On Friday November 18 2011 08:50:38 Anders Logg wrote: > On Fri, Nov 18, 2011 at 08:36:01AM -0800, Johan Hake wrote: > > On Friday November 18 2011 02:29:18 Anders Logg wrote: > > > On Thu, Nov 17, 2011 at 11:02:30AM -0800, Johan Hake wrote: > > > > On Thursday November 17 2011 10:55:34 Anders Logg wrote: > > > > > Would it be a good idea to implement our own assert (dolfin_assert) > > > > > which would use an #ifdef to stay out of place when not building > > > > > with the DEBUG flag enabled. > > > > > > > > > > The point would be for it to call dolfin_error instead of just > > > > > segfaulting. That way it's very easy to see where it comes from > > > > > since it will generate a trace in Python. > > > > > > > > > > I'm debugging some unit tests now which give > > > > > > > > > > ......python: > > > > > /home/logg/scratch/src/dolfin/1.0.x-logg/dolfin/la/EpetraVector.cpp > > > > > :306 > > > > > > > > > > : virtual void > > > > > : dolfin::EpetraVector::get_local(dolfin::Array<double>&) > > > > > > > > > > const: Assertion `x' failed. > > > > > > > > > > and I have no idea from which test this comes from. > > > > > > > > > > We had our own assert before bug decided to use the standard C++ > > > > > assert. > > > > > > > > > > Any objections to me adding a dolfin_assert and replacing all > > > > > asserts? > > > > > > > > That would be great! > > > > > > In progress... dolfinreplace works great! :-) > > : > > :) > > > > Would be cool if we could apply some regexp to that script too! > > Yes. It ended up replacing dolfin_assert by dolfin_dolfin_assert... :-)
He, he. I always run dolfin_replace using -s (simulate) to test what happens first. Johan > -- > Anders _______________________________________________ Mailing list: https://launchpad.net/~dolfin Post to : dolfin@lists.launchpad.net Unsubscribe : https://launchpad.net/~dolfin More help : https://help.launchpad.net/ListHelp
