I guess I should remove the "signal on novalue" that is at the beginning of the macro, except that I would consider it to be bad programming technique if I did so. Chuckie may like having self-initializing symbols, but I would rather let the interpreter tell me about my mistakes. In this case, the test macro that I used to try the various versions did not have a routine labeled "novalue:", so the "signal on novalue" caused the syntax error. In any event, having gotten correct results from the EXTRACT command was a happenstance caused by a fortunate selection of errors to have in the code.
-----Original Message----- From: VM/ESA and z/VM Discussions [mailto:[EMAIL PROTECTED] On Behalf Of Alan Altmark Sent: Tuesday, November 22, 2005 11:45 AM To: [email protected] Subject: Re: REXX Syntax On Tuesday, 11/22/2005 at 01:06 CST, Brian Nielsen <[EMAIL PROTECTED]> wrote: > You misunderstand Alan's post, or he was not clear enough. The only > syntax checking REXX will do is that the line is a valid REXX expression > (and all that entails). The question mark will not cause a REXX syntax > error because it is a valid constant string. Some other special > characters will cause REXX syntax errors because they are REXX operators. > Neither the question mark nor the period are REXX operators. Technically, a question mark is a symbol, not a constant string. That is, you can assign a value to ?, but you cannot assign a value to the period. The ? symbol has an initial value of '?'. But you're right: Neither is a Rexx operator and so there is no syntax error. Alan Altmark z/VM Development IBM Endicott
