On Fri, Feb 03, 2006 at 09:55:49PM -0700, Warner Losh wrote.. > From: Scott Long <[EMAIL PROTECTED]> > Subject: Re: [releng_6 tinderbox] failure on sparc64/sparc64 > Date: Fri, 03 Feb 2006 19:40:42 -0700 > > > Warner Losh wrote: > > > From: [EMAIL PROTECTED] (Dag-Erling Smørgrav) > > > Subject: Re: [releng_6 tinderbox] failure on sparc64/sparc64 > > > Date: Fri, 03 Feb 2006 11:53:38 +0100 > > > > > > > > >>Kris Kennaway <[EMAIL PROTECTED]> writes: > > >> > > >>>COPTFLAGS=-O -pipe according to the tinderbox logs. > > >> > > >>Hmm, yes, apparently it only uses -O2 on HEAD. > > > > > > > > > Can we not have special flags for tinderbox builds? It make > > > pre-commit testing a big pita. How about just -O on both head and in > > > RELENG_6? The kernel make files have special magic to disable the > > > parts of -O2 that are known to be bad because tinderbox uses -O2, > > > despite efforts in the past to stop the practice. > > > > > > Warner > > > > > > > > > > There is value in testing -O2, since enabling that is a good long-term > > goal. What might be nice is to run tinderboxes with all default > > compiler settings, and then once or twice a week to a special run that > > has the more experimental flags. > > My point is that it is unreasonable to get bitched at for tinerbox > breakages that don't show up when building lint because the tinderbox > person is too stubborn to not use non-standard flags.
I would think that the tinderboxes should run 100% the same flags as what normal release builds use. Nothing more, nothing less. -- Wilko Bulte [EMAIL PROTECTED] _______________________________________________ freebsd-stable@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-stable To unsubscribe, send any mail to "[EMAIL PROTECTED]"
