snip...

> What I am suggesting is the following:
>  1. The first diagnostic should always be clear and complete.
>  2. If it is not possible to continue because the runtime is damaged,
>    we should stop immediately.
>  3. If it is possible to continue without risk of further unclear
>    cascading errors, we should continue.
>  3. If it is not possible to continue without risk of further unclear
>    cascading errors, we should stop immediately.
> I think there is agreement on 1, 2 and 4.  Perhaps the disagreement
> is on whether 3 is possible and practical.

That sounds right. Maybe in determining that the runtime is damanged
and therefore we should stop in 2 we will get an understanding of
whether there are scenarios where we consider the runtime to be
undamaged, i.e. 3.

Simon

Reply via email to