Re: cvs commit: src/libexec/rtld-aout Makefile dynamic.hmd-prologue.c rtld.1aout rtld.c shlib.c shlib.h support.c support.h mdprologue.S

From: Giorgos Keramidas <keramida_at_FreeBSD.org>
Date: Tue, 11 Jan 2005 16:50:27 +0200
On 2005-01-11 09:31, Tom Rhodes <trhodes_at_freebsd.org> wrote:
>On Tue, 11 Jan 2005 16:15:14 +0200 Ruslan Ermilov <ru_at_freebsd.org> wrote:
>>On Tue, Jan 11, 2005 at 03:54:11PM +0200, Giorgos Keramidas wrote:
>>> On 2005-01-11 14:17, Lukas Ertl <le_at_FreeBSD.org> wrote:
>>> >On Tue, 11 Jan 2005, Tom Rhodes wrote:
>>> >> Log:
>>> >> Remove a.out runtime linker.  It doesn't build and was removed from the
>>> >> build over two years ago by peter.
>>> >
>>> > This breaks buildworld in sbin/ldconfig.
>>>
>>> A repocopy of shlib.[ch] and support.[ch] from src/libexec/rtld-aout
>>> into src/sbin/ldconfig and the following patch in src/sbin/ldconfig
>>> fixed the build of ldconfig here:
>>>
>>> % Index: Makefile
>>> % ===================================================================
>>> % RCS file: /home/ncvs/src/sbin/ldconfig/Makefile,v
>>> % retrieving revision 1.19
>>> % diff -u -r1.19 Makefile
>>> % --- Makefile    23 Feb 2004 20:13:53 -0000      1.19
>>> % +++ Makefile    11 Jan 2005 13:48:11 -0000
>>> % _at__at_ -2,11 +2,8 _at__at_
>>> %
>>> %  PROG=  ldconfig
>>> %  SRCS=  elfhints.c ldconfig.c shlib.c support.c
>>> % -LDDIR?= ${.CURDIR}/../../libexec/rtld-aout
>>> % -CFLAGS+=-I${LDDIR} -DFREEBSD_AOUT
>>> % +CFLAGS+=-I${.CURDIR}
>>> %  WARNS?=        0
>>> %  MAN=   ldconfig.8
>>> %
>>> % -.PATH: ${LDDIR}
>>> % -
>>> %  .include <bsd.prog.mk>
>>>
>>> Note, however, that I haven't had the chance to test the resulting
>>> ldconfig binary.  I need my workstation to do other work and cannot boot
>>> it into a test install.
>>
>> The resulting binary has an identical MD5 checksum.
>
> The cvs admins suck.  I'll restore only the files needed to
> fix the build.  Then!  I'll look into really de-orbiting
> the a.out support ...

I've already posted a request for a repocopy.

Please, give them some time, so we don't have to do CVS reposurgery and
other black magic stuff later to restore the CVS tree to a 'canonical'
state *and* avoid messing up the CVSup updates of everyone.

- Giorgos
Received on Tue Jan 11 2005 - 14:50:30 UTC