On Sat, 26 Aug 2006 15:30:29 +1200 Andrew Thompson <[EMAIL PROTECTED]> wrote:
> On Fri, Aug 25, 2006 at 07:31:20PM -0700, Steve Kargl wrote: > > On Sat, Aug 26, 2006 at 12:08:54PM +1000, Peter Jeremy wrote: > > > On Thu, 2006-Aug-24 18:44:02 -0400, Alexander Kabaev wrote: > > > >personally. The mistake you made just shows how important it is > > > >to do a buildworld _from freshly checked out sources_ after > > > >every major vendor import. > > > > > > I'm not sure how doing this would have prevented the breakage. It > > > might have allowed Guido to notice the breakage more quickly. > > > > > Doing a fresh checkout and a "make universe" > > > (since "make buildworld" is not always sufficient to catch > > > breakage) may not have been that much faster. > > > > Huh? If Guido had done a "make buildworld" or "make universe" > > before he committed the source, then the problem would never have > > existed. > > Thats not true in this case. The problem was that a new file hadnt > been cvs added, the file was sitting in his tree so it built > correctly, no amount of "make buildworld" or "make universe" would > have caught that, let it go. > > > Andrew For the last time, what part of _freshly checked out sources_ needs special explanation? Before that needs even more explanation, I'll even take it further: the test should be done on local repo copy first before 'real' one. Even without an extensive test above, a simple 'cvs update' before commit can be used to reveal all files that probably should either be added to CVS or removed. It isn't hard to spot all files prefixed with '?' in cvs output. The original breakage in -current was caused by a bug that simple compile test would have found. The compile test that clearly had not been tried. Between that and RELENG6 breakage I find it hard to believe that due attention has been paid doing ipfilter update work. -- Alexander Kabaev
signature.asc
Description: PGP signature
