Oppolzer - Informatik / Blog

is it possible that in the second case the differents parts of the outcome of
the macro are evaluated separately, before the whole macro has terminated? that
is: on every single ANSWER statement?

That would maybe explain it:

maybe the first part

          R = 'ZAKQTRCP IF_PRINT_TRACE_LEVEL ';

is processed by the macro processor again, without the remaining parameters,
which are still missing at this time ...

maybe this leads to the "unmatched quote" error and then to the other errors
that follow ???

Is this the defined behaviour of ANSWER ?

If so, it is all clear, and I have to take care that always complete macro calls
are processed by ANSWER ...

Kind regards

Bernd



Am 11.02.2014 20:57, schrieb Bernd Oppolzer:
> Hello all,
>
> I have some difficulties with a PL/1 macro (preprocessor function)
> I wrote. The macro tries to activate another macro by producing an
> output string containing the call of the other macro using ANSWER
> statements.
>
> When I do it like this, all works well:
>
>  %XTRACEB: PROC ();
>
>           DCL R CHAR;
>
>           R =      'ZAKQTRCP IF_PRINT_TRACE_LEVEL ';
>           R = R !! '   (TRACE_LEVEL_DIAGNOSTIC) ';
>           R = R !! '   L(''['' !! PGMNAME !! ''] ';
>           R = R !! '** START UPRO '' !! ';
>           R = R !! 'PROCNAME () !! '' **'') ';
>           R = R !! 'ENDIF_TRACE;';
>           ANSWER (R) SKIP;
>
>           %END;
>
> I get the coding that the ZAKQTRCP macro provides.
>
> When I do it like this
>
>  %XTRACEE: PROC ();
>
>           DCL R CHAR;
>
>           R = 'ZAKQTRCP IF_PRINT_TRACE_LEVEL ';
>           ANSWER (R) SKIP;
>           R = '   (TRACE_LEVEL_DIAGNOSTIC) ';
>           ANSWER (R) SKIP;
>           R = '   L(''['' !! PGMNAME !! ''] ';
>           ANSWER (R) SKIP;
>           R = '** ENDE UPRO '' !! ';
>           ANSWER (R) SKIP;
>           R = 'PROCNAME () !! '' **'') ';
>           ANSWER (R) SKIP;
>           R = 'ENDIF_TRACE;';
>           ANSWER (R) SKIP;
>
>           %END;
>
> I get the following messages:
>
>
> MACRO (Built:20100917) Messages
> Message       Line.File Message Description
> IBM3983I S    2802.0    Premature end-of-source in scan.
> IBM3961I S    2802.0    End-of-source has been encountered after an
>                         unmatched quote.
> IBM3961I S    2802.0    End-of-source has been encountered after an
>                         unmatched quote.
> IBM3997I S    2802.0    Internal preprocessor error: no WHEN clause
>                         satisfied within EXP_GET_STR_LEN
> No Compiler Messages
>
>
> and the Compiler ends without producing anything.
>
> The line 2802 refers to the source line of the first macro call XTRACEE.
> Compilation stops after processing this line. I have no idea what
> EXP_GET_STR_LEN is; maybe an internal procedure name of the preprocessor.
> It is in any case no name of my program.
>
> The difference IMO is (apart from minor differences between the two macros)
> that in the first case the whole text is built in variable R and there is only one
> ANSWER statement; in the second case, there are many ANSWER statements -
> which should be possible - in other situations, it worked without problem.
>
> Of course, the solution was to code both macros like the first one. But anyway:
> what do you think of this? Is this an error in the preprocessor that should be
> resolved? Or: do I use some sort of strange feature that is not supported?
>
> The compiler is EP V3.9, AFAIK - didn't copy the release information,
> but: see the build date in 2010, above.
>
> Thank you, kind regards
>
> Bernd
>
>
>