Ruediger Pluem wrote:
On 01/21/2007 04:09 PM, Scott MacVicar wrote:
We did use mod_expires but the Expires header was being passed on to the
client, mod_headers didn't appear to be able to unset this during tests.
This is true, but what is the problem with passing the Expires header to the
client? If you want to prevent the clients from requesting / revalidating the
resource
frequently you can simply set an expiration date far in the future (about a
year).
True this would work but requires configuration on each of our web
servers, I'd also need to find an equivalent for our non-Apache
machines. I guess I prefer having a centralised point in which to
control the caching behaviour.
The reason for our desire to cache is that a version number is used in
the query string and incremented when appropriate on the resource.
So from my limited understanding of your environment it seems to make sense to
set
a long expiration time on this resource as a new version of a response will have
a different query string.
Most browsers seem to ignore RFC2616 13.9 in regards to the query string
being present in the URI.
What do you mean by this? That browsers cache the response even if no Expires
header is present in the response?
Internet Explorer and Firefox both seem to cache content that have a
query string when there is no expires header, according to the RFC which
Apache is following there should be no caching.
Apologies for not getting the gist of the thread.
No need to apologize at all for contributing to this discussion.
Reagards,
Scott
