>Number:         5959
>Category:       standards
>Synopsis:       compiler allows something language probably shouldn't?
>Confidential:   no
>Severity:       serious
>Priority:       low
>Responsible:    gnats-admin (GNATS administrator)
>State:          open
>Class:          sw-bug
>Submitter-Id:   net
>Arrival-Date:   Wed Aug 12 16:35:01 1998
>Last-Modified:
>Originator:     Andrew Brown
>Organization:
none
>Release:        1.2 (yeah, i know, i gotta upgrade)
>Environment:

System: NetBSD noc 1.2 NetBSD 1.2 (null) #4: Fri Oct 24 01:27:28 EDT 1997 root@noc:/usr/src/sys/arch/i386/compile/null i386

>Description:

in c and c++ code, rvalues cannot be assigned to, only lvalues
can. ie:

	/* "6" is the rvalue (obviously), "foo" is the lvalue (since
	it's on the left), and the entire statement yields an rvalue
	(since you can assign from it, but not to it) */

	int foo;
	foo=6;

an "assignment" usually yields an rvalue (unless you write a c++ class
that returns a non-const reference from operator=(), which is yet
another way to make your code confusing :), ie:

	int foo, bar;
	foo=bar=6;

is legal code.  c++ introduces the concept of references as a way of
getting around using pointers.  they're kinda neat, but also kinda
confusing (cf. http://www.daemon.org/c++-hoax.html).  i was under the
(possible mistaken) impression that rvalues could not be passed by
reference, but i just noticed that i'm doing it quite frequently.
ie:

	void function(int &t) { t=4; }
	int foo;
	...
	function(foo=3);
	printf("%d\n",foo);

will yield "4" as the output, in effect, assigning to an rvalue.  and
int is certainly not a c++ class.

is this the correct behavior?  it doesn't strike me as such, from what
i know of the language.  perhaps it's a "feature" in the language?  or
a "bug" in the compiler?

>How-To-Repeat:

write code that does this thing (that might be wrong?).

>Fix:

contact the people who make the compiler?  or define the language?  or
not?  just please set my mind at rest?
>Audit-Trail:
>Unformatted: