Re: [AD] 4.3 error handling

[Peter Wang <tjaden@xxxxxxxxxx>]

On 2006-06-08, Chris <chris.kcat@xxxxxxxxxx> wrote:
> On Thursday 08 June 2006 18:12, Peter Wang wrote:
> > Found this in the info manual:
> >
> >    When you perform a non-local exit, accessible objects generally
> > retain whatever values they had at the time `longjmp' was called.  The
> > exception is that the values of automatic variables local to the
> > function containing the `setjmp' call that have been changed since the
> > call to `setjmp' are indeterminate, unless you have declared them
> > `volatile'.
> 
> Non-static local variables are supposed to be on the stack, aren't they? 

No, local variables can be in registers.

> According to the man page:
> 
> setjmp() saves the stack context/environment in env for later use by 
> longjmp(). The stack context will be invalidated if the function which called 
> setjmp() returns.
> 
> Which sounds to me like the local variables will be preserved /as long as/ the 
> function that called setjmp doesn't return.. and since you don't prematurely 
> exit a try block, and the end of the try block automatically clears the last 
> setjmp point, this can't happen.

I think what they're trying to say is that the stack frame will have at
least a return address that points to the previous stack frame.  Once
you return from the function you called setjmp from, that address could
be pointing to pretty much anything.

Peter