PocketMatrix

TCHAR *x=new TCHAR[123];
delete x;

In my memory, when I'm using the new [] operator, I think we are supposed to use delete [] operator. So I suspect that there is a leak in that.


For the rest, everything looks fine.

Yes you shall use the delete[] but I don't think that this will result in a leak. It just means that destructors won't be called but since TCHAR don't have a destructor...

I also had a quick look at the code and could not find anything that would result in a leak.

You could test to see if there is any difference if you use STLPort instead.

I believe behaviour is undefined if you use new[] but not delete[].

STLPort 4.5.3
also leaked...
even with Pejo's ... virtual ~IObject() {};

is there a method other than clear() that frees up everything vector allocates?

or guess i'll move on and use a primitive linked list
i'll probably save a few kbs from not creating many vector templates

How about you fix the reinterpret_cast<> and the delete [] problems before blaming STLport? Or did you fix them?

You don't have anything to do to clear a vector, it will take care of itself.

yes, i removed the tchar[] delete.. and used static_cast as you suggested.

it resulted in two leaks

1304 bytes long and 320 bytes long,

i had a _CrtSetBreakAlloc(45) ... this one is inside the push_back method


note:
vc++ only produced one leak 64 bytes long...
i used a simple array instead of vector, no leak is produced...

i'll check other ways to reproduce the leak without reinterpret_cast or static_cast