Re: On-Error

[Barbara Morris <bmorris@xxxxxxxxxx>]

rob=tksaTn4SAz0AvxtiuMwx3w@public.gmane.org wrote:
>
> This will make the logic kind of confusing
> Monitor;
>   read (e) myfile;
>   Select;
>   When %status=NoError; // Good read
>     ...
>   When %status=EndOfFile; // Break out of loop
>     ...
>   When %status=RecordNotFound; // Strange
>     ...
>   EndSl;
>   // Check for other errors
>   On-Error TriggerErrBef; // Trigger error, maybe we don't have access?
>     ...
>   On-Error FileClosed; // Oops, need to open the file
>     ...
>   On-Error;  // Catch all for other errors
>     ...
> EndMon;
>

Rob, in this case, maybe a select -would- be better.

MONITOR wasn't intended to replace (e).  It was more intended to handle
the cases where the *PSSR is the only handler, for the cases where you
want to continue.  Just code a calculation in a MONITOR without having
to check in advance that it won't crash (thereby avoiding the trip to
the *PSSR for an allowable error).

It can also simplify exception handling for I/O opcodes in a
subprocedure; if you just want to fail the subprocedure if any error at
all occurs, you can use a big MONITOR and not have to bother with (e)
and %error-checking on every I/O.

>
> Why didn't they allow the sub 100's to be trapped for?
>

MONITOR was intended for exception handling.  I can see that being able
to monitor for status codes between 1 and 99 could be useful. But I
don't think that you should be able to monitor for a status of 0.

By the way, it was probably a typo, but if you code read(e), the (e)
will handle all the file errors, so your on-errors will never fire.