Correct. As far as I'm concerned, adding the additional header with the uncompressed content-length when content-length would otherwise have been added makes this interop just fine: In cases where the uncompressed content-length is unknowable, nothing changes in terms of interop. In cases where the uncompressed content-length is knowable, nothing changes in terms of interop. -=R On Fri, Mar 14, 2014 at 3:33 PM, Julian Reschke <julian.reschke@gmx.de>wrote: > On 2014-03-14 23:04, William Chan (陈智昌) wrote: > >> It's unfortunate, but I agree that given the interop issues, I don't >> think we can require this. >> > > HTTP/1.1 already REQUIRES the ability to process chunked encoding. What > we're talking about is adding gzip. Streaming/buffering shouldn't be any > more issue than it already is in 1.1. > > > ... > > Best regards, Julian >
Received on Friday, 14 March 2014 22:45:53 UTC