[development] Aggregator to nodes

Boris Mann boris at bryght.com
Thu Oct 5 05:27:19 UTC 2006


On 10/4/06, Rob Barreca <rob at electronicinsight.com> wrote:
> Khalid B wrote:
> > - aggregator2 had a lot of momentum earlier this year, and is 4.7
> > compatible
> > but not branched for 4.7 so far.
> >
> I've used aggregator2 quite a bit in the past with great success in 4.7.
> Recently, I've had a couple clients try to use agg2 and say that it was
> incredible buggy. Following the issue queue, the author seems to have
> gone AWOL and the module is in disarray. It's got a lot of potential and
> with some direction and love I think it could be very worthwhile again.
>
> There were talks of moving all RSS parsing code to an API module which
> makes a lot of sense. I think for agg2 to be successful again, the RSS
> code should be externalized and a few of the major bugs would need to be
> squashed.

There are a bunch of us that have been looking at this recently.

Please see https://svn.bryght.com/dev/wiki/DrupalFeedParsing for a summary.

budda's feedparser has the cleanest code and looks like a good path
forward. However, I would like to see the "feed enabling" approach
that David Ayre took under our direction with the "feed_node" module
-- see https://svn.bryght.com/dev/browser/feed_node

The other big thing is the ability to take different actions based on
the namespace. For example, examining the feed, seeing it has GeoRSS
extensions, and location-enabling nodes that are based on it.

So, the idea behind feed node is that there are many things in core
aggregator that can be improved. In working with it, we'll also be
contributing core patches back. We may end up maintaining a backported
/ patched aggregator for 4.7 (...5.0) as we continue to try and
improve core as well.

Two of the obvious things that comes to mind for core patches are:
* paging the admin/aggregator and aggregator/sources pages
* making block creation per feed optional, similar to how the quotes
module does it

agg2 was never an elegant bit of code. The leech / leech_news
successor has even more convoluted code. My vote is for helping budda
with feedparser, at this point.

-- 
Boris Mann
Vancouver 778-896-2747
San Francisco 415-367-3595
Skype borismann
http://www.bryght.com


More information about the development mailing list