[vtk-developers] New build error in 64 bits

Sean McBride sean at rogue-research.com
Wed Jun 20 12:17:27 EDT 2007


On 6/20/07 11:55 AM, David Cole said:

>By very recently, they mean as in the email that started this thread:
>
>The error is caused by a recent change in VTK:
><http://public.kitware.com/cgi-bin/viewcvs.cgi/IO/vtkDataReader.cxx?
>r1=1.143&r2=1.144>
>
>Line 1167 and 1187:
>vtkTypeInt64 *ptr = ((vtkTypeInt64Array *)array)-
>>WritePointer(0,numTuples*numComp);
>vtkTypeUInt64 *ptr = ((vtkTypeUInt64Array *)array)-
>>WritePointer(0,numTuples*numComp);

Precisely.  I do wish our nightly 4-way Universal Build could be
submitted to kitware's public dashboard, then the above commit would
have triggered those lovely emails ("There are build errors") but until
10.5 is out, we have to continue using our private internal DART2 server.


On 6/20/07 11:46 AM, Brad King said:

>A temporary solution is to follow your proposal to consider long long
>before long when looking for a 64-bit integer type.  Let's go with that
>for now.

OK.  We can implement that here, confirm that it works, and commit it. 
I hope it will not break other platforms/configs, but I don't think it
will.  I'll wait until first thing in the morning to commit.  Is that ok?

>The real problem is that the vtkTypeInt64Array class is defined
>differently depending on the architecture.  Instead of generating the
>class it should make the selection at preprocessing time.  Some kind of
>hack will have to be put in to the Tcl wrappers to make the same
>selection at preprocessing time when compiling the wrappers.

I think this is beyond me.  :)  Should I file a bug for this, so that it
is in someone's queue?

Thanks to all,

-- 
____________________________________________________________
Sean McBride, B. Eng                 sean at rogue-research.com
Rogue Research                        www.rogue-research.com 
Mac Software Developer              Montréal, Québec, Canada





More information about the vtk-developers mailing list