Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!know!news.cs.indiana.edu!att!ucbvax!dog.ee.lbl.gov!elf.ee.lbl.gov!torek
From: torek@elf.ee.lbl.gov (Chris Torek)
Newsgroups: comp.lang.c
Subject: ?: syntax (was Problem with #define'd macro...)
Message-ID: <11807@dog.ee.lbl.gov>
Date: 4 Apr 91 22:25:23 GMT
References: <18146@crdgw1.crd.ge.com> <18200@crdgw1.crd.ge.com> <10903@ncar.ucar.edu> <18234@crdgw1.crd.ge.com>
Reply-To: torek@elf.ee.lbl.gov (Chris Torek)
Organization: Lawrence Berkeley Laboratory, Berkeley
Lines: 17
X-Local-Date: Thu, 4 Apr 91 14:25:23 PST

In article <18234@crdgw1.crd.ge.com> volpe@camelback.crd.ge.com
(Christopher R Volpe) writes:
[given	a ? b = 1 : c = 1  or equivalent]
>GCC ... with -ansi [and -pedantic] ... never finds the syntax
>violation.

True, but it does produce a diagnostic, so I think it is doing the
right thing.  It is no surprise that gcc's parser accepts the
expression: gcc uses a single yacc (well, bison) grammar and has
extensions, so it must use some other error for `-ansi -pedantic'
violations that otherwise fit its extensions (as in this case).
I am sure the standard does not require the diagnostic for the
above to be `syntax error'.  `Invalid lvalue in assignment' is
not the best error message, though.
-- 
In-Real-Life: Chris Torek, Lawrence Berkeley Lab CSE/EE (+1 415 486 5427)
Berkeley, CA		Domain:	torek@ee.lbl.gov