c++/1632: Invalid and incomprehensible error message

c++/1632: Invalid and incomprehensible error message

[wolfgang.bangerth]

>Number:         1632
>Category:       c++
>Synopsis:       Invalid and incomprehensible error message
>Confidential:   no
>Severity:       serious
>Priority:       medium
>Responsible:    unassigned
>State:          open
>Class:          rejects-legal
>Submitter-Id:   net
>Arrival-Date:   Fri Jan 12 08:16:02 PST 2001
>Closed-Date:
>Last-Modified:
>Originator:     Wolfgang Bangerth
>Release:        unknown-1.0
>Organization:
>Environment:
Yesterday's CVS (2000/01/11)
>Description:
With yesterday's CVS I got the following message:
deal.II/deal.II> cat t.cc
  struct X;

  template <int dim> struct Y {
      X* f (int i = 0) const;
  };

  template <>
  X* Y<2>::f (int i) const {
    return f(i);
  };
deal.II/deal.II> c++ -c t.cc -o /dev/null
  t.cc: In method `X* Y<dim>::f(int) const [with int dim = 2]':
  t.cc:4: prior parameter's size depends on `const Y<2>* const this'

Besides the fact that I have no clue what the compiler wants to
tell me (what is that "prior parameter"?), the message is bogus
and goes away if (among other possibilities) I
- remove the default value of "i"
- remove the "const"ness of the function.
Seems as if something's thoroughly wrong.

Regards
  Wolfgang
>How-To-Repeat:

>Fix:

>Release-Note:
>Audit-Trail:
>Unformatted:
>From ritzert@linux.globe-tec.de Sun Apr 01 00:00:00 2001
From: Michael Ritzert <ritzert@linux.globe-tec.de>
To: nobody@gcc.gnu.org
Cc: gcc-prs@gcc.gnu.org
Subject: Re: c++/1683: failure during bootstrap (current CVS)
Date: Sun, 01 Apr 2001 00:00:00 -0000
Message-id: <20010117201600.2189.qmail@sourceware.cygnus.com>
X-SW-Source: 2001-q1/msg00424.html
Content-length: 296

The following reply was made to PR c++/1683; it has been noted by GNATS.

From: Michael Ritzert <ritzert@linux.globe-tec.de>
To: gcc-gnats@gcc.gnu.org, nobody@gcc.gnu.org
Cc:  
Subject: Re: c++/1683: failure during bootstrap (current CVS)
Date: Wed, 17 Jan 2001 21:07:42 +0100

 Working again.
 
>From nathan@gcc.gnu.org Sun Apr 01 00:00:00 2001
From: nathan@gcc.gnu.org
To: nobody@gcc.gnu.org
Cc: gcc-prs@gcc.gnu.org
Subject: Re: c++/643
Date: Sun, 01 Apr 2001 00:00:00 -0000
Message-id: <20010227103602.13142.qmail@sourceware.cygnus.com>
X-SW-Source: 2001-q1/msg01784.html
Content-length: 969

The following reply was made to PR c++/643; it has been noted by GNATS.

From: nathan@gcc.gnu.org
To: gcc-gnats@gcc.gnu.org, mbrubake@bucknell.edu, nobody@gcc.gnu.org
Cc:  
Subject: Re: c++/643
Date: 27 Feb 2001 10:29:55 -0000

 Synopsis: Not using the best matches when multiple functions with different parameters are available
 
 State-Changed-From-To: open->closed
 State-Changed-By: nathan
 State-Changed-When: Tue Feb 27 02:29:55 2001
 State-Changed-Why:
     not a bug. The choice is ambiguous (and -pedantic will tell
     you this). but there is a gnu extension which accepts
     rejects the ambiguous choice with the worst conversion.
     In this case the choice for the first call is
     an integral conversion (signed -> unsigned), and bool
     conversion (pointer->bool). The other choice has
     a user conversion (string constant-> string), which is
     worse than the other.
 
 http://gcc.gnu.org/cgi-bin/gnatsweb.pl?cmd=view&pr=643&database=gcc