > latter file is a PAM configuration file, so modifying it
> >> > was a policy violation.  Most packages have been modified
> >> > accordingly, but if you have trouble with some programs,
> >> > please file bugs against them.  
> >> 
> >> localeconfig.postinst does write /etc/environment, is this a bug?  
> >> 
> >> If yes, which package is supposed to handle the file, then?  It isn't
> >> registered with dpkg, no other maintainer script except localeconfig's
> >> handles it, and it isn't in /var/lib/ucf/hashfile, either.
> >
> > You Cc: Progeny folks, so it looks like you believe that localeconf
> > is the culprit; maybe the contents of this file can shed some light.
> > This package is unmaintained for years, and is mostly useless nowadays.
> > I just filed #378941.
> 
> Thank you, indeed I thought that localeconfig was wrong, but I wasn't
> sure.  I've written an e-mail to [EMAIL PROTECTED] about the progeny
> packages.  
> 
> Anyway, my other question is not answered:  Which packages does handle
> /etc/environment?  There must be one that creates it upon installations.

This is because in your original mail, I thought that 'the file' was
referring to localeconf.postinst and not /etc/environment ;)

The latter is a PAM configuration file, so I would say that libpam-runtime
'owns' it.  But this file does not have to exist, it is optional, this is
why libpam-runtime does not create it.
It used to be created by d-i when user performs an installation in a
language different than English.  When the locales package have been
modified, Christian and I decided to let d-i create both files
(/etc/environment and /etc/default/locale) until most packages have
been patched.  Maybe it is time to finish this transition, and modify
d-i to no more create /etc/environment, this can surely be done after
d-i beta3.

Denis


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Re: greylisting on debian.org?

[Claus Färber]

Thomas Bushnell BSG <[EMAIL PROTECTED]> schrieb/wrote:
> So the meaning of 4xx is "temporary local problem".

RFC 2822 says "4yz Transient Negative Completion reply" (p. 42). The
standard also encourages the re-use of existing error codes for
"slightly different situations" (p. 43).

Claus



-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Re: Measuring "should I greylist?" false positive rate [was: greylisting on debian.org?]

[Claus Färber]

Pierre Habouzit <[EMAIL PROTECTED]> schrieb/wrote:
> Yeah, master took 31 minutes to deliver the mail to the next SMTP
> server, and I'm an admin on it (that's the mail server I was speaking
> of earlier in the thread) and it was its first submission.

In other words, not using greylisting delays legitimate mail?

Claus



-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Re: Using the SSL snakeoil certificate

[Milan P. Stanic]

On Thu, Jul 20, 2006 at 11:24:34AM +0200, Martin Schulze wrote:
> For example:
> 
>   Dovecot uses .
> 
>   This is a symbolic link to  if
>   the above file or link does not exist during configuration of
>   dovecot.
> 
> That way, the admin can easily replace the symlink with a real
> certificate if they want per-service certificates.
> 
> If, however, they want to have one real certificate for everything,
> they can replace the snakeoil certificate like Martin Pitt proposed.
 
Sorry if I misunderstand something, but is it okay to call it snakeoil
if it is real certificate? I like to say that the symbolic links for
per-service certificate shouldn't point to something called snake-oil.

Just my opinion.


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Re: Bug#378876: Bug#378872: ITP: libsvm0 -- LibSVM is a machine-learning library providing Support Vector Machines in C++

[Rudi]

On Thu, Jul 20, 2006 at 12:55:23AM +0100, martin f krafft wrote:
> also sprach Rudi Cilibrasi <[EMAIL PROTECTED]> [2006.07.19.1353 +0100]:
> > * Package name: libsvm0
> > * Package name: libsvm0-dev
> > * Package name: libsvm0-tools
> > * Package name: libsvm-doc
> > * Package name: libsvm-ruby1.8
> 
> These all look like packages you want to generate from a single
> source package. Also, libsvm0-dev is a bad name, IMHO, because the
> ABI, which the 0 identifies, is not relevant for the -dev package,
> unless the 0 also specifies the API version.
> 
> I suggest you merge the five bugs into one as ITPs are generally per
> source package.

Dear Martin,

Thanks for pointing this problem out to me.  I am new at packaging and
made several similar errors in other packages as well at first.  I have
followed your advice and closed down all the old packages and opened
a new one called libsvm.  I have also renamed libsvm0-dev to libsvm-dev.
Thanks for the info!  Cheers,

Rudi Cilibrasi


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Re: Bits from the Package Tracking System

[Raphael Hertzog]

On Wed, 19 Jul 2006, Matt Zimmerman wrote:
> > Denis, pissed off
> > PS: No, I am not joking
> 
> Raphael ensured us that this would be an opt-in notification, and we agreed
> to provide the data feed for it under this assumption.  If there was a
> configuration error which caused this not to be the case, please try not to
> overreact, be patient and allow it to be corrected.

I did what I said in the announce (ie initialize the "derivatives" keyword
with the people who had 'already opt-in' for the cvs keyword).

I had this discussion with Denis on IRC before and it's precicely for
people like him that I added the "keywordall" command.

There's nothing else to add. I did what's best im my "PTS maintainer"
point of view. And since I did the work, I decided following my opinion.

Cheers,
-- 
Raphaël Hertzog

Premier livre français sur Debian GNU/Linux :
http://www.ouaza.com/livre/admin-debian/


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Bug#379051: ITP: libsvm -- LibSVM is a machine-learning library providing Support Vector Machines in C++

[Rudi Cilibrasi]

Package: wnpp
Severity: wishlist
Owner: Rudi Cilibrasi <[EMAIL PROTECTED]>


* Package name: libsvm
  Version : 2.8.2
  Upstream Author : Chih-Jen Lin   <[EMAIL PROTECTED]>
* URL : http://www.csie.ntu.edu.tw/~cjlin/libsvm/
* License : GPL
  Programming Lang: C++
  Description : LibSVM is a machine-learning library providing Support 
Vector Machines in C++

Support Vector Machines have become as popular and powerful as Neural
Networks and have outpaced them in certain application domains.  LibSVM
provides a fast, simple, and easy to use interface to training, testing,
and using for prediction your own custom SVM's.

-- System Information:
Debian Release: testing/unstable
  APT prefers testing
  APT policy: (500, 'testing'), (1, 'experimental')
Architecture: i386 (i686)
Shell:  /bin/sh linked to /bin/bash
Kernel: Linux 2.6.15-1-k7
Locale: LANG=en_US.UTF-8, LC_CTYPE=en_US.UTF-8 (charmap=UTF-8)


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Re: Using the SSL snakeoil certificate

[tony mancill]

On Thu, Jul 20, 2006 at 11:24:34AM +0200, Martin Schulze wrote:

> Hence, I propose to stay with virtual per-service certificates, but to
> link them to the common snakeoil certificate from ssl-certificates
> during configuration and only if there is no other setting.
> 
> For example:
> 
>   Dovecot uses .
> 
>   This is a symbolic link to  if
>   the above file or link does not exist during configuration of
>   dovecot.
> 
> That way, the admin can easily replace the symlink with a real
> certificate if they want per-service certificates.
> 
> If, however, they want to have one real certificate for everything,
> they can replace the snakeoil certificate like Martin Pitt proposed.

This would be a great improvement.  I'd suggest one more level of
symlinks.  Have the individual services symlink to
/etc/ssl/certs/ssl-cert-site.pem, which is then symlinked to
ssl-cert-snakeoil.pem.  When/if the local admin installs an actual
site-wide certificate, updating the one ssl-cert-site.pem symlink will
update all of the individual services using the the site cert, and the
snakeoil cert is still available if you ever need to fail back to it.

tony


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Bug#379054: ITP: lisaac -- Lisaac is the first object-oriented language base on prototype

[picca frederic]

Package: wnpp
Severity: wishlist
Owner: picca frederic <[EMAIL PROTECTED]>


* Package name: lisaac
  Version : 0.84
  Upstream Author : Benoit Sonntag <[EMAIL PROTECTED]>
* URL : http://isaacos.loria.fr/li.html
* License : Cecill (compatible GPL)
  Programming Lang: (C, Lisaac)
  Description : Lisaac is the first object-oriented language base on 
prototype

The ideas in Lisaac are mostly inspired by Smalltalk (all values are
objects), Self (prototype-based) and Eiffel (design by contract)
features:
 * pure object language
 * very fast (like C code)
 * dynamic and multiple inheritance
 * dynamic definition slot
 * static typing (invariant)
 * genericity type
 * auto-cast type system
 * programming by contract
 * interrupt manager
 * include C code facilities

-- System Information:
Debian Release: testing/unstable
  APT prefers testing
  APT policy: (500, 'testing'), (500, 'stable')
Architecture: i386 (i686)
Shell:  /bin/sh linked to /bin/bash
Kernel: Linux 2.6.15-1-686
Locale: [EMAIL PROTECTED], [EMAIL PROTECTED] (charmap=UTF-8)


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Re: Bug#378876: Bug#378872: ITP: libsvm0 -- LibSVM is a machine-learning library providing Support Vector Machines in C++

[martin f krafft]

also sprach Rudi <[EMAIL PROTECTED]> [2006.07.20.2125 +0100]:
> Thanks for pointing this problem out to me.  I am new at packaging and
> made several similar errors in other packages as well at first.  I have
> followed your advice and closed down all the old packages and opened
> a new one called libsvm.  I have also renamed libsvm0-dev to libsvm-dev.
> Thanks for the info!  Cheers,

You should have merged the bugs into one and gone from there. Also,
about the -dev package name, please read the Debian Library
Packaging Guide (I can't get at the URL right now, search for it)
and decide for yourself. There are reasons why one would want
a version number in the -dev package name...

-- 
Please do not send copies of list mail to me; I read the list!
 
 .''`. martin f. krafft <[EMAIL PROTECTED]>
: :'  :proud Debian developer and author: http://debiansystem.info
`. `'`
  `-  Debian - when you have better things to do than fixing a system
 
"never try to explain computers to a layman.
 it's easier to explain sex to a virgin."
-- robert heinlein
 
(note, however, that virgins tend to know a lot about computers.)


signature.asc
Description: Digital signature (GPG/PGP)

someone using apmd on ppc please test patch for #222635

[Aníbal Monsalve Salazar]

Hello,

Someone using apmd on ppc please test patch for #222635 [0].

[0] http://bugs.debian.org/222635

Best Regards,

Aníbal Monsalve Salazar
-- 
http://v7w.com/anibal


signature.asc
Description: Digital signature

Re: Bug#378876: Bug#378872: ITP: libsvm0 -- LibSVM is a machine-learning library providing Support Vector Machines in C++

[Luca Capello]

Hello!

On Fri, 21 Jul 2006 00:29:46 +0200, martin f krafft wrote:
> Also, about the -dev package name, please read the Debian Library
> Packaging Guide (I can't get at the URL right now, search for it)

http://www.debian.org/doc/developers-reference/ch-best-pkging-practices.en.html#s-bpp-libraries

and then

http://www.netfort.gr.jp/~dancer/column/libpkg-guide/libpkg-guide.html

Thx, bye,
Gismo / Luca


pgprJSJ6OIvwj.pgp
Description: PGP signature

Configuration file shadowed?

[Manoj Srivastava]

Hi,

Today, when upgrading my unstable box, I got a debconf message
 from the package tex-common, to wit:
“The mechanism of TeX font caching has changed, requiring the
 installation of /usr/share/texmf/web2c/mktex.cnf.  On your
 system, /etc/texmf/web2c/mktex.cnf exists with possible local
 changes, and makes the new file invisible to TeX.  Font
 caching might not work until the setting of MT_FEATURES has
 been transferred to your configuration file.”

What does this imply? Why are configuration files (I am
 assuming that something called mktex.cnf is actually a configuration
 file) being installed in /usr/share/? What exactly is the usewr
 supposed to do, anyway? If nothing, why is there a debconf notice?

manoj
-- 
The girl who swears no one has ever made love to her has a right to
swear. Sophia Loren
Manoj Srivastava   <[EMAIL PROTECTED]>  
1024D/BF24424C print 4966 F272 D093 B493 410B  924B 21BA DABB BF24 424C

orphaning gitweb

[Andres Salomon]

Hi,

I'm going to orphan gitweb; I haven't used it in a long time, and I've
been doing a poor job of keeping it up-to-date.  It doesn't have any
bugs open on it; it just needs the occasional update (and the one patch
I've done for it should be fed upstream if they're willing to take it).

Who wants it?

(Please CC me; I'm not subscribed to -devel.)


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Work-needing packages report for Jul 21, 2006

[wnpp]

The following is a listing of packages for which help has been requested
through the WNPP (Work-Needing and Prospective Packages) system in the
last week.

Total number of orphaned packages: 332 (new: 23)
Total number of packages offered up for adoption: 76 (new: 1)
Total number of packages requested help for: 23 (new: 0)

Please refer to http://www.debian.org/devel/wnpp/ for more information.



The following packages have been orphaned:

   aptconf (#379037), orphaned today
 Description: Debconf infrastructure for setting up apt sources
 Installations reported by Popcon: 184

   atari800 (#379022), orphaned today
 Description: Atari emulator for X/curses/SDL
 Installations reported by Popcon: 49

   cbrowser (#378796), orphaned 2 days ago
 Description: a C/C++ source code indexing, querying and browsing
   tool
 Installations reported by Popcon: 270

   configlet (#379031), orphaned today
 Description: alternative debconf configuration interface: core API
 Reverse Depends: configlet-frontends python-configlet
 Installations reported by Popcon: 229

   cscope (#378802), orphaned 2 days ago
 Description: Interactively examine a C program source
 Reverse Depends: cbrowser kscope
 Installations reported by Popcon: 963

   cthugha (#378946), orphaned yesterday (non-free)
 Description: an oscilloscope on acid
 Installations reported by Popcon: 62

   discover (#379043), orphaned today
 Description: hardware identification library
 Reverse Depends: bootcd-mkinitrd discover gibraltar-bootsupport
   libdiscover-dev xdebconfigurator
 Installations reported by Popcon: 1729

   discover-data (#379044), orphaned today
 Description: Data lists for Discover hardware detection system
 Reverse Depends: discover-udeb gibraltar-bootsupport libdiscover2
   mkinitrd-cd
 Installations reported by Popcon: 1821

   etherconf (#379032), orphaned today
 Description: debconf interface to Ethernet device configuration
 Installations reported by Popcon: 215

   ggi-doc (#378919), orphaned yesterday
 Description: General Graphics Interface project documentation
 Installations reported by Popcon: 23

   gnome-tasksel (#379038), orphaned today
 Description: GNOME interface to Debian tasks
 Installations reported by Popcon: 162

   gtm (#379000), orphaned today
 Description: Multiple files transfer manager
 Installations reported by Popcon: 228

   kernel-patch-nfs-swap (#378928), orphaned yesterday
 Description: patch to linux to enable swapping over nfs
 Installations reported by Popcon: 8

   libggigcp (#378921), orphaned yesterday
 Description: GGI Color and Palette Manager extension
 Reverse Depends: libggigcp0-dev
 Installations reported by Popcon: 15

   libggiwmh (#378924), orphaned yesterday
 Description: GGI Window Manager Hints extension development package
 Reverse Depends: libggiwmh0-dev libggiwmh0-target-x
 Installations reported by Popcon: 23

   libgii (#378925), orphaned yesterday
 Description: General Input Interface runtime libraries
 Reverse Depends: freesci heroes-ggi libggi-samples libggi-target-aa
   libggi-target-emu libggi-target-fbdev libggi-target-glide
   libggi-target-lcd823 libggi-target-monotext libggi-target-svgalib
   (11 more omitted)
 Installations reported by Popcon: 4071

   localeconf (#379033), orphaned today
 Description: debconf interface to locale configuration
 Installations reported by Popcon: 760

   mozilla-locale-zh-cn (#378751), orphaned 2 days ago
 Description: Mozilla Simplified Chinese (China) language/region pack

   mozilla-locale-zh-tw (#378752), orphaned 2 days ago
 Description: Mozilla Traditional Chinese (Taiwan) language/region
   pack

   picax (#379039), orphaned today
 Description: Tool for creation of aptable media
 Installations reported by Popcon: 9

   python-parted (#379034), orphaned today
 Description: Python bindings for GNU Parted
 Installations reported by Popcon: 20

   svgalib4libggi (#378927), orphaned yesterday
 Description: SVGAlib wrapper library for LibGGI
 Installations reported by Popcon: 16

   timezoneconf (#379035), orphaned today
 Description: debconf interface to system timezone, date, and time
 Reverse Depends: gnustep-base-common
 Installations reported by Popcon: 1056

309 older packages have been omitted from this listing, see
http://www.debian.org/devel/wnpp/orphaned for a complete list.



The following packages have been given up for adoption:

   mypasswordsafe (#378540), offered 3 days ago
 Description: Easy-to-use password manager
 Installations reported by Popcon: 73

75 older packages have been omitted from this listing, see
http://www.debian.org/devel/wnpp/rfa