[vtkusers] Thanks + Help!

[John Biddiscombe]

Thanks to all of you who posted me results and sorry for not replying 
individually.

I received 13 mails containing "It works" or similar. Results cover a range 
of platforms and operating systems.

Mine is the only one that crashes. I'm stepping through the code with the 
debugger and sure enough the
mean /= count;

line in vtkImageQuantizeRGBToIndex is dividing zero by zero and throwing an 
exception as no pixels have a certain colour (histogram bin is empty it 
seems - but I've not examined the algorithm fully yet). (I have the correct 
vtk.tif image - checked with one reply poster)

If anyone can suggest why everyone but me gets a valid output, please feel 
free to post a suggestion. (rounding error of some kind?)

I'm running Win2000Pro on an AMD Athlon 1.2GHz with vtk updated from CVS 
yesterday or the day before and the code is all fine. Possibly I'm the only 
one using a Borland compiler, but I'm not convinced it should make any 
difference. Optimizations are off.

ttfn

JB
PS. This enquiry arose because a regression test fails on my box, but is 
fine everywhere else and I need it to work so I can fix another piece of code!