Thanks for your response. I checked the beta 1 code and it appears that the core approach to reading a char/byte at a time from the InputStream is still there.
It looks like there have been tons of structural changes, so perhaps things will be better. I had been wondering why BufferedInputStreams weren't being used to improve
performance (assuming that the cpu consumption I discovered would be improved by buffering). Perhaps there are important design considerations that mandate
InputStreams from the Commons point of view. But I should probably post these questions to the Commons Httpclient list.
So you're able to use the latest webdavclient lib from CVS with HttpClient beta1/2 with no modifications? Or are the modifications to get these to work together in your submitted patch?
I'm confused about Webdavresource.java having an import of:
import org.apache.util.HttpURL;
and the presence of a org.apache.commons.httpclient.HttpURL in the beta 1 distribution.
Thanks, Tony On Wednesday, July 2, 2003, at 12:28 PM, Eric Johnson wrote:
Tony,
I've been working with the latest CVS snapshot of the webdavclient in conjunction with the latest HttpClient with great success.
I submitted a patch last week that would make it work even better, but based on the slow response, committers on the list do not appear to be paying much attention. Hopefully, your interest will prod them to pay more attention!
I've been actively contributing to both the Slide and HttpClient lists to improve the behavior of both together, and there seem to be others out there doing the same thing.
I'm not surprised that you're seeing significant performance issues with the older HttpClient, and I'm afraid that your best bet will be upgrading.
