Re: Crash or return forbidden value?
Shawn Pringle wrote:
>
> CChris wrote:
> >
> > Currently, the interpreter always checks the validity of any index or slice,
> > and errors out whenever a test fails.
> >
> > My point is: there is no need to duplicate the testing made by the
> > interpreter,
> > and it should do the testing. Just allow RTFatal() to do something smarter,
> > like skipping the call or branching to some place. This would make the user
> > code more efficient (no duplicate tests) and cleaner, as part of the error
> > handling
> > will simply be in the backand, where it is already.
> >
> > CChris
>
> What you are talking about here is building an Exception system into the
> interpreter.
Exactly. In an earlier post, I was showing how Eiffel does it - simpler, but
less flexible.
It would be analogous to try/throw/catch and Exceptions like
> in Java or C++. Perhaps exceptions could be less invasive than adding C++
> syntax if say some new builtin routines such as a register_exception_routine
> were added. Routines that are passed to that function would get called
> in case of exceptions. There would have to be a special return mechanism
> for these functions. They could return to the offending routine's caller.
Either the calling statement (retry) or the next one. Both have separate uses.
>
> You can put in some kind of side effect like syntax keyword:
> onexception (s[6] = 4) {
> ..
> }
>
>
> On the other hand, perhaps compiled code could remove the second check
> after an if statement.
>
> s[6] = 5 -- check is done to make sure s is a sequence and 6 is a valid index
>
> if sequence(s) and length(s) > 7 then
> s[6] = 5 -- no check is made to make sure s is a sequence for the
> -- the user did it. Additionally 7 and 6 are constants
> -- so index checking could be eliminated at compile time
> -- too.
> end if
>
> Shawn Pringle
CChris
|
Not Categorized, Please Help
|
|
