short-circuit logic about gmqcc HOT 7 CLOSED

I was wondering how a short circut XOR would work considering that each operand needs to be evaluated inorder for the xor semantics require an evaluation. The semantics of SCL logic dictate that only sub-expressions of a full expression be evaluated to some extent to short the logic, if they dont evaluate to a truth value to simply stop the evaluation. If we added a short-circut XOR operator, perhaps "^^". How would you implement SCL logic on it. I'm certian it isn't possible at all, the evaluation would need to take place, which means it cannot map 1:1 with operator logic. Any thoughts? I really want a SCL XOR :P

from gmqcc.

Simply not possible, because the result always depends on both operands. (False and X) = False, (True or X) = True, but (False xor X) = X and (True xor X) = Not(X), there's always a dependency here.

What WOULD make sense however is when we want perl-like results of logical operations. div0 mentioned he'd like that: x && b would result in a ? b : a, in other words, if a is true, it yields b, which doesn't necessarily have to be a boolean value. Eg. (1 && 5) doesn't yield 1 but 5, or (ent_a && ent_b) yields the entity ent_b if ent_a != world. In this case, the XOR operator could work like described above: (a ? !b : b), however this could become a bit uncomfortable since !b is always a boolean (ie. float in QC), but b itself could be an entity, so we'd need to implicitly cast false to world in such a construct.

from gmqcc.

since short circut evaluation cannot be implement on a logical XOR, we should still implement an operator for it, something along the lines of ^^, I cannot digress how much I despise the idea of !a != !b, which when truncated for whitespace looks insane "!a!=!b". I'd only allow this operator for -std=gmqcc, since I don't suspect another standard of the language implements it, but I know div would generally agree with I suspect; his perl-likeness of perl-like operators is insane, fteqcc even implements a backwards bitwise assignment operator, with negation, In C you'd use something like a&=b, in fteqcc it's a&=b, which is just insane. I suspect -std=fteqcc will need to implement this operator for fteqcc conformancy. Please open a ticket on this matter

from gmqcc.

Nice catch, I completely forgot about div's &~= operator...
We now need 2 sets of tests: one for -fshort-logic, and one for this with an additional -fperl-logic.

from gmqcc.

I feel as if we should call -fperl-logic, f-omgwtfiswrongwithyou-logic

from gmqcc.

It's perfectly natural, many languages use it. shells, LUA, python, ...

from gmqcc.

added tests

from gmqcc.