Prior to this APAR, I was forced to do the SETA against the equate before the
macro call and then pass the resolved value into the macro i.e.
&L SETA equ_name
MacroName LEN=&L
The ability to do the SETA inside the macro really simplifies our code, but
it's implementation seem to be somewhat flaky. Sometimes it work, other times
it get errors.
TestSETA TITLE 'Test LEN macro'
MACRO
BLAH &FLD
AGO .AAm040
AIF (D'&FLD).AAm030
MNOTE 4,'&FLD is not defined'
AGO .AAm040
.AAm030 ANOP ,
MNOTE 0,'&FLD is defined'
.AAm040 ANOP ,
&L SETA &FLD
MNOTE 0,'Length of &FLD is &L'
MEND ,
IHASAVER ,
TestSETA RSECT
BLAH SAVF4SA_Len
BLAH Saver_Len
END
This code above assembles correctly, but if you comment out the AGO on line 4,
it fails assembly???
Robert Ngan
DXC Luxoft
-----Original Message-----
From: IBM Mainframe Assembler List <[email protected]> On Behalf
Of Jonathan Scott
Sent: Friday, November 17, 2023 16:22
To: [email protected]
Subject: Re: Macro variable attributes
> Is there a way to know the real value of the LEN below when I use a
> variable instead of a value?
>
> MACRO
> &LABEL WORK &LEN=0
> &LABEL DS XL(&LEN)
> MNOTE *,.LEN=&LEN.
> MEND
If you have a recent maintenance level of HLASM, including APAR
PH34116 from February 2021, you can use SETA to obtain the value of an equated
symbol which has been defined and resolved. (Some parameter cases did not work
and required a further fix for APAR PH50923).
So in this case, if you know that &LEN contains a single symbol which has been
defined and resolved before the macro is invoked, you can simply use something
like:
&N SETA &LEN
If it contains a more complex expression, you can use EQU to create a dummy
variable equal to the expression and then use that in a SETA expression,
provided it has been resolved.
It is possible to test whether a symbol is defined using the defined attribute
D'&LEN but unfortunately it is not currently possible to test whether it has
been resolved (which may not be the case if the EQU expression contains forward
references), so I do not think it is possible to write code which will assemble
without error but which will only conditionally extract the value if it has
already been resolved.
Jonathan Scott, HLASM
IBM Hursley, UK
