Re: [RFC] Alternate approach to keeping convenience variables

[Andrew STUBBS <andrew dot stubbs at st dot com>]

Jim Blandy wrote:
> On 1/4/06, Andrew STUBBS <andrew.stubbs@st.com> wrote:
>
> > One point though - why is it ok to leak the memory? This seems like bad
> > practice to me - I mean, types can be arbitrarily large and, even if
> > they are typically small, the variables may be rewritten many many times
> > (particularly by scripts with loops). Are you saying that they will be
> > actually leaked or just left for some sort of garbage collection?
>
>
> No, it'll really be leaked.  But the leakage only occurs when symbol
> files get unloaded from memory, so it would only affect scripts that
> load and unload symbol files in a loop.  I agree it's not optimal, but
> I think the alternative is GC.
>
> Maybe we should think about that.  Has anyone ever tried running GDB
> using the Boehm collector's replacements for malloc and free?

Perhaps you could modify copy_type_recursive such that it creates a 
cleanup chain specific to the internal variable. Then call do cleanups 
in set_internal_var.

Might be tricky with shared type copies though. Just a thought.