Subject: Re: Bug in g++?
To: None <firstname.lastname@example.org>
From: David Gilbert <email@example.com>
Date: 12/25/1996 11:19:27
>>>>> "Paul" == Paul Boven <firstname.lastname@example.org> writes:
Paul> Hi everyone, Yes, I am *still* trying to build octave for the
Paul> Sparc-NetBSD. It compiles with hardly a problem, but the linker
Paul> doesn't like things at all. This weekend I have access to an
Paul> intel-machine running NetBSD-1.2, just like I do. Although t
Paul> doesn't have swapspace and diskspace enough on it to finish
Paul> things, I have been able to compare object-files. I don't know
Paul> alas if this NetBSD-version would yield better results because
Paul> of these differences.
I've just gotten through build Hylafax-4p1, and it was quite a
dredful experience. Firstly, there were a number of places where the
right include files or symbols were not defined for NetBSD (I will be
back-submitting all that stuff), but one thing that I can't pawn off
on them is the ___pure_virutal problem.
At the end of this message is a chunk of C++ code that when
compiled (gcc -o pure pure.c++) will come up with a symbol that is not
defined '___pure_virtual' with the stock NetBSD compiler. I'm using a
late 1.1-current here.
Now... I also have my own install of GCC (because I want to
play with fcc et. al.) and when I compile the same code with it,
everything works as expected.
virtual void doit(int, int) = 0;
class child : public parent
virtual void doit(int, int);
void child::doit(int a, int b)
i1 = a;
i2 = b;
parent *bar = &foo;
|David Gilbert, PCI, Richmond Hill, Ontario. | Two things can only be |
|Mail: email@example.com | equal if and only if they |
|http://www.pci.on.ca/~dgilbert | are precisely opposite. |