Re: gnubol: "Illegal" Next Sentence

[RKRayhawk@aol.com]

In a message dated 12/6/99 8:58:43 PM EST, wmklein@ix.netcom.com writes:

<< The topic of "illegal" Next Sentence (i.e. a Next Sentence in an IF with an
 END-IF or a SEARCH with an END-SEARCH) is one of those areas that gets into
 "religious wars" among COBOL people. >>

If we resolve these specific situations to a unique error code, and 
externalize the severity level, then we can address the divergent code base 
issues with user modifiable error/diagnostic libraries.

Initial gens of these libraries can reflect our best guess of standard 
compliance combined with preferences where allowed.  Alternate gens of these 
error/diagnostic libraries might provide certain obvious alternatives, like 
most liberal on all dicey issues, and most conservative on those issues.  The 
file could be modifiable. 

If we project a diagnostic library that say 
  nnnn-W NEXT SENTENCE okay but you are warned
the user could replace it with simply one change
  nnnn-E NEXT SENTENCE okay but you are warned
or they could rescript the text as well, as
  nnnn-E NEXT SENTENCE not accepted here

We could anticipate some of these areas of special interest in the field and 
gloss the text to support easy change of the -W/-E without changing the text.

Syntax and semantics only would know the nnnn error number. If they need the 
-E and -W, they need to hear it from the error/diagnostics function.

This applies to recoverable situations. Generally syntax should pass 
everything to semantics anyway. But reinforcement generally will lead to 
syntax accessing the -W/-E attribute to tag the construct and to commute the 
error condition up in recurse rule structures.

Bob Rayhawk

--
This message was sent through the gnu-cobol mailing list.  To remove yourself
from this mailing list, send a message to majordomo@lusars.net with the
words "unsubscribe gnu-cobol" in the message body.  For more information on
the GNU COBOL project, send mail to gnu-cobol-owner@lusars.net.