Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10.1 6/24/83; site pur-ee.UUCP
Path: utzoo!linus!decvax!genrad!grkermit!masscomp!clyde!floyd!harpo!ihnp4!inuxc!pur-ee!iuvax!apratt
From: apratt@iuvax.UUCP
Newsgroups: net.micro.pc
Subject: Lattice CXERRs - (nf)
Message-ID: <1362@pur-ee.UUCP>
Date: Wed, 11-Jan-84 22:40:02 EST
Article-I.D.: pur-ee.1362
Posted: Wed Jan 11 22:40:02 1984
Date-Received: Fri, 13-Jan-84 05:39:12 EST
Sender: notes@pur-ee.UUCP
Organization: Electrical Engineering Department , Purdue University
Lines: 27

#N:iuvax:1500005:000:805
iuvax!apratt    Jan 11 17:12:00 1984

<sacrifice>

Lattice C has a compiler error which I intend to report: certain errors with
the conditional operator are not trapped and result in CXERR:18 (I think 18 is
right). The following code will cause the error:
main()
{
	int c;
	c = 1 + (2 ? 1 : );
}

(note the blank field following the :).  Clearly this is a badly-formed
expression, but the fact that it generates a CXERR is (by definition) a
compiler error.

	Another strange item: using the constant 0 caused a CXERR in the second
pass (of an entirely different program), while substituting a variable
(initialized to zero) compiled properly. Has anyone else had this problem? I
haven't studied it thoroughly enough to report on it yet.
			Your friendly neighborhood compiler watchdog...
						-- Allan Pratt
					...ihnp4!inuxc!iuvax!apratt