Fix off-by-one in path finding logic

[StephenCWills]

This fixes an issue where items and objects can no longer be interacted with unless the player is on top of them.

In the following conditional statement, the call to GetDistance() always returned zero. The reason is that the logic thought that a path of length 1 exceeds the maximum path length of 1 that was passed in via the second argument.

devilutionX/Source/controls/plrctrls.cpp

Lines 171 to 174 in 20c0a8d

	if (targetPosition != futurePosition && GetDistance(targetPosition, 1) == 0) {
	// Don't check the tile we're leaving if the player is walking
	continue;
	}

This resolves #7681

[Trihedraf]

That does fix it in game. But now the tests are broken 🫠

[glebm]

The failing test says:

	// Longest possible path is currently 24 steps meaning tiles 24 units away are reachable
	Point startingPosition { 56, 56 };
	CheckPath(startingPosition, startingPosition + Displacement { 24, 24 }, std::vector<std::string>(24, "↘"));

	// But trying to navigate 25 units fails
	CheckPath(startingPosition, startingPosition + Displacement { 25, 25 }, {});

However, the longest possible path is actually 25, so I think the test is wrong (or we can reduce the maximum path length for monsters to 24, I think that might be closer to previous behaviour)

[StephenCWills]

I kept the monster path length at 25 because I do not think it particularly matters. Assuming the save format supports path lengths of 25, I think it's reasonable to consider this a bugfix.

I restricted the path tests to 24 so we wouldn't have to change the current test suite. It's still a good test to demonstrate we are appropriately failing when we reach the maximum path length.

[ephphatha]

The failing test says:

	// Longest possible path is currently 24 steps meaning tiles 24 units away are reachable
	Point startingPosition { 56, 56 };
	CheckPath(startingPosition, startingPosition + Displacement { 24, 24 }, std::vector<std::string>(24, "↘"));

	// But trying to navigate 25 units fails
	CheckPath(startingPosition, startingPosition + Displacement { 25, 25 }, {});

However, the longest possible path is actually 25, so I think the test is wrong (or we can reduce the maximum path length for monsters to 24, I think that might be closer to previous behaviour)

This was a flaw in the vanilla pathfinding code. The array that stored pathfinding steps had room for 25 entries, but vanilla logic would never use the final entry. https://github.com/diasurgical/devilution/blob/6b7135cec37e9cf0f6740d7e1a1ec880f89e73ff/Source/path.cpp#L78 should've been <= instead to allow for up to 25 steps instead of limiting to 24 or fewer. The test was set up to document that restriction but now that it's been fixed changing the test is fine.