> I can't see that being viable. What would be the benefit of caching only part of the content received by ATS from the origin server? I understand the benefit of caching part of an object because that is all the origin server sent but dropping part of data available seems a bit odd.
The purpose of my previous questions was to find out if it is possible to implement caching of range requests as separate objects from a plugin (with minimal changes to the core). As I understand from you currently for any requested range we *must* cache the whole resource referenced by the range. In order to allow caching of ranges from a plugin there should be a way for the plugin to tell this to the core and also for a plugin to combine several cache objects to serve a request. As I understand from you at least the last part is not so straight forward to implement with the current plugin API. Does it make some sense now?
