[DO NOT MERGE] Test build with eigen 5#1903
Conversation
|
See conda-forge/eigen-feedstock#47 (comment) . We can skip PCL for now. |
Temporarily disable PCL in YCM additional CMake arguments.
|
Ok, this is easy to fix: |
https://github.com/search?q=org%3Aami-iit%203.2.92&type=code 26 projects affected : ( |
|
The problem affects also manif: |
|
Now casadi is failing: https://github.com/casadi/casadi/blob/aa46f563b55017a2c4e16829164491d117c75a0a/cmake/FindEigen3.cmake#L4 needs either to be updated or removed (it is probably not useful anymore as all the Eigen3 installation I am aware of ship a |
Fix proposed upstream in artivis/manif#335 . |
Remove temporary disabling of PCL in build options.
|
After almost reaching the leaf of the dependency tree, we finally reached an actually code incompatibility: |
Related to https://gitlab.com/libeigen/eigen/-/merge_requests/826/diffs#10a394f8d6cafac3a0b5c595a09e93071cf661f4 . |
|
Ok, all the fixes were merged in all the repos we mantain, so we need to:
|
To have a release compatibile with Eigen3 5.0.0 . xref: robotology/robotology-superbuild#1902 xref: robotology/robotology-superbuild#1903 xref: conda-forge/eigen-feedstock#47
To have a release compatibile with Eigen3 5.0.0 . xref: robotology/robotology-superbuild#1902 xref: robotology/robotology-superbuild#1903 xref: conda-forge/eigen-feedstock#47
|
The conda ci started to fail because it is installing eigen5, here is the temporary workaround: Also the conda generation workflow: And the Pixi ci: |
To test #1902, xref: conda-forge/eigen-feedstock#47 .