David R. MacIver - Latest Comments in Unsigned comparison in Java/Scala

Re: Unsigned comparison in Java/Scala

Anon — Tue, 05 Aug 2008 18:23:28 -0000

So it returns true if "i >= j"?

Re: Unsigned comparison in Java/Scala

Tony Garnock-Jones — Mon, 04 Aug 2008 22:28:51 -0000

(Oh dear. Less-than/greater-than munging obliterated the fully spelled-out version at the end of the previous comment.)

Re: Unsigned comparison in Java/Scala

Tony Garnock-Jones — Mon, 04 Aug 2008 22:27:59 -0000

Here's another way of reading it:

"(i < 0) ^ (j < 0)" reads as "i and j have different signs".

So the whole thing reads "i less than j, xor i and j have different signs", so

- either i j, and i and j have differing signs

Re: Unsigned comparison in Java/Scala

david — Mon, 04 Aug 2008 18:01:18 -0000

A truth table is just a way of laying out the answers. It doesn't omit the burden of actually proving each case, which is what I did there.

Re: Unsigned comparison in Java/Scala

Rich Dougherty — Mon, 04 Aug 2008 17:55:43 -0000

> (there’s probably a nicer way to see that it works, but I’m blanking on what it is)

A truth table?

Rich

Re: Unsigned comparison in Java/Scala

david — Mon, 04 Aug 2008 14:00:19 -0000

Of course it does. But this doesn't use a ternary operator.

Re: Unsigned comparison in Java/Scala

chris — Mon, 04 Aug 2008 13:57:09 -0000

actually the ternary operator DOES result in a branch, it's ust not based on an "IF" in the language, but for the compile/JVM, it's same. No matter, this is still an elegant solution. I HATE C/C++, but times like this particular one make me think "macro." Still I'll take java anyday over C/C++