You could have fooled me.

Are you honestly an idiot?
Your callback gives incorrect result both for sort and stable_sort.
See updated previous post for a demonstration.

That's my default manner of speaking. Shame on me.
I don't outright flame people with personal attacks and I do try to help and answer people's questions, so bear with me.

For what it's worth, I find the documentation of STL's sort() very ambiguous on how the compare function should behave.
Libc's qsort() seems more sane to me as it specifies that the comparison functions should handle explicitely the 'equal' case. Most comparison functions I've written for qsort were simply return obj2->y - obj1->y;

I think I just did that.

EDIT: And I appologize for being a prick. I should have corrected you in a more calm manner ... got the rage in me, it's a sickness.

That's actually true. Thankfully, it comes with an example. But to be honest, that's not even the point. The point was that the stability depends on the algorithm, not on the comparison function. Granted, it won't work if you don't provide it with the callback it expects, but that's obvious and besides the point.

I honestly don't understand what you're talking about. I hope you're just trolling... Correct me if I'm wrong, but you said that this callback:

bool compare_foos(Foo *a, Foo *b) {
   return a->bar <= b->bar;
}

Can be used with std::sort to achieve stable sorting.
My example demonstrates that it doesn't work with neither std::sort nor std::stable_sort. There's nothing wrong with my array. I copy it into a vector then sort the vector by using the integer part as the key. The fractional part is used to differentiate between the numbers to see if they end up in the original relative order. Clearly, they don't. Feel free to hack up your own example and post it here so we could all observe how it also fails.

Come on, guys. You're better than that. Somebody take a serious look at it and explain to him why he's wrong.

[EDIT]

To be fair, this example isn't even mine. It's from the STL documentation that demonstrates stable sorting. I just replaced their callback with yours.

Ok, I'm wrong. I had my definitions wrong. By 'stable sort' I actually meant a sort that will return the same results every time. Hell, I may even be wrong about that, but I remember having different results returned when using < and not <= when implementing A* one time. So it's not a stable sort, and I'm not even sure that < will return different results at any time. I was sure that was the case though. Anyway, appologies for posting incorrect information.

Ok dude, think what you want. I don't really care. If you can't accept an appology then you don't deserve one.

I was being completely honest. I think you just have a hard time seeing any merit in anyone but yourself. You win on technical terms, but I have no doubt that I walk away the better person.

I was right about one thing: asshat. And that's not a compliment even though you wear it like a badge of pride. Don't feel sorry for me, my shit has been sorted for a long time now. Good luck with yours.

Another thread completely destroyed because of petty arguments, sigh

Yes. The only way if you want to use the STL sort functions.

Predictable.

Let's rewind and see who threw the first punch...