qpOASES updated to version 3.2.0

[gabrielenava]

I updated qpOASES to version 3.2.0 with the hope of solving this issue. I did two different commits:

• https://github.com/gabrielenava/qpOASES/commit/407c77adaba8397d7281c2c32584b878dc070cff is the original qpOASES 3.2.0 as downloaded.

• https://github.com/gabrielenava/qpOASES/commit/267c1c0b8c1f49caa07879a66926c2abfc726b5c are local modifications to comply with our workflow.

I would like to proper test the new qpOASES before merging.

[gabrielenava]

I tested the update in simulation. Nothing changed from the controllers point of view. Still, tests on the robot are missing.

[traversaro]

We typically use [WIP] for avoiding accidental merges.

[gabrielenava]

Tested on iCubGenova02. No issues to report. The behaviour of the robot was (qualitative) as before the update. I think we can merge this PR. Only question: is qpOASES used only for the balancing controllers?

[traversaro]

+1 to merging this PR.

Only question: is qpOASES used only for the balancing controllers?

https://github.com/search?l=Matlab&q=org%3Arobotology+qpOASES&type=Code
https://github.com/search?l=Matlab&q=org%3Arobotology-playground+qpOASES&type=Code

As OpenSOT uses another fork, I guess the answer is yes.

[gabrielenava]

Ok, merging then!