Hi to all beta testers! Last two weeks I was dealing with updating feeds. There are two major problems: a) updating is slow b) updating is not reliabe For testing I have used Hal's OPML file (over 60 feeds) - very good for testing. The current implementation of download/update is done via WinInet API. It is very easy to use, but very slow. So I tried to replace this API with my own. I succeeded. The result is: - 4x times quicker update - pRSSreader is about 3KB larger But there are some problems. If I used Hal's test file, about 10 feeds was updated sucessfully, the rest was not. Not good. So I tried to figure out where is the problem. I tried similar implemetation with WinInet then with MFC wrapper of WinInet. The former one had similar result as my own implemetation, the latter one was a bit more successful. Every updates were run with cradled PDA and connection via ActiveSync on desktop PC. On monday I tried my own implementation on WiFi connection and it runs perfectly - fast and reliable. I have made test with Egress, if it is able to update Hal's OPML file and the result was pretty same as with pRSSreader. Thus, there must be some error in ActiveSync which disallows a lot of subsequent connections, because if I updated site by site manually (not via Update All), the reliability grows up (nearly 100%). Now, I have to make decision which implementation should I use and I would like to know your opinions. Here are pros (+) and cons (-): WinInet: + quite reliable with ActiveSync + smaller pRSSreader's binary + debugged - slow - hard to extent in future (SSL, digest authentication), but probably possible My implementation: + fast + easy to extent + quite reliable with ActiveSync + full control on what is happening - bigger pRSSreader's binary - not full implementation of HTTP - will need lot of tests to work properly (will need your help) -- David -- pRSSreader development mailing list prssr-devel-ml@xxxxxxxxxxxxx //www.freelists.org/archives/prssr-devel-ml