[Insight-developers] Valgrind need attention

[Williams, Norman K]

As I explained in direct e-mail to Bill, I spent a couple of days staring
at the FEM code and couldn't find the problems causing the remaining FEM
valgrind leak reports.  Then I threw up my hands and Vince Magnotta has
been putting some hours in and came to similar conclusion.


The problem is that memory that is allocated by the ::New() method, and
only ever assigned to smart pointers, is being reported as leaked. Which
means that there are 3 possibilities:

1. Somehow, somewhere, the reference count is getting double-incremented.
2. There really is a leak somewhere that's eluded both Vince & I.
3. Valgrind is wrong.

My last exchange with Bill he indicated that he believed the leak was
real.  If that's the case, it's eluded 2 reasonably intelligent &
experienced programmers for a minimum of 40 man hours.  If someone else
has any insights they would be extremely welcome.

On 7/28/11 3:48 PM, "Nicholas Tustison" <ntustison at gmail.com> wrote:

>I'll get to it right away.
>Nick
>
>On Jul 28, 2011, at 4:09 PM, Bill Lorensen wrote:
>
>
>FEM and BSpline checkins have numerous valgrind defects. There are also a
>bunch of doxygen related defects. We need to zero these out.
>
>http://www.cdash.org/CDash/viewDynamicAnalysis.php?buildid=1373394
>
>Bill
>
>
>
>
>



________________________________
Notice: This UI Health Care e-mail (including attachments) is covered by the Electronic Communications Privacy Act, 18 U.S.C. 2510-2521, is confidential and may be legally privileged.  If you are not the intended recipient, you are hereby notified that any retention, dissemination, distribution, or copying of this communication is strictly prohibited.  Please reply to the sender that you have received the message in error, then delete it.  Thank you.
________________________________