Provide useful submission feedback for incremental solving process

[marcellussiegburg]

(Some?) Students claimed that they would like to get immediate feedback on Autotool tasks. In order to distribute bonus points useful anyway, the idea is to “punish” multiple submissions in this scenario. Furthermore, tasks should provide useful feedback that does not (implicitly) tell the correct solution.

[jvoigtlaender]

A current example of this is in LogicTasks.Semantics.Resolve:

logic-tasks/src/LogicTasks/Semantics/Resolve.hs

Lines 333 to 334 in 381f8cf

	when printFeedbackImmediately $ do
	stepsGraded

in partialGrade' and

logic-tasks/src/LogicTasks/Semantics/Resolve.hs

Lines 353 to 354 in 381f8cf

	unless printFeedbackImmediately $ do
	recoverFrom stepsGraded

in completeGrade'