Because http/1.1 support, default "auto" behaviour of -m option is not
desirable, since it is used as HTTP pipelining, and it is not used in
practice.

fix typos in documents.

We modeled max_header_fields and header_field_buffer limit from Apache
configuration directives.  In Apache, they are only applied to request
header fields, while we applied both request and response.  Since
nghttpx is used as reverse proxy and backend server is relatively
"trusted", this commit removes the application to response header
fields.

Return NGHTTP2_ERR_INVALID_ARGUMENT from nghttp2_submit_headers() if
given stream ID and pri_spec->stream_id are the same (thus trying to
depend on itself).

Also return NGHTTP2_ERR_INVALID_ARGUMENT from nghttp2_submit_request()
and nghttp2_submit_headers() with stream_id == 1, when new stream ID
equals to pri_spec->stream_id.

Previously, these cases are not checked, and just sent to peer.

The naming convention in asio lib does not use get_something.

header value should not be inp_strlower

http header keys are case-insensitive, but header values are case-sensitive, so it should not be changed.

With the presence of idle stream related API (e.g.,
nghttp2_create_idle_stream()), it is more predictable for client to
create idle streams with its dependency to another idle stream.
Previously, we didn't create complete parent idle stream in this case.
Now we create idle streams as we do on server side.

This commit includes backward incompatible change, since we change
private field in public API class.

Previously we scheduled the transmission of response HEADERS using
priority tree in the belief that it allows more better utilization of
bandwidth for prioritized streams.  But to reduce the overhead of
reconstruction of priority queue when connection level flow control
window is depleted, we just don't check priority tree in this case.
This means that response HEADERS frames are not sent even though they
are not flow controlled.  This could waste bandwidth.  To improve this
situation, we stop scheduling response HEADERS with priority tree for
now.  Now they are just sent in the order they submitted.  The
response body DATA continued to be scheduled with priority tree as
before.

return session_inflate_handle_invalid_stream(...) case is for streams
for INITIAL state, but this is rare case.  In general, we'd like to
reduce RST_STREAM transmission, and it is suffice to ignore this frame
for now.