Re: [eigen] 3.2.3 for the end of the week.

In addition to the aforementioned unused typedef, I've hit this warning, with MSVC 12 and warnings cranked up:

eigen\src/Geometry/Transform.h(233): warning C4127: conditional _expression_ is constant

Currently working around it by using a header to push and pop warning state when including eigen.

Also, there's no mention of Isometry as a mode in the Geometry/Transform docs. That's probably not necessary to try to squeeze in a release, since I use the docs on the web site anyway...

Ryan

On Thu, Dec 11, 2014 at 4:07 AM, Christoph Hertzberg <chtz@xxxxxxxxxxxxxxxxxxxxxxxx> wrote:

Addendum:
We could add -DEIGEN_NO_EIGEN2_DEPRECATED_WARNING when compiling the eigen2support tests.

On 11.12.2014 11:01, Christoph Hertzberg wrote:

On 10.12.2014 14:35, Gael Guennebaud wrote:

the 3.2 branch received a lot of fixes since the latest 3.2.3 release,
and
it's time for a 3.2.3 release.

There are some problems with metis_support, mpreal_support and
opengl_support (all of them had been fixed in devel, I think).
Also with gcc-4.8 there are lots of warnings like:

typedef ‘Scalar’ locally defined but not used [-Wunused-local-typedefs]

we could either disable those or remove these typedefs.

Furthermore, gcc-4.4 has lots of warnings:

/.../eigen3.2/Eigen/src/Core/ProductBase.h:92: warning: comparison of
unsigned _expression_ < 0 is always false

These are kind of silly, as they result from template parameters and
intended to be known at compile time. Eigen 3.3 seems not to use that
mechanism anymore, so I can't figure out if we had some work-around for
that at one point.

I don't have MKL, so I can't test anything related to that (btw, there
are some MKL related bugs, partially over a year old ...)

Christoph

--
----------------------------------------------
Dipl.-Inf., Dipl.-Math. Christoph Hertzberg
Cartesium 0.049
Universität Bremen
Enrique-Schmidt-Straße 5
28359 Bremen

Tel: +49 (421) 218-64252
----------------------------------------------