[bugfix] Break deadlock for one must/can't be blocked pair.

[Grath]

Handle the case where a 'must be blocked if able' creature is attacking but there aren't enough available creatures to block it due to Menace or other effects which increase getMinBlockedBy.

This bug resulted in me winning a game of Commander which I should not have won on Friday, because an opponent was forced to concede due to being unable to declare blockers.

[xenohedron]

Is this something that can be picked up by the test framework or not?

[Grath]

It is not, it isn't forced to block and it throws an UnsupportedOperationException if you tell it to block with just the one creature.

[JayDi85]

Tests added by b2279a8, related test for PR in bd1f6a4 (test_MustBeBlockedWithMenace_low_blockers_auto)

P.S. Whole blockers logic has many potential problems:

• it use auto-fix of blocker configuration with AI related code inside -- it's will be hard to integrate AI simulations (Documentation about developing an AI player? #10154) for declare blocker step without rework (human declare block + AI declare code + configuration checking code);
• there are some bad use cases with restriction/require effects -- so auto-fix blockers configuration fails on it (see some ignored tests), e.g. AI can use all blockers instead required only.