Debian Bug report logs - #700591
[pbuilder] Add support for shared memory mounting point

[Message part 1 (text/plain, inline)]
Package: pbuilder
Version: 0.213
Severity: normal
Tags: patch

Dear mantainer,
               I've attached a patch that adds support for shared
memory. Just creates a mounting point at /run/shm (if exists, otherwise
uses /dev/shm as monting point) as tmpfs. I've found this constraint
when running the test cases one of my packages.

Could you check and apply it, please?

-- 
TiN

[0001-Added-support-for-shared-memory-mounting-point.patch (text/x-patch, attachment)]
[signature.asc (application/pgp-signature, attachment)]
[Message part 1 (text/plain, inline)]
The package which requires shared memory for test cases is python-doit.
When I tried without the patch attached before, It fails and the output
build is attached. I've isolated the python line that breaks in the test
cases. I add it to see if it helps clarify.

"from multiprocessing.synchronize import Lock; Lock()"

If pbuilder does not have shared memory mounted it fails with the output
"OSError: [Errno 38] Function not implemented".

-- 
TiN

[doit_0.20.0-1_amd64.build (text/plain, attachment)]
[signature.asc (application/pgp-signature, attachment)]
Agustin Henze dixit:

>               I've attached a patch that adds support for shared
>memory. Just creates a mounting point at /run/shm (if exists, otherwise
>uses /dev/shm as monting point) as tmpfs. I've found this constraint
>when running the test cases one of my packages.

apr-util also needs this, and appending this to pbuilderrc for when
building such packages suffices:

BINDMOUNTS="$BINDMOUNTS /run/shm"

bye,
//mirabilos
-- 
13:37⎜«Natureshadow» Deep inside, I hate mirabilos. I mean, he's a good
guy. But he's always right! In every fsckin' situation, he's right. Even
with his deeply perverted taste in software and borked ambition towards
broken OSes - in the end, he's damn right about it :(! […] works in mksh

[Message part 1 (text/plain, inline)]
On 02/14/2013 07:47 PM, Thorsten Glaser wrote:
> Agustin Henze dixit:
> 
>>               I've attached a patch that adds support for shared
>> memory. Just creates a mounting point at /run/shm (if exists, otherwise
>> uses /dev/shm as monting point) as tmpfs. I've found this constraint
>> when running the test cases one of my packages.
> 
> apr-util also needs this, and appending this to pbuilderrc for when
> building such packages suffices:
> 
> BINDMOUNTS="$BINDMOUNTS /run/shm"

Hi Thorsten,
            that's sounds "a little" easier :). Ok, I think that
/run/shm should be add by default into BINDMOUNTS option.

Best regards,

-- 
TiN

[signature.asc (application/pgp-signature, attachment)]
Agustin Henze dixit:

>            that's sounds "a little" easier :). Ok, I think that
>/run/shm should be add by default into BINDMOUNTS option.

There’s probably a reason it’s not default…
but I’ll leave that to the pbuilder/cowbuilder
people to discover or explain.

bye,
//mirabilos
-- 
„nein: BerliOS und Sourceforge sind Plattformen für Projekte, github ist
eine Plattform für Einzelkämpfer“
	-- dieses Zitat ist ein Beweis dafür, daß auch ein blindes Huhn
	   mal ein Korn findet, bzw. – in diesem Fall – Recht haben kann

I think the reason it's not done yet is that pbuilder predates /run/shm and we need to be updated...

 At Fri, 15 Feb 2013 16:36:36 +0000 (UTC),
Thorsten Glaser wrote:
> 
> Agustin Henze dixit:
> 
> >            that's sounds "a little" easier :). Ok, I think that
> >/run/shm should be add by default into BINDMOUNTS option.
> 
> There’s probably a reason it’s not default…
> but I’ll leave that to the pbuilder/cowbuilder
> people to discover or explain.
> 
> bye,
> //mirabilos
> -- 
> „nein: BerliOS und Sourceforge sind Plattformen für Projekte, github ist
> eine Plattform für Einzelkämpfer“
> 	-- dieses Zitat ist ein Beweis dafür, daß auch ein blindes Huhn
> 	   mal ein Korn findet, bzw. – in diesem Fall – Recht haben kann
> 
> _______________________________________________
> Pbuilder-maint mailing list
> Pbuilder-maint@lists.alioth.debian.org
> http://lists.alioth.debian.org/cgi-bin/mailman/listinfo/pbuilder-maint

Junichi Uekawa dixit:

>I think the reason it's not done yet is that pbuilder predates
>/run/shm and we need to be updated...

No, it was not done for /dev/shm either. /me smells security issues
or something like that, though I never understood what GNU libc uses
this for.

> At Fri, 15 Feb 2013 16:36:36 +0000 (UTC),
>Thorsten Glaser wrote:

http://www.afaik.de/usenet/faq/zitieren/

bye,
//mirabilos
-- 
13:37⎜«Natureshadow» Deep inside, I hate mirabilos. I mean, he's a good
guy. But he's always right! In every fsckin' situation, he's right. Even
with his deeply perverted taste in software and borked ambition towards
broken OSes - in the end, he's damn right about it :(! […] works in mksh

At Wed, 27 Feb 2013 22:30:58 +0000 (UTC),
Thorsten Glaser wrote:
> 
> Junichi Uekawa dixit:
> 
> >I think the reason it's not done yet is that pbuilder predates
> >/run/shm and we need to be updated...
> 
> No, it was not done for /dev/shm either. /me smells security issues
> or something like that, though I never understood what GNU libc uses
> this for.

I would imagine shmget etc will probably create temporary file there
and mmap. If someone depends on /run/shm and is just using that for
creating temp files, that's bad.

(I don't particularly like the fact that simple builds requiring
directories outside of your home directory other than /tmp to be
writeable)

> 
> > At Fri, 15 Feb 2013 16:36:36 +0000 (UTC),
> >Thorsten Glaser wrote:
> 
> http://www.afaik.de/usenet/faq/zitieren/
> 
> bye,
> //mirabilos
> -- 
> 13:37⎜«Natureshadow» Deep inside, I hate mirabilos. I mean, he's a good
> guy. But he's always right! In every fsckin' situation, he's right. Even
> with his deeply perverted taste in software and borked ambition towards
> broken OSes - in the end, he's damn right about it :(! […] works in mksh
> 

Junichi Uekawa dixit:

>(I don't particularly like the fact that simple builds requiring
>directories outside of your home directory other than /tmp to be
>writeable)

I agree with that (in fact, IMHO even writing to something rooted
under $HOME is not nice).

In BSD ports, there is some attempt using the systrace mechanism,
flawed that it is it at least helps unless someone wants to acti‐
vely circumvent it, to prevent writing outside one’s own tree.

bye,
//mirabilos
-- 
<Natureshadow> Wieso gehen die Nachrichten jetzt alle auf mein Handy?
<Natureshadow> Jabber ist scheiße.
<mirabilos> tys (told you so)
<Natureshadow> Ja, aber ich will jetzt verstehen, warum das so ist!

Source: pbuilder
Source-Version: 0.214

We believe that the bug you reported is fixed in the latest version of
pbuilder, which is due to be installed in the Debian FTP archive.

A summary of the changes between this version and the previous one is
attached.

Thank you for reporting the bug, which will now be closed.  If you
have further comments please address them to 700591@bugs.debian.org,
and the maintainer will reopen the bug report if appropriate.

Debian distribution maintenance software
pp.
Junichi Uekawa <dancer@debian.org> (supplier of updated pbuilder package)

(This message was generated automatically at their request; if you
believe that there is a problem with it please contact the archive
administrators by mailing ftpmaster@debian.org)

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Format: 1.8
Date: Sat, 02 Mar 2013 16:47:04 +0900
Source: pbuilder
Binary: pbuilder pbuilder-uml
Architecture: source amd64 all
Version: 0.214
Distribution: unstable
Urgency: low
Maintainer: Debian pbuilder maintenance team <pbuilder-maint@lists.alioth.debian.org>
Changed-By: Junichi Uekawa <dancer@debian.org>
Description: 
 pbuilder   - personal package builder for Debian packages
 pbuilder-uml - user-mode-linux version of pbuilder
Closes: 694311 700591
Changes: 
 pbuilder (0.214) unstable; urgency=low
 .
   [ Junichi Uekawa ]
   * update documentation about CCACHEDIR (closes: #694311)
   * update the release target to work with wagner, and release.
 .
   [ Agustin Henze ]
   * Added support for shared memory mounting point (closes: #700591)
Checksums-Sha1: 
 b4942f4eaa34a437fa72b502e1ec13c0c29e7dad 1187 pbuilder_0.214.dsc
 fc8f74bf5c6b29aa761226a52d38090f11b8f474 939371 pbuilder_0.214.tar.gz
 d62a7a5d2780f9c1c06b07487a6b1c810d27abb9 44028 pbuilder-uml_0.214_amd64.deb
 9b785fa03521e34cd429ff1c2e099aa4393b468c 390986 pbuilder_0.214_all.deb
Checksums-Sha256: 
 cd758b19faeea1b5344e71341d551e916ce5c233265eeefa764d4de54841a997 1187 pbuilder_0.214.dsc
 0056439777edea8278c6f9dd95d42c6ac8f6cfe2235bcea6fbcebb790098d034 939371 pbuilder_0.214.tar.gz
 c0abd8b5a3b1d0c66faff0d6e81916a56c7a4bba5a776ef906d02144455c1fe4 44028 pbuilder-uml_0.214_amd64.deb
 91088ac7f1f91e12eb2a5ebca63410f60b33baa3cc73fec730b1ed8a44bc5d2f 390986 pbuilder_0.214_all.deb
Files: 
 2af142cd7ff0a36fbba0f1b03d8f782e 1187 devel extra pbuilder_0.214.dsc
 a1cc0683a301484d29ca29aa19c8160a 939371 devel extra pbuilder_0.214.tar.gz
 7a5930d04541d69a99bdee47126027d3 44028 devel extra pbuilder-uml_0.214_amd64.deb
 50e7f604300ad324c867e0e9bd76dd32 390986 devel extra pbuilder_0.214_all.deb

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.12 (GNU/Linux)

iD8DBQFRMbCW2Dd9TugeVcERAjgLAJ9Sw60VriP/1ZF1vNQ9qK5SLtBVPgCfYi22
FgpEFCZaEXTKEA4P0FvW9K8=
=vgSb
-----END PGP SIGNATURE-----

Debian Bug report logs - #700591 [pbuilder] Add support for shared memory mounting point

Debian Bug report logs - #700591
[pbuilder] Add support for shared memory mounting point