Re: Using %paddr(GetInnerDoc) in http_url_post_xml and thereafter http_parse_xml_stmf

[Scott Klement <sk@xxxxxxxxxxxxxxxx>]

hello Jan,

You have coded the following directive to HTTPAPI:

    http_xmlReturnPtr(*ON);

This tells HTTPAPI that you do NOT want to receive a simple VARYING 
field for your "value" parameter.  Instead, you want to receive a data 
structure in the following format:

     D Element_Val_t   ds                  qualified
     D                                     based(Template)
     D    buf                          *
     D    len                        10i 0

This advantage to this method is that you can handle XML elements that 
contain more than 65535 characters in their data.  The disadvantage is 
that the preceding data structure isn't as simple to work with as an RPG 
VARYING string.

Anyway... you seemed to understand this perfectly when you wrote the
GetInnerDoc() routine.  You seemed to understand quite clearly that you 
had to use this data structure, and use it appropriately.

However, when you coded your Incoming() procedure, you seemed to have 
forgotten this...   you coded a simple VARYING string, and this is why 
you are getting "garbage" in the parameter.  You've coded the parameter 
as a VARYING string instead of the data structure noted above.

Two possible fixes:

ALTERNATIVE #1:

Before your second XML parse, you could do this:

        http_xmlReturnPtr(*OFF);

        rc = http_parse_xml_stmf( filename
                               : HTTP_XML_CALC
                               : *null
                               : %paddr(Incoming)

Because you turned off the "return pointer" feature, HTTPAPI will go 
back to giving you a simple VARYING field, so if you choose that 
alternative, you can keep your incoming() procedure as-is.

ALTERNATIVE #2:

You can change your incoming() procedure to have the following prototype:

     D Incoming        PR
     D   userdata                      *   value
     D   depth                       10I 0 value
     D   name                      1024A   varying const
     D   path                     24576A   varying const
--> D   value                             likeds(Element_Val_t) <--
     D   attrs                         *   dim(32767)
     D                                     const options(*varsize)

Then you'd have to change your logic accordingly.

I hope you understand that these two fixes are /alternatives/ to each 
other.  Please don't try to do both of them at the same time (or you'll 
end up with similar garbage in the parameter)

Good luck


janottar@xxxxxxxxxx wrote:
>    Hi Scott.
>    Thanks for your interest in my problems.
>    Here is my test program that gives garbage in "value":
[SNIP]
>    D Element_Val_t   ds                  qualified
>    D                                     based(Template)
>    D    buf                          *
>    D    len                        10i 0
> 
>    D GetInnerDoc     PR
>    D   filename                    50a   varying
>    D   depth                       10I 0 value
>    D   name                      1024A   varying const
>    D   path                     24576A   varying const
>    D   value                             likeds(Element_Val_t)
>    D   Attrs                         *   dim(32767)
>    D                                     const options(*varsize)
> 
>    D Incoming        PR
>    D   userdata                      *   value
>    D   depth                       10I 0 value
>    D   name                      1024A   varying const
>    D   path                     24576A   varying const
>    D   value                    65535A   varying const
>    D   attrs                         *   dim(32767)
>    D                                     const options(*varsize)
[SNIP]
>      http_xmlReturnPtr(*ON);
[SNIP]
>       rc = http_parse_xml_stmf( filename
>                              : HTTP_XML_CALC
>                              : *null
>                              : %paddr(Incoming)
-----------------------------------------------------------------------
This is the FTPAPI mailing list.  To unsubscribe, please go to:
http://www.scottklement.com/mailman/listinfo/ftpapi
-----------------------------------------------------------------------