RE: My dilemma with VARYING was ->RE: Error in %trim BIF with V5R3?

There's always the dilemma of defining and enforcing contracts.  If you
can't trust the people who call your procedures to do it correctly, then
either you aren't communicating those requirements correctly, or they're not
very good programmers.  Neither option bodes well for long term software
quality.

However, there is a sort of halfway position.  Just check the last position
of the passed string.  If it's blank, trim it.  I think something like this
will work:

If (%len(var) > 0) and ((%subst(var): %len(var): 1) = ' ');
  %len(var) = %len(%trim(var));
Endif;

Joe

From: albartell

CatPath += ('cat=' + %char(CatArr(i)) + '&');

I like using this syntax along with VARYING fields (for all of the obvious
reasons), but the thing I struggle with is when I need to be ensured that
something does not have trailing blanks (e.g. concatenating a IFS stream
file path). This seems to be the case more often than not.  A lot (nearly
all) of my development is done with RPG ILE service program sub procs. So
I
am constantly providing interfaces to other programmers.  I am not so
worried about them implementing the interface correct (i.e. using VARYING
in
their programs), but more so that they aren't ensuring the absence of
blanks
in strings before they pass them on to sub procs I create. Sure I could
just
insert a few more %TRIM's here and there to bullet proof my programs, but
then I just lost some of the benefit I was trying to gain.