Re: [head tinderbox] failure on amd64/amd64

On Thu, Aug 14, 2008 at 04:11:25AM +0200, Dag-Erling Sm??rgrav wrote:
> FreeBSD Tinderbox <tinderbox_at_freebsd.org> writes:
> > TB --- 2008-08-13 18:55:00 - tinderbox 2.3 running on freebsd-current.sentex.ca
> > TB --- 2008-08-13 18:55:00 - starting HEAD tinderbox run for amd64/amd64
> > TB --- 2008-08-13 18:55:00 - cleaning the object tree
> > TB --- 2008-08-13 18:55:53 - cvsupping the source tree
> > TB --- 2008-08-13 18:55:53 - /usr/bin/csup -r 3 -g -L 1 -h localhost -s /tinderbox/HEAD/amd64/amd64/supfile
> > TB --- 2008-08-13 18:56:00 - building world (CFLAGS=-O -pipe)
> > TB --- 2008-08-13 18:56:00 - cd /src
> > TB --- 2008-08-13 18:56:00 - /usr/bin/make -B buildworld
> > TB --- 2008-08-13 18:56:00 - WARNING: /usr/bin/make returned exit code  1 
> > TB --- 2008-08-13 18:56:00 - ERROR: failed to build world
> > TB --- 2008-08-13 18:56:00 - tinderbox aborted
> > TB --- 2.38 user 4.29 system 60.01 real
> 
> The following code in src/Makefile:
> 
> STARTTIME!= LC_ALL=C date
> CHECK_TIME!= find ${.CURDIR}/sys/sys/param.h -mtime -0
> .if !empty(CHECK_TIME)
> .error check your date/time: ${STARTTIME}
> .endif
> 
> incorrectly assumes that the build cannot possibly start less than a
> second after sys/sys/param.h is updated.
> 
"-mtime -0" interacts badly with the TIME_CORRECT() macro
when mtime and run-time happen to be the same second.
While I think it's a bug in the find(1) code that needs to
be fixed, using "-mtime -0s" seems to avoid the macro and
thus works as expected.

Can you do me a favour and commit it for me please?


Cheers,
-- 
Ruslan Ermilov
ru_at_FreeBSD.org
FreeBSD committer