[Teammetrics-discuss] Web Framework

Andreas Tille andreas at an3as.eu
Thu May 24 20:30:16 UTC 2012

On Fri, May 25, 2012 at 12:36:01AM +0530, Vipin Nair wrote:
> Hi Sukhbir,
> >    Why do you want to use 1.4? By this I mean, what specific
> > functionality is missing from 1.3 that you want to use 1.4 _and_ that
> > you cannot do without?

Same question on my side.  Currently we have on blends.debian.net:

$ LC_ALL=C apt-cache policy python-django
  Installed: (none)
  Candidate: 1.3.1-2~bpo60+1
  Version table:
     1.3.1-2~bpo60+1 0
        501 http://backports.debian.org/debian-backports/ squeeze-backports/main i386 Packages
     1.2.3-3+squeeze2 0
        500 http://mirrors.gandi.net/debian/ squeeze/main i386 Packages
        500 http://security.debian.org/ squeeze/updates/main i386 Packages
> First of all, its more convenient when not to have to use a virtualenv/VM. :)

I do not see any need for a VM when we could install 1.3.1-2~bpo60+1 -
at least if this is sufficient.
> If backport is not preferred, I'll have to use Django 1.2 as 1.3 is
> also a backport in squeeze.

Backports are fine but currently there is no 1.4 backport.

> I have not used django enough to specifically pin point a particular
> feature. I have read the release notes for both 1.3 and 1.4 and here
> are some features I would love to have.
> >From 1.3 release notes[1]:
> 1) Logging: I use this heavily and would love to have it.
> 2) Uniitest2 module : I have not used this before but was planning to
> follow a TDD methodology.

So that's fine for me - just tell me if you want me to install it on
blends.debian.net (and what other modules might be needed).
> >From 1.4 release notes[2]:
> 1) Selenium: I haven't used this before but would love to have in
> browser tests as well, again a part of TDD.

I can not tell how important this feature for the GSoC project might
be.  If there is no strong reason I'd like to go with the features
of 1.3 and upgrade once 1.4 is backported.
> As I said earlier, even 1.3 is a backport. :)

Backports are no problem at all.  The UDD clone is running on a
backported postgresql-9.1 and this is a cruxial piece of software.

> But as Enrico pointed
> out in an another thread, the requirement for being acceptable on a
> debian.org host is having the framework in squeeze+backports and asked
> me to target 1.3 then.


> But in our last conversation he did say we
> could use 1.4 if the expected deployment time is around 2-3 months and
> if its not a backport by then, I should remind him. I also offered my
> help to make a 1.4 backport :) So I do think we are safe using it. And
> as you said, we should wait for Andreas' comments.

I'd suggest to go with the latest backport.  So if you really need 1.4
you should push the backporting work (and by doing so becoming involved
in learning packaging techniques). 

Kind regards



More information about the Teammetrics-discuss mailing list