On 11/3/06, James M Snell <[EMAIL PROTECTED]> wrote:
Calling release closes the underlying http connection as well as the InputStream. Because of the incremental parsing model, we can only call release when we're done parsing the document. Release also has the side effect of causing the http client to lose all of the headers returned with the response.
Perhaps it makes sense to have a "Just pull it all down and parse it" mode that reads all the data from the request and parses it fully, copying the headers into our own data structure, etc. Then there would be an alternative interface that allows you to avoid pulling all of the data into memory. -garrett
