Added IKFast factory boilerplate

[rjoomen]

This adds IKFast boilerplate code as suggested by @marip8.

[Levi-Armstrong]

Just so this code gets tested could we updated the ikfast test plugins to utilize this?

[rjoomen]

Yes, that's a good idea. And the docs could then be simplified a lot (also, there's a number of errors/typos in the IKFast solver/plugin section, but those would be gone then).

[rjoomen]

BTW, I also had the idea to enhance the IKFast plugin boilerplate to use a discretization 'resolution' parameter for each free joint (making use of the joint limits from the URDF) instead of the list of exact joint values that is required currently. Because if you have for example a free joint with a 210 degree range and want to sample it in one degree steps, it requires a long and impractical list of values (I tested this).

[Levi-Armstrong]

Sounds good to me.

[rjoomen]

Would you mind to already merge, and I'll add the free joint enhancements later?

[Levi-Armstrong]

Waiting on the joint enhancement is good, but I would like to update the unit tests to leverage this so this code is unit tested before merging.

[rjoomen]

I updated the unit tests.

[Levi-Armstrong]

It looks like clang-tidy is not happy

[Levi-Armstrong]

@rjoomen I think this code can be removed and should solve the clang-tidy issue.

[Levi-Armstrong]

@rjoomen I think this code can be removed and should solve the clang-tidy issue.

Actually you may be able to just remove the -DIKFAST_HAS_LIBRARY from here.

[rjoomen]

Now all that's left are clang-tidy issues in the solvers and external/ikfast.h. Shouldn't these have been suppressed by TESSERACT_COMMON_IGNORE_WARNINGS_PUSH?

[Levi-Armstrong]

I just pushed a commit which should fix the clang-tidy issue which was tested locally.

[rjoomen]

Great, thanks!