On Tue, 6 Dec 2011 20:56:11 +0200, karj wrote:
Hi,
I have come across a problem with a flash application that's using
cache
killer.
For example all the request URL's ends with
liveofday.xml_nocache=1322784085349
liveofday.xml_nocache=1322784100431
Sigh. This kind of URL is the direct end product of admin configuring
"private", and "no-store" objects to be force-cached all over the
network, along with ignoring "no-cache" and expiry headers. Corrupting
the information itself in the name of speed or cost.
This should be a huge sign that the website authors *do* know a bit
about caching and that people absolutely should not cache this content,
yes?
/end rant
Before you go any further run those URLs through the tool at redbot.org
and see how knowledgeable the authors actually are about caching.
NP: a file called "liveofday" implies that the content is constantly
changing. Particularly on an XML data feed, caching it is probably NOT a
good idea.
Is there a way for squid to cache all the xml files ignoring the
nocache
numerical value.
The fact that you call them files (with an 's') indicates that you know
full well that they are different content. Why pretend that one file is
exactly binary-equal to all the others that will *ever* exist at that
URL?
I don't want to use an external program such us a perl re-writer.
No there is no way to do it (yet) without the external helper program.
In fact, no way to do it without the Squid-2.7 storeurl_rewrite
feature.
Amos
