Re: Please requeue axiom
- To: Kurt Roeckx <Q@ping.be>
- Cc: email@example.com
- Subject: Re: Please requeue axiom
- From: Camm Maguire <firstname.lastname@example.org>
- Date: 07 Oct 2004 16:58:53 -0400
- Message-id: <email@example.com>
- In-reply-to: <20040909163547.GA6975@ping.be>
- References: <E1C4keG-0002nLfirstname.lastname@example.org> <20040907201410.GA892@ping.be> <email@example.com> <20040908221040.GA4800@ping.be> <firstname.lastname@example.org> <20040909163547.GA6975@ping.be>
Greetings, and please excuse my late reply.
Kurt Roeckx <Q@ping.be> writes:
> On Wed, Sep 08, 2004 at 06:17:15PM -0400, Camm Maguire wrote:
> > OK, thanks! Would be interested if each of your failed attempts was
> > with gcl-2.6.5-1, and if you succeed now with gcl 2.6.5-2. If so,
> > then there is a problem with binutils 2.15 on amd64.
> It build properly using 2.6.5-2 and it's now uploaded. I've
> tried again with 2.6.5-1 too, and it failed.
> What would be wrong with binutils in that case? Could you file a
> bug report for it?
The only difference between 2.6.5-1 and 2.6.5-2 is that the former
uses the system bfd library, and the latter compiles a local copy from
the gcl source (pre 2.15). gcl uses
bfd_get_relocated_section_contents to relocate compiled lisp objects
into its running image, and bfd hash functions to get symbol addresses
from its saved image. I can't recall the precise location of the
original failure, but one of these is implicated. Should be
straightforward to find with gdb and a local -g compiled copy of
binutils-2.15, but I'm running short of time right now.
Camm Maguire email@example.com
"The earth is but one country, and mankind its citizens." -- Baha'u'llah