execute_control_command may not remove its cleanups

[Daniel Jacobowitz]

On Thu, Feb 19, 2004 at 01:21:52PM -0500, Dave Allan wrote:
> > > However, it seems from code inspection and the gdb internals
> > > documentation that the call to do_cleanups ought to be unconditional. 
> > > Does that seem right?  
> > 
> > No, instead, the cleanup chain should always have an item on it.  If
> > make_cleanup is not called then old_chain will remain NULL, and
> > do_cleanups (NULL) means "do all cleanups", not "do nothing".  It looks
> > to me like command_handler is responsible for there always being a
> > cleanup on the chain:
> >   old_chain = make_cleanup (null_cleanup, 0);
> > but maybe I'm mistaken about that; it's a bit far down the tree.
> 
> I definitely understand that do_cleanups(NULL) will do all cleanups
> which is not what's wanted here.  The call is do_cleanups(old_chain),
> though, so if there are cleanups on the chain already, they are
> preserved.  The problem isn't the do_cleanups call, it's the fact that
> the do_cleanups call is conditional.  The solution is to remove the if
> (old_chain) statement and always do the cleanup.  
> 
> Given what's stated in the docs, that a function must always remove the
> cleanups it creates, it would seem to me that regardless of the state of
> cleanup_chain at the beginning of execute_control_command, whether it's
> NULL or contains cleanups, we want to get back to that state before we
> return.  
> 
> Looking at what cleanups execute_control_command puts on cleanup_chain,
> that is correct.  Either one or two cleanups are put on the chain where
> arg is an automatic variable and function is free_current_contents.  If
> these cleanups aren't done before the stack frame is destroyed,
> something undefined will later be freed when the cleanups are done.  

Think about this again.  Both of those cleanups are conditionally
created.  If neither of them is created, old_chain will still be NULL.
This will lead to running cleanups prematurely.  If the cleanup chain
is non-empty, things work OK.

The alternative is null_cleanup.

-- 
Daniel Jacobowitz
MontaVista Software                         Debian GNU/Linux Developer