At Mon, 10 Apr 2006 12:29:52 +0000 (GMT), Thomas Chust wrote: > > On Mon, 10 Apr 2006, DaishiKato wrote: > > > [...] Thanks. This looks good. I would implement it in > > http:send-request, though. > > Hello, > > I just decided against patching the http-client internals themselves for > the "proof of concept" example. Writing a new version of http:GET was less > invasive.
Alright. So, here is my patch. Three notes: - I wonder if (ensure-local-connections) is a good way; since threads are chaper than connections, shouldn't we reuse connection among threads? - My rough test shows some delay with keep-alive connections; somehow (read-string) is slow with the length specified. What's the problem? - In order to support persistent connection, the support of Transfer-Encoding: chunked would be required; but I didn't come up with a nice implementation idea for http:GET. Any comments? one more; it is a client responsibility to use http:send-request, and input port has to be read properly accrding to Content-Length or Transfer-Encoding; otherwise, the upcoming request that reuses the connection would meet a trouble. Thanks, Daishi
http_egg.patch
Description: Binary data
_______________________________________________ Chicken-users mailing list [email protected] http://lists.nongnu.org/mailman/listinfo/chicken-users
