Re: More XLPARSE problems

Scott,

No, I'm calling the program from the command line, through a CMD.

And I use system('ADDENVVAR ...') simply because I didn't know that 
putenv() exists.  But I'll definitely look at it now!

How does one start the JVM? Doesn't that happen automagically when calling 
your first java program?

Thanks,

Peter Colpaert
Application Developer
Massive - Kontich, Belgium
-----
Yoda of Borg are we.  Futile is resistance, assimilated will you be.
-----




Scott Klement <rpg400-l@xxxxxxxxxxxxxxxx>
Sent by: rpg400-l-bounces@xxxxxxxxxxxx
15/07/2005 22:31
Please respond to RPG programming on the AS400 / iSeries

 
        To:     RPG programming on the AS400 / iSeries <rpg400-l@xxxxxxxxxxxx>
        cc: 
        Subject:        Re: More XLPARSE problems


> If CLASSPATH is stored at job level (which is what I thought), then how
> come the program gives a 'Class not found' or similar error if I add the
> variable from the command line before running the program interactively,
> but does find the class if the classpath is setup using a system() call
> within the program?

I don't know?  On my system, it works just fine when the variable is set 
prior to calling the program.

Wait... you aren't calling the program through QShell, are you? That would 

make it behave differntly.  Just call it with the CALL command.

Why are you using system() to set the variable?  Why not use putenv()? 
How are you starting the JVM?

> "The excel record that HSSF is not familiar with" always appears to be 
the
> record below the last relevant row in my sheet.  Strangely enough, if I
> create a sheet that has twice the number of rows, it's again the row
> following the last "filled in" one.

I've never encountered the error, I'm just guessing that it means that 
it's a record type that HSSF doesn't understand.  Please note that an 
Excel record is not the same as "one row of cells".

> As for stopping the parsing, do I understand correctly that I can simply
> have the subprocedure return *ON if it encounters a "blank" line?  For
> instance, if column 0 is always filled in as character, have the 
charcell
> procedure return *ON if Value is *blanks?

Returning *ON will end the parsing, but I don't know how to detect that 
you're at the end of the spreadsheet.  If Excel always writes the cells so 

that the blank ones come after the zero ones, then yes, your method should 

work.

Unfortunately, I haven't experimented with it, so you'll have to 
experiment and see what works.