[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: glibc & libio
From: |
Marcus Brinkmann |
Subject: |
Re: glibc & libio |
Date: |
Sun, 20 Jan 2002 01:39:55 +0100 |
User-agent: |
Mutt/1.3.25i |
On Sat, Jan 19, 2002 at 05:56:06PM -0500, Roland McGrath wrote:
> The build didn't complain, but looking at the binaries they have undefined
> dynamic symbols for
>
> __register_frame_info
> __register_frame_info_bases
> __deregister_frame_info
> __deregister_frame_info_bases
>
> and maybe a couple others, that don't have any symbol version assignment
> and I can't figure out where they're defined.
>
> But my gcc-3 cross-compiler build may be bad or incomplete too.
I have a build on stampede now, in the subhurd. I see the same. The
functions are in libgcc.a. Those functions seem to be the new Dwarf2
support that this whole junk is about. I take this from the configure
check (see configure.in, libc_cv_gcc_dwarf2_unwind_info).
As far as I grok it, the framestate.c stuff is for the case where libgcc.a
has not the __register_frame_info interface, and as such is used to support
older gcc's (2.95.x). If we decide to drop gcc 2.95 and just work with gcc 3,
I think we are fine with just leaving framestate out. At least that is what
I figure from the data, maybe we can get Jakub to verify this for us.
Then it would be a policy decision if we wait for someone to fix the
troublesome case gcc 2.95.x + oldest abi 2.2.5, or if we just go for gcc 3.
Thanks,
Marcus
--
`Rhubarb is no Egyptian god.' Debian http://www.debian.org address@hidden
Marcus Brinkmann GNU http://www.gnu.org address@hidden
address@hidden
http://www.marcus-brinkmann.de