This is the mail archive of the
binutils@sourceware.cygnus.com
mailing list for the binutils project.
Re: A proposed patch for undefined_symbol ()
- To: Koundinya K <knk at dde dot dk>
- Subject: Re: A proposed patch for undefined_symbol ()
- From: "H . J . Lu" <hjl at valinux dot com>
- Date: Wed, 1 Mar 2000 10:35:53 -0800
- Cc: binutils at sourceware dot cygnus dot com
- References: <20000301095946.A8291@valinux.com> <38BD7C2A.A1EF668C@dde.dk>
On Wed, Mar 01, 2000 at 01:23:06PM -0700, Koundinya K wrote:
> Hi hjl,
> Normally when -Bsymbolic is uses when building a shared object (library) , the
> linker should bind references to global symbols to their definitions within the
> object, if defintions are available. Normally, since references to global
> symbols within shared objects are not bound until run time, even if
> definitions are available, so that definitions of the same symbol in an
> executable or other shared objects can
> override the object's own definition. The linker should issue warnings for
That is not true for -Bsymbolic. If a definition for a global symbol
is available in DSO, it will be used in DSO no matter what. It is what
-Bsymbolic is used for.
> undefined symbols unless there is an option in the Linker to force a fatal error if
> any undefined symbols remain at
My patch doesn't change the undefined symbol warning with -Bsymbolic
when building DSO. What it does is not to abort if undefined symbol
is ok.
> the end of the link. So I feel we need to have a separate option in the Linker to
> address this.
>
We have --noinhibit-exec to ignore all errors. But I don't
think it is appropriate for -Bsymbolic.
H.J.