Another not-very-appealing possibility is to iterate over the grammar pool before and after parsing the document or root schema and look at the change. This assumes the grammar pool is not used by multiple threads.
Jeff ----- Original Message ----- From: "Bob Foster" <[EMAIL PROTECTED]> To: <[EMAIL PROTECTED]> Sent: Monday, December 06, 2004 12:12 PM Subject: Re: Problem with grammar pool > Yes, you could do that. Then if you had seen a sequence of schemas and > later one of these changed you could just throw away everying in the > sequence preceding the changed schema. Assuming that which schema > changes is somewhat random, you will, on average, throw away half of the > cache related to an instance document for every change. Yes, this is > better than throwing away the entire cache on each change (and the cache > may hold unrelated sequences of schemas), but not very satisfactory. --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
