[vtk-developers] bug 8513, updating freetype; need Windows dev to help

Sean McBride sean at rogue-research.com
Tue Dec 15 14:38:55 EST 2009


On 12/15/09 12:22 PM, Moreland, Kenneth said:

>Sean,
>
>This morning I noticed that a bunch of the ParaView tests were failing
>on blackrose that were not failing yesterday.
>
>http://www.cdash.org/CDash/viewTest.php?onlyfailed&buildid=493486
>
>I eventually noticed that all the tests were in some way printing
>characters on the screen.  Using a debugger, I discovered that an
>internal freetype object named clazz, which holds a bunch of function
>pointers for various font manipulations, appears to be set to garbage. 
>See the screen captures to see the stack trace right before the crash
>and the odd values stored for function pointers.  The crash occurs when
>it tries to run the request_size function.
>
>I ran the same test on my Mac and saw that the function pointers were
>set to valid functions.  For example, the request_size function is set
>to T1_Size_Request (defined in the freetype source file t1objs.c).  I
>observe that the names of most of the functions defined in the clazz
>object are not mangled.  Could it be on blackrose that somehow pointers
>from a different library are somehow interfering with function pointers
>to the local library?

Ugh.

I noticed from your screenshots that you were running in 64 bit.  Did
you run in 64 bit on your Mac too?  I wonder if the pointers are being
truncated by being converted to a 32bit 'int' somewhere.  Could you test
in 32 bit on blackrose and in 64 bit on your Mac?

-- 
____________________________________________________________
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