pysparse-developers

[Pysparse-developers] Python 2.5 and memory management

[Adam C P. IV <haz...@de...>]

Greetings,

First, let me introduce myself: I'm the maintainer of the Debian package
of pysparse.  The package takes a somewhat different approach from
upstream: instead of building the superlu, umfpack and amd sources in
the pysparse package, it links to existing Debian packages for those.

This is an "FYI" on Debian bug 468991, viewable at
http://bugs.debian.org/468991 .  The bug submitter asserts that the
memory management model of python 2.5 has changed, such that objects
created using PyObject_NEW should use PyObject_Del in their destructors,
not PyMem_DEL.

According to Thomas Viehmann, this will not cause a segfault in
pysparse, but should be cleaned up eventually.  I figured you should be
aware of this as a maintenance "To-Do" item.

I'll leave this bug open until pysparse changes.

Thanks,
-Adam
-- 
GPG fingerprint: D54D 1AEE B11C CE9B A02B  C5DD 526F 01E8 564E E4B6

Engineering consulting with open source tools
http://www.opennovation.com/

Re: [Pysparse-developers] Python 2.5 and memory management

[Daniel W. <dan...@gm...>]

Hi Adam,

I'll look into this in the near future. Thanks for taking care of the
deb packaging.

I do have one related question for you. How does one test deb packages without a
clean system or root access? There doesn't seem to be an obvious way unpack and
install locally using apt-get. Any clues?

Cheers

On Mon, Mar 24, 2008 at 4:17 PM, Adam C Powell IV <haz...@de...> wrote:
> Greetings,
>
>  First, let me introduce myself: I'm the maintainer of the Debian package
>  of pysparse.  The package takes a somewhat different approach from
>  upstream: instead of building the superlu, umfpack and amd sources in
>  the pysparse package, it links to existing Debian packages for those.
>
>  This is an "FYI" on Debian bug 468991, viewable at
>  http://bugs.debian.org/468991 .  The bug submitter asserts that the
>  memory management model of python 2.5 has changed, such that objects
>  created using PyObject_NEW should use PyObject_Del in their destructors,
>  not PyMem_DEL.
>
>  According to Thomas Viehmann, this will not cause a segfault in
>  pysparse, but should be cleaned up eventually.  I figured you should be
>  aware of this as a maintenance "To-Do" item.
>
>  I'll leave this bug open until pysparse changes.
>
>  Thanks,
>  -Adam
>  --
>  GPG fingerprint: D54D 1AEE B11C CE9B A02B  C5DD 526F 01E8 564E E4B6
>
>  Engineering consulting with open source tools
>  http://www.opennovation.com/
>
>
>  -------------------------------------------------------------------------
>  This SF.net email is sponsored by: Microsoft
>  Defy all challenges. Microsoft(R) Visual Studio 2008.
>  http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/
>  _______________________________________________
>  Pysparse-developers mailing list
>  Pys...@li...
>  https://lists.sourceforge.net/lists/listinfo/pysparse-developers
>



-- 
Daniel Wheeler

Re: [Pysparse-developers] Python 2.5 and memory management

[Andreas K. <li...@in...>]

On Dienstag 25 März 2008, Daniel Wheeler wrote:
> Hi Adam,
>
> I'll look into this in the near future. Thanks for taking care of the
> deb packaging.
>
> I do have one related question for you. How does one test deb packages
> without a clean system or root access? There doesn't seem to be an obvious
> way unpack and install locally using apt-get. Any clues?

Hit enter on a deb in midnight commander (mc).

:)

Andreas

Re: [Pysparse-developers] Python 2.5 and memory management

[Daniel W. <dan...@gm...>]

On Tue, Mar 25, 2008 at 10:16 AM, Andreas Klöckner
<li...@in...> wrote:

>
>  Hit enter on a deb in midnight commander (mc).
>
>  :)

That much seems good, but what about dependencies?

I'd like to set up a secondary "/usr" type directory in my local area
and specify
apt-get to download and install everything in that area. There doesn't seem to
be command line options for apt-get that allows that. Anyhow, this is
probably the
wrong list for this discussion.



-- 
Daniel Wheeler

Re: [Pysparse-developers] Python 2.5 and memory management

[Adam C P. IV <haz...@de...>]

Hello Dan, it's good to hear from you.

On Tue, 2008-03-25 at 10:06 -0400, Daniel Wheeler wrote:
> Hi Adam,
> 
> I'll look into this in the near future. Thanks for taking care of the
> deb packaging.

You're welcome.

> I do have one related question for you. How does one test deb packages without a
> clean system or root access? There doesn't seem to be an obvious way unpack and
> install locally using apt-get. Any clues?

A .deb package is an ar archive with a tarball of package files, a
tarball of control files, and a brief version file.  You can use "ar x
package.deb" which will give you these three, and then unpack the
package files and try to use them.  But you don't get the
system-integrating scripts etc.

As far as I know there's no good way to unpack and use a .deb in one's
home directory.

Do you have sufficient access to make a chroot environment?

-Adam

> On Mon, Mar 24, 2008 at 4:17 PM, Adam C Powell IV <haz...@de...> wrote:
> > Greetings,
> >
> >  First, let me introduce myself: I'm the maintainer of the Debian package
> >  of pysparse.  The package takes a somewhat different approach from
> >  upstream: instead of building the superlu, umfpack and amd sources in
> >  the pysparse package, it links to existing Debian packages for those.
> >
> >  This is an "FYI" on Debian bug 468991, viewable at
> >  http://bugs.debian.org/468991 .  The bug submitter asserts that the
> >  memory management model of python 2.5 has changed, such that objects
> >  created using PyObject_NEW should use PyObject_Del in their destructors,
> >  not PyMem_DEL.
> >
> >  According to Thomas Viehmann, this will not cause a segfault in
> >  pysparse, but should be cleaned up eventually.  I figured you should be
> >  aware of this as a maintenance "To-Do" item.
> >
> >  I'll leave this bug open until pysparse changes.
> >
> >  Thanks,
> >  -Adam
-- 
GPG fingerprint: D54D 1AEE B11C CE9B A02B  C5DD 526F 01E8 564E E4B6

Engineering consulting with open source tools
http://www.opennovation.com/

Re: [Pysparse-developers] Python 2.5 and memory management

[Daniel W. <dan...@gm...>]

On Wed, Mar 26, 2008 at 8:18 AM, Adam C Powell IV <haz...@de...> wrote:

>  > I do have one related question for you. How does one test deb packages without a
>  > clean system or root access? There doesn't seem to be an obvious way unpack and
>  > install locally using apt-get. Any clues?
>
>  A .deb package is an ar archive with a tarball of package files, a
>  tarball of control files, and a brief version file.  You can use "ar x
>  package.deb" which will give you these three, and then unpack the
>  package files and try to use them.  But you don't get the
>  system-integrating scripts etc.
>
>  As far as I know there's no good way to unpack and use a .deb in one's
>  home directory.
>
>  Do you have sufficient access to make a chroot environment?

Yes I do. I haven't tried it though. It may be the correct way to
create a sandbox for testing
packaging. I'll try it out and let you know. Thanks!

-- 
Daniel Wheeler

Re: [Pysparse-developers] Python 2.5 and memory management

[Adam C P. IV <haz...@de...>]

On Fri, 2008-03-28 at 10:18 -0400, Daniel Wheeler wrote:
> On Wed, Mar 26, 2008 at 8:18 AM, Adam C Powell IV <haz...@de...> wrote:
> 
> >  > I do have one related question for you. How does one test deb packages without a
> >  > clean system or root access? There doesn't seem to be an obvious way unpack and
> >  > install locally using apt-get. Any clues?
> >
> >  A .deb package is an ar archive with a tarball of package files, a
> >  tarball of control files, and a brief version file.  You can use "ar x
> >  package.deb" which will give you these three, and then unpack the
> >  package files and try to use them.  But you don't get the
> >  system-integrating scripts etc.
> >
> >  As far as I know there's no good way to unpack and use a .deb in one's
> >  home directory.
> >
> >  Do you have sufficient access to make a chroot environment?
> 
> Yes I do. I haven't tried it though. It may be the correct way to
> create a sandbox for testing
> packaging. I'll try it out and let you know. Thanks!

Just FYI: if you want to make a chroot, debootstrap is your friend...

-Adam
-- 
GPG fingerprint: D54D 1AEE B11C CE9B A02B  C5DD 526F 01E8 564E E4B6

Engineering consulting with open source tools
http://www.opennovation.com/