Add methods to get list of names of supported parameters for each solver

[traversaro]

This permits to get all the parameters supported by a solver, prerequisite for #8 .

[S-Dafarra]

Looks good to me. The only remark, would be not to throw errors in case a parameter is not found, as future versions of the solvers might add parameters that are not yet listed as supported

[traversaro]

The only remark, would be not to throw errors in case a parameter is not found, as future versions of the solvers might add parameters that are not yet listed as supported

Sorry, it is not clear to me how this is related to the current PR. Is this suggesting a change here or is a general remark?

[S-Dafarra]

The only remark, would be not to throw errors in case a parameter is not found, as future versions of the solvers might add parameters that are not yet listed as supported

Sorry, it is not clear to me how this is related to the current PR. Is this suggesting a change here or is a general remark?

Not directly, but since the list of parameters has been specified explicitly, I thought to mention it anyhow. For example, I noticed the pattern in the test

qpsolvers-eigen/tests/QPTest.cpp

Line 44 in 7059f5c

REQUIRE(std::find(parametersNames.begin(), parametersNames.end(), "this_is_not_a_supported_parameter") == parametersNames.end());

and I thought it might be problematic if used also in the actual code. But again, it is not related to the code as is right now

[traversaro]

The only remark, would be not to throw errors in case a parameter is not found, as future versions of the solvers might add parameters that are not yet listed as supported

Sorry, it is not clear to me how this is related to the current PR. Is this suggesting a change here or is a general remark?

Not directly, but since the list of parameters has been specified explicitly, I thought to mention it anyhow. For example, I noticed the pattern in the test

qpsolvers-eigen/tests/QPTest.cpp

Line 44 in 7059f5c

REQUIRE(std::find(parametersNames.begin(), parametersNames.end(), "this_is_not_a_supported_parameter") == parametersNames.end());

and I thought it might be problematic if used also in the actual code. But again, it is not related to the code as is right now

Yes, that was just a way to check that the vector did not contained something that was not a parameter. I thought this_is_not_a_supported_parameter was a safe bet for a non-existing parameter name, but if we ever add a solver with a parameter named this_is_not_a_supported_parameter we can change the test.