[theforum] peo proposal

Elfur Logadóttir elfur at elfur.is
Tue Nov 11 13:41:53 CST 2008


.:| >> I propose that we
.:| >> 1) shift it to Drupal, using a similar set of templates to (the
.:| new)
.:| >> weo, wherein adding new feeds is cake 2) actively seek new feeds
.:| >
.:| >
.:| > +1
.:| >
.:| > /elfur
.:| 
.:| -1  sysadmin-related objection
.:| 
.:| Elfur, more than anyone else I know, ought to be aware of just how
.:| much of a clusterfuck that can turn into. :)

Hehe, you mean that my project ain't different? [1]
I thought that the problems I had were because I was running cron every
10mins updating half of 150+ feeds every 30 minutes (the other half in the
span of 1-6hrs) plus hacking the feed module to publish further information
about the feed in the blocks I placed on the front page.

So, if evolt.org is planning that ... I say: don't :)

.:| Drupal's aggregator is slow, memory-heavy, uses an error-laden XML
.:| parser, and is prone to very nasty race conditions during its cron
.:| runs, which in turn can cause other timed functions like search
.:| indexing to fail.

True, this is the default aggregator module. I'm atm testing FeedAPI in D6
to see if that behaves better (am still building the base so test not really
active yet).

.:| As a happy medium I'd be open to putting it together and not actually
.:| linking it, to see if it brings anything down, before a public
.:| release.  (One hopes that this module has improved in Drupal 6
.:| compared to previous versions.)

Perhaps we should combine our tests then :)
The FeedAPI running certain amount of feeds with specific criteria vs. the
core aggregator module.


Cheers
Elfur 


[1] Here I was thinking that I was special.




More information about the theforum mailing list