VW7.3.1 windows XP, shiftDown problem ...

> I am trying to test for shiftDown.  My code is
>
>    someWindow sensor shiftDown
>
> This works 100% of the time -- except yesterday and today on one
> developers laptop,
> on which it has worked all the time before.
>
> Now if I trace the results, it tests 3 times following the press of a
> specific button,
> 2 get false, the 3rd gets true (shift IS down).
>
> However, if I connect to his PC using a RAdmin (a remote access thingy),
> it works perfectly
> when "I" hold down shift on MY PC.
>
> Anyone seen anything like this before??
>

> Dennis Smith wrote:
>> I am trying to test for shiftDown.  My code is
>>
>>    someWindow sensor shiftDown
>>
>> This works 100% of the time -- except yesterday and today on one
>> developers laptop,
>> on which it has worked all the time before.
>>
>> Now if I trace the results, it tests 3 times following the press of a
>> specific button,
>> 2 get false, the 3rd gets true (shift IS down).
>>
>> However, if I connect to his PC using a RAdmin (a remote access
>> thingy), it works perfectly
>> when "I" hold down shift on MY PC.
>>
>> Anyone seen anything like this before??
>>
>
>

>
> This is a well-known multithreading bug: The Views that make up your
> application get queued Events which do not record the state of the
> various shift keys, so if you received such an event and then ask the
> sensor (which delegates to InputState default) for the state of the
> shift key at a later time this may be off by several queued messages.
> Such queued messages could be exposure events or other events being sent
> to any of your windows by the OS (or even by another application).
>
> Having said all this I've never run into this problem except:
> 1) when debugging application code that does such a test
> 2) when releasing the shift key (during a long-running operation) before
> the #shiftDown is executed.
>
>
> R
> -
>

> -----Original Message-----
> From: Dennis Smith [mailto:[hidden email]]
> Sent: Wednesday, May 03, 2006 11:53 AM
> To: Reinout Heeck
> Cc: VWNC,
> Subject: Re: VW7.3.1 windows XP, shiftDown problem ...
>
>
>
> Reinout Heeck wrote:
> > Dennis Smith wrote:
> >> I am trying to test for shiftDown.  My code is
> >>
> >>    someWindow sensor shiftDown
> >>
> >> This works 100% of the time -- except yesterday and today on one
> >> developers laptop,
> >> on which it has worked all the time before.
> >>
> >> Now if I trace the results, it tests 3 times following the press of a
> >> specific button,
> >> 2 get false, the 3rd gets true (shift IS down).
> >>
> >> However, if I connect to his PC using a RAdmin (a remote access
> >> thingy), it works perfectly
> >> when "I" hold down shift on MY PC.
> >>
> >> Anyone seen anything like this before??
> >>
> >
> >
> None of this would surprise me -- what does surprise me is that we have
> been doing this for
> a few years with quite a number of different PCs and laptops (including
> this one), and this is the first
> time we have had a problem, and the problem is repeatable -- on that
> machine only.
>
> However, thinking more about it -- I will bet we have a problem with the
> GUI processes (window processes)
> because also on that laptop pressing buttons (on a window) seems to miss
> presses sometimes -- and I have seen
> this before -- my dirty solution being to rebuild the image from our
> base image (via store).
>
>
> >
> > This is a well-known multithreading bug: The Views that make up your
> > application get queued Events which do not record the state of the
> > various shift keys, so if you received such an event and then ask the
> > sensor (which delegates to InputState default) for the state of the
> > shift key at a later time this may be off by several queued messages.
> > Such queued messages could be exposure events or other events being sent
> > to any of your windows by the OS (or even by another application).
> >
> > Having said all this I've never run into this problem except:
> > 1) when debugging application code that does such a test
> > 2) when releasing the shift key (during a long-running operation) before
> > the #shiftDown is executed.
> >
> >
> > R
> > -
> >
>
> --
> Dennis Smith                        [hidden email]
> Cherniak Software Development Corporation       +1 905.771.7011
> 400-10 Commerce Valley Dr E                Fax: +1 905.771.6288
> Thornhill, ON Canada L3T 7N7    http://www.CherniakSoftware.com

> I am trying to test for shiftDown.  My code is
>
>    someWindow sensor shiftDown
>
> This works 100% of the time -- except yesterday and today on one
> developers laptop,
> on which it has worked all the time before.
>
> Now if I trace the results, it tests 3 times following the press of a
> specific button,
> 2 get false, the 3rd gets true (shift IS down).
>
> However, if I connect to his PC using a RAdmin (a remote access thingy),
> it works perfectly
> when "I" hold down shift on MY PC.
>
> Anyone seen anything like this before??
>