On Thu, 2005-09-01 at 16:09 +0200, Oleg Kalnichevski wrote: > On Wed, Aug 31, 2005 at 08:22:17PM -0400, John D'Ausilio wrote: > > So .. my company makes a portal for network management apps. One of the > > vendors we support has a product which implements a broken webserver .. > > it stuffs an extra CRLF after the header (so instead of 2, there's 3). > > THis is clearly in violation of HTTP1.1, but getting the vendor to fix > > his product is virtually impossible. > > > > While analyzing the problem, it occurs to us that it might be sensible, > > in the case of a non-persistent connection (and with "strict" turned > > off) to arrange httpclient to read to end-of-stream instead of > > content-length (which would fix our problem in a less hacky way then the > > other alternatives). > > > > If this doesn't cause too much heartache for the dev team I'd be happy > > to submit a patch .. if someone has a real issue I'd like to hear why. > > > > John, > > There are hundreds if not thousands of CGI scripts out there that > produce broken HTTP responses. All of them are broken in all sorts of > wonderful ways. We cannot be possibly be expected to provide workarounds > for all of them in the stock version of HttpClient. Consider extending > relevant HttpMethod classes and overriding the #readResponseBody method > in order to implement an application specific workaround for the problem > > Oleg >
Oleg, Point (and advice) taken .. thanks for your time. John D --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
