Date: Sun, 1 Mar 1998 16:18:33 -0800 (PST)
Message-Id: <199803020018.QAA20629@adit.ap.net>
Mime-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
To: Ralph Gesler <rgesler AT pacificnet DOT net>
From: Nate Eldredge <eldredge AT ap DOT net>
Subject: Re: Problem with C++ and function
Cc: Steve Marton <istvan AT outer-net DOT com>, djgpp AT delorie DOT com
Precedence: bulk

At 10:38  3/1/1998 -0800, Ralph Gesler wrote:
>
>
>Nate Eldredge wrote:
>
>> At 07:12  2/27/1998 GMT, Steve Marton wrote:
>> >I'm trying to point to a member function of a class. But for some
>> >reason, its address &Class::f is always 1. Even in Class::f1(){&f==1}.
>> >That's the same for any member function. How come? How can I point to
>> >it?
>>
>> Did you find this out by doing something like `cout << &f'? If so, the
>> problem is not in the address at all, but is a known bug in the routine that
>> prints it out. Try `cout << (unsigned)&f', or actually check the result of
>> `&f == (void *)1' in the code.
>>
>> Nate Eldredge
>> eldredge AT ap DOT net
>
>Thanks for your response to the original message. I was having the exact
problem
>you alluded to, and your work-around was the answer.  Is there somewhere that a
>list of known bugs in DJGPP and/or the libraries can be found? I checked have
>checked the DJGPP info and no known bugs are listed.

The DJGPP web site has a bug list, I believe. Look at:

http://www.delorie.com/djgpp/

The `cout' bug will go away with the release of gcc 2.8 (the DJGPP version
is in alpha now) and its accompanying libstdc++.

Nate Eldredge
eldredge AT ap DOT net