PRBT IkFast patch from robostack

[tylerjw]

Description

Modified version of this: https://github.com/RoboStack/ros-humble/blob/main/patch/ros-humble-moveit-resources-prbt-ikfast-manipulator-plugin.patch

Quoting @traversaro

This seems a use of variable length array (that is not supported on MSVC) and hack to implement a static assert that can be substituted with a C++11's static_assert ?

I used a std::vector as we tend to use C++ types to avoid raw new/delete here in MoveIt.

On the version of ikfast changing in that static_assert. I searched for copies of the ikfast.h header, and it is in two places in our repo, and in both places, the version is 61. When I set the version to what it was in that hack, it failed, complaining that the version is 61 so the old test wasn't working anyway, and we've been compiling with version 61.

[traversaro]

static_assert ?\n\nI used a std::vector as we tend to use C++ types to avoid raw new/delete here in MoveIt.

As long as a dynamic memory allocation is ok in this context, this seems ok.

[codecov]

Codecov Report

All modified lines are covered by tests ✅

Comparison is base (f998962) 50.78% compared to head (035df7a) 50.80%.

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #2395      +/-   ##
==========================================
+ Coverage   50.78%   50.80%   +0.02%     
==========================================
  Files         386      386              
  Lines       31972    31972              
==========================================
+ Hits        16235    16239       +4     
+ Misses      15737    15733       -4     

see 3 files with indirect coverage changes

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[tylerjw]

As long as a dynamic memory allocation is ok in this context, this seems ok.

I don't understand how it could have worked before without dynamic memory allocation so I have to assume it is ok. C variable length array was using dynamic memory too.

[traversaro]

C variable length array was using dynamic memory too.

As far as I know (but I do not know in-depth) VLA at least in GCC use the stack, see https://gcc.gnu.org/onlinedocs/gcc/Variable-Length.html (the referenced function alloca allocates on the stack, see https://man7.org/linux/man-pages/man3/alloca.3.html).