[Gambas-user] Searching an error in the whole class

Benoit Minisini gambas at ...1...
Thu May 15 12:16:55 CEST 2008


On jeudi 15 mai 2008, Rolf-Werner Eilert wrote:
> Benoit Minisini schrieb:
> > On mercredi 14 mai 2008, Rolf-Werner Eilert wrote:
> >> Hi,
> >>
> >> Is it possible to let the program jump to an error routine if an error
> >> occurs SOMEWHERE in the class?
> >>
> >> I cannot find the reason for an error in one of my apps the students
> >> use. This error occurs especially when lots of french letters and/or
> >> apostrophies are written, but it is not predictable. So I cannot try it
> >> for myself (what I did, but it didn't run into an error...)
> >>
> >> Of course, they use a compiled version of the program. So I have to log
> >> the error into a file and search for the reason. But with TRY etc. I can
> >> only catch an error within one specific function... Since I don't know
> >> where the error occurs, it would be nice to let the program jump into
> >> some "global" error routine.
> >>
> >> Thanks for your hints!
> >>
> >> Rolf
> >
> > Why can't you know where the error occurs? When you compile the program,
> > check the check-box that tell the compiler to keep debuggin information
> > inside the executable, and you will get the location of the error. You
> > can even get the backtrace on the standard error output!
> >
> > Regards,
>
> You were right, as usual :-) Just wanted to let you know, I
> forgot/didn't see the checkbox with the "keep debug information" thing.
> So at last I found the place the trouble began, and I hope I could get
> rid of it.
>
> By the way, does this have any influence on speed, efficiency etc. of
> the app being interpreted? After all, there is a checkbox...
>
> Regards,
>
> Rolf
>

Debugging information takes place (you can do the comparison). As for speed, a 
breakpoint is generated for each line of code, but it does nothing. So you 
maiy not notice anything.

In the future, I plan to remove the useless breakpoint in that specific case.

Regards,

-- 
Benoit Minisini




More information about the User mailing list