Re: Fwd: Re: need help updating imgsizer for python2.1

2002-01-16 Thread Peter S Galbraith
Sorry for taking so long to reply.

dman <[EMAIL PROTECTED]> wrote:
 
> | ] The package currently depends on python1.5.  What should it depend on
> | ] when fixed?
> | 
> | python
> 
> This must be versioned though.  
> python >= (2.1) , python << (2.2)
> will get you the current default version.

Considering that the package would then depend on python 2.1, I can
simply apply the following patch:

1c1
< #!/usr/bin/python1.5
---
> #!/usr/bin/python
60c60
< import sys, os, getopt, cmp, string, re, urllib, commands
---
> import sys, os, getopt, filecmp, string, re, urllib, commands
258c258
< if cmp.cmp(file, tempfile):
---
> if filecmp.cmp(file, tempfile):

Right?
This works for me.  No real need to write the script so it work on
either 1.5 or 2.1 if the package depends on 2.1.  I haven't missed
something obvious, have I?

So for the trivial questions.  Please CC me in replies. Although I will
look at the mailing list archives for replies, I am not subscribed to
the list.

Peter




Re: Bug#128957: *.py[co] files are architecture independent

2002-01-16 Thread Matthias Klose
Matt Zimmerman writes:
> On Sun, Jan 13, 2002 at 10:00:23PM +0100, Matthias Klose wrote:
> 
> > - wajig uses /usr/bin/python as interpreter and therefore should
> >   depend on "python (>= 2.1), python (<< 2.2)". Same for the build
> >   dependency.
> 
> Why is python (<< 2.2) necessary?  apt-listchanges, for example, was
> developed for and tested with both python2.1 and python2.2, so I used a
> dependency of python (>= 2.1).

looking at the package, it's probably not necessary. But why should it
run with any python version? The package is mostly tested with
Debian's default python version, which is (and will be) 2.1 for woody.

apt-listchanges is a special case, because it alway runs as root and
does not rely on third party python packages which are not avilable
for some python version. Assume, that you run apt-listchanges with
1.5, then the .py file get recompiled for python-1.5 (if you do NOT
run as root) every time, assuming they were compiled with 2.x before.
There is currently no way to switch off python's side effect of trying
to write compiled files.

For apt-listchanges this doesn't look very harmful, but consider the
current packaging of the mailman package (where I disagree with the
mailman maintainer ;-):

- mailman does not depend on a particular python version (because it
  runs with versions from 1.5 and up).

- assume, you install mailman, having python-1.5 as your default
  interpreter. The python files get compiled for 1.5.

- you install python-2.1.

- mailman is a cgi application, running with www-data.www-data, so
  now every time you call the cgi, thousends of lines are recompiled,
  but not saved.

- so instead of choosing one python version, the mailman maintainer
  gives the user the "freedom" to choose one python version, and
  charges him with excessive runtime.

- In an ideal world, we would have a recompilation of python files for
  every application an an own python interpreter (alternative) for each
  application. But we currently don't have ...

In the case of apt-listchanges you should take care of removing any
compiled *.py files in the postrm.

Hope this helps ...

Matthias




J.T. Sterlings Daily Special - January 16, 2002

2002-01-16 Thread listadmin
Title: Welcome to J.T. Sterlings








  
J.T. Sterlings Daily Specials - January 16, 2002Our Daily Specials change once every day at Midnight, Eastern Time.

You are subscribed to J.T. Sterlings Daily Special mailings.


Visit us at www.jtsterlings.com or click below to place your order.



  



  
  Item
21611
Jewelry box with a roomy pull-out drawer as well as storage room at the top! Side handles. 9 3/4" x 6 1/4" x 6" high. $59.95 Regular Price.
Regular Price $59.95
Sale Price $43.76
Today's Special Price $39.38
You save 10%

Click
Here To Order!
 


  

  



  
  Item
21704
Faux marcasite earrings. Simulated onyx on a variety of six classic shapes. Hypo-allergenic posts. $3.95 Pair Regular Price.
Regular Price $29.95
Sale Price $21.86
Today's Special Price $19.67
You save 10%

Click
Here To Order!
 


  

  



  
  Item
21729
Chinese lions of mythology protect the home and family from evil spirits. 5" high. Simulated wood base. $24.95 Pair Regular Price.
Regular Price $24.95
Sale Price $18.21
Today's Special Price $16.39
You save 10%

Click
Here To Order!
 


  

  



  
  Item
21730
Creatures of the deep adorn this most unusual sculpture... it's a 3-dimensional collectors' item. 8 1/2" high. $24.95 Regular Price.
Regular Price $24.95
Sale Price $18.21
Today's Special Price $16.39
You save 10%

Click
Here To Order!
 


  

  



  





You are receiving this special offer because you have provided permission to receive third party email communications regarding special online promotions or offers.  Any third-party offers contained in this email are the sole responsibility of the offer originator. 
Copyright © 2001 J.T. Sterlings - 5700 Memorial Highway Suite 206, Tampa FL 33615. All Rights Reserved.
J.T. Sterlings does not condone the use of unsolicited email (spam).  If you do not wish to receive any further messages from J.T. Sterlings, please follow the instructions below to unsubscribe.  


---
You are currently subscribed to dailyspecial as: debian-python@lists.debian.org
To unsubscribe send a blank email to:
 [EMAIL PROTECTED]







R&D framework funding

2002-01-16 Thread A pocketfull of euros



See www.apocketfullofeuros.com for an 
unofficial guide to EU R&D framework 
funding


Re: math.exp bug on alpha?

2002-01-16 Thread Gregor Hoffleit
* Bastian Kleineidam <[EMAIL PROTECTED]> [020111 11:41]:
> On Thu, Jan 10, 2002 at 09:28:30PM +0300, Mikhail Sobolev wrote:
> > On Thu, Jan 10, 2002 at 06:29:58PM +0100, Bastian Kleineidam wrote:
> > > On Thu, Jan 10, 2002 at 05:25:43PM +, Luigi Ballabio wrote:
> > > > on alpha debian, python 2.1.1-8:
> > > > 
> > > Works for me, no aborts. (i386, python 2.1.1-8)
> > Just compare the platform... :)
> I did. So its a platform specific bug.

Cf. http://bugs.debian.org/106421.

Gregor