libgcj/1443: symbol `__vt_Q34java4lang5Class' is multiply defined

[daniel.sadinoff@gs.com]

>Number:         1443
>Category:       libgcj
>Synopsis:       symbol `__vt_Q34java4lang5Class' is multiply defined
>Confidential:   no
>Severity:       critical
>Priority:       medium
>Responsible:    tromey
>State:          closed
>Class:          sw-bug
>Submitter-Id:   net
>Arrival-Date:   Wed Dec 20 12:19:55 PST 2000
>Closed-Date:    Mon Sep 04 14:50:01 PDT 2000
>Last-Modified:  Mon Sep  4 15:00:00 PDT 2000
>Originator:     Danny Sadinoff
>Release:        libgcj-2.95.1
>Organization:
>Environment:
solaris-2.6, gcc-2.95.1
>Description:
Build gcc-2.95.1 from scratch, then attempted to build libgcj. 

Got the following error during link of libgcj.so: 

make[2]: Entering directory `/export/opt3/sadind-home/build/gcj/sparc-sun-solaris2.6/libjava'
rm -fr .libs/libgcj.la .libs/libgcj.* .libs/libgcj.*
ld: fatal: symbol `__vt_Q34java4lang5Class' is multiply defined:
	(file java/lang/Class.lo and file java/lang/natClass.lo);
ld: fatal: symbol `__vt_Q34java4lang6Object' is multiply defined:
	(file java/lang/Object.lo and file java/lang/natObject.lo);
ld: fatal: File processing errors. No output written to .libs/libgcj.so.0.0.0
>How-To-Repeat:
mkdir builddir; cd builddir; ../libgcj-2.95.1/configure; make
>Fix:

>Release-Note:

>Audit-Trail:

Formerly PR libgcj/63


From: Tom Tromey <tromey@cygnus.com>
To: daniel.sadinoff@gs.com
Cc: java-gnats@sourceware.cygnus.com
Subject: Re: libgcj/63: symbol `__vt_Q34java4lang5Class' is multiply defined
Date: Fri, 8 Oct 1999 08:45:26 -0700

 In my build, "__vt_Q34java4lang5Class" is defined only in Class.o and
 not natClass.o.
 
 Are you using the C++ compiler that you built when you built the Java
 compiler?  If so, can you find out why it is putting the vtable into
 natClass.o?  It shouldn't.
 
 FWIW, I haven't heard of this problem from anybody else.
 
 What version of "make" are you using?
 
 There is a remote possibility, I think, that your "Class.h" and
 "Object.h" files were built by gcjh (most headers we use are generated
 from the .class files, but these two should *not* be).  You can see if
 this is the case by looking for them in the build tree (they should
 not be there).  If the first line of either file says something about
 being machine generated, then this is the problem.
 
 Tom
State-Changed-From-To: open->feedback
State-Changed-By: tromey
State-Changed-When: Sat Apr 22 19:35:31 2000
State-Changed-Why:
    Sorry for the long delay on this PR.
    Did you get my followup message?
    Did it work for you?
    Please reply, even if to tell me you aren't interested.
    I'll just close the PR...

From: tromey@cygnus.com
To: daniel.sadinoff@gs.com, java-gnats@sourceware.cygnus.com,
  tromey@cygnus.com
Cc:  
Subject: Re: libgcj/63
Date: 23 Apr 2000 02:35:32 -0000

 Synopsis: symbol `__vt_Q34java4lang5Class' is multiply defined
 
 State-Changed-From-To: open->feedback
 State-Changed-By: tromey
 State-Changed-When: Sat Apr 22 19:35:31 2000
 State-Changed-Why:
     Sorry for the long delay on this PR.
     Did you get my followup message?
     Did it work for you?
     Please reply, even if to tell me you aren't interested.
     I'll just close the PR...
 
 http://sourceware.cygnus.com/cgi-bin/gnatsweb.pl?cmd=view&pr=63&database=java
State-Changed-From-To: feedback->closed
State-Changed-By: green
State-Changed-When: Mon Sep  4 14:50:01 2000
State-Changed-Why:
    No feedback from Daniel.  Time to close this one.

From: green@cygnus.com
To: daniel.sadinoff@gs.com, java-gnats@sourceware.cygnus.com,
  tromey@cygnus.com
Cc:  
Subject: Re: libgcj/63
Date: 4 Sep 2000 21:50:01 -0000

 Synopsis: symbol `__vt_Q34java4lang5Class' is multiply defined
 
 State-Changed-From-To: feedback->closed
 State-Changed-By: green
 State-Changed-When: Mon Sep  4 14:50:01 2000
 State-Changed-Why:
     No feedback from Daniel.  Time to close this one.
 
 http://sources.redhat.com/cgi-bin/gnatsweb.pl?cmd=view&pr=63&database=java
>Unformatted: