std::vector[vector.size() - 1]

myVector.end() - 1;

Quote:

Original post by Dave myVector.end() - 1;

vector::end() returns an invalid iterator. You can't legally use it to reach a valid one.

*edit: Ignore me, I have been living a lie.

CM

[Edited by - Conner McCloud on August 16, 2006 2:29:04 PM]

Quote:

Original post by Conner McCloud Quote: Original post by Dave myVector.end() - 1; vector::end() returns an invalid iterator. You can't legally use it to reach a valid one. CM

So that's why none of my code ever works. Bugger.

I'M A NOOB!

My code is a mess. I used to use vec.end[-1], then I discovered vec.back() [smile]

Out of interest (Because I'm too lazy to look for myself), what does std::vector do if you call back() and it's empty? The same as accessing an invalid index with operator[] or .at()?

Quote:

Original post by Evil Steve Out of interest (Because I'm too lazy to look for myself), what does std::vector do if you call back() and it's empty? The same as accessing an invalid index with operator[] or .at()?

a.back() returns *--a.end() for vector a [ISO/IEC 14882:1998(E) 23.1.1 Table 68]. If the vector is empty, the results will be undefined since in that case a.begin == a.end(), and one-before-the-beginning of a vector is undefined.

Quote:

Original post by Conner McCloud Quote: Original post by Dave myVector.end() - 1; vector::end() returns an invalid iterator. You can't legally use it to reach a valid one. CM

That's not true. It's perfectly fine to go from end to another iterator through iterator operations. This holds for all container types.

Quote:

Original post by Dave Quote: Original post by Conner McCloud Quote: Original post by Dave myVector.end() - 1; vector::end() returns an invalid iterator. You can't legally use it to reach a valid one. CM So that's why none of my code ever works. Bugger. I'M A NOOB!

In that case, it probably does work. Which is why its so important not to get into the habit of doing it. A more complex structure [say, std::set] may not work that way.

*edit: Or maybe I'm mistaken...

CM

I would consider vector::end()-1 to be suspect code as well, but it seems that microsoft doesn't have a problem with it.

Quote:

Original post by Polymorphic OOP Quote: Original post by Conner McCloud Quote: Original post by Dave myVector.end() - 1; vector::end() returns an invalid iterator. You can't legally use it to reach a valid one. CM That's not true. It's perfectly fine to go from end to another iterator through iterator operations. This holds for all container types.

I stand corrected, and have edited my post accordingly. I was under the impression that i-- required i be dereferenceable, but apparently it does not.

CM