Compiler issue using Eigen Wrappers about optim

[ts826848]

Can confirm tip-of-HEAD Eigen works.

Looks like Eigen::all was (is? will be?) introduced in Eigen 3.4:

Slicing and Indexing

This page presents the numerous possibilities offered by operator() to index sub-set of rows and columns. This API has been introduced in Eigen 3.4.
...
The symbol Eigen::all representing the whole set of respective rows or columns in increasing order.

Interestingly, the docs were committed 2018-11-09, and the change moving Eigen::placeholders::all to Eigen::all was committed 2018-09-15, so seems this has been baking for quite a while.

[qacwnfq]

Indeed this Error does not happen with a newer version of Eigen. (Issue happens with 3.3.7 but works with 3.4.99. Note that 3.3 is still the stable release and newer versions are only obtainable from Eigens source repo.)

[titovmaxim]

Looks like duplicate of #29. Which -D options do you you for compilation? Please check if Eigen is properly included from optim_options.hpp under #ifdef OPTIM_ENABLE_EIGEN_WRAPPERS. Are you sure you have Eigen sources nearby?

[kfavre]

I compile my code using the configure file with the option -l eigen which leads to -D option being -DOPTIM_ENABLE_EIGEN_WRAPPERS
The path to Eigen library seems ok (I use Eigen in several other projects without trouble).
The includes in optim_options.hpp also seem fine.

[kfavre]

Quick update:
it looks like the root cause of the problem is in the unconstrained/nm.cpp file (which displays the error in the console)
I could compile the optim lib properly, using eigen, by removing it. As I do not need the Nelder-Mead algorithm I did not look any further in the issue, so it is more of a walk around than a solution.
The tutorial example works fine.

[kthohr]

My guess:

• you need to update your copy of Eigen,
• your include path for Eigen is not set correctly, and/or
• there is an issue with the compiler flags you are passing (which may or may not be related to the include path for Eigen).

I cannot replicate the error when using the latest version of Eigen.