Make floor-control more robust.

Make floor-control more robust.

Check all injected mouse positions, not just the first one, to deal with
possibility that mouse-injection API might fail.

Also rename recent_remote_mouse_positions_ to injected_mouse_positions_ to
emphasize these are the events that were (attempted to be) injected into the
host.

BUG=None
TEST=Use buggy Mac client to send invalid mouse position to Linux host (move
mouse off-screen.


Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=96124

[Jamie, 2011-08-09 19:42:44]

LGTM with some wording nits, and one debug log left in the code.

http://codereview.chromium.org/7604007/diff/1/remoting/host/client_session.cc
File remoting/host/client_session.cc (right):

http://codereview.chromium.org/7604007/diff/1/remoting/host/client_session.cc...
remoting/host/client_session.cc:101: LOG(INFO) << "Queue too big";
If anything, the problem is that the queue is too small in this case :) I don't
mind adding the log, but could we say something along the lines of "Injected
mouse positions queue full" instead.

http://codereview.chromium.org/7604007/diff/1/remoting/host/client_session.cc...
remoting/host/client_session.cc:125: // mouse recording API might skip one or
two positions).  In this case, the
I think this can be phrased better. Perhaps "This is because we assume the local
input monitor will receive mouse position events in the order in which they were
injected, if at all--out-of-sequence events are not expected and can be
discarded."

http://codereview.chromium.org/7604007/diff/1/remoting/host/client_session.cc...
remoting/host/client_session.cc:133: << ") != front of queue, treating as local
mouse-movement";
Remove

http://codereview.chromium.org/7604007/diff/1/remoting/host/client_session.h
File remoting/host/client_session.h (right):

http://codereview.chromium.org/7604007/diff/1/remoting/host/client_session.h#...
remoting/host/client_session.h:117: // mouse events from the recording API are
echoes of injected positions, or
I wouldn't refer to it as the recording API. It's called the local input monitor
elsewhere.

[Lambros, 2011-08-09 20:35:42]

http://codereview.chromium.org/7604007/diff/1/remoting/host/client_session.cc
File remoting/host/client_session.cc (right):

http://codereview.chromium.org/7604007/diff/1/remoting/host/client_session.cc...
remoting/host/client_session.cc:101: LOG(INFO) << "Queue too big";
On 2011/08/09 19:42:45, Jamie wrote:
> If anything, the problem is that the queue is too small in this case :) I
don't
> mind adding the log, but could we say something along the lines of "Injected
> mouse positions queue full" instead.

Done.  I didn't mean to keep that log in.  But I've moved it now to VLOG(1)
since it might become spammy :)

http://codereview.chromium.org/7604007/diff/1/remoting/host/client_session.cc...
remoting/host/client_session.cc:125: // mouse recording API might skip one or
two positions).  In this case, the
On 2011/08/09 19:42:45, Jamie wrote:
> I think this can be phrased better. Perhaps "This is because we assume the
local
> input monitor will receive mouse position events in the order in which they
were
> injected, if at all--out-of-sequence events are not expected and can be
> discarded."
Have a look at my new wording, see if it's any clearer.  Also got rid of
first-person "we"...

http://codereview.chromium.org/7604007/diff/1/remoting/host/client_session.cc...
remoting/host/client_session.cc:133: << ") != front of queue, treating as local
mouse-movement";
On 2011/08/09 19:42:45, Jamie wrote:
> Remove

Done.

http://codereview.chromium.org/7604007/diff/1/remoting/host/client_session.h
File remoting/host/client_session.h (right):

http://codereview.chromium.org/7604007/diff/1/remoting/host/client_session.h#...
remoting/host/client_session.h:117: // mouse events from the recording API are
echoes of injected positions, or
On 2011/08/09 19:42:45, Jamie wrote:
> I wouldn't refer to it as the recording API. It's called the local input
monitor
> elsewhere.

Done.

[Jamie, 2011-08-09 20:57:11]

LGTM

[commit-bot: I haz the power, 2011-08-10 03:30:21]

Change committed as 96124