Re: bug

["Martin v. Loewis" <martin AT mira.isdn.cs.tu-berlin.de>]

> i have a amdk6-2 3dnow! 400Mhz and i experienced problesm compiling my SuSE
> 6.0 kernel.
> make suddenly stops with several different errors.
> i also had problems with the processor under windows.
> when i slow the speed down to 300Mhz it works just fine!
> what could i do?

I think you should return the computer to the vendor, and have it
replaced with one that actually works.

> does the new version of gcc help?

Unlikely.

Regards,
Martin
>From martinATmira.isdn.cs.tu-berlin.de Mon Sep 13 23:45:00 1999
From: "Martin v. Loewis" <martin AT mira.isdn.cs.tu-berlin.de>
To: andrew.bell AT bigfoot.com
Cc: bug-gcc AT gcc.gnu.org
Subject: Re: Virtual Mem Func Ptr Call Fails When Base Class Has No Vtable
Date: Mon, 13 Sep 1999 23:45:00 -0000
Message-id: <199909140637.IAA20455@mira.isdn.cs.tu-berlin.de>
References: <99091108362702.28114@bellfarm>
X-SW-Source: 1999-09/msg00378.html
Content-length: 860

> The following causes various faults (I've seen SEGV, BUS, ILL) on my system. 
> If the function Sub::test is non-virtual, or Base contains some virtual
> function, all is fine:

Thanks for your bug report. The bug is in your code, however: Without
the cast

   BaseMemCb cb = (BaseMemCb) &Sub::test;

g++ says (correctly)

a.cc: In function `int main()':
a.cc:21: converting `void (Sub::*)()' to `void (Base::*)()' is a contravariance violation
a.cc:21: conversion to `void (Base::*)()' from `void (Sub::*)()'
a.cc:21: type `Sub' is not a base type for type `Base'
a.cc:21:    in pointer to member conversion

Please see your C++ book on why this cast is invalid.

Strictly speaking, this is a reinterpret_cast; using the pointer
results in undefined behaviour. g++ choses to crash the code, which is
conforming to the C++ standard.

Hope this helps,
Martin
>From martinATmira.isdn.cs.tu-berlin.de Mon Sep 13 23:51:00 1999
From: "Martin v. Loewis" <martin AT mira.isdn.cs.tu-berlin.de>
To: oliva AT lsd.ic.unicamp.br
Cc: Stefan.Eilemann AT simpack.de, bug-gcc AT gnu.org
Subject: Re: Internal compiler error
Date: Mon, 13 Sep 1999 23:51:00 -0000
Message-id: <199909140648.IAA20509@mira.isdn.cs.tu-berlin.de>
References: <XFMail.990913124632.Stefan.Eilemann@simpack.de> <199909140627.IAA20402@mira.isdn.cs.tu-berlin.de> <orn1uq9eu9.fsf@cupuacu.lsd.ic.unicamp.br>
X-SW-Source: 1999-09/msg00379.html
Content-length: 310

> > without preprocessor output, there is not much we can do.
> 
> But there was an attachment with preprocessor output bzip2ed.

Yes, I was replying to the wrong report. On this report, I can't
comment much - it seems to be architecture specific, as I couldn't
reproduce it on i386.

Sorry about that,
Martin
>From olivaATlsd.ic.unicamp.br Tue Sep 14 00:03:00 1999
From: Alexandre Oliva <oliva AT lsd.ic.unicamp.br>
To: "Martin v. Loewis" <martin AT mira.isdn.cs.tu-berlin.de>
Cc: andrew.bell AT bigfoot.com, bug-gcc AT gcc.gnu.org
Subject: Re: Virtual Mem Func Ptr Call Fails When Base Class Has No Vtable
Date: Tue, 14 Sep 1999 00:03:00 -0000
Message-id: <oriu5e9dqq.fsf@cupuacu.lsd.ic.unicamp.br>
References: <99091108362702.28114@bellfarm> <199909140637.IAA20455@mira.isdn.cs.tu-berlin.de>
X-SW-Source: 1999-09/msg00380.html
Content-length: 2111

On Sep 14, 1999, "Martin v. Loewis" <martin@mira.isdn.cs.tu-berlin.de> wrote:

> Without the cast

>    BaseMemCb cb = (BaseMemCb) &Sub::test;

> g++ says (correctly)

> a.cc:21: converting `void (Sub::*)()' to `void (Base::*)()' is a contravariance violation

> Please see your C++ book on why this cast is invalid.

Err...  The C++ Standard says [expr.static.cast]:

9 An rvalue of type "pointer to member of D of type cv1 T" can  be  con-
  verted  to  an  rvalue of type "pointer to member of B of type cv2 T",
  where B is a base class (clause _class.derived_)  of  D,  if  a  valid
  standard  conversion  from  "pointer  to  member  of  B  of type T" to
  "pointer to member of D of type T" exists (_conv.mem_), and cv2 is the
  same cv-qualification as, or greater  cv-qualification  than,  cv1.11)
  The  null  member  pointer value (_conv.mem_) is converted to the null
  member pointer value of the destination type.  If class B contains the
  original member, or is a base or derived class of the class containing
  the original member, the resulting pointer to  member  points  to  the
  original  member.   Otherwise,  the  result  of the cast is undefined.
  [Note: although class B need not  contain  the  original  member,  the
  dynamic  type of the object on which the pointer to member is derefer-
  enced must contain the original member; see _expr.mptr.oper_.  ]

Since the class B is a base of the class containing the original
member, the old-style cast should succeed as a static_cast, and should
not invoke undefined behavior.


I seem to recall some wording in the Standard (or some earlier WP)
that would make the *use* of this pointer-to-member undefined unless
it was declared as virtual in the case class, but I can't find it any
more.  Can anybody?  Or was it removed, and the code should work?

-- 
Alexandre Oliva http://www.dcc.unicamp.br/~oliva IC-Unicamp, Bra[sz]il
oliva@{dcc.unicamp.br,guarana.{org,com}} aoliva@{acm.org,computer.org}
oliva@{gnu.org,kaffe.org,{egcs,sourceware}.cygnus.com,samba.org}
** I may forward mail about projects to mailing lists; please use them