Re: _Rreadnc() and _Rwrited() from RPG

Hi Scott,
Thank you for the indications about the signal() use. I will discard it.

I have tried also the rtncode=y option on the _Ropen() and it didn't change
the issue.
It is not explicitly described on the doc but what we have is:
"This parameter is only valid for database and DDM files."
Display file are unfortunately not included.


Hi Chuck, I tried with CEEHDLR and it works, you find the solution! Thanks
a lot.

Thank you Mark and all for your help.

Pascal


2014/1/8 <c400-l-request-Zwy7GipZuJhWk0Htik3J/[email protected]>

> Send C400-L mailing list submissions to
>         c400-l-Zwy7GipZuJhWk0Htik3J/[email protected]
>
> To subscribe or unsubscribe via the World Wide Web, visit
>         http://lists.midrange.com/mailman/listinfo/c400-l
> or, via email, send a message with subject or body 'help' to
>         c400-l-request-Zwy7GipZuJhWk0Htik3J/[email protected]
>
> You can reach the person managing the list at
>         c400-l-owner-Zwy7GipZuJhWk0Htik3J/[email protected]
>
> When replying, please edit your Subject line so it is more specific
> than "Re: Contents of C400-L digest..."
>
>
> Today's Topics:
>
>    1. Re: _Rreadnc() and _Rwrited() from RPG (Scott Klement)
>    2. Re: _Rreadnc() and _Rwrited() from RPG (Pascal Polverini)
>    3. Re: _Rreadnc() and _Rwrited() from RPG (Mark S Waterbury)
>    4. Re: _Rreadnc() and _Rwrited() from RPG (Mark S Waterbury)
>    5. Re: _Rreadnc() and _Rwrited() from RPG (CRPence)
>
>
> ----------------------------------------------------------------------
>
> message: 1
> date: Tue, 07 Jan 2014 13:22:08 -0600
> from: Scott Klement <[email protected]>
> subject: Re: [C400-L] _Rreadnc() and _Rwrited() from RPG
>
> Pascal,
>
> signals are, as far as I can tell, unrelated to what you're doing.
>
> Signals are usually used for asynchronous notifications, for example,
> your program might be running along processing things when another job
> wants to interrupt it and tell it to quit.  A signal would be ideal here.
>
> SIGIO is usually used in sockets programming when you want your program
> to be notified of data arriving on the socket, while it's doing
> something else.
>
> There are a few cases in unix programming where signals are used for
> errors (those that are normally fatal) but, it's not the normal means of
> error handling -- and the _Ropen/_Rread/etc functions are not Unix
> functions, they are for IBM i only -- so it would not make sense for
> them to use signals for error handling.
>
> Anyway...
>
> There is a parameter to _Ropen() that controls whether programs are sent
> escape messages or not.   Take a look at the 'rtncode' parameter, here:
> http://pic.dhe.ibm.com/infocenter/iseries/v7r1m0/topic/rtref/ropen.htm
>
> Not 100% sure how it interacts with Rreadnc() since I've never used the
> recio functions with subfiles, personally.
>
> -SK
>
>
> On 1/7/2014 12:38 PM, Pascal Polverini wrote:
> > Hi Buck,
> >
> > I get an escape message.
> > I/O error CPF5037 was detected in file MYDSPF(C G D F).
> >
> > I have attached all the prototypes used for RPG (Rxxx_H.txt).
> >
> > I have also tried to add a signal() on my code but I don't know if this
> is
> > relevant or coded properly. This (or the way it is coded) doesn't change
> > anything
> >
> > Line added at the very beginning of the code:
> > signal(SIGIO : %paddr(SignalHandler));
> >
> > Here the procedure:
> > P signalHandler   b
> > D signalHandler   pi
> > D  sig                          10i 0 Value
> >   /free
> >    dsply ('signalHandler called');
> >    return;
> >   /end-free
> > P signalHandler   e
> >
> > To test it, you simply need to run a READC without having any record
> > changed in your subfile (Normally a simple Enter on a SFL display to a
> pgm
> > that uses / checks for readc).
> >
> > Pascal
> >
> >
> >
>
>
> ------------------------------
>
> message: 2
> date: Tue, 7 Jan 2014 20:37:33 +0100
> from: Pascal Polverini <[email protected]>
> subject: Re: [C400-L] _Rreadnc() and _Rwrited() from RPG
>
> Hi Mark,
>
> I have tried with
>
> on-error *ALL;
>
>
> but still have the same result.
>
> I have the feeling that the C function _Rreadnc() can throw an escape msg
> directly and out of the RPG "monitor" instruction control...
>
> Pascal
>
>
> ------------------------------
>
> message: 3
> date: Tue, 07 Jan 2014 14:50:02 -0500
> from: Mark S Waterbury <[email protected]>
> subject: Re: [C400-L] _Rreadnc() and _Rwrited() from RPG
>
> Hi, Pascal:
>
> See Scott Klement's recent reply.  I think all of these C _Rxxxx
> functions have the ability to use a "return code" versus issuing an
> *ESCAPE message, so you really need to research that approach.
>
> Hope that helps,
>
> Mark S. Waterbury
>
>  > On 1/7/2014 2:37 PM, Pascal Polverini wrote:
> > Hi Mark,
> >
> > I have tried with
> >
> > on-error *ALL;
> >
> >
> > but still have the same result.
> >
> > I have the feeling that the C function _Rreadnc() can throw an escape msg
> > directly and out of the RPG "monitor" instruction control...
> >
> > Pascal
>
>
>
> ------------------------------
>
> message: 4
> date: Tue, 07 Jan 2014 14:57:50 -0500
> from: Mark S Waterbury <[email protected]>
> subject: Re: [C400-L] _Rreadnc() and _Rwrited() from RPG
>
> Hi, again, Pascal:
>
> Also, I just noticed that this was posted in the C400-L list -- but this
> topic should really be posted in the RPG400-L list, since you are trying
> to do this using ILE RPG IV ...  you might get more people over there
> who have used some of these APIs, who can respond with better answers ...
>
> By the way, since you are using ILE RPG IV, why not just use regular ILE
> RPG IV I/O statements, rather than using the C _Rxxx APIs? =-O
>
> All the best,
>
> Mark S. Waterbury
>
>  > On 1/7/2014 2:37 PM, Pascal Polverini wrote:
> > Hi Mark,
> >
> > I have tried with
> >
> > on-error *ALL;
> >
> >
> > but still have the same result.
> >
> > I have the feeling that the C function _Rreadnc() can throw an escape msg
> > directly and out of the RPG "monitor" instruction control...
> >
> > Pascal
>
>
>
> ------------------------------
>
> message: 5
> date: Wed, 08 Jan 2014 01:53:22 -0800
> from: CRPence <CRPbottle-/[email protected]>
> subject: Re: [C400-L] _Rreadnc() and _Rwrited() from RPG
>
> On 07-Jan-2014 10:38 -0800, Pascal Polverini wrote:
> >
> > I get an escape message.
> > I/O error CPF5037 was detected in file MYDSPF(C G D F).
> >
> > <<SNIP>>
>
>    The CPF5037 is almost surely sent as a /notify/ message, for which an
> ILE Condition Handler likely is required to avoid that RPG run-time
> handler from intercepting the condition [and sending the inquiry if
> necessary error handling is not in effect for what I infer is a 1299
> condition; error RNX1299 with inquiry RNQ1299].  Spool the joblog
> [and\or review the actual logged CPF5037 with the F1=Help and with
> F9=Details and\or output from F6=Print] to verify that is the case, and
> to see the originator.  Do the same for the [default] reply.
>
>    Some doc references that might help.  Do not /get hung-up on/ the
> references to CL in the drill-down or supporting text in my
> snippets-only doc references; visit each link for complete details:
>
> http://pic.dhe.ibm.com/infocenter/iseries/v7r1m0/topic/rbam6/enmsg.htm
> IBM i 7.1 Information Center -> Programming -> Control language ->
> Messages -> Commands used to send messages from a CL program
> _Escape and notify messages_
> "...
> A notify message tells the caller about a condition under which
> processing can continue. The calling program or procedure can monitor
> for the arrival of the notify message and handle the condition it
> describes. If the caller is an Integrated Language Environment?
> procedure, it can perform the following functions:
>
>     ? It can handle the condition.
>     ? It can send a reply back to the caller.
>     ? It can allow the sending procedure to continue processing.
>
> If the caller is an original program model (OPM) program and is not
> monitoring for the message, the sender receives a default reply. If the
> caller is an ILE procedure, then the message percolates to the control
> boundary. When finding no monitor, the system returns a default reply to
> the sender. The sender then resumes processing.
>   ..."
>
>
>
> http://pic.dhe.ibm.com/infocenter/iseries/v7r1m0/topic/rzasc/sc092507345.htm
> IBM i 7.1 Information Center -> Programming -> Programming languages ->
> RPG -> ILE RPG Programmer's Guide -> Debugging and Exception Handling ->
> Handling Exceptions
> _ILE Condition Handlers_
> "ILE condition handlers are exception handlers that are registered at
> run time using the Register ILE Condition Handler (CEEHDLR) bindable
> API. They are used to handle, percolate or promote exceptions. The
> exceptions are presented to the condition handlers in the form of an ILE
> condition. You can register more than one ILE condition handler. ILE
> condition handlers may be unregistered by calling the Unregister ILE
> Condition Handler (CEEHDLU) bindable API.
>
> There are several reasons why you might want to use an ILE condition
> handler:
>
>    ? You can bypass language-specific handling by handling the exception
> in your own handler.
>
>      This enables you to provide the same exception handling mechanism
> in an application with modules in different ILE HLLs.
>   ..."
>
>
> http://pic.dhe.ibm.com/infocenter/iseries/v7r1m0/topic/apis/CEEHDLR.htm
> _Register a User-Written Condition Handler (CEEHDLR) API_
> "...
> The Register a User-Written Condition Handler (CEEHDLR) API registers a
> user-written condition handler for the current call stack entry.
>   ...
>
> _ILE Condition Handler Interface_
>
> Following is a description of the interface that the system uses to
> communicate with ILE condition handlers.
>   ...
>
> result_code (output)
>
>      This field contains the instructions from the condition handler to
> the system regarding the actions that the system should take.
>
>      ILE condition handlers get control for all *ESCAPE, *STATUS,
> *NOTIFY, and function check messages. Not all result code actions are
> valid for all types of messages.
>
>      If the message is handled by the ILE condition handler, the
> result-code action is not performed.
>
>      ...
>      Valid result codes are:
>
>      Resume
>          This result code can be used for all exception types.
>          10  Resume at the resume cursor, and handle the condition, as
> follows:
>          ...
>              *NOTIFY    The default reply is sent and the message
> appears in the job log.
>   ..."
>
> http://pic.dhe.ibm.com/infocenter/iseries/v7r1m0/topic/rbam6/ntmsg.htm
> IBM i 7.1 Information Center -> Programming -> Control language ->
> Messages -> Monitoring for messages in a CL program or procedure
> _Monitoring for notify messages_
> "...
> Notify messages are sent to tell your procedure or program of a
> condition that is not typically an error. By monitoring for notify
> messages, you can specify an action different from what you would
> specify if the condition had not been detected. Very few IBM-supplied
> commands send notify messages.
>
> Monitoring for and handling notify messages is similar to monitoring for
> and handling escape messages. The difference is in what happens if you
> do not monitor for and handle notify messages. ...
>   ..."
>
>
>
> http://pic.dhe.ibm.com/infocenter/iseries/v7r1m0/topic/rzasc/sc092507348.htm
> IBM i 7.1 Information Center -> Programming -> Programming languages ->
> RPG -> ILE RPG Programmer's Guide -> Debugging and Exception Handling ->
> Handling Exceptions
> _Problems when ILE CL Monitors for Notify and Status Messages_
> "...
> ... if the ILE RPG procedure writes a record to a printer file and the
> actual printer file has a shorter record length that was declared in the
> RPG procedure, notify message CPF4906 is sent to the RPG procedure. The
> RPG exception handling percolates this message which causes the default
> reply of 'I' to ignore the message. This should allow the output
> operation to continue normally, and the RPG procedure should proceed to
> the next instruction.
>   ...
> If you encounter this problem, you have two possible ways to avoid it:
>   ...
>     2. Enable an ILE condition handler in the RPG procedure. In the
> handler, if the message is one that you want to ignore, indicate that
> the message should be handled. Otherwise, indicate that it should be
> percolated.
>   ...
>      Figure 147 shows an example of a ILE condition handler that ignores
> CPF4906.
>   ..."
>
>
> --
> Regards, Chuck
>
>
> ------------------------------
>
> --
> This is the Bare Metal Programming IBM i (AS/400 and iSeries) (C400-L)
> digest list
> To post a message email: C400-L-Zwy7GipZuJhWk0Htik3J/[email protected]
> To subscribe, unsubscribe, or change list options,
> visit: http://lists.midrange.com/mailman/listinfo/c400-l
> or email: C400-L-request-Zwy7GipZuJhWk0Htik3J/[email protected]
> Before posting, please take a moment to review the archives
> at http://archive.midrange.com/c400-l.
>
>
>
> End of C400-L Digest, Vol 12, Issue 5
> *************************************
>
-- 
This is the Bare Metal Programming IBM i (AS/400 and iSeries) (C400-L) mailing list
To post a message email: C400-L-Zwy7GipZuJhWk0Htik3J/[email protected]
To subscribe, unsubscribe, or change list options,
visit: http://lists.midrange.com/mailman/listinfo/c400-l
or email: C400-L-request-Zwy7GipZuJhWk0Htik3J/[email protected]
Before posting, please take a moment to review the archives
at http://archive.midrange.com/c400-l.