qualified field names include the DS

demailaddrs.address is the field name.
Phil

and in an array
demailaddrs(1).address

Thank you Phil,
Unfortunately I’m getting the same errors. Here is the code where I’m trying to fill the data:

     For x = 1 to 20;
       select;
         when x > 4;
           leave;
         when x = 1;
           toaddrs.emailaddrs(x).e_type = 'TO ';
           toaddrs.emailaddrs(x).name   = *blanks;
           toaddrs.emailaddrs(x).address= 'yehudav@bagir.com';
         when x = 2;
           toaddrs.emailaddrs(x).e_type = 'TO ';

toaddrs(x).to_type = ‘TO ‘;

From the structure above toaddrs is the name of a qualified data structure that is like the data structure emailaddrs. And the subfield names seemed to have changed from the structure above the the code sample.
Phil

*RNF7030 30 596700

Statement 596700 is a long, long way down the source for an RPG procedure. I wouldn’t expect it to refer to a D-spec, so I assume it’s a C-spec.

Can you show us that emailaddrs is defined within the same scope as statement 596700?

Tom

Phil,
You are correct but it was wrong typing from my side. The e_type is to_type.

Tom,
I’m not really understanding what you mean. In my original question I included the code both for the ExtPgm(‘SNDEMAIL’) and the DS for emailaddrsDS. Anyway here are the codes around statement 596700:

   If  %scan('CS':SRNMBR)<>0 and stsemail<>'1';
     fromaddr = '*CURRENT';
     fromname = '*CURRENT';
     srnumcs  = srnmbr;

     For x = 1 to 20;
       select;
         when x = 1;
           toaddrs.emailaddrs(x).to_type = 'TO ';
           toaddrs.emailaddrs(x).name   = *blanks;
           toaddrs.emailaddrs(x).address= 'yehudav@bagir.com';
         when x = 2;
           toaddrs.emailaddrs(x).to_type = 'TO ';
           toaddrs.emailaddrs(x).name   = *blanks;

           ...
       EndSl;
     EndFor;

     Monitor;
       sndemail (subject:htmlmsg:toaddrs:fromaddr:fromname);
       stsemail ='1';
     on-error;
       stsemail = ' ';
     Endmon;
  
    Endif;

Is it what you mentioned Tom?

I think this is what you want:

     For x = 1 to 20;
       select;
         when x > 4;
           leave;
         when x = 1;
           toaddrs(x).to_type = ‘TO ‘;
           toaddrs(x).name   = *blanks;
           toaddrs(x).address= ‘yehudav@bagir.com’;
         when x = 2;
           toaddrs(x).to_type = ‘TO ‘;

Tom asked about the scope of the declarations.
You’ve shown us the ‘D’ specs
Are they at the beginning of the program–making then global declarations
Or are they at the beginning of this procedure–making them local declarations.
Or are they at the beginning of another procedure — making them invisible to the procedure that contains this code.
Phil

There are many possible problems. The error message shows that there are almost 6000 source lines at least, so it’s impossible to tell what actual errors there are. With that many source lines, especially when a PR shows that procedures are being prototyped, some errors could be explained by “scope” problems. There is no way for us to see what the scope of the definitions is by just seeing snippets. I wanted to be sure that “scope” hadn’t been overlooked, so I only added the comment in case more needed to be looked at.

However, my comment was added after Phil’s original comment.

His comment indicates a very likely source of the error, but the source lines that had the errors weren’t shown. The RNF7030 error message in a listing will only show the lowest level name rather than the qualified name, so we can’t tell about an error without both the error message and the source statements at the same time. We can’t tell if the source line actually included qualification unless the source with the error is shown.

Since emailaddrs is a qualified DS, you can’t reference the subfields without qualification. That was the point of Phil’s first comment from what I read.

Now, after that, I don’t see where toaddrs is defined. We’re now missing the D-specs for it. So far, you have only shown it referenced under the PR. But where is it defined? You can’t use the name without defining it first. The PR doesn’t define it; it only reserves a parameter interface for it. Whatever you use for that parameter when you call the proc has to meet the requirements of the parameter, but the name under the PR is not relevant.

Also, it looks as if you expect the PR parameter to reference an area of memory that can hold up to 20 copies of the emailaddrs DS. And the source shows emailaddrs with an index. But emailaddrs isn’t an array, so an index doesn’t really belong there. (It might be valid; I haven’t tried it like that. But it won’t matter until a D-spec is created for toaddrs.)

Maybe that’s enough rambling to help clear up some confusion. Ask about parts that need more discussion.

Tom

Thank you Tom the explanation.
First of all this program is a monster and I shouldn’t write it in that way.

Everything is declared in the D spec as global variables.

0122.00 D subject         s             80a
0123.00 D htmlmsg         s           5000a
0124.00 D toaddrs         s           3106a
0125.00 d srnumcs         s              8a
0126.00 D fromaddr        s            100a
0127.00 D fromname        s            100a

0124.00 D toaddrs s 3106a

Okay, so toaddrs is neither a DS nor an array. Since it’s not a DS, it can’t be qualified and it won’t have any subfields. It’s just a long character field.

Consider changing this to a DS with likeds(emailaddrs) dim(20).

Tom

Tom, thank you for the suggestion, but how could I do that?

Fixed, thank you all for your valuable suggestions.
YuVa

Sometimes the easiest thing to do is to code a few lines into a test member to see what happens. I have a few dozen members in different languages that focus on different things. This is one that I have used to play around with DSs:

     D ds1             ds
     D  v1                            3
     D  v2                            3

     D ds2             ds                  qualified
     D  v3                            3
     D  v4                            3

     D ds3             ds                  likeds( ds2 )
      /free

       v1        = 'ABC' ;
       v2        = 'DEF' ;
       v3        = 'GHI' ;
       v4        = 'JKL' ;
       ds3.v4    = 'JKL' ;
       ds3.v5    = 'JKL' ;

       *inlr = *on ;
       return ;
      /end-free

With little examples like that, you can always add or remove a line, add or remove a keyword, just to see what the compiler does. Sometimes it just takes seeing the error messages from a compile to help make things clearer. Other times you might scan down to the cross-reference in the listing to get clues.

Tom

“We know that he did not come through the door, the window, or the chimney. We also know that he could not have been concealed in the room, as there is no concealment possible. When, then, did he come?”
Arthur Conan Doyle, The Sign of the Four, 1890