[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Ftpapi] Optional XML Labels



Jon,

Yes, sorry...   that slipped my mind for the moment.

But, I last gave this presentation in September, which was before the ON-EXIT and nested DS support were added, that is why they're not in the slides on my site.

Next time I'm asked to speak on this subject again, I will certainly add both features to the slides. Plus whatever else is available by then.

-SK


On 3/8/2017 4:22 PM, Jon Paris wrote:
Nice presentation Scott - couple of things there that had slipped by me. One that you are missing - or at least haven't added yet - the ability to directly code nested DS which will be available any minute now. Huge help when coding DS for XML-INTO because the declaration looks like the XML in terms of the "shape" like so:

*Dcl-DS  Divisions  Dim(5) Qualified;
      DivCode  Zoned(2);
      Dcl-DS  Departments  Dim(10);
        DeptCode  Zoned(3);
        Dcl-DS  Products  Dim(99);
          ProdCode     Zoned(5);
          MonthsSales  Packed(9:2)  Dim(12);
        End-DS;
      End-DS;
    End-DS;
*


Jon Paris

www.partner400.com <http://www.partner400.com>
www.SystemiDeveloper.com

On Mar 8, 2017, at 4:24 PM, Scott Klement <sk@xxxxxxxxxxxxxxxx <mailto:sk@xxxxxxxxxxxxxxxx>> wrote:

Thomas,

You don't have to define the null indicator field. Please don't think that my example is the only way to do things! You can do this, for example:

*dcl-s myfield char(123) nullind;**

**%nullind(myfield) = *on;**
**if %nullind(myfield) = *on;**
**   // do something**
**endif;*

Actually, that might be nicer than having the separate null indicator for a standalone field ;) You can also do stuff like specify a data structure as the null indicator for another data structure. Where the null data structure has indicators for each corresponding field in the other data structure. Please read more about it in the RPG Reference.

Also, you can read the slides from a presentation I sometimes give about what's new in RPG here:
http://www.scottklement.com/presentations/#WHATSNEW

I use the chart in that presentation quite frequently to remember which releases the different features were available in. Due to the fact that many of them were PTFed to previous releases, it can be hard to remember them all, so I find that the chart is handy.

The presentation is missing the "ON-EXIT" feature that was very recently added to RPG because I haven't added it yet. If you are interested in ON-EXIT, you can read about it here:
http://ibm.biz/RPG_ON_EXIT_Section

I hope that these newer features are useful for you in WSDL2RPG (I had to say this so that this e-mail was on-topic for the mailing list, haha.)

-SK

On 3/8/2017 3:13 PM, Thomas Raddatz wrote:
Hi Scott,

Thank you for having enlightened me. The NULLIND keyword is promising
and it makes me think about whether or not it makes sense that I try to
change WSDL2RPG right now or wait for 7.3.

But I wonder why we have to define the null indicator field. From my
point of view (I am a lazy developer) that is something the compiler
should do under the cover as it already does (I assume) for external
data structures with null-capable fields.

Why must it be so often so complicated?

Regards,

Thomas.

Am 08.03.2017 um 21:34 schrieb Scott Klement:
Thomas,

I think he's referring to your statement that RPG doesn't support
declaring null-capable fields.  However, I think Jon is wrong about 7.2
-- this was a 7.3 feature!  But, RPG does support null-capable fields,
you can read about it under "What's New in 7.3" in the RPG Reference,
search the page for "NULLIND":

https://www.ibm.com/support/knowledgecenter/ssw_ibm_i_73/rzasd/rpgrelv7r3.htm



An example would be:

*ctl-opt alwnull(*usrctl);**

**dcl-s myfield_null ind;**
**dcl-s myfield char(50) NULLIND(myfield_null);**
**
myfield_null = *on;   // myfield is now null**
*
(could also use *%nullind(myfield) = ***on*)

etc.

So in 7.3, RPG does support any field being null-capable, so you could
potentially add that support to WSDL2RPG, that it uses those null
indicators when a SOAP message supports nillable fields to indicate that
null to the RPG program.  And when generating the response, you could
check the null indicators and generate the XML according to whether it
is/isn't null.

At least, I think that's what Jon was trying to say. (He didn't provide
much detail.)

-SK

On 3/8/2017 2:08 PM, Thomas Raddatz wrote:
Jon,

Sorry, but what do you think would not work earlier that V7.2? The 4MB
buffer? It is heap storage dynamically allocated by WSDL2RPG and
controlled by the following procedure call:

    hSoapPart =
       MessageContext_getSoapPart(hMsgCtx: 1024*1024*4); // 4 MB cache

Regards,

Thomas.

Am 08.03.2017 um 20:47 schrieb Jon Paris:
It would not work for anything earlier than V7.2 I think Thomas - but
the current support for creating your own null capable fields in RPG
might lead to a possible solution.


Jon Paris

www.partner400.com
www.SystemiDeveloper.com

On Mar 8, 2017, at 2:31 PM, Thomas Raddatz
<thomas.raddatz@xxxxxxxxxxx> wrote:

Christian,

Sadly WSDL2RPG does not support NULL values, which equals a dropped
element. The actual problem is that RPG does not support NULL values
and
it is not possible to define "special" values for recognizing a NULL
value. For strings *BLANKS may work, but for example "0" or any other
value does not work for numeric values. So actually you need an
indicator for each nillable field. I did not yet enhanced WSDL2RPG to
support NULL values, because that is a hell lot of work.

Therefore please change the generated stub by hand to meet your needs.

Here are some hints:

a) WSDL2RPG generates a "Serializer" procedure for each data structure of the request message. So if you do not want to generate a whole data
structure, just return the serializer procedure immediately on a
condition of your choice.

b) Inside the serializer procedure parts of the XML are assigned to
field "g_requestBuffer". Then the content of "g_requestBuffer" is
append
to the request message buffer by calling
OutputStream_appendGlobalRequestBuffer(). Afterwards the value of the
XML element is "marshalled" by a type specific "Marshaller", e.g.
Marshaller_toStringV6().

c) Just for your information: The request message buffer is a 4MB
memory
block (default value) to cache the request message. When the message
exceeds 4MB it is dumped to an IFS file.

d) It is up to you to define the conditions when a field (XML element)
or data structure (XML complex type) is not added to the request
message.

e) You can safely add indicator field to the *_t template variables to
indicate whether or not something shall be serialized. Make sure to
compile everything of the stub after having changed a *_t field.

Regards,

Thomas.

Am 08.03.2017 um 14:16 schrieb larsenvalverde@xxxxxxxxx:
Hello.

I’m using WSDL2RPG to generate the stubs to connect to a webservice.

I have a problem.

The XML that I have to send (generated by WSDL2RPG), have a lot of
“optional labels”.  When I don’t use them, the result XML have a
lot of sentences like this:


…
    </LABEL1>
    </LABEL2>
…

(because I cannot send the information of the LABEL1, LABEL2, …).

If I send this to the webservice, the server give me an error,
because I’m not sending the information of that “optional” labels.

The question is… have I had to manually change the stubs generated
by WSDL2RPG, or there another way to manage them??


Christian.

--
_______________________________________________
Ftpapi mailing list
Ftpapi@xxxxxxxxxxxxxxxxxxxxxx
http://scottklement.com/mailman/listinfo/ftpapi

--
_______________________________________________
Ftpapi mailing list
Ftpapi@xxxxxxxxxxxxxxxxxxxxxx
http://scottklement.com/mailman/listinfo/ftpapi




--
_______________________________________________
Ftpapi mailing list
Ftpapi@xxxxxxxxxxxxxxxxxxxxxx
http://scottklement.com/mailman/listinfo/ftpapi