Thanks for catching and fixing this.  I suppose it's a holdover habit from C/PETSc programming: 
PetscNew memzeros the allocated memory, so struct members of pointer type are effectively zero-initialized.  
I guess I always assumed that pointers are zero-initialized by default.  Looking at the C++11 standard 
(or the closest free proxy I could find: http://www.open-std.org/jtc1/sc22/wg21/docs/papers/2012/n3337.pdf)
it appears (Section 8.5 paragraphs 11 and 6) that, absent an explicit initializer, pointers are left with an indeterminate 
value.  Hence, this bug.

Dmitry. 




On Sat, Oct 27, 2012 at 12:24 AM, Roy Stogner <roystgnr@ices.utexas.edu> wrote:

On Wed, 24 Oct 2012, Dmitry Karpeev wrote:

let me know if this causes any kind of grief 

Leaving those pointers uninitialized gave me a segfault; it should be
fixed in r6225.
---
Roy