Interesting code bug at work

[jzig]

I just fixed a bug at work that formed an interesting puzzle. We have a quick printf function at work that handles formatting certain data types faster than microsoft printf (which is really slow). This code:

printf("%d", -2147483648)

produced this result:

-(

See if you can figure out why!

Comments

[jzig]

Nah, it's all integer stuff. Oh, and that specific number is the only one that failed, all other numbers work.

[kvance]

I think it went something like this:

The printer sees the negative bit, so it prints a '-' and then negatizes the number.
Being -2^31, this doesn't go well. Now besides being the wrong number, it's still negative.
So when it's time to calculate the ASCII code of the first digit by adding it to ASCII '0', subtraction is performed instead, going to ASCII '('.
Then it divides by 10, checks if it's still > 0 (oops, nope, it's negative!) and stops printing.

[kvance]

When I was writing that, I was thinking "negatize... there must be a better word for that." I suppose some people might say "negate" ;)

[jzig]

Yup, kev wins the prize! Negating -2^31 and assigning the value to a signed 32 gives you... -2^31 it turns out :)

The other satisfactory answer is "because -( looks like a sad cyclops"

[kvance]

Heh. Sad cyclops is totally what I would have guessed next.