Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!mnetor!seismo!rutgers!clyde!bellcore!ulysses!allegra!alice!bs
From: bs@alice.UUCP
Newsgroups: comp.lang.c++
Subject: Re: cfront: internal error
Message-ID: <6348@alice.UUCP>
Date: Sat, 15-Nov-86 11:05:44 EST
Article-I.D.: alice.6348
Posted: Sat Nov 15 11:05:44 1986
Date-Received: Mon, 17-Nov-86 08:56:13 EST
References: <644@comp.lancs.ac.uk>
Organization: AT&T Bell Laboratories, Murray Hill NJ
Lines: 23
Summary: Please be specific



stephen @ Department of Computing at Lancaster University, UK. writes:

> One of my colleagues had a program which caused cfront to fail with an internal
> error as its first error message.  The problem was caused by the fact that he
> put a ; after every }.  Now, in my view, cfront should make up its mind either
> to accept this or not.

(1) Clearly internal errors should not happen so there is a cfront bug.

(2) Cfront should implement C++, not ``make its mind up'' about what to accept.

(3) There are (confusingly enough) cases in C++ (as in C) where a ; after a }
	is illegal. For example:

	      ; if (a) { f1(); g1(); }; else { f2(); g2(); };
				      ^			    ^
			syntax error--^                ok --^

(3) This message is not precise enough to allow anyone to avoid or fix the
	problem. Please give us the benefit of your experienc; please don't
	just tease us with ``there is a bug somewhere''.