[merge] MVC progress bars, and indication of network traffic

[John Arbash Meinel]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1


...

>     martin>          :param relpath: The relative path to the file
>     martin>          """
>     martin> -        return self.get(relpath).read()
>     martin> +        f = self.get(relpath)
>     martin> +        try:
>     martin> +            return f.read()
>     martin> +        finally:
>     martin> +            f.close()
> 
> 
> EErk ! I'm pretty sure you just closed the http socket here (the
> http transport doesn't redefine get_bytes) ! If no test is
> failing there, we need one. On the other hand, I'm not sure you
> need that part in your patch.
> 

Unless something has changed, HTTPTransport.get() does:

code, response_file = self._get(relpath, None)
# FIXME: some callers want an iterable... One step forward, three steps
# backwards :-/ And not only an iterable, but an iterable that can be
# seeked backwards, so we will never be able to do that.  One such
# known client is bzrlib.bundle.serializer.v4.get_bundle_reader. At the
# time of this writing it's even the only known client -- vila20071203
  return StringIO(response_file.read())


Mostly I'm pointing that we don't return the raw socket, we wrap it in a
StringIO.

Originally we needed that because "gzip.py" does a seek() on its input,
but we changed that in "tuned_gzip.py"

So it is possible you might change it at some point in the future,
though I think not dealing in a raw socket may work out best for
everyone. (For example, making sure all the bytes are consumed before
the next request is sent.)

John
=:->
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.9 (Cygwin)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org

iEYEARECAAYFAklJE6wACgkQJdeBCYSNAAP6/ACdEVsgkGdDrRRvnVqJ55ImLC8O
/BkAn1xtbkoaAOoa5OFqQBDTYwxi2x0l
=xdr6
-----END PGP SIGNATURE-----