[Laszlo-dev] DHTML runtime safety vs. efficiency

[Max Carlson]

+1 for me.  Any less overhead when debugging is a good thing!

P T Withington wrote:
> Now that we have fixed http://jira.openlaszlo.org/jira/browse/LPP-8222, 
> by providing a way for errors that a user intentionally throws to get 
> past compiler.catcherrors, I propose that compiler.catcherrors should 
> default to `true` when you are debugging.  This will alleviate the need 
> for OL users to resort to the browser debugger for debugging their OL 
> code, as the built-in debugger will catch, report, and ignore any 
> runtime errors that are not explicitly permitted using `#pragma 
> "throwsError=true"`.
> 
> My motivation for this proposal is 
> http://jira.openlaszlo.org/jira/browse/LPP-8479, were we would remove 
> the runtime overhead of two function calls to evaluate any potentially 
> error-signalling reference.
> 
> [If you do not default catcherrors to true in the debug case, the error 
> will continue to be thrown and will be caught by the browser debugger 
> (if any), but that will also result in your program halting, making it 
> difficult to debug more than one bug at a time.
> 
> Comments?

-- 
Regards,
Max Carlson
OpenLaszlo.org