This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Re: IA32 bootstrap error
- From: Richard Henderson <rth at redhat dot com>
- To: Richard Kenner <kenner at vlsi1 dot ultra dot nyu dot edu>
- Cc: gcc at gcc dot gnu dot org
- Date: Sun, 30 Dec 2001 16:01:15 -0800
- Subject: Re: IA32 bootstrap error
- References: <10112302303.AA07030@vlsi1.ultra.nyu.edu>
On Sun, Dec 30, 2001 at 06:03:51PM -0500, Richard Kenner wrote:
> It's one thing for it to be required that they *work together*, but a
> vastly different one that the latest version of each is required for
> GCC. We always used to require that GCC work with an older version of
> each of these tools and I think that's still critical.
GCC will still work with older versions of binutils, and I agree
that GCC should continue to at least bootstrap with them.
GDB is a slightly different issue, in that it is not required to
bootstrap. For instance, I do not think we should work around bugs
in older versions of GDB. Essentially, I think that's what the
dwarf2 issue you are experiencing is -- a failure to implement the
entire dwarf2 spec.
And as for the specific question at hand, "Should we have switched
to default to dwarf2 for elf," I would even consider this a special
case in that we *absolutely must* have this work for gcc 3.1. The
new libstdc++ simply cannot be debugged with stabs, as stabs cannot
represent namespaces.
> I think the SC needs to address this issue.
Perhaps.
> ... and not everybody who is on workstations is in a position to
> constantly get the "latest and greatest" of everything.
I wouldn't say "constantly". I update gdb irregularly. Mostly
when I encounter problems and want to see if it's been fixed in
the meantime. In practice, about once a month.
r~