Re: [eigen] (vec1*vec2.transpose())*vec3 on 2.0.5/6

[ Thread Index | Date Index | More lists.tuxfamily.org/eigen Archives ]


2009/9/29 Rhys Ulerich <rhys.ulerich@xxxxxxxxx>:
>>> On 2.0.5 and 2.0.6 I've noticed that
>>>    (vec1*vec2.transpose())*vec3
>>> errors out at compile time
>
>> It's definitely not legit, you found a bug, let me investigate it...
>
> When you're investigating, would you please also check that it behaves
> correctly for
>    (vec1*vec1.transpose())*vec2
> where the outer product is vec1 against itself?  That's the original
> use case where I found the problem.  The GNU 4.3.1 and Intel 10.1
> compilers both complained in a similar way.

Here your example indeed triggers a compilation error with the
development branch;
but it compiles and runs normally with the 2.0 branch, with 2.0.5, and
with 2.0.6.  (And yes, I uncommented the nasty line).

I tried with GCC 4.4 and GCC 4.1.

Can you show me the compilation errors that you get?

Benoit



Mail converted by MHonArc 2.6.19+ http://listengine.tuxfamily.org/