Part of twisted.web2.channel.http View Source View In Hierarchy
|Method||rawDataReceived||Handle incoming content.|
|Method||headerReceived||Store this header away. Check for too much header data (> channel.maxHeaderLength) and abort the connection if so.|
|Method||splitConnectionHeaders||Split off connection control headers from normal headers.|
Split off connection control headers from normal headers. The normal headers are then passed on to user-level code, while the connection headers are stashed in .connHeaders and used for things like request/response framing. This corresponds roughly with the HTTP RFC's description of 'hop-by-hop' vs 'end-to-end' headers in RFC2616 S13.5.1, with the following exceptions: * proxy-authenticate and proxy-authorization are not treated as connection headers. * content-length is, as it is intimiately related with low-level HTTP parsing, and is made available to user-level code via the stream length, rather than a header value. (except for HEAD responses, in which case it is NOT used by low-level HTTP parsing, and IS kept in the normal headers.