- From: ak.miller at auckland.ac.nz (Andrew Miller)
- Subject: [cellml-discussion] CellML API 1.0rc4 (release candidate for CellMLAPI 1.0) out
- Date: Fri, 24 Nov 2006 10:39:17 +1300
David Nickerson wrote:
>
The attached model hierarchy causes a segfault when using the CellML2C
>
application (built on a Fedora Core 5 machine) with the
>
import-test.xml model. The problem is the messed up encapsulation
>
hierarchy in m1.xml - if the parent component is changed from x -> y
>
(as it should be) then the expected code is generated.
>
>
(This example is definitely a bug in the model, but is probably good
>
to catch segfaults where possible).
I agree. This turned out to be a stack overflow (infinite recursion)
caused by treating a std::list sentinel as a VariableDisjointSet. I have
now added a check for this case (revision 918 on trunk, it will
eventually be put on the branch if no problems arise on the trunk).
Best regards,
Andrew
Archive powered by MHonArc 2.6.18.