[vtk-developers] Re: Problems with vtkTimeStamp

[Will Schroeder]

Nick-

I have to say that I am very pleased that VTK ran long enough that the time 
stamp actually wrapped around :-) I'd like to believe that this is because 
VTK is getting very robust and not because there are too many complicated 
cascades of modified time going on...

I'm going to float this past the VTK developers and see what we think the 
best course of action is... I appreciate the feedback.

Will

At 01:11 PM 6/3/2005, Thomas, NJ \(Nick\) wrote:
>Dear Kitware,
>
>I'm mailing this directly to you since i'm having problems getting through to
>the news group.
>
>I've just encountered some problems with vtkTimeStamp for long
>simulations (well actually it took me ages to track it down to the timestamp).
>Basically i find that filters stop updating after the time stamp has
>wrapped back to zero since executetime will be a large number and
>modified time will be a small number.
>Having just looked at the latest version of you manual i see that you say:
>
>void vtkTimeStamp::Modified  (     )
>
>    Set this objects time to the current time. The current time is just a 
> monotonically increasing unsigned long integer. It is possible for this 
> number to wrap around back to zero. This should only happen for processes 
> that have been running for a very long time, while constantly changing 
> objects within the program. When this does occur, the typical consequence 
> should be that some filters will update themselves when really they don't 
> need to.
>
>But indeed the opposite seems to be the case (at least for me), i.e. 
>filters not updating themselves when they really ought -
>this is a big problem since it effectively limits the maximum number of 
>modified events that can happen.
>
>Is there anything i can do to change this behaviour easily? I'm guessing 
>that changing the time stamp to an unsigned INT64 would
>be a big job...?!? I'm not so au fait with these things... is it possible 
>to change some compiler options?
>
>Regards,
>Nick Thomas
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://public.kitware.com/pipermail/vtk-developers/attachments/20050603/ed8a4829/attachment.html>