Maybe we can consider this in 4.0 as well? (Forming relative urls
throughout the webapp with absolute url usage in feeds where required).
--a.
----- Original Message -----
From: "rob" <[EMAIL PROTECTED]>
To: <roller-dev@incubator.apache.org>
Sent: Friday, February 02, 2007 7:26 AM
Subject: Re: absolute vs. relative urls in roller
Allen Gilliland wrote:
you are right that it's a bit confusing at times and it would be nice to
not confuse things by offering lots of relative vs. absolute options. The
reason absolute urls are needed is because so much of blog content goes
into feeds which are not read back on the same host as the site, so
absolute urls are necessary to ensure everything is properly referenced.
If you have relative urls to anything in a feed that resource will not
get properly located.
k, now I understand why it exists, although is this still true for atom
feeds? We are only going to be supporting atom feeds and while I know this
is a problem for RSS I think atom has addressed it.
As for your problem, I don't see how absolute urls are causing problems.
What kind of deployment configuration do you have such that setting the
absolute site url doesn't ensure that everything works properly?
If we set the site url then everything does work properly. However, the
deployment where we were having problems has an external (internet) url
and an internal (intranet) url. Folks wanted to test the deployment
hitting only the internal url before bringing up the external proxies.
They would actually switch back and forth depending on whether the proxies
were up and constantly changing the setting was getting to be a pain.
Roller usage at IBM forms one part of a suite and the other members of the
suite weren't having this problem as they used relative urls throughout.
Rob
