Previously when we do pause because of out of buffer, we just throw
away unread data in connection buffer.  This just broke protocol
framing, and I saw occasional FRAME_SIZE_ERROR.  This commit fix this
issue by remembering how much data read, and in the next iteration, we
process remaining data.

This commit fixes the bug that streams get stuck if stream gets some
DATA, and stream->closed becomes true at the same time.  Previously,
in this condition, after we processed DATA, we are going to try to
read data from underlying transport, but there is no data, and gets
EAGAIN.  There was no code path to evaludate stream->closed.

... as it was only used from there.

With the "drained" functionality we can get here slightly asynchronously
so the stream have have been closed but there is pending data left to
read.

... as it does for pipelining when we're multiplexing, as we need the
different buffers to store incoming data correctly for all streams.

No need to wait for our "spot" like for pipelining

... which is necessary since the socket won't be readable but there is
data waiting in the buffer.

... from the connection struct. The stream one being the 'struct HTTP'
which is kept in the SessionHandle struct (easy handle).

lookup streams for incoming frames in the stream hash, hashing is based
on the stream id and we get the SessionHandle for the incoming stream
that way.

Once we know we are HTTP/2 enabled we know the server can multiplex.

... and do not blacklist any.

All the details mentioned here are better documented in man pages

This file was removed in commit fd137786

Previously we counted all connections to a specific host name and that
would be used for the CURLMOPT_MAX_HOST_CONNECTIONS check for example,
while servers on different port numbers are normally considered
different "origins" on the web and should thus be considered different
hosts.

All the existing Curl_bundle* functions were only ever used from within
the conncache.c file, so I moved them over and made them static (and
removed the Curl_ prefix).

This avoids unnecessary dynamic allocs and as this also removed the last
users of *hash_alloc() and *hash_destroy(), those two functions are now
removed.

avoids extra dynamic allocation

... by using plain structs instead of pointers for the connection cache,
we can avoid several dynamic allocations that weren't necessary.

This ensures an alternate address is not used.
Does not apply to proxy tunnel.