Re: [lmi] 004e:err:tooltips:TOOLTIPS_Timer How did this happen?

[Greg Chicares]

On 3/28/18 12:58 AM, Vadim Zeitlin wrote:
> On Tue, 27 Mar 2018 22:23:13 +0000 Greg Chicares <gchicares@sbcglobal.net> 
> wrote:
> 
> GC> Below I attach a complete log of a 'wx_test' run, just in case that's
> GC> helpful somehow, but the only interesting thing I notice is the seven
> GC> occurrences of:
> GC>   004e:err:tooltips:TOOLTIPS_Timer How did this happen?
> GC> If no one else sees this without 'wine', I'll just dismiss it as yet
> GC> another 'wine' oddity.
> 
>  Well, it's definitely a Wine message (see
> https://github.com/wine-mirror/wine/blob/0a72708126f5b8fc06698172d973bb529944c3bf/dlls/comctl32/tooltips.c#L2081
> for where it's given in its sources), so it definitely can't be seen when
> not running Wine. But it could still indicate that we're doing something
> wrong, e.g. forgetting to relay some message to the tooltip. However I
> don't really see anything wrong in wxMSW code and I suspect that perhaps
> Wine code just doesn't deal with the automated testing scenario in which
> some messages are generated much faster than usual (the expiration delay
> for this timer is 40ms which is a really long time when events are being
> simulated programmatically...).

I saw it again today:

/opt/lmi/src/lmi[0]$/opt/lmi/src/lmi/gui_test.sh
NOTE: starting the test suite
018d:err:tooltips:TOOLTIPS_Timer How did this happen? [repeated eleven times]
SUCCESS: 21 tests successfully completed.
NOTE: 4 tests were skipped

If I first move the mouse pointer to an xfce "panel", and then
run the GUI test, everything works as intended and that weird
diagnostic does not appear.

My hypothesis is that it's caused by running the GUI test
without first moving the mouse pointer to a place on the xfce
screen that will not be occupied by the lmi GUI under wine
emulation--otherwise, the mouse pointer could be hovering
over what becomes a tooltip "mouse-over" location.

Yet another reason to use xvfb.