5 ms·
Django 1.2 roadmap and schedule
- ubernostrum 17y ago(if anyone has questions, feel free to throw them at me here; I'll put on my release-manager hat and do my best to give useful answers)
- alrex021 17y ago> Multi-database support Its about time that it will officially, out of the box, support this important feature. :) [edit] direct link to feature list http://code.djangoproject.com/wiki/Version1.2Features http://code.djangoproject.com/wiki/Version1.2Features
- jacobian 17y agoI completely agree! If you want to help us make sure this happens, please check out the multidb branch, give it a spin, and report any bugs: http://code.djangoproject.com/svn/django/branches/soc2009/multidb/ http://code.djangoproject.com/svn/django/branches/soc2009/mu....
- silkodyssey 17y agoI like the idea of a smarter template if tag but after having looked at the snippets page it doesn't look quite like what I was hoping for. I was looking for something more the tornado framework's implementation. I prefer tornado's template system in general. It's much closer to actual python in terms of syntax and the ability to pass functions from a request handler to the template is sweet!
- simonw 17y agoOne of the stated goals of Django's template system is to be "safe" - to ensure that template developers can't cause too much damage. The Tornado model essentially allows arbitrary Python code to be evaluated, which isn't necessarily a bad thing but would break the contract Django has already established. I think the smarter {% if %} statement proposal is much closer to Django's philosophy.
- lecha 17y ago> Support for non-relational databases: Tokyo Cabinet/Tyrant, MongoDB, Cassandra, Google Bigtable (AppEngine, #10192), Amazon SimpleDB, CouchDB, etc Is the intent to create some-kind of common framework that will work with more than one DB? Or to somehow select/prioritize one particular DB and support that?
- jacobian 17y agoKeep in mind: these are proposed features -- as in, proposed by anyone with enough smarts to edit a wiki page -- not necessarily ones that'll make it into 1.2. That proposal, in particular, is a Big Honking Deal with basically no concrete proposal, specification, or working code. In other words, the chance that Django will grow non-relational database support by March is... not so good.