#3189 portald silly use of RSS bites us

[Jamie McCarthy]

portald has always built up sectional slashboxes by fetching and
processing the RSS feed that open_backend.pl writes out. It's a
silly way to generate that list of sectional stories:

* it adds an additional point of failure (if the webheads are down
or briefly unreachable, the attempt fails)

* it adds delay (open_backend only runs periodically, so portald
isn't seeing the freshest data)

* it adds hits in the accesslog (a site with a lot of sections or strict
limitations can potentially ban its own portald)

Now there's a new problem:

* we're getting &from=rss appended to all the links in the sectional
slashboxes.

So the from=rss traffic stats for the last week aren't going to be
all that useful...

Can you rewrite the way portald does sectional slashboxes by
having it make the same call to the API that open_backend.pl does,
and parsing that data directly? Thanks.

[Chris Nandor]

Logged In: YES
user_id=3660

For now, I am just stripping the from=rss. The rest, there's no good
simple way to do it.

[Rob Malda]

Logged In: YES
user_id=288856

I just want to make sure this gets done by the next code refresh...

[Chris Nandor]

Logged In: YES
user_id=3660

The RSS stripping is done, was done for last refresh.

[Chris Nandor]

Logged In: YES
user_id=3660

This is largely solved by now using file:// URLs. Still would be nice to abstract out
this code.