[Python-modules-team] Bug#894560: pygame: Don't drop python2 package

[Dominik George]

Control: tags -1 + pending

Hi,

the next upload in a couple of days will re-introduce the python 2 package.

-nik


signature.asc
Description: PGP signature
___
Python-modules-team mailing list
Python-modules-team@lists.alioth.debian.org
http://lists.alioth.debian.org/cgi-bin/mailman/listinfo/python-modules-team

[Python-modules-team] Accepted pygame 1.9.3+dfsg2-1 (source) into unstable

[Dominik George]

-BEGIN PGP SIGNED MESSAGE-
Hash: SHA256

Format: 1.8
Date: Sat, 31 Mar 2018 22:11:00 +0200
Source: pygame
Binary: python3-pygame python-pygame-doc
Architecture: source
Version: 1.9.3+dfsg2-1
Distribution: unstable
Urgency: medium
Maintainer: Debian Python Modules Team 
<python-modules-team@lists.alioth.debian.org>
Changed-By: Dominik George <naturesha...@debian.org>
Description:
 python-pygame-doc - SDL bindings for games development in Python 
(documentation)
 python3-pygame - SDL bindings for games development (Python 3)
Closes: 861249 861250 861252 861253 861499 871910
Changes:
 pygame (1.9.3+dfsg2-1) unstable; urgency=medium
 .
   [ Ondřej Nový ]
   * d/control: Set Vcs-* to salsa.debian.org
   * d/copyright: Use https protocol in Format field
   * d/watch: Use https protocol
   * d/changelog: Remove trailing whitespaces
 .
   [ Dominik George ]
   * Update dependency on timgm6mb-soundfont. (Closes: #871910)
   * Depend on FreeSans font. (Closes: #861499)
   * Update my mail address in Uploaders.
   * Remove Python 2 build.
   * Bumb Standards-Version and compat level.
 + Remove default variables from d/rules.
   * Fix arithmetic errors on some platforms.
 (Closes: #861250, #861252, #861253, #861249)
   * Fix unicode in Sphinx output.
   * Fix orig tarball, which contained files it shouldn't.
 + Bump upstream dfsg version for that.
Checksums-Sha1:
 303d9e453a04d4ec97e76eba8d06543e6bdd9ed6 2656 pygame_1.9.3+dfsg2-1.dsc
 3082c11d8d09e8c3fe14e5ac7c91e8a7d60e061c 2659415 pygame_1.9.3+dfsg2.orig.tar.gz
 4f0a8ab36850dad4ba0459b10708cc2f336b47c9 25212 
pygame_1.9.3+dfsg2-1.debian.tar.xz
 bcb5fb0b1ebb5e6a75cf1509b76d5f7ad4bd4e79 13714 
pygame_1.9.3+dfsg2-1_source.buildinfo
Checksums-Sha256:
 94b3abaee07af6b3d05422e86845b2e13c7d4f8ca990a53ccaea62eb3ec7ce0a 2656 
pygame_1.9.3+dfsg2-1.dsc
 e393c0a1a8907104804088cc2f4d1bfb52e93fa9fa8bb251099c12810f57f5a7 2659415 
pygame_1.9.3+dfsg2.orig.tar.gz
 62e5883864777d28faba4bee0a88e714b55ccddf5f35a0a3086190866bb8061e 25212 
pygame_1.9.3+dfsg2-1.debian.tar.xz
 354fbbe113cdb2221bdad3ed4b213d338a87ba184ca6cabdc45d0943e27c0ea8 13714 
pygame_1.9.3+dfsg2-1_source.buildinfo
Files:
 724228d3b211ef958b0d099303ad1c8d 2656 python optional pygame_1.9.3+dfsg2-1.dsc
 776dbd9deab7d0a147d7e1116c8d7e3e 2659415 python optional 
pygame_1.9.3+dfsg2.orig.tar.gz
 3315cfd4030f032a9293405055fc468d 25212 python optional 
pygame_1.9.3+dfsg2-1.debian.tar.xz
 9af60fb27bfdf058f1647bd6650846df 13714 python optional 
pygame_1.9.3+dfsg2-1_source.buildinfo

-BEGIN PGP SIGNATURE-
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=PYgq
-END PGP SIGNATURE-


___
Python-modules-team mailing list
Python-modules-team@lists.alioth.debian.org
http://lists.alioth.debian.org/cgi-bin/mailman/listinfo/python-modules-team

[Python-modules-team] Bug#894560: pygame: Don't drop python2 package

[Dominik George]

Hi,

> Your latest upload drops the python-pygame pacakge but I count 35
> packages in Debian Testing that depend on it. Please restore the
> python-pygame package until all those packages no longer depend on it.
> 
> Even without a bug being filed, I believe the broken dependency issue
> would have prevented this package from migrating to Testing.

So, what's the use of that?

I do not target for testing - I target for buster and for sid.  Neither
pygame nor its dependencies, using Python 2, will survive the buster release
cycle, so there is no point in forcefully keeping it around in testing.  And
as you said: Non of this will transition until the problem is solved (either
by removal of the dependencies or their switch to Python 3).

All that will happen with and without the Python 2 version of Pygame.

-nik


signature.asc
Description: PGP signature
___
Python-modules-team mailing list
Python-modules-team@lists.alioth.debian.org
http://lists.alioth.debian.org/cgi-bin/mailman/listinfo/python-modules-team

[Python-modules-team] Bug#867558: flask-ldapconn FTBFS: build dependencies python-ldap3/python3-ldap3 are only available in more recent versions

[Dominik George]

Control: tags -1 + upstream

>  builddeps:flask-ldapconn : Depends: python-ldap3 (< 2.0~) but it is not 
> going to be installed
> Depends: python3-ldap3 (< 2.0~) but it is not 
> going to be installed

ldap3 version 2 requires a complete rewrite. Upstream tracks this at
https://github.com/rroemhild/flask-ldapconn/20

-nik


signature.asc
Description: PGP signature
___
Python-modules-team mailing list
Python-modules-team@lists.alioth.debian.org
http://lists.alioth.debian.org/cgi-bin/mailman/listinfo/python-modules-team

Re: [Python-modules-team] pygame_1.9.3+dfsg-1_i386.changes REJECTED

[Dominik George]

Hi Chris,

On Sat, Feb 11, 2017 at 10:00:10PM +, Chris Lamb wrote:
> 
> Quite a few missing references, including at least src/freetype/ft_layout.c 
> etc
> (I stopped looking there).
> 
>  -- Chris Lamb <la...@debian.org>  Sat, 11 Feb 2017 21:13:55 +

I prepared a new upload, having done a full license review and updated
d/copyright.

It will enter NEW today or tomorrow, hopefully.

Cheers,
Nik

-- 
PGP-Fingerprint: 3C9D 54A4 7575 C026 FB17  FD26 B79A 3C16 A0C4 F296

Dominik George · Hundeshagenstr. 26 · 53225 Bonn
Phone: +49 228 92934581 · https://www.dominik-george.de/

Teckids e.V. · FrOSCon e.V.
Fellowship of the FSFE · Piratenpartei Deutschland
Opencaching Deutschland e.V. · Debian Maintainer

LPIC-3 Linux Enterprise Professional (Security)


signature.asc
Description: PGP signature
___
Python-modules-team mailing list
Python-modules-team@lists.alioth.debian.org
http://lists.alioth.debian.org/cgi-bin/mailman/listinfo/python-modules-team

Re: [Python-modules-team] pygame 1.9.3 upload (was: Re: pygame_1.9.3+dfsg-1_source.changes REJECTED)

[Dominik George]

On Fri, Jan 27, 2017 at 01:32:59AM -0800, Vincent Cheng wrote:
> (please keep the team's mailing list cc-ed)

I bounced the mail there right after I sent it…

> Do you know whether pygame 1.9.3 introduces any backwards-incompatible
> API changes?

Now that's about being a dependency, not being a reverse dependency, as
you said before ;).

And no, both 1.9.2 and 1.9.3 are pure bugfix releases.

> If yes, we would want to treat this like any other
> library transition, i.e. defer it until the next stable release. What
> I want to avoid at this point is to be uploading library packages that
> break its reverse dependencies/build-dependencies and causes a bunch
> of new RC bugs. Have you had a chance to verify at least some of the
> packages listed below to see if they would be negatively affected by
> this upload?
> 
> […]

Yes, I actually did that, as well as some other pygame projects not in
Debian to verify.

Cheers,
Nik

-- 
PGP-Fingerprint: 3C9D 54A4 7575 C026 FB17  FD26 B79A 3C16 A0C4 F296

Dominik George · Hundeshagenstr. 26 · 53225 Bonn
Mobile: +49-1520-1981389 · https://www.dominik-george.de/

Teckids e.V. · FrOSCon e.V.
Fellowship of the FSFE · Piratenpartei Deutschland
Opencaching Deutschland e.V. · Debian Maintainer

LPIC-3 Linux Enterprise Professional (Security)


signature.asc
Description: PGP signature
___
Python-modules-team mailing list
Python-modules-team@lists.alioth.debian.org
http://lists.alioth.debian.org/cgi-bin/mailman/listinfo/python-modules-team

Re: [Python-modules-team] pygame 1.9.3 upload (was: Re: pygame_1.9.3+dfsg-1_source.changes REJECTED)

[Dominik George]

> Well, I do have a sponsor, but apparently, he failed to correctly
> re-sign my prepared files yesterday ;). Will get him to re-upload
> today.

Oh, and then it was refused again because of the -doc package… now
seriously, what's the use of *that* policy (source only uploads not
allowed to NEW)?

-nik

-- 
PGP-Fingerprint: 3C9D 54A4 7575 C026 FB17  FD26 B79A 3C16 A0C4 F296

Dominik George · Hundeshagenstr. 26 · 53225 Bonn
Mobile: +49-1520-1981389 · https://www.dominik-george.de/

Teckids e.V. · FrOSCon e.V.
Fellowship of the FSFE · Piratenpartei Deutschland
Opencaching Deutschland e.V. · Debian Maintainer

LPIC-3 Linux Enterprise Professional (Security)


signature.asc
Description: PGP signature
___
Python-modules-team mailing list
Python-modules-team@lists.alioth.debian.org
http://lists.alioth.debian.org/cgi-bin/mailman/listinfo/python-modules-team

Re: [Python-modules-team] pygame 1.9.3 upload (was: Re: pygame_1.9.3+dfsg-1_source.changes REJECTED)

[Dominik George]

Hi,

>Thanks for preparing an update for pygame!
>
>I haven't yet had a chance to take a close look at your changes, but
>just from skimming your changelog, I have a few comments. Have you had
>a chance to send your newly added patches upstream?

Not yet, but I will.

> Also, would you consider targeting experimental?

No, for various reasons...

> I don't
>think now is an appropriate time to attempt to push packages into
>unstable/testing, especially since pygame is a reverse-dep for many
>other packages.

Hmm... So, I see no issues with that. Dependencies haven't changed (apart from 
sphinx and the font package), the python3 package has the same dependencies as 
the python2 package, and I don't see how it would impact other packages badly.

I was reminded of pygame being old and not having a python3 package yesterday, 
and as I said on IRC, rumour has it the release team might with a very small 
chance let it through.

Pygame is a very important package in education, for example at Teckids, so I 
wanted to at least try.

Again, I do not see negative influence, whether it migrates or not.

>Unless you already have a sponsor, I can take a closer look and
>sponsor your package if you'd like?

Well, I do have a sponsor, but apparently, he failed to correctly re-sign my 
prepared files yesterday ;). Will get him to re-upload today.

I do think the changes really make the package fit for sid or even stable.

Cheers,
Nik

___
Python-modules-team mailing list
Python-modules-team@lists.alioth.debian.org
http://lists.alioth.debian.org/cgi-bin/mailman/listinfo/python-modules-team

[Python-modules-team] Bug#848287: [debian-mysql] Fwd: osmalchemy is marked for autoremoval from testing

[Dominik George]

Hi,

> If I understand the issue here, this is nothing to do with MariaDB being or
> not being a drop-in for MySQL. The problem seems to be this patch in the
> Debian packaging:
> 
>   
> https://github.com/ottok/mariadb-10.1/blob/master/debian/patches/mdev-8375-passwordless-root-via-socket-auth.patch
> 
> The idea is to make the default install of package mariadb-server-10.1 use
> socket authentication for the root user, which seems fine. But the patch
> seems completely wrong. Rather than adding needed functionality to enable
> postinst to setup socket auth, instead it hardcodes this decision into
> mysql_install_db, which breaks other users.
> 
> So it has nothing to do with MySQL vs. MariaDB, such patch could just as
> well have been made against MySQL packaging, with same bad consequences. It
> is simply a bug / unintended consequence of an addition to debian/patches/,
> and simply needs to be fixed. Feel free to correct me if I'm wrong?
> 
> Suggestion for fixing: Add options --auth-root-socket and
> --auth-root-nopasswd to mysql_install_db. Echo a corresponding
> "SET @auth_root_socket=1" or "SET @auth_root_nopasswd=1" down the
> mysqld_install_cmd_line pipe. Then in mysql_system_tables_data.sql choose
> one or the other contents for the user table like this:
> 
>   REPLACE INTO tmp_user_nopasswd ...
>   INSERT INTO tmp_user_socket ...
>   INSERT INTO user SELECT * FROM tmp_user_nopasswd WHERE @had_user_table=0 
> and @auth_root_nopasswd=1;
>   INSERT INTO user SELECT * FROM tmp_user_socket WHERE @had_user_table=0 and 
> @auth_root_socket=1;
> 
> This way, mariadb-server-10.1 postinst can use
> mysql_install_db --auth-root-socket. And ruby-mysql2 can use
> mysql_install_db --auth-root-nopasswd. And if --auth-root-nopasswd is made
> the default, then existing users can work fine without any changes.
> Sounds reasonable?

That's exactly what this is all about.

Thanks for putting it in other words and probably doing better at that
than me!

-nik

-- 
PGP-Fingerprint: 3C9D 54A4 7575 C026 FB17  FD26 B79A 3C16 A0C4 F296

Dominik George · Hundeshagenstr. 26 · 53225 Bonn
Mobile: +49-1520-1981389 · https://www.dominik-george.de/

Teckids e.V. · FrOSCon e.V.
Fellowship of the FSFE · Piratenpartei Deutschland
Opencaching Deutschland e.V. · Debian Maintainer

LPIC-3 Linux Enterprise Professional (Security)


signature.asc
Description: PGP signature
___
Python-modules-team mailing list
Python-modules-team@lists.alioth.debian.org
http://lists.alioth.debian.org/cgi-bin/mailman/listinfo/python-modules-team

[Python-modules-team] Bug#848287: Fwd: osmalchemy is marked for autoremoval from testing

[Dominik George]

> The mysql->mariadb switch comes with the unfortunate cost of excluding a
> few packages from stretch that require oracle's mysql.
> 
> > Do we need the tech-ctte to get this settled?
> 
> Good luck overriding the security team.

It's not about overriding the security team, it's simply about the MySQL
maintainers fixing #848616 [1] or providing reasonable amount of
information on how to migrate packages. Their claim that mariadb can
simply be used as a drop-in replacement does not hold true, and I am not
the only package maintainer stumbling over that.

This package does not require Oracle's MySQL, it requires a root user
that can authenticate to the database and use it in its entirety, a
feature which was deliberately broken in mariadb.

Unfortunately, the MySQL maintainers do not see an issue with that and
ignore threads about it on their mailing list.

So, whether a drop-in replacement in a transition should break a basic
API for everyone clearly is something for the tech-ctte to decide. I am
not talking about removing Oracle's MySQL.

-nik

[1] https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=848616

-- 
PGP-Fingerprint: 3C9D 54A4 7575 C026 FB17  FD26 B79A 3C16 A0C4 F296

Dominik George · Hundeshagenstr. 26 · 53225 Bonn
Mobile: +49-1520-1981389 · https://www.dominik-george.de/

Teckids e.V. · FrOSCon e.V.
Fellowship of the FSFE · Piratenpartei Deutschland
Opencaching Deutschland e.V. · Debian Maintainer

LPIC-3 Linux Enterprise Professional (Security)


signature.asc
Description: PGP signature
___
Python-modules-team mailing list
Python-modules-team@lists.alioth.debian.org
http://lists.alioth.debian.org/cgi-bin/mailman/listinfo/python-modules-team

[Python-modules-team] Bug#848287: Fwd: osmalchemy is marked for autoremoval from testing

[Dominik George]

Hi,

obviously, you raised the severity of the bug mentioned below to serious 
without providing any justification.

I am still waiting for…

…some official decision that this huge change will be made during the freeze 
without a transition,
…the MySQL/MariaDB maintainers to fix their packages.

Concerning the latter, both upstream mariadb and the maintainers here claim 
that mariadb is a drop-in replacement for mysql, but it turns out it isn't as 
they decided to change the default behaviour in a very incompatible way.

The maintainers do not appear to be able to advise others on the unadvertised 
breakage they created by crippling the UNIX socket in mariadb by default.

Right now, dropping MySQL support everywhere seems to be the only viable 
solution.

Do we need the tech-ctte to get this settled?

Cheers,
Nik


 Ursprüngliche Nachricht 
Von: Debian testing autoremoval watch 
Gesendet: 12. Januar 2017 05:39:08 MEZ
An: osmalch...@packages.debian.org
Betreff: osmalchemy is marked for autoremoval from testing

osmalchemy 0.1.+2-2 is marked for autoremoval from testing on 2017-02-10

It (build-)depends on packages with these RC bugs:
848287: python-testing.mysqld: (build-)depends on mysql-{client,server}

___
Python-modules-team mailing list
Python-modules-team@lists.alioth.debian.org
http://lists.alioth.debian.org/cgi-bin/mailman/listinfo/python-modules-team

[Python-modules-team] Bug#848287: python-testing.mysqld: (build-)depends on mysql-{client, server}

[Dominik George]

Dear MySQL/Mariadb maintainers,

> > I was able to identify the issue, but need help fixing it.
> > 
> > The problem is that the testing module uses the default root user of the
> > newly created database, and it uses the UNIX socket, and that has
> > peercred authentication by default in MariaDB.
> > 
> > I tried the following to disable peercred for the socket:
> > 
> > $ cat >init.sql
> > USE mysql;
> > UPDATE user SET plugin='' WHERE User='root';
> > FLUSH PRIVILEGES;
> > 
> > $ mysqld … --initialize-insecure --init-file=init.sql
> > 
> > But it still does not allow connecting a non-root user as root through
> > the UNIX socket.
> > 
> > Any help appreciated.
> 
> Cc'ing pkg-mysql-maint, maybe someone there can help.

As I would like to fix this issue before the full freeze, can you give
any hints on how to circumvent this change in default behaviour of the
mariadb "drop in replacement"?

Kind regards,
Nik

-- 
PGP-Fingerprint: 3C9D 54A4 7575 C026 FB17  FD26 B79A 3C16 A0C4 F296

Dominik George · Hundeshagenstr. 26 · 53225 Bonn
Mobile: +49-1520-1981389 · https://www.dominik-george.de/

Teckids e.V. · FrOSCon e.V.
Fellowship of the FSFE · Piratenpartei Deutschland
Opencaching Deutschland e.V. · Debian Maintainer

LPIC-3 Linux Enterprise Professional (Security)


signature.asc
Description: PGP signature
___
Python-modules-team mailing list
Python-modules-team@lists.alioth.debian.org
http://lists.alioth.debian.org/cgi-bin/mailman/listinfo/python-modules-team

[Python-modules-team] Bug#848287: python-testing.mysqld: (build-)depends on mysql-{client, server}

[Dominik George]

Hi,

I was able to identify the issue, but need help fixing it.

The problem is that the testing module uses the default root user of the
newly created database, and it uses the UNIX socket, and that has
peercred authentication by default in MariaDB.

I tried the following to disable peercred for the socket:

$ cat >init.sql
USE mysql;
UPDATE user SET plugin='' WHERE User='root';
FLUSH PRIVILEGES;

$ mysqld … --initialize-insecure --init-file=init.sql

But it still does not allow connecting a non-root user as root through
the UNIX socket.

Any help appreciated.

-nik

-- 
PGP-Fingerprint: 3C9D 54A4 7575 C026 FB17  FD26 B79A 3C16 A0C4 F296

Dominik George · Hundeshagenstr. 26 · 53225 Bonn
Mobile: +49-1520-1981389 · https://www.dominik-george.de/

Teckids e.V. · FrOSCon e.V.
Fellowship of the FSFE · Piratenpartei Deutschland
Opencaching Deutschland e.V. · Debian Maintainer

LPIC-3 Linux Enterprise Professional (Security)


signature.asc
Description: PGP signature
___
Python-modules-team mailing list
Python-modules-team@lists.alioth.debian.org
http://lists.alioth.debian.org/cgi-bin/mailman/listinfo/python-modules-team

[Python-modules-team] Bug#848287: python-testing.mysqld: (build-)depends on mysql-{client, server}

[Dominik George]

Control: severity -1 important

Hi,

> We're aiming at dropping mysql from testing/stretch because of concerns from 
> the
> security team. See email threads in debian-release@ and pkg-mysql-maint@.

ok, that changes things.

It's very unfortunate that this was not really announced anywhere.

Still, lowering severity as this, per se, is not an RC bug. This package
depends on mysql-server, and this is reality, and there is nothing that
prevents it from doing so if it does for a good reason.

In the case that mysql-server is removed from testing, so will this
package, anyway.

I am trying to port it to mariadb, though.

-nik

-- 
PGP-Fingerprint: 3C9D 54A4 7575 C026 FB17  FD26 B79A 3C16 A0C4 F296

Dominik George · Hundeshagenstr. 26 · 53225 Bonn
Mobile: +49-1520-1981389 · https://www.dominik-george.de/

Teckids e.V. · FrOSCon e.V.
Fellowship of the FSFE · Piratenpartei Deutschland
Opencaching Deutschland e.V. · Debian Maintainer

LPIC-3 Linux Enterprise Professional (Security)


signature.asc
Description: PGP signature
___
Python-modules-team mailing list
Python-modules-team@lists.alioth.debian.org
http://lists.alioth.debian.org/cgi-bin/mailman/listinfo/python-modules-team

[Python-modules-team] Bug#848287: python-testing.mysqld: (build-)depends on mysql-{client, server}

[Dominik George]

Hi,

> You should have kept me in Cc if you wanted me to see your reply and have a
> chance to reply.

sorry, I just cannot remember that the Debian BTS does not automatically
notify the submitter ☹.

So, I have several problems:

 * I cannot find the formal transition for this.
 * I cannot see that mysql-server will be removed from stretch.

As a matter of sad fact, testing.mysqld does not work with mariadb right
now.

What is wrong with depending on mysql-server, given that mysql-server
will still exist although mariadb will become the default?

-nik

-- 
PGP-Fingerprint: 3C9D 54A4 7575 C026 FB17  FD26 B79A 3C16 A0C4 F296

Dominik George · Hundeshagenstr. 26 · 53225 Bonn
Mobile: +49-1520-1981389 · https://www.dominik-george.de/

Teckids e.V. · FrOSCon e.V.
Fellowship of the FSFE · Piratenpartei Deutschland
Opencaching Deutschland e.V. · Debian Maintainer

LPIC-3 Linux Enterprise Professional (Security)


signature.asc
Description: PGP signature
___
Python-modules-team mailing list
Python-modules-team@lists.alioth.debian.org
http://lists.alioth.debian.org/cgi-bin/mailman/listinfo/python-modules-team

[Python-modules-team] Bug#848287: python-testing.mysqld: (build-)depends on mysql-{client, server}

[Dominik George]

Control: severity -1 important

Hi,

> Your package (build-)depends on mysql-server/client. Since we're
> transitioning to mariadb as the default mysql provider, you should
> switch your build dependencies and dependencies to something like:
> 
> default-mysql-server | virtual-mysql-server, default-mysql-client | 
> virtual-mysql-client
> 
> I have seen in your override that you have forwarded this upstream.
> I am filing this anyway to keep track of this along with the rest of the
> packages.

Lowering priority to important (you said "should").

As we are in a transition freeze and nothing along that lines will
happen to stretch, I do not see why this should remove
python-testing.mysqld and all packages that depend on it from testing.

-nik

-- 
PGP-Fingerprint: 3C9D 54A4 7575 C026 FB17  FD26 B79A 3C16 A0C4 F296

Dominik George · Hundeshagenstr. 26 · 53225 Bonn
Mobile: +49-1520-1981389 · https://www.dominik-george.de/

Teckids e.V. · FrOSCon e.V.
Fellowship of the FSFE · Piratenpartei Deutschland
Opencaching Deutschland e.V. · Debian Maintainer

LPIC-3 Linux Enterprise Professional (Security)


signature.asc
Description: PGP signature
___
Python-modules-team mailing list
Python-modules-team@lists.alioth.debian.org
http://lists.alioth.debian.org/cgi-bin/mailman/listinfo/python-modules-team

[Python-modules-team] Bug#841726: ipython3: importing something from ipython has side effects

[Dominik George]

Hi,

> > Certainly, a module breaking the core Python interpreter qualifys for
> > "makes unrelated software break"?
> A module breaking the core Python interpreter only after being imported? 
> Do you have any other scenarios apart from "import all modules" ones?

Well, not technically, but, as already mentioned, "importing all modules" is
what completely normal things in core Python do when you don't expect it,
like help("modules").  Look at this:

8>< -- Console log -- ><8
nik@portux ~ % python3
Python 3.5.2+ (default, Sep 22 2016, 12:18:14) 
[GCC 6.2.0 20160927] on linux
Type "help", "copyright", "credits" or "license" for more information.
>>> help("modules")

Please wait a moment while I gather a list of all available modules...

[…] Warnings from other misbehaving modules […]

[…] list of modules […]

Enter any module name to get more help.  Or, type "modules spam" to search
for modules whose name or summary contain the string "spam".

In : <-- ipdb shell here - and no Python help shell anymore!
8>< -- Console log -- ><8

It is arguable that importing all modules is bad, but it is a sad fact that
it is necessary to gather enough information about a module, and it is
another sad fact that Python has to do it in things like its online help
system.  While this certainly is something that has to be changed, I think
that module developers have to cope with the fact right now and do their
best to not wreak havoc in such a way as the one depicted above.

Cheers,
Nik

-- 
PGP-Fingerprint: 3C9D 54A4 7575 C026 FB17  FD26 B79A 3C16 A0C4 F296

Dominik George · Hundeshagenstr. 26 · 53225 Bonn
Mobile: +49-1520-1981389 · https://www.dominik-george.de/

Teckids e.V. · FrOSCon e.V.
Fellowship of the FSFE · Piratenpartei Deutschland
Opencaching Deutschland e.V. · Debian Contributor

LPIC-3 Linux Enterprise Professional (Security)


signature.asc
Description: PGP signature
___
Python-modules-team mailing list
Python-modules-team@lists.alioth.debian.org
http://lists.alioth.debian.org/cgi-bin/mailman/listinfo/python-modules-team

[Python-modules-team] Bug#841726: ipython3: importing something from ipython has side effects

[Dominik George]

>Control: severity -1 wishlist

Huh? The definition of the critical severity is very clear.

Certainly, a module breaking the core Python interpreter qualifys for "makes 
unrelated software break"?

pgpwzXynlxrS6.pgp
Description: PGP signature
___
Python-modules-team mailing list
Python-modules-team@lists.alioth.debian.org
http://lists.alioth.debian.org/cgi-bin/mailman/listinfo/python-modules-team

[Python-modules-team] Bug#841726: ipython3: importing something from ipython has side effects

[Dominik George]

> There's no need to check this specific instance, it's just one example of a
> potential side-effects that mean importing everything is, alas, by design
> broken. Don't forget all the other import-time monkey-patching, etc. too.

Totally agreed.

-nik

-- 
PGP-Fingerprint: 3C9D 54A4 7575 C026 FB17  FD26 B79A 3C16 A0C4 F296

Dominik George · Hundeshagenstr. 26 · 53225 Bonn
Mobile: +49-1520-1981389 · https://www.dominik-george.de/

Teckids e.V. · FrOSCon e.V.
Fellowship of the FSFE · Piratenpartei Deutschland
Opencaching Deutschland e.V. · Debian Contributor

LPIC-3 Linux Enterprise Professional (Security)


signature.asc
Description: PGP signature
___
Python-modules-team mailing list
Python-modules-team@lists.alioth.debian.org
http://lists.alioth.debian.org/cgi-bin/mailman/listinfo/python-modules-team

[Python-modules-team] Bug#841726: Acknowledgement (ipython3: importing something from ipython has side effects)

[Dominik George]

Control: reassign -1 python3-ipdb 0.10.1-1

Oh, I was actually wrong in that this is IPython's mistake…

The root cause is ipdb, which causes misbehaviour in both the core
interpreter and IPython ;)!

Sorry, IPython.

-nik

-- 
PGP-Fingerprint: 3C9D 54A4 7575 C026 FB17  FD26 B79A 3C16 A0C4 F296

Dominik George · Hundeshagenstr. 26 · 53225 Bonn
Mobile: +49-1520-1981389 · https://www.dominik-george.de/

Teckids e.V. · FrOSCon e.V.
Fellowship of the FSFE · Piratenpartei Deutschland
Opencaching Deutschland e.V. · Debian Contributor

LPIC-3 Linux Enterprise Professional (Security)


signature.asc
Description: PGP signature
___
Python-modules-team mailing list
Python-modules-team@lists.alioth.debian.org
http://lists.alioth.debian.org/cgi-bin/mailman/listinfo/python-modules-team

[Python-modules-team] Bug#841726: ipython3: importing something from ipython has side effects

[Dominik George]

Hi,

> > Besides speech-dispatcher, against which I also reported a bug, I do not
> > have any packages on my system that show such behaviour.
> 
> Oh, really?  But surely some of them adjust sys.path (probably the most
> common?) as well as importing shared libraries and running stuff like
> ``apt_pkg.init_config()``.
> 
> The side-effects might be more subtle but they are still, alas, side-
> effects.

Yep.  That's not good either, but it doesn't make unrelated software, let
alone the core interpreter, misbehave (i.e. take control from it).

I wouldn't open an RC bug against python-apt because it calls init_config(),
but I did against speech-dispatcher because it configures argparse and
consumes sys.args and I did against ipython bacause it takes control over
parts of the core interpreter.

-nik

-- 
PGP-Fingerprint: 3C9D 54A4 7575 C026 FB17  FD26 B79A 3C16 A0C4 F296

Dominik George · Hundeshagenstr. 26 · 53225 Bonn
Mobile: +49-1520-1981389 · https://www.dominik-george.de/

Teckids e.V. · FrOSCon e.V.
Fellowship of the FSFE · Piratenpartei Deutschland
Opencaching Deutschland e.V. · Debian Contributor

LPIC-3 Linux Enterprise Professional (Security)


signature.asc
Description: PGP signature
___
Python-modules-team mailing list
Python-modules-team@lists.alioth.debian.org
http://lists.alioth.debian.org/cgi-bin/mailman/listinfo/python-modules-team

[Python-modules-team] Bug#841726: ipython3: importing something from ipython has side effects

[Dominik George]

> Not at all.  This is the only way to get a list of all available Python
> packages, and it is even what core Python's help("modules") does.

And no, I do not like that either ☹.  But it's a matter of sad fact.

-- 
PGP-Fingerprint: 3C9D 54A4 7575 C026 FB17  FD26 B79A 3C16 A0C4 F296

Dominik George · Hundeshagenstr. 26 · 53225 Bonn
Mobile: +49-1520-1981389 · https://www.dominik-george.de/

Teckids e.V. · FrOSCon e.V.
Fellowship of the FSFE · Piratenpartei Deutschland
Opencaching Deutschland e.V. · Debian Contributor

LPIC-3 Linux Enterprise Professional (Security)


signature.asc
Description: PGP signature
___
Python-modules-team mailing list
Python-modules-team@lists.alioth.debian.org
http://lists.alioth.debian.org/cgi-bin/mailman/listinfo/python-modules-team

[Python-modules-team] Bug#841726: ipython3: importing something from ipython has side effects

[Dominik George]

Hi,

> Many, many Python packages have import-time side-effects that are
> far more suble than this. Whilst this a regrettable state of affairs,
> fixing them all is just not possible anymore.

Well, this bug report is about this specific bug, not all bugs of this kind.

> 
> A script that imports all packages installed on a system as you
> suggest seems to be the problem, rather than ipython itself.

Not at all.  This is the only way to get a list of all available Python
packages, and it is even what core Python's help("modules") does.  So, when
ipython is installed, running help("modules") in a pure python3 interpreter
modifes the execution environment for good, in such a way that it drops to
an IPython shell at any point as it likes.

Besides speech-dispatcher, against which I also reported a bug, I do not
have any packages on my system that show such behaviour.

-nik

-- 
PGP-Fingerprint: 3C9D 54A4 7575 C026 FB17  FD26 B79A 3C16 A0C4 F296

Dominik George · Hundeshagenstr. 26 · 53225 Bonn
Mobile: +49-1520-1981389 · https://www.dominik-george.de/

Teckids e.V. · FrOSCon e.V.
Fellowship of the FSFE · Piratenpartei Deutschland
Opencaching Deutschland e.V. · Debian Contributor

LPIC-3 Linux Enterprise Professional (Security)


signature.asc
Description: PGP signature
___
Python-modules-team mailing list
Python-modules-team@lists.alioth.debian.org
http://lists.alioth.debian.org/cgi-bin/mailman/listinfo/python-modules-team

[Python-modules-team] Bug#841726: ipython3: importing something from ipython has side effects

[Dominik George]

Package: ipython3
Version: 2.4.1-1
Severity: critical
Justification: breaks unrelated software

-BEGIN PGP SIGNED MESSAGE-
Hash: SHA256

Importing ipython in any other Python script has side effects.  This is
sometimes done, e.g.  in situations where a script wants to list all
packages installed on a system, like so:

  import pkgutil
  for i in pkgutil.walk_packages(onerror=lambda x: None):
pass

Python imports all packages discovered by that once, and simply importing
something must not have any side effects.

However, Python drops into an IPython shell afterwards and also behaves
oddly in other ways.

- -- System Information:
Debian Release: stretch/sid
  APT prefers unstable
  APT policy: (500, 'unstable'), (500, 'testing'), (500, 'stable')
Architecture: amd64 (x86_64)
Foreign Architectures: i386

Kernel: Linux 4.7.0-1-amd64 (SMP w/4 CPU cores)
Locale: LANG=de_DE.UTF-8, LC_CTYPE=de_DE.UTF-8 (charmap=UTF-8)
Shell: /bin/sh linked to /bin/lksh
Init: systemd (via /run/systemd/system)

Versions of packages ipython3 depends on:
ii  python3-decorator  4.0.6-1
ii  python3-pkg-resources  28.0.0-1
ii  python3-simplegeneric  0.8.1-1
pn  python3:any

ipython3 recommends no packages.

Versions of packages ipython3 suggests:
pn  ipython3-notebook   
pn  ipython3-qtconsole  
pn  python3-zmq 

- -- no debconf information

-BEGIN PGP SIGNATURE-
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=20Ku
-END PGP SIGNATURE-

___
Python-modules-team mailing list
Python-modules-team@lists.alioth.debian.org
http://lists.alioth.debian.org/cgi-bin/mailman/listinfo/python-modules-team

[Python-modules-team] Maintenance of new Flask

[Dominik George]

Hi DPMT,

I would like to package and maintain some Flask extensions. Flask, a web 
micro-framework, and some extensions for it are already available in Debian 
and under team maintenance of the DPMT.

I would initially like to add these two extensions:

  Flask-Restless
  Flask-Compress
  Flask-LDAPConn

Also, I would like to help fix open bugs in existing Flask packages.

I would base my packages off of existing ones in order to enable the team to 
co-maintain the packages.

Apart from these new packages, I maintain a few other packages in Debian, in 
cooperation with several DDs.

If the DPMT is willing to co-maintain the new packages, I would like to keep 
the package repositories and the like in DPMT from the beginning.

Please tell me if it is ok to ask for team membership on Alioth ☺!

Cheers,
Nik

-- 
PGP-Fingerprint: 3C9D 54A4 7575 C026 FB17  FD26 B79A 3C16 A0C4 F296

Dominik George · Mobil: +49-1520-1981389

Teckids e.V. · FrOSCon e.V. · OpenRheinRuhr e.V.
Fellowship of the FSFE · Piratenpartei Deutschland
Opencaching Deutschland e.V. · Debian Contributor

LPIC-3 Linux Enterprise Professional (Security)

signature.asc
Description: This is a digitally signed message part.
___
Python-modules-team mailing list
Python-modules-team@lists.alioth.debian.org
http://lists.alioth.debian.org/cgi-bin/mailman/listinfo/python-modules-team

[Python-modules-team] Bug#695050: RFS: python-webob/1.1.1-1.1 [RC] [NMU] -- fix for RC bug #695050

[Dominik George]

-BEGIN PGP SIGNED MESSAGE-
Hash: SHA256

Package: sponsorship-requests
Severity: important

Dear mentors,

I am looking for a sponsor for my NMU for package python-webob. The 
upload would fix RC bug #695050.

Package name: python-webob
Version : 1.1.1-1.1
Section : python

It builds those binary packages:

   python-webob - Python module providing WSGI request and response objects

Please note that the package produces two lintian warnings, but those were 
already in the previous version and were not introduced with the bugfix. 
To ease the work for the release team (concerning the freeze exception to 
be made for the bugfix) and to not touch the maintainer's work more than 
necessary for the bugfix, I ignored these warnings.

To access further information about this package, please visit the 
following URL:

  http://mentors.debian.net/package/python-webob

Alternatively, one can download the package with dget using this command:

  dget -x 
http://mentors.debian.net/debian/pool/main/p/python-webob/python-webob_1.1.1-1.1.dsc

Changes since the last upload:

python-webob (1.1.1-1.1) unstable; urgency=low

  * Non-maintainer upload.
  * Fix response header bug reported by Mehdi Abaakouk (Closes: #695050).

 -- Dominik George n...@naturalnet.de  Tue, 04 Dec 2012 00:09:48 +0100

I propose uploading to DELAYED/5 to give the maintainer some more time to 
respond to the bug report themselves, but to squash RC bugs I'd really 
love to see the upload happen so things can go on in a few days!

Regards,
 Dominik George

- -- 
* mirabilos is handling my post-1990 smartphone *
mirabilos Aaah, it vibrates! Wherefor art thou, daemonic device??

PGP fingerprint: 2086 9A4B E67D 1DCD FFF6  F6C1 59FC 8E1D 6F2A 8001
-BEGIN PGP SIGNATURE-
Version: GnuPG v2.0.19 (GNU/Linux)

iQFOBAEBCAA4BQJQwNonMRpodHRwczovL3d3dy5kb21pbmlrLWdlb3JnZS5kZS9n
cGctcG9saWN5LnR4dC5hc2MACgkQWfyOHW8qgAEnvQf/a/PIhqXxFn1gTBRemmve
6bIefNma/4wwbtsyZfLWBerFrz+Bp19Q9Z9wAYr2a73UWVFn7nMwWpgqACW0l719
T9nv0FYqe9lLUUBo1PXu9ufigwbUaGmh5zYGWbY1Hve85rIds1jCZTRwOumYRkHE
21yu649PlDxKAgZPetZhB+hn1vJ6wrxz7Pr/p/EM80xzdzNd90/OltJCCSJkTUL0
lFY9nnKRB1YrSp5xHiRhgqUDRa2NEYOsyIdnxRexaD1xrJQ/lwYVfTQOCIiPpse0
9YOsjfc35MPf/p3tnivaYtJUDscKmr40rkAA7Tv+lmBEMznQOM02nIDD9ggH3m6u
eQ==
=tKKt
-END PGP SIGNATURE-

___
Python-modules-team mailing list
Python-modules-team@lists.alioth.debian.org
http://lists.alioth.debian.org/cgi-bin/mailman/listinfo/python-modules-team

[Python-modules-team] Bug#695050: python-webob: webob last stable version 1.1.1 response header bug

[Dominik George]

-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1

tags 695050 + patch pending
severity 695050 important
kthxbye

Hi,

I ported the patch from Ubuntu to the Debian package and prepared an NMU 
should the team appreciate that. The following files will be of interest:

Source package: http://shore.naturalnet.de/~nik/python-webob_1.1.1-1.1.dsc
Binary package: http://shore.naturalnet.de/~nik/python-webob_1.1.1-1.1_all.deb

Mehdi, can you please test the package in your test case and report if 
that really fixes the bug for you? If it does, the Python modules team 
will either fix the bug themselves or I will RFS the package.

Attached is the debdiff for this change.

I chose to downgrade the severity because this bug does not affect all use 
cases of python-webob. Please note that important bugs still qualify for 
wheezy freeze exceptions.

Cheers,
Nik

- -- 
* mirabilos is handling my post-1990 smartphone *
mirabilos Aaah, it vibrates! Wherefor art thou, daemonic device??

PGP fingerprint: 2086 9A4B E67D 1DCD FFF6  F6C1 59FC 8E1D 6F2A 8001
-BEGIN PGP SIGNATURE-
Version: GnuPG v2.0.19 (GNU/Linux)

iQFOBAEBAgA4BQJQvTq9MRpodHRwczovL3d3dy5kb21pbmlrLWdlb3JnZS5kZS9n
cGctcG9saWN5LnR4dC5hc2MACgkQWfyOHW8qgAG3RQf+J2UUm8eaHyp9bbQ94bED
GB8JhPNR7o24Hap2fZAiq5hjSKN/cMz/We3KMB/5e7MThdYeZUOCqTaocb/fUymD
2MISV7/Ql7TGeOdZ5hisLEbzai8LTI00DB3T/iKVuvei3M5FTcEi3VD+4JX2qen/
ggszA3h9vt2VJUDPOObi2P8e8XhB9vjKkuqq3hqNimK2i4a6+cLsrLmvsec/s11d
FgRVirIvdjabeLsOBwfT0Yv5pJTRPG3Sz9v0rc0xRZxngJnDg9dm4fS+fB2d4HCm
rH8rIcuiDZYsSeOoGCyACgbB6izVVnLe8DZeAL9gG0woc/02S3yr098/A8lDCIGT
JA==
=rKmS
-END PGP SIGNATURE-diff -Nru python-webob-1.1.1/debian/changelog python-webob-1.1.1/debian/changelog
--- python-webob-1.1.1/debian/changelog	2011-09-19 19:53:36.0 +0200
+++ python-webob-1.1.1/debian/changelog	2012-12-04 00:11:36.0 +0100
@@ -1,3 +1,10 @@
+python-webob (1.1.1-1.1) unstable; urgency=low
+
+  * Non-maintainer upload.
+  * Fix response header bug reported by Mehdi Abaakouk (Closes: #695050).
+
+ -- Dominik George n...@naturalnet.de  Tue, 04 Dec 2012 00:09:48 +0100
+
 python-webob (1.1.1-1) unstable; urgency=low
 
   * New upstream release
diff -Nru python-webob-1.1.1/debian/patches/head_bugfix_lp_920197.patch python-webob-1.1.1/debian/patches/head_bugfix_lp_920197.patch
--- python-webob-1.1.1/debian/patches/head_bugfix_lp_920197.patch	1970-01-01 01:00:00.0 +0100
+++ python-webob-1.1.1/debian/patches/head_bugfix_lp_920197.patch	2012-12-04 00:09:11.0 +0100
@@ -0,0 +1,71 @@
+--- a/tests/test_exc.py
 b/tests/test_exc.py
+@@ -55,6 +55,7 @@
+ from webob.exc import HTTPInsufficientStorage
+ from webob.exc import HTTPExceptionMiddleware
+ from webob import exc
++from webob.exc import status_map
+ 
+ from nose.tools import eq_, ok_, assert_equal, assert_raises
+ 
+@@ -250,6 +251,43 @@
+ exc.newstyle_exceptions = False
+ assert_equal( excep(environ,start_response), [] )
+ 
++def test_HTTPOk_head_of_proxied_head():
++# first set up a response to a HEAD request
++HELLO_WORLD = Hi!\n
++CONTENT_TYPE = application/hello
++def head_app(environ, start_response):
++An application object that understands HEAD
++status = '200 OK'
++response_headers = [('Content-Type', CONTENT_TYPE),
++('Content-Length', len(HELLO_WORLD))]
++start_response(status, response_headers)
++
++if environ['REQUEST_METHOD'] == 'HEAD':
++return []
++else:
++return [HELLO_WORLD]
++
++def verify_response(resp, description):
++assert_equal(resp.content_type, CONTENT_TYPE, description)
++assert_equal(resp.content_length, len(HELLO_WORLD), description)
++assert_equal(resp.body, '', description)
++
++req = Request.blank('/', method='HEAD')
++resp1 = req.get_response(head_app)
++verify_response(resp1, first response)
++
++# Copy the response like a proxy server would.
++# Copying an empty body has set content_length
++# so copy the headers only afterwards.
++resp2 = status_map[resp1.status_int](request=req)
++resp2.body = resp1.body
++resp2.headerlist = resp1.headerlist
++verify_response(resp2, copied response)
++
++# evaluate it again
++resp3 = req.get_response(resp2)
++verify_response(resp3, evaluated copy)
++
+ def test_HTTPMove():
+ def start_response(status, headers, exc_info=None):
+ pass
+--- a/webob/exc.py
 b/webob/exc.py
+@@ -314,11 +314,12 @@
+ return resp(environ, start_response)
+ 
+ def __call__(self, environ, start_response):
+-if self.body or self.empty_body:
++is_head = environ['REQUEST_METHOD'] == 'HEAD'
++if self.body or self.empty_body or is_head:
+ app_iter = Response.__call__(self, environ, start_response)
+ else:
+ app_iter = self.generate_response(environ, start_response)
+-if environ['REQUEST_METHOD'] == 'HEAD':
++if is_head