On Fri, Sep 21, 2012 at 7:30 AM, Jeremy Nicoll - ml sourceforge <jn.ml.sfrg.25@letterboxes.org> wrote:
Rick McGuire <object.rexx@gmail.com> wrote:

>This is working correctly.  The function name uses normal Rexx rules for
>name resolution.  If the call name is coded as a symbol, then the name
>being called is ALWAYS the uppercase version, since all symbols have as an
>eventual value the uppercase value of the symbol.  So calls to foo, Foo,
>and FOO are all calls to a function named "FOO".  This all happens before
>the search for an external function ever begins, so that code is not even
>aware of the original case.  The instruction itself has already folded the
>name to uppercase by execution time, so the original case is no longer
>available.  If a quoted string is used, then the original case is

Sure; I wasn't asking that.  Why does the ref manual say that a lower-case
name search will be tried?

Because it IS part of the search order. But that really only applies to the searches using file extensions.  This is not done with the "raw" name because it resulted in too many false positives with non-rexx programs (this was actually added at one point, but it had to be backed out).  In the poster's situation, he's created his files without extensions, so only the matching case is used.  If he renamed them to have a ".rex" extension, then there is much less case insensitivity involved. 


Jeremy Nicoll - my opinions are my own

Got visibility?
Most devs has no idea what their production app looks like.
Find out how fast your code is with AppDynamics Lite.
Oorexx-devel mailing list