Tweak timing display

- Remove elapsed time. Space is at a premium here, and this is somewhat
redundant with the rate figure. We should display complete timing information
somewhere in the detailed flow view.
- Tone down the colour. Reserve highlights for stuff that should really pop out
to the user.
- Make rate calculation more acurate. Include header sizes. Use response start
and end time, rather than request end and response end. This means that we show
actual transfer rates, not including DNS requests and so forth.

libmproxy/console/common.py

@@ -144,9 +144,7 @@ def raw_format_flow(f, focus, extended, padding):
         if f["resp_ctype"]:
             resp.append(fcol(f["resp_ctype"], rc))
         resp.append(fcol(f["resp_clen"], rc))
-
-        resp.append(fcol(f["resp_et"], "time"))
-        resp.append(fcol(f["resp_rate"], "highlight"))
+        resp.append(fcol(f["resp_rate"], rc))
 
     elif f["err_msg"]:
         resp.append(fcol(SYMBOL_RETURN, "error"))
@@ -190,15 +188,15 @@ def format_flow(f, focus, extended=False, hostheader=False, padding=2):
         else:
             contentdesc = "[no content]"
 
-        delta = f.response.timestamp_end - f.request.timestamp_start
-        rate = utils.pretty_size(len(f.response.content) / delta)
+        delta = f.response.timestamp_end - f.response.timestamp_start
+        size = len(f.response.content) + f.response.get_header_size()
+        rate = utils.pretty_size(size / delta)
 
         d.update(dict(
             resp_code = f.response.code,
             resp_is_replay = f.response.is_replay(),
             resp_acked = f.response.reply.acked,
             resp_clen = contentdesc,
-            resp_et = "{0:2.0f}ms".format(delta * 1000),
             resp_rate = "{0}/s".format(rate),
         ))
         t = f.response.headers["content-type"]