Storing flushes in the refs vector doesn't let us recognize when the
remote is empty, as we'd always introduce at least one element into
it. These flushes aren't necessary, so we can simply ignore them.

This tells the remote to send us any tags that point to objects that
we are downloading.

This lets us notify the user of what the remote end is doing while we
wait for it to start sending us the packfile.

Instad of each transport having its own function and logic to get to
its refs, store them directly in transport.

Leverage the new gitno_buffer to make the parsing and storing of the
refs use common code and get rid of the git_protocol struct.

This allows us to add capabilitites to both at the same time, keeps
them in sync and removes a lot of code.

gitno_buffer now uses a callback to fill its buffer, allowing us to
use the same interface for git and http (which uses callbacks).

Recognize and report server-side error messages

When e.g. a repository isn't found, the server sends an error saying
so. Put that error message in our error buffer.