Re: A patch for default version and archive

["H . J . Lu" <hjl at valinux dot com>]

On Mon, Nov 13, 2000 at 05:41:51PM -0800, Ulrich Drepper wrote:
> Nick Clifton <nickc@redhat.com> writes:
> 
> > Uli - do you still have any objections to this patch ?  If so, please
> > could you explain what they are ?
> 
> The whole concept of adding "versioning" information in .o files is
> against the design.  What you do is give the functions different names
> which can be done cleaner with the asm() renaming in prototypes.
> Versioning is there to record and enforce dependencies between DSOs
> and binaries.  Using the same mechanism for something else is
> misleading and possibly dangerous if it cannot be maintained in the
> same way if something is changes.  This is simply not thought through
> so that one could be sure no problems arise if this hack is not used.
> 

They are 2 separate issues here. The one you were referring to here
looks like my symbol versioning archive proposal to glibc, which has
nothing to do with binutils. I will send an email to the glibc mailing
list later.

Let's just address binutils here. The symbol versioning should always
be used with care regardless if my patch is checked in or not. My
point is the current linker doesn't do what is expected. If people
think symbol versioning in an archive is desirable, this small
problem certainly won't stop them.


-- 
H.J. Lu (hjl@valinux.com)