From mboxrd@z Thu Jan  1 00:00:00 1970
Return-Path: <gcc-prs-return-16819-listarch-gcc-prs=gcc.gnu.org@gcc.gnu.org>
Received: (qmail 5776 invoked by alias); 4 Apr 2002 02:49:49 -0000
Mailing-List: contact gcc-prs-help@gcc.gnu.org; run by ezmlm
Precedence: bulk
List-Archive: <http://gcc.gnu.org/ml/gcc-prs/>
List-Post: <mailto:gcc-prs@gcc.gnu.org>
List-Help: <mailto:gcc-prs-help@gcc.gnu.org>
Sender: gcc-prs-owner@gcc.gnu.org
Received: (qmail 5740 invoked by uid 61); 4 Apr 2002 02:49:47 -0000
Date: Wed, 03 Apr 2002 18:49:00 -0000
Message-ID: <20020404024947.5739.qmail@sources.redhat.com>
To: alian@cpan.org, bkoz@gcc.gnu.org, gcc-bugs@gcc.gnu.org,
  gcc-prs@gcc.gnu.org, nobody@gcc.gnu.org
From: bkoz@gcc.gnu.org
Reply-To: bkoz@gcc.gnu.org, alian@cpan.org, bkoz@gcc.gnu.org,
  gcc-bugs@gcc.gnu.org, gcc-prs@gcc.gnu.org, nobody@gcc.gnu.org,
  gcc-gnats@gcc.gnu.org
Subject: Re: libstdc++/6062: Coredump when use dlsym with C++ objects with gcc3
X-SW-Source: 2002-04/txt/msg00282.txt.bz2
List-Id: <gcc-prs.sourceware.org>

Synopsis: Coredump when use dlsym with C++ objects with gcc3

Responsible-Changed-From-To: unassigned->bkoz
Responsible-Changed-By: bkoz
Responsible-Changed-When: Wed Apr  3 18:49:47 2002
Responsible-Changed-Why:
    Mine.
State-Changed-From-To: open->closed
State-Changed-By: bkoz
State-Changed-When: Wed Apr  3 18:49:47 2002
State-Changed-Why:
    This is not a bug. Instead, there are several errors in both the C++ code and the usage of the dynamic loader. I'm including the fixes sources, output from gcc-3.1 on linux, and the modified shell script used to build it.
    
    best,
    benjamin
    
    build.sh:
    c++ -rdynamic -g -fPIC -Wall -shared tc_lib2.cc -o libtc2.so 
    c++ -rdynamic -g -fPIC -Wall tc_prog.cc -L. -ltc2 -ldl -o prog
    
    
    #include <stdio.h>
    
    namespace TestN 
    {
      class Interface 
        {
        public:
          void fcn2(void); 
        };
    }
    
    tc_lib2.cc:
    
    #include "tc_lib2.h"
    
    extern "C" 
    {
      TestN::Interface*
      InstanceFactory()
      { return new TestN::Interface(); }
    }
    
    namespace TestN 
    {
      void 
      Interface::fcn2(void) 
      { printf("in fcn2\n"); }
    }
    
    
    tc_prog.cc:
    #include <dlfcn.h>
    #include <stdio.h>
    #include <stdlib.h>
    #include <iostream.h>
    #include "tc_lib2.h"
    
    int main() 
    {
      TestN::Interface* instance;
      void *lib;
      
      cout << "in fcn1, dlopening lib2..." << endl;
    
      //  int flag = DL_LAZY; // bsd
      // int flag = RTLD_NOW; // linux
      int flag = RTLD_LAZY; // linux try 2
      lib = dlopen("libtc2.so", flag); 
      if (!lib)
        {
          cout << "can't dlopen lib2: " << dlerror() << endl;
          exit(2);
        }
      
      printf("dlsymming instance...\n");
      instance = (TestN::Interface*) dlsym(lib, "InstanceFactory");
      const char* msg = dlerror();
      if (msg != NULL) 
        {
          cout << "can't dlsym instance: " << msg << endl;
          exit(3);
        }
      
      printf("calling fcn2...\n");
      instance->fcn2();
      printf("back from fcn2\n");
    
      dlclose(lib);
      return 0;
    }
    
    
    
    %./prog
    in fcn1, dlopening lib2...
    dlsymming instance...
    calling fcn2...
    in fcn2
    back from fcn2
    
    
    I'm not quite sure why this was coring for you before. At first I thought it might be that NULL string, but of course, that was not it as...
    
    
    
    #include <iostream.h>
    
    int main()
    {
      cout << "testing..." << NULL << endl;
      return 0;
    }
    
    
    is totally fine.
    
    At this point, I don't really care either.
    
    best,
    benjamin

http://gcc.gnu.org/cgi-bin/gnatsweb.pl?cmd=view%20audit-trail&database=gcc&pr=6062