In my opinion, all templating should be add-ons in BFG. Since we're not
constrained to use any particular templating in BFG, we should figure out
what we want to use and suck it in as a dependency of our app, rather than
BFG. I do think this basic story is a little strange: "You can use widget X
with our framework, but you don't have to. Whether you use it or not,
though, we'll suck it in as a dependency."
My 2 cents,
On Sun, Apr 26, 2009 at 11:42 PM, Chris McDonough <chr...@plope.com> wrote:
> Malthe Borch just made a release of chameleon.core and chameleon.zpt that
> removes any dependency on lxml. Woo hoo, thanks Malthe! The benefits of
> it should be possible to install BFG on all systems much more easily and it
> should be reasonable to use BFG on Windows and GAE.
> However, to move to this version of Chameleon in BFG, we have to do some
> and make some decisions.
> chameleon.genshi still depends on lxml. Currently BFG has c.genshi
> support, but
> when we move to the new chameleon version, we'll need to package
> chameleon.genshi support outside bfg itself in an add-on package (which
> still depend on lxml). Or we make chameleon.genshi not depend on lxml.
> The xlst support in BFG still depends on lxml. When we move to the new
> chameleon version, we'll need to package repoze.bfg.xlst support outside
> itself in an add-on package (which will still depend on lxml).
> The non-lxml versions of chameleon.core rely on the "etree" module, which
> only present in Python 2.5+. We'll either need to make 2.5 the lowest
> BFG version or we'll need to change chameleon.core to fall back to an
> elementtree module for the same functionality.
> Any opinions welcome.
> - C
> Repoze-dev mailing list
Repoze-dev mailing list