On Wed, Dec 08, 2010 at 06:59:52PM +0100, Mathieu Bridon wrote:
Hi Luke,
On Tue, 2010-12-07 at 15:30 -0500, Luke Macken wrote:
> Hey all,
>
> I just threw all of my bodhi v2.0 notes on to the wiki:
>
>
https://fedoraproject.org/wiki/Bodhi/2.0
>
> Please feel free to chime in with your questions, comments, ideas,
> criticism, etc.
I have a question about the chosen template engine.
The page says that Bodhi2 will use Mako. Is there any reason for this
choice? (except for the obvious fact that it's actually fast and well
maintained upstream, which Genshi isn't)
If that's the only reason, did you consider Kajiki? [1]
I see real benefits in using an XML-based template engine, and using
Kajiki would allow to mostly reuse the Bodhi1 Genshi templates
(migration seems to be trivial [2]) while having performance roughly
similar to that of Mako.
Note that I don't want to bikeshed indefinitely on this, I was just
wondering what the reason for choosing Mako was.
And given the huge list of stuff to be implemented for Bodhi2, it seems
like everything that can be reused from Bodhi1 would be a nice help in
getting there.
I am definitely still open to considering different template engines for
Bodhi v2.0. I'm friends with Rick, the creator of Kajiki, and TG2
recently added support for it. I had to patch it out of our package,
since Kajiki isn't packaged yet for Fedora. I am definitely open to
considering it, or Jinja2 for that matter (which is apparently faster
than Mako in a variety of ways (there is also chamelon.genshi, which is
supposed to be pretty fast as well)). I have yet to see any Kajiki
benchmarks, so I'm not sure if there is a huge win with it yet, although
your point about re-using Bodhi1's template is good (even though we
would still need to port the Kid-isms to Genshi/Kajiki).
Personally, the only thing I care about with a templating engine is
speed. Strict XML compliance or not, I just want it to be fast :)
Other than that, I only see awesome ideas for the future of Bodhi,
both
from a Fedora maintainer and a downstream distributor point of view.
Excellent, I'm glad to hear it. Please feel free to add any ideas that
you have as well.
Is the page supposed to contain only the high-level goals of Bodhi 2
or
can we start polluting it with ideas for the implementation details?
That page started as just a braindump. I think for each feature, we
should create a ticket in bodhi's trac, reference it on the wiki, and
discuss the implementation details/ideas in the ticket.
luke