On Tue, 20 Sep 2005, Phil Bogle wrote:
* The x-always-cache directive overrides all other response headers that would otherwise cause Squid not to cache a response (e.g. Expires). Be careful with how you use this header since it could cause unexpected caching.
Doesn't the existing refresh_pattern overrides do the job as well? Or is there any specific reason you need this on a per-request basis rather than based on what is being requested?
* The x-cache-key directive overrides the actual URL used to fetch the content for purposes of determining cache hits. It is typically used to ignore session query string parameters that would otherwise prevent caching.
To be safe this logics needs to be moved into the proxy rather than the client (crawler in your case).
One way I see this done "proper" is similar to the redirector interface we have today, but instead of rewriting the URL to be requested only the cache key URL is rewritten.
Regards Henrik
