On 2/24/14, Saint Germain <[email protected]> wrote:
> Hello,
>
Hi !
:)
[...]
>>
>> What features might the integration provide?
>>
>
> Potentially a lot of things:
[...]
> - database (use of Django ORM)
> - database migration (South)
[...]
a few comments about this based on my personal opinions (rather than
absolute truth) as well :
- Database migration tools should be merged (eventually if not already)
into django core
* see [1]_
- we better look at SQLAlchemy as ORM
* ... and DB migrations powered by Alembic
[...]
> However it seems that it has the same benefits of Python: it has a huge
> repo of available apps which cover a lot of needs.
sure !
> The problems with Django are pretty well advertised: it is an
> integrated framework with its own ORM so it may not be
> flexible/powerful enough for some needs.
>
... but ...
> So basically what I have in mind is: the biggest problem to install
> Bloodhound is not Bloodhound itself, it's the whole infrastructure
> behind (have a secured server, apache/nginx, uwsgi/gunicorn,
> PostgreSQL/MySQL, etc...).
> If you have a running Django website, you already have all this. So if
> the Django admins can just install Bloodhound as an app, we may see
> adoption increased quickly.
>
> In addition, as you said, it may be possible for Bloodhound to rely on
> Django components (templating system,
please don't do this by default ... Django templates have a lot of bad
issues , IMO we better keep it going using Genshi .
disclaimer : I build myself Django web sites powered by django-genshi
rather than built-in template system .
> database,
this may be interesting , yes , though any other ORM would be fine ,
especially one not tied to a particular framework .
> internationalization,
afaict, babel is great (I see your point though ;)
> URL management,
in this field Routes framework is the killer , and it is possible to
integrate it with BH by using the web bootstrap handlers deployed in
blood-hound.net (e.g. sub-domain deployments) scheduled to be released
shortly after BH=0.8
(I see your point though ... and the only thing neede is a custom
bootstrap handler ;)
> etc.). However I don't know yet how goods are the
> current Bloodhound/Trac internal components compared to Django.
>
They both are good . It'd be nice to make them interoperate (e.g. Trac
authenticators implementing Django auth handlers ;)
[...]
.. [1]
https://www.djangoproject.com/weblog/2013/mar/22/kickstarting-schema-migrations-django/
.. [2]
https://www.kickstarter.com/projects/andrewgodwin/schema-migrations-for-django
--
Regards,
Olemis - @olemislc
