Use Translator versions of problem graders instead of PGanswermacros.pl versions.

[drgrice1]

Instead of the ENDDOCUMENT method setting the PROBLEM_GRADER_TO_USE flag to the avg_problem_grader or std_problem_grader methods from the PGanswermacros.pl file (in the case that macro is loaded, or generally if those methods are defined), the PROBLEM_GRADER_TO_USE environment variable value is simply transferred to the corresponding flag at that time. Then the code in WeBWorK::PG that already exists for this takes care of choosing the WeBWorK::PG::Translator versions of those methods instead. If the install_problem_grader method is used to set the problem grader to code, that is still used as before.

The std_problem_grader code is cleaned up and synchronized in both the WeBWorK::PG::Translator package and in the PGTanswermacros.pl file, although the code in the latter will no longer be used.

The POD for both the std_problem_grader and avg_problem_grader methods is also copied into the WeBWorK::PG::Translator package.

This is the last thing that is needed to make everything in the PGanswermacros.pl either deprecated or unneeded.

This should work with all existing problems since the methods in the PGanswermacros.pl file and the methods in the WeBWorK::PG::Translator package are functionally the same. Note that those problems that load the PGanswermacros.pl macro (typically via loading PGstandard.pl) and then set the grader via install_problem_grader(~~&std_problem_grader) or install_problem_grader(~~&avg_problem_grader) will still be using the methods from the PGanswermacros.pl file.

If instead that macro is NOT loaded, and the grader is set with install_problem_grader('std_problem_grader') or
install_problem_grader('avg_problem_grader'), then the WeBWorK::PG::Translator versions will be used. That should be considered the modern way of doing this. Of course setting the avg_problem_grader is pointless since that is the default. The sample problems that mention using the std_problem_grader have been updated to use this approach. You should note that this approach of using the string instead of the method has actually always worked, and is how this should have been done along. It is much nicer than using the ~~& construct.