RE: [open-cobol-list] discussion: implemention of truncation COBOL-WORD to 30 (31) chars ?

[William M. Klein]

IBM issues (I think) the message

  IGYXX0025-E     Name "?" was invalid.  It was processed as "?".

Like Micro Focus, this is an ERROR message, not a warning (or informational)
message.  In my opinion, this should not be "accepted" in either STD, MF, or
IBM modes (i.e. like the Standard and MF and IBM it should issue an ERROR
message) 

> -----Original Message-----
> From: address@hidden 
> [mailto:address@hidden On 
> Behalf Of Thomas Biehler
> Sent: Tuesday, March 30, 2004 9:07 AM
> To: address@hidden
> Cc: Keisuke Nishida
> Subject: [open-cobol-list] discussion: implemention of 
> truncation COBOL-WORD to 30 (31) chars ?
> 
> Hi all,
> 
> i would like to know what the member of this mailinglist
> think about a (automatic; implicit) truncation
> of a to long COBOL-WORD (identifier ...)
> to the maximal allowed length from the standard. 
> 
> for the maximal allowed length of a COBOL-WORD see in the standard:
>   -  ANSI INCITS 23-1985 (R2001)  § 4.2.2.1 ==>  30 Characters
>   -  ISO/IEC 2002(E) §8.3.1.1  ==> 31 Characters.
> 
> The truncation is Micro-Focus "normal" behavior  (for the 
> 1985 - standard!) .
> (1006-E  --> Error ---> automatic correctur ,
>   see the following excerpt from the error message guide )
> 
> Micro Focus:
> ======================================================
> 1006  COBOL word contains more than 30 characters. Word truncated
> 
>       The name that you have specified is longer than 30 characters.
>       Your COBOL system treats this as a name consisting 
>       of the first 30 characters of your original name
> ======================================================. 
> 
> I have attached a patch to the current cvs-version from pplex.l 
> which truncates to 30 chars (1985 standard!) as a proof of concept.
> 
> I am porting a "MF-application", which contains a small amount 
> of to long identifiers. Changing the non-allowed identifiers 
> / programs 
> would not generate big effort to me! 
> I could also change the MF-compiler configuration!
>   (=> CHANGE-MESSAGE ...  1006-E to 1006-S --> severe error!)
> 
> How the too long identifier come into existence ?
>  -->   the "to long identifier" are generated from a database; 
>         with a additional prefix they are to long!
> 
> I really want to know what other cobol users think about this. 
> (And perhaps needs, if anybody also port a "Micro-Focus Application")
> 
> You can see the attached patch is very small! 
> Easy to implement!
> 
> Should this behavior be implemented in open-cobol? 
> In a similar way?  (with a warning ?)
> As a MF-Cobol behavior only? (with a special configuration option ?!)
> 
> Or are there any good reasons against this?
> (except the "nonconforming standard" argument)
> 
> Comments and suggestions are welcome!
> 
> Thomas
>