[Grok-dev] Re: collective.namedfile + blob

[Martijn Faassen]

On Thu, Feb 28, 2008 at 10:49 PM, Dirceu Pereira Tiegs
<dirceutiegs at gmail.com> wrote:
> On Thu, Feb 28, 2008 at 5:59 PM, Dirceu Pereira Tiegs
>  <dirceutiegs at gmail.com> wrote:
>  >  BTW, after adding z3c.blobfile support to collective.namedfile, I'll
>  >  need to test it with Plone 3 too, because I have to see if it have any
>  >  incompatibilities of ZODB versions, etc.
>
>  It seems that adding z3c.blobfile to collective.namedfile dependecies
>  will cause the install of the newest ZODB 3.*; at least the
>  plone.app.blob [1] documentation says that, if you want blob support
>  on plone, ZODB 3.8 will be installed. I don't know if it's a good idea
>  to do this.

Hm, are you sure? In combination with Grok this shouldn't happen, as we pin the
ZODB down with versions in buildout. The only thing that z3c.blobfile
requires according
to its setup.py is ZODB3, so that isn't a conflict.

We just want to make sure that this works in the Zope 2 context that
collective.namedfile
is mostly used in. I imagine Zope 2 uses the non-egg ZODB, and pulling
in an egg would get confusing...

>  1 - http://plone.org/products/plone.app.blob
>
>  Should I implement the NamedBlobFile and NamedBlobImage fields and
>  widgets on megrok.form? It will take just a few lines of python code
>  and some ZCML. Martijn, what do you think?

If this is the easiest to make progress, go ahead. I think what we
should at least try is make
the widget in namedfile more flexible so you can specify the factory
by subclassing, instead of
having it always create a NamedFile object. The amount of code in
megrok.form can then stay very minimal.

By the way, I think this communication should at very least go back to
grok-dev, so I'll cc-ed this in here.

Regards,

Martijn