[Spice-devel] [PATCH 3.12/12] Various minor style changes to worker and cursor channel

Marc-André Lureau mlureau at redhat.com
Fri Oct 30 09:21:08 PDT 2015


Hi

----- Original Message -----
> > > @@ -446,10 +446,11 @@ void cursor_channel_process_cmd(CursorChannel
> > > *cursor,
> > > RedCursorCmd *cursor_cmd,
> > >          return;
> > >      }
> > >  
> > > -    if (red_channel_is_connected(&cursor->common.base) &&
> > > (cursor->mouse_mode == SPICE_MOUSE_MODE_SERVER ||
> > > -                                   cursor_cmd->type !=
> > > QXL_CURSOR_MOVE ||
> > > cursor_show)) {
> > > -        red_channel_pipes_new_add(&cursor->common.base,
> > > new_cursor_pipe_item,
> > > -                                  (void*)cursor_item);
> > > +    if (cursor->mouse_mode == SPICE_MOUSE_MODE_SERVER
> > > +        || cursor_cmd->type != QXL_CURSOR_MOVE
> > > +        || cursor_show) {
> > > +        red_channel_pipes_new_add(&cursor->common.base,
> > > +                                  new_cursor_pipe_item,
> > > cursor_item);
> > >      }
> > 
> > Why silently remove the red_channel_is_connected check?
> 
> Yes, I had the same question and was going to comment on it as well.
> Marc-Andre, if you're reading this, do you remember?

I don't remember. It looks quite a useless check to me, if process_cmd() is called, it suggests the channel is connected, isn't it? Furthermore, such a check could possibly exists in red_channel_pipes_new_add(). And even if it didn't, I suppose that wouldn't be a big deal to free it later. But anyway, it looks safer to add back if we don't know why it's changed.


More information about the Spice-devel mailing list