Hi Wenbo, This looks like a reasonable request. Since we are in agreement on the spec, I have logged https://bugs.openjdk.java.net/browse/JDK-8218554 to track it. The issue has a link to this mail thread for reference.
best regards, -- daniel On 02/02/2019 00:27, Wenbo Zhu wrote:
On Fri, Feb 1, 2019 at 2:58 PM Daniel Fuchs <daniel.fu...@oracle.com <mailto:daniel.fu...@oracle.com>> wrote:Hi Wembo, On 01/02/19 19:50, Wenbo Zhu wrote: > 1) clarify in the API javadoc that chunked encoding is always applied > even with Connection: close Chunked encoding is always applied if 0 is passed to sendResponseHeaders (this is a bit counter-intuitive, but 0 means chunked coding and -1 means Content-Length: 0 or no content - depending on response code, and n > 0 means Content-Length: n).Yeah, it could be better ... but I can live with this, 0 meaning "content-length unknown when response headers are generated".Whether chunked coding is applied or not bears no relationship with Connection: close whatsoever. On the client side - if we have 200 with no Content-Length header and no Transfer-Encoding then we don't expect the content to be chunked. We simply drain the bytes until EOF is reached. Is this what you mean by always applied? Yes. best regards, -- daniel
