Re: END_STREAM flag and trailing headers

On 10 May 2014 00:36, Julian Reschke <julian.reschke@gmx.de> wrote:
> The use case is pushing responses for safe requests that *do* have q request
> body, such as a directory update (PROPFIND).

Yes, this is a perfectly valid use case, but we didn't find enough
support for it.  That's all.

The counterarguments were, if I recall:

1. as requests increase in complexity, the odds that a server
correctly synthesizes a request that the client can subsequently use
diminishes
2. adding a request body represents a non-trivial increase in protocol
complexity; this would probably require an additional state in the
promised stream that would be used for sending the request entity

Other than that, there isn't much more to be said.  I'm sure that if
circumstances change significantly, we can reopen the issue.

Received on Monday, 12 May 2014 17:05:51 UTC