Re: (no subject)

[Hans Boldt <boldt@xxxxxxxxxx>]

Bob Cozzi wrote:
> Cool, now we just need to know if %trimL maps to that MI instruction,
> and if it does not, then we can do performance tests to see which one is
> faster.

I'd be interested in seeing performance numbers too.  But notice
that the string parameter to function triml is a null-terminated
string.  Using OPTIONS(*STRING) in RPG will cause the string value
to be moved into a compiler-generated temp for the length of the
string, which itself ain't cheap either.

Also, it would appear that the triml function just takes a single
character as the character to trim.  Using the idiom
"%SUBST(STR:1:%CHECKR('XYZ':STR))" would today allow you to trim any
of a *set* of characters from the end of a string.  (Trimming from
the left end of a string isn't as straight-forward, though, since
you'd have to test for 0 returned from %CHECK() - %SUBST() doesn't
like 0 as a start position.)

As I suggested earlier, perhaps performance could be improved for
some worst case scenarios at the expense of some of the best case
scenarios.  On the whole, given other priorities and alternatives,
I'm not sure if it's really worthwhile.

Cheers!  Hans