More robust test for multi material portions

[bombasticSlacks]

Summary

None

Purpose of change

Should de-flake the recent test in #55405

Describe the solution

test now tests against the identical armor but if it were broken. So if they aren't basically identical the code is working.

Also to make this pass I fixed a potential divide by 0

Describe alternatives you've considered

Testing

If it passes it should be good.

Additional context

[BrettDong]

Seems the UBSan test fails because of a division-by-zero error at

Cataclysm-DDA/src/item_factory.cpp

Lines 786 to 793 in 7fa132e

	for( armor_portion_data &it : obj.armor->data ) {
	for( part_material &mat : it.materials ) {
	// scale the value of portion covered based on how much total is covered
	// if you proportionally only cover 5% of the arm but overall cover 50%
	// you actually proportionally cover 10% of the armor
	mat.cover = static_cast<float>( mat.cover ) / ( static_cast<float>( it.coverage ) / 100.0f );
	}
	}

[bombasticSlacks]

Seems the UBSan test fails because of a division-by-zero error at

Thanks for letting me know I'll PR a fix.

EDIT: Ahh I see it's just happening with this new armor, so I'll fix it here actually thanks

[BrettDong]

The test still fails
!(5.9342f == Approx( 5.9210000038 ))

[bombasticSlacks]

Alright I think the issue was the armor values were small enough that since damage is bottomed/rounded to an int when applied to hp in a lot of cases the small differences between the two armors weren't enough to get to the next whole number. The increased thickness should effect it enough to make this reliable.