From:	Riddler02 AT aol DOT com
Message-ID:	<7a.178f849.25d6f9ab@aol.com>
Date:	Sat, 12 Feb 2000 13:00:11 EST
Subject:	Linking to my library STILL sucks!
To:	djgpp AT delorie DOT com
MIME-Version:	1.0
X-Mailer:	AOL 4.0 for Windows 95 sub 38
Reply-To:	djgpp AT delorie DOT com

itmiller AT dera DOT gov DOT uk (Ian Miller) writes:

> Have you tried writing a *really* simple library? Containing one class with
> just a constructor and destructor that both just write to cout, say. Then
> you could try to link a tiny program to that library and at least confirm
> that you can do this at all.

I just tried this out.  here's what it looks like:

LIBTEST.A consists of 2 files, a header and a source:
    //TEST.H
    class test_class
    {
    void test_message(void);
    };

    //TEST.CPP
    #include <iostream.h>
    #include "test.h"
    void test_class::test_message(void)
    {
    cout << "My library doesn't work!";
    }

This library builds correctly.  Now here is my simple library-testing program 
("test" is checked off in RHIDE's library menu):

    //TESTPROG.CPP
    #include <iostream.h>
    #include "test.h"

    void main()
    {
    test_class test_object;
    test_object.test_message();
    }

This program compiles correctly with the command:   
    gcc -g -c testprog.cpp -o testprog.o

And LINKS correctly too with the command:
    gcc -o testprog.exe testprog.o -ltest -lstdcx
(note that "-lstdcx" SHOULD say "-lstdcxx", but I changed the name to be 1 
character shorter to deal with the long file name problem - it works fine now 
though)

So now I know that a simple library will work - there must be something wrong 
with my code.  nick AT arcticmail DOT com suggested that the order in which the 
libraries are linked is important - object files, user libraries, then the 
standard libraries, but I checked and this is the order I do it in.  Anybody 
have any suggestions on where to start looking in my code? :)  thanx again 
for all your support

-Kevin
    

- Raw text -