Re: Is the dangling else a syntax bug?

Actually COBOL is not LALR(k) for any k, because of issues in the READ
statement where the syntax depends on the organization of the file.


EJP


David Thompson wrote:
> > However, in natural languages, omitting a comma is rarely a cause for
> > misunderstanding, whereas in programming languages it is almost always
> > considered a syntactic error. Few languages make the semicolon
> > optional (I don't consider line-oriented languages in that group); ...
>
> If we consider any statement terminator not just semicolon, FWIW COBOL
> uses period followed by a space to terminate a sequence of one or more
> statements, which is almost isomorphic to terminating a statement but
> being omittable on all but the last in a "sentence" or paragraph.
> (COBOL85 adds END-IF etc. to specifically terminate nestable
> statements, but still allows the old way.) But all statements begin
> (at least) with a reserved keyword avoiding any ambiguity; in fact it
> wouldn't surprise me if COBOL is LL(1), though I haven't checked. And
> of course it tried unusually hard to "look like" natural language, at
> least in PROCEDURE DIVISION.

Return to the comp.compilers page.
Search the comp.compilers archives again.