List problems

With all due respect if the error is not with the instantiation of the list then why did you paste it? [grin]

Paste the usage of that list please.

Dave

With the provided information, this is just a wild guess, but are you maybe dereferencing end() ? AFAIK, the iterator returned actually points one element past the end of the list. Someone please correct me if I'm wrong.

Quote:

Original post by Anatol With the provided information, this is just a wild guess, but are you maybe dereferencing end() ? AFAIK, the iterator returned actually points one element past the end of the list. Someone please correct me if I'm wrong.

Yes, it points to one element past the end of the list. Dereferencing .end() would be like dereferencing a null pointer.

Quote:

Original post by deathkrush Quote: Original post by Anatol With the provided information, this is just a wild guess, but are you maybe dereferencing end() ? AFAIK, the iterator returned actually points one element past the end of the list. Someone please correct me if I'm wrong. Yes, it points to one element past the end of the list. Dereferencing .end() would be like dereferencing a null pointer.

Yeah pretty much.
If you actually want an iterator pointing to the last element in the list, use:

CurrentNPCs.rbegin()

Just make sure that you already know that the list isn't empty, or check the resulting iterator against CurrentNPCs.rend() before using it.

The usage of the list is just CurrentPCs.end(). I'm not dereferencing it at all. However, my list might be empty... is that likely to cause a problem?

It doesn't matter if the list is empty. Something like this works fine:

list<int> *l = new list<int>;l->end();

Without code, everybody's just going to be taking wild stabs in the dark. Provide the actual code that causes the problem, and the specific error you recieve, and we can actually start making educated statements.

CM

Why not show the actual code that causes the problem, like the first responder suggested?

All right, here is the code which causes the error.

If your list is empty, I believe (not positive) that begin() is equal to end(), so when you start your loop and immediately increment your iterator, you will go past end(), to garbage. Your equality check will fail and your code will try to call update on the iterator, which will explode.

Try this:

for (PCs_iterator = CurrentPCs->begin(); PCs_iterator != CurrentPCs->end(); ++PCs_iterator)PCs_iterator->Update();

Try



(Or, something I find a bit easier to read)



-HQ.

Or

std::for_each(CurrentPCs->begin(), CurrentPCs->end(), std::mem_fun_ref(&npc_object::Update));

As has been seen, writing loops manually is somewhat error-prone. ;)

just to reiterate the problem:

end() is a logical point one past the last item in the list - accessing and comparing it is legal, dereferencing it is not. The item one past end() is undefined garbage, invalid (undefined) to access in any way.

in an empty list list.begin() == list.end() ... so ++list.begin() is invalid for an empty list.

I've attempted to use your solutions, and while I think you've probably saved me from some problems down the line, it doesn't seem to have been the root of my current problem. Not only does your fix not work, but out of sheer frustration I re-wrote my code so that it never once used std::list.end(), and I'm still getting the same error. The reason I didn't want to post the code that causes the error earlier was that I wasn't sure if it WAS the code that caused the error. The actual exception is thrown somewhere in the middle of list.h, and although I did try to use a breakpoint to verify that that was where the error occured, I'm not sure that's what caused it, since fixing this didn't change anything and the breakpoint just had me stepping through millions of lines of assembly (an exaggeration) which I didn't understand. The actual error I'm getting is this:

Unhandled exception at 0x004190f4 in Tiling_Tryouts.exe: 0xC0000005: Access violation reading location 0xcdcdcddd.

And it occurs in:



which is in list.h.

What does your callstack look like when it crashes? What portion of your code was calling into list code?

Quote:

Original post by EmrldDrgn Not only does your fix not work, but out of sheer frustration I re-wrote my code so that it never once used std::list.end(), and I'm still getting the same error.

The facts are that your original code was wrong and would do bad things in any program. Whereas the code others have posted to help you is correct and does not in itself have any bugs at all. To those familiar with STL, problems like this are obvious and easy to correct. Failing to show appreciation (or worse) for the useful help given, will not help your popularity or speediness of fixing the whole problem. You'd be hard-pressed to find many people giving better advice than Zahlman, so you should consider yourself lucky he is willing to help, and take his advice (and many others here) at every opportunity. I'm sure what you meant to say was: "Thankyou for your help so far, but I am having furthur problems..."[smile]

You've asked for help, and you certainly seem to need it! How about posting a decent amount of (your) code so that we can really get to the bottom of things for you? Where it crashes doesn't matter as it may be other bits of code that are corrupting the data structure. Any bit of (your) code that interacts with the std::list is worth having us take a look at. You'll probably learn much more than you expect!

Anyway, 0xCDCDCDCD is a debug value used to show up uninitialised variables. So there's a good chance this will come down to an uninitialised variable problem.

Quote:

Original post by iMalc I'm sure what you meant to say was: "Thankyou for your help so far, but I am having furthur problems..."[smile]

That's exactly what I meant to say... the wording of my post was... unfortunate. I'm grateful to everyone who's helping me, and as I said above, you've probably saved me major problems down the line.

I'll post all the code I can find that interacts with std::list in the morning, when I can get at my development computer, but I wanted to correct what I was saying now, when I first noticed it. I'm sorry if I seem ungrateful. I'll look for an uninitialized variable in my code somewhere, as well.

Hey that's okay man. I'm sure we can all see your good intentions.
We'll get to the bottom of this soon for ya.

Just a thought, have you tried turning your warning level up to W4 instead of the default W3? It sometimes finds bugs for ya! Don't necessarily feel obliged to change your code to prevent all warnings at that level, but do check them out in case some of them do happen to point to something that might not be quite right.