Possible order-of-operations problem with recent -CURRENT's buildworld

I saw a problem building today's -CURRENT on an SMP box using -j8,
but did not see it on the same system (w/ same sources) without -j;
I also did not see it on a UP system with -j4 (with sources whose
differences should be restricted to things like patches for the "an"
driver, since the UP system is my laptop).

I did the "make buildworld" (& friends) within "script", so I have all of
the output.  However, it's about 14.5 MB, so sending all of it anywhere
would seem ... unkind, at best.  :-}

The part just before the "make -j8 buildworld" stopped looked like:

===> gnu/usr.bin/cvs/cvsbug
sed -e "s,_at_VERSION_at_,1.11.5-FreeBSD,g" /usr/src/gnu/usr.bin/cvs/cvsbug/../../../../contrib/cvs/src/cvsbug.in > cvsbug
gzip -cn /usr/src/gnu/usr.bin/cvs/cvsbug/../../../../contrib/cvs/man/cvsbug.8 > cvsbug.8.gz
===> gnu/usr.bin/cvs/doc
makeinfo --no-split -I /usr/src/gnu/usr.bin/cvs/doc -I /usr/src/gnu/usr.bin/cvs/doc/../../../../contrib/cvs/doc /usr/src/gnu/usr.bin/cvs/doc/../../../../contrib/cvs/doc/cvs.texinfo  -o cvs.info
makeinfo --no-split -I /usr/src/gnu/usr.bin/cvs/doc -I /usr/src/gnu/usr.bin/cvs/doc/../../../../contrib/cvs/doc /usr/src/gnu/usr.bin/cvs/doc/../../../../contrib/cvs/doc/cvsclient.texi  -o cvsclient.info
gzip -cn cvsclient.info > cvsclient.info.gz
gzip -cn cvs.info > cvs.info.gz
1 error
*** Error code 2
1 error
*** Error code 2
1 error


I re-started the build without -j in order to try to pinpoint the error;
it completed successfully.  :-}

I am tracking (both -STABLE and) -CURRENT daily, using a private mirror
of the FreeBSD CVS repository.  Its recent update history is:

freebeast(5.1-BETA)[6] tail /var/log/cvsup-history.log
CVSup begin from cvsup13.freebsd.org at Mon May 12 03:47:15 PDT 2003
CVSup ended from cvsup13.freebsd.org at Mon May 12 03:54:08 PDT 2003
CVSup begin from cvsup13.freebsd.org at Tue May 13 03:47:15 PDT 2003
CVSup ended from cvsup13.freebsd.org at Tue May 13 03:53:56 PDT 2003
CVSup begin from cvsup13.freebsd.org at Wed May 14 03:47:16 PDT 2003
CVSup ended from cvsup13.freebsd.org at Wed May 14 03:54:20 PDT 2003
CVSup begin from cvsup13.freebsd.org at Thu May 15 03:47:16 PDT 2003
CVSup ended from cvsup13.freebsd.org at Thu May 15 03:54:05 PDT 2003
CVSup begin from cvsup13.freebsd.org at Fri May 16 03:47:16 PDT 2003
CVSup ended from cvsup13.freebsd.org at Fri May 16 03:54:09 PDT 2003
freebeast(5.1-BETA)[7] 

(The laptop's CVS repository gets mirrored from the one updated by the
above process, so I have some reason to believe that the sources are
mostly synchronized.)

Oh:  as evidence that it did, in fact, work:
freebeast(5.1-BETA)[8] uname -a
FreeBSD freebeast.catwhisker.org 5.1-BETA FreeBSD 5.1-BETA #27: Fri May 16 09:58:48 PDT 2003     root_at_freebeast.catwhisker.org:/common/S4/obj/usr/src/sys/FREEBEAST  i386
freebeast(5.1-BETA)[9] 

I did not see the problem yesterday, so that would imply that the change
arrived at cvsup13 some time between Thu May 15 03:47:16 PDT 2003 and
Fri May 16 03:54:09 PDT 2003 -- or so I would think.

I'd appreciate suggestions for identifying the problem.

Thanks,
david
-- 
David H. Wolfskill				david_at_catwhisker.org
Based on what I have seen to date, the use of Microsoft products is not
consistent with reliability.  I recommend FreeBSD for reliable systems.