Re: c++/6331: g++ 3.1 looses const qualifiers

[Jason Merrill <jason@redhat.com>]

The following reply was made to PR c++/6331; it has been noted by GNATS.

From: Jason Merrill <jason@redhat.com>
To: Mark Mitchell <mark@codesourcery.com>
Cc: "gcc-bugs@gcc.gnu.org" <gcc-bugs@gcc.gnu.org>,
	"gcc-gnats@gcc.gnu.org" <gcc-gnats@gcc.gnu.org>
Subject: Re: c++/6331: g++ 3.1 looses const qualifiers
Date: Wed, 24 Apr 2002 18:18:04 +0100

 >>>>> "Mark" == Mark Mitchell <mark@codesourcery.com> writes:
 
 > Shucks, we were duplicating work; I was working on that too.
 
 Oops.  I was guessing that you hadn't started on it yet.
 
 > When I tried the change you made to build_modify_expr, I got a failure in
 > this g++.benjamin/14664-1.c, which tests that we cannot assign a string
 > literal to an array of characters (because the array of characters is
 > const).  This is probably a goofy test; what should be interesting is
 > whether we can assign the element types and it looks like the logic
 > for creating this test was that EDG issues an error -- but about array
 > assignment in general.
 
 I agree.
 
 > Anyhow, did you change the test, or did it still pass for you?
 
 Neither; I failed to notice the failure in my testrun.  Double oops.
 
 The original bug report that the testcase is derived from complained about
 the error; the customer said that either the extension should be dropped or
 the testcase should be accepted.  Neither of those came to pass.
 
 I wouldn't object to making the pedwarn unconditional.
 
 Jason