Re: libgcj/65: Uninitialized global variables

[Tom Tromey <tromey at cygnus dot com>]

The following reply was made to PR libgcj/65; it has been noted by GNATS.

From: Tom Tromey <tromey@cygnus.com>
To: Patrik Hagglund <patha@ida.liu.se>
Cc: Tom Tromey <tromey@cygnus.com>,
        Java Gnats Server <java-gnats@sourceware.cygnus.com>
Subject: Re: libgcj/65: Uninitialized global variables 
Date: Thu, 7 Oct 1999 15:13:55 -0700

 Please continue to CC java-gnats, so that this information is logged.
 
 >>>>> "Patrik" == Patrik Hagglund <patha@ida.liu.se> writes:
 
 >> This is the famous Solaris bug.
 
 Patrik> Wait a minute. Do you mean that semantics of the code is
 Patrik> well-defined?
 
 I believe so.  Those variables should be initialized before main() is
 run -- at the same time C++ static constructors are run (since this is
 what they are).
 
 They are not run correctly due to a linker bug, or a gcc/linker
 interaction problem (I don't know the details).  Using GNU binutils
 fixes this.
 
 Patrik> binutils 2.9.1 has known bugs on this platform. We recommend
 Patrik> to use the vendor tools (Sun as, Sun ld) until these have been
 Patrik> fixed.
 
 I don't know what version other people are using, but we have heard
 reports that using the binutils "ld" fixes the problem.  Perhaps
 people are using a binutils snapshot.
 
 Tom