> I understood the original question to be how to avoid OMVS
Not exactly. (I love OMVS!) For the original requirement I am going to take
the liberty of quoting here @Tony Harminc's explanation of the problem.
> Back 10+ years ago our problem was that we'd ship our product to a new
> customer, or an existing customer would install it on a new image, and
> they'd overlook our (well documented with examples) requirement for a
> userid with a UNIX (OMVS) segment. So we'd get running, open config
> files, attach various subtasks, initialize a log, etc. etc.and then at
> some point we'd go to open a socket, and Bang! We'd issue a lame
> message about a TCP/IP return code, and keep retrying at intervals.
> Eventually the customer would call support, and someone would work
> from the TCP/IP error code to the actual problem and then walk them
> through fixing up the userid.
Prodded by that thread I want to add some call to our code for the purpose
of explicitly diagnosing the lack of an OMVS segment. The initialization
sequence for a complex product is of course complicated and I wanted to get
the test call in the right place: as late in the initialization as possible
so that as many other diagnostic facilities are available as possible
("printed" listing, for example) but before any function that would require
OMVS and spoil the whole intent of this exercise. I can run an experiment
but of course a particular situation may not exercise every code path
through initialization.
I am wondering if this is not worthy of an RCF to the C library reference.
Here is a significant prerequisite for some fraction of the calls and it is
at best documented by implication and folklore.
Charles
-----Original Message-----
From: IBM Mainframe Discussion List [mailto:[email protected]] On
Behalf Of David Griffiths1
Sent: Wednesday, March 8, 2017 6:29 AM
To: [email protected]
Subject: Re: Which C library functions imply dub?
I understood the original question to be how to avoid OMVS and you almost
certainly need OMVS for printf. For Metal C I wrote my own printf subset
that calls WTO - you at least have varargs.
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN
