Re: memory debugging and C++ (program keeps crashing)
On Mon, Sep 30, 2002 at 01:41:29AM -0400, Glenn Maynard wrote:
> On Sun, Sep 29, 2002 at 10:13:57PM -0700, Marc Singer wrote:
> > Indeed. I mean that the memory is allocated and freed through the
> > same mechanisms. Unlike other operating systems, there is usually
> > only one heap per process. In the absence of constructors and
> > destructors, the calls are interchangeable.
> Emphasizing that this isn't portable may not even work on later versions
> of g++,
There is no guarantee, but it will tend to work.
> and probably won't work at all for delete.
Look. I'm not recommending this behavior. I strongly recommend folks
to be parallel in their implementation. There is no *good* reason not
to. My point has been that there is only one heap.
> (This was implied, and may be obvious to most of us, but should be said
> explicitly to avoid confusion.)
> Glenn Maynard