On 2016-01-05 15:26, Tony Harminc wrote:
> On 5 January 2016 at 16:25, Paul Gilmartin
> <[email protected]> wrote:
>> But I think (I may never have tried) that SMP/E is GOFF-savvy.
> 
> I'd say it's object-format-agnostic. It has no cause that I can think
> of to be looking at the insides of what follows a ++MOD statement.
> Hmmm... I suppose a GOFF module could contain the string ++MOD, and
> that could get wrapped so it's in column 1.
>
(Or just "++".)  I gotta try that.  If it breaks, I'll SR it.

> Where have we met this one before?
> 
For example NETDATA (TSO TRANSMIT) format is not SYSIN-friendly
because "//" can land in column 1.  In one instance, I looped
through "DLM=??" values until I found a safe one for a particular
NETDATA.  I could uuencode it -- uuencode is SYSIN-friendly.

And GIMDTS is intended to protect against "++" in data.  But if
the input is F-80 and contains no "++", GIMDTS simply copies to
output.  So one can not use GIMDTS to protect JCL elements in
a PTF in SYSIN.  Why!?  (And GIMDTS wouldn't solve the GOFF
hazard.)  For that matter, if a data element, by happenstance,
starts with an apparent GIMDTS prologue, SMP/E RECEIVE will try
to extract it and almost certainly fail.  But GIMDTS won't
transform it if it contains no "++".

Your IBM geniuses at work.

-- gil

Reply via email to