This is the mail archive of the
gcc-bugs@gcc.gnu.org
mailing list for the GCC project.
c++/9878: [Regression 3.3/3.4] error: non-lvalue in unary `&' wrongly issued
- From: aj at suse dot de
- To: gcc-gnats at gcc dot gnu dot org
- Cc: matz at suse dot de
- Date: 27 Feb 2003 14:52:42 -0000
- Subject: c++/9878: [Regression 3.3/3.4] error: non-lvalue in unary `&' wrongly issued
- Reply-to: aj at suse dot de
>Number: 9878
>Category: c++
>Synopsis: [Regression 3.3/3.4] error: non-lvalue in unary `&' wrongly issued
>Confidential: no
>Severity: serious
>Priority: medium
>Responsible: unassigned
>State: open
>Class: rejects-legal
>Submitter-Id: net
>Arrival-Date: Thu Feb 27 14:56:00 UTC 2003
>Closed-Date:
>Last-Modified:
>Originator: Andreas Jaeger
>Release: unknown-1.0
>Organization:
>Environment:
Linux x86 and x86-64, gcc 3.3 from 20030215 and GCC 3.4
from 20030215.
>Description:
This small C++ program:
void f(void)
{
short x = 0;
const int &y = x;
}
fails to compile with:
gee:~/tmp:[0]$ /opt/gcc/3.4-devel/bin/g++ -c t.C
t.C: In function `void f()':
t.C:4: error: non-lvalue in unary `&'
gee:~/tmp:[1]$ /opt/gcc/3.3-devel/bin/g++ -c t.C
t.C: In function `void f()':
t.C:4: error: non-lvalue in unary `&'
GCC 3.2 compiles this fine.
IMHO this is valid: according to 8.5.3[#5], the reference is bound to a temporary holding the value of np_ptr_F. Replaceing short by double makes it compile (with two warnings).
>How-To-Repeat:
compile the small test program.
>Fix:
>Release-Note:
>Audit-Trail:
>Unformatted: