https://gcc.gnu.org/bugzilla/show_bug.cgi?id=120786
James K. Lowden <jklowden at gcc dot gnu.org> changed:
What |Removed |Added
----------------------------------------------------------------------------
Assignee|unassigned at gcc dot gnu.org |jklowden at gcc dot
gnu.org
Last reconfirmed| |2025-06-23
Status|UNCONFIRMED |ASSIGNED
Ever confirmed|0 |1
--- Comment #1 from James K. Lowden <jklowden at gcc dot gnu.org> ---
Per ISO:
8.3.1 General
The lexical elements are character-strings and separators.
A character-string is a character or a sequence of contiguous characters that
forms a COBOL word, a literal, or a picture character-string. A
character-string is delimited by separators.
"NOT" and "=" are COBOL words; "NOT=" is not. It's interesting that MF accepts
it, because they say they don't.
https://www.microfocus.com/documentation/visual-cobol/vc90/DevHub/HRLHLHPDF60A.html
says, "A space must precede and follow each reserved word comprising the
relational operator."
This PR has two possible resolutions:
1. a better error message, "NOT requires a separating space", and/or
2. accept "NOT=" as "NOT =".
If #2, what is the rule? The parser could accept e.g. NOTEQUAL, which today is
a user-defined word. Is there a non-Latin rule that algebraic operators have
implied separator space? That would mean
NOT<=-1
would be acceptable.
NOT is a versatile word in COBOL. The error message is due to the fact that
the gcobol parser does not recognize "NOT =" as two tokens; rather, the scanner
reads "NOT =" and returns a single NE token. (That arrangement likewise
rejects NOT NOT =).
