cvs commit: src/usr.sbin/config configvers.h

[John Baldwin]

On Friday 04 November 2005 11:32 am, Ruslan Ermilov wrote:
> On Fri, Nov 04, 2005 at 09:15:49AM -0700, M. Warner Losh wrote:
> > In message: <20051104120642.GL29387 at submonkey.net>
> >
> >             Ceri Davies <ceri at submonkey.net> writes:
> > : On Fri, Nov 04, 2005 at 12:56:01PM +0200, Ruslan Ermilov wrote:
> > : > On Fri, Nov 04, 2005 at 09:34:36AM +0000, Ceri Davies wrote:
> > : > > On Fri, Nov 04, 2005 at 09:18:58AM +0200, Ruslan Ermilov wrote:
> > : > > > On Fri, Nov 04, 2005 at 04:16:44AM +0000, Warner Losh wrote:
> > : > > > > imp         2005-11-04 04:16:44 UTC
> > : > > > >
> > : > > > >   FreeBSD src repository
> > : > > > >
> > : > > > >   Modified files:
> > : > > > >     usr.sbin/config      configvers.h
> > : > > > >   Log:
> > : > > > >   since nocpu isn't used in the kernel config base, we don't
> > : > > > > need to bump the version.  Peter Wemm, John Baldwin and I
> > : > > > > hammered this out after the last time I needlessly incremented
> > : > > > > the version.
> > : > > > >
> > : > > > >   Revision  Changes    Path
> > : > > > >   1.38      +1 -1      src/usr.sbin/config/configvers.h
> > : > > >
> > : > > > Our users wanted to *use* the "nocpu" directive in their configs
> > : > > > that are based on GENERIC (GENERIC is included), hence I bumped
> > : > > > config(8).  Otherwise, one day some would scream that "nocpu"
> > : > > > doesn't work for them, because they're using the older version
> > : > > > of config(8).
> >
> > I will agree that bumping the config version will force users to
> > install a new config.  At the same time, a new config cannot be used
> > for old kernels, like the ones in 6.0.  Since this is the only
> > incompatibility between 6.0 and current, it seemed a little
> > gratuitous.
> >
> > There's no documentation about what features are in what versions of
> > config.  If you don't bump the version, then we can MFC these changes
> > and users can use config in 6.x.  A nit is that there's no nocpu in
> > config.5.

config.5 is already a different version though.

> > : > > I'd agree that the Makefiles don't need to be bumped, as they don't
> > : > > actually require the new config, but config has new functionality
> > : > > and the version should probably be bumped in the binary.
> > : >
> > : > You cannot just bump it in one place, the version requirement doesn't
> > : > understand >=, only =.
> > :
> > : Oh.  Insert obvious next question here. :)
> > : What happens to those people who have already installed 600004 then?
> >
> > They will need to install a new config if they want to build a new
> > kernel.  Maybe it is time to change the versioning so that multiple
> > versions will work.
>
> This is all not needed if you're using an official upgrade method:
> buildworld/buildkernel, which will take care of updating config(8)
> for you.

I never do that for test kernels as it is too tedious. :)  Also, for my test 
kernels it is very handy to have the build directory shared over NFS just 
like my test trees so that if the machine crashes, I still have access to the 
kernel.debug on my NFS server.  I think the more useful thing would be to get 
config to understand >= for it's version tests though so that version bumps 
aren't quite so painful.

-- 
John Baldwin <jhb at FreeBSD.org>  <><  http://www.FreeBSD.org/~jhb/
"Power Users Use the Power to Serve"  =  http://www.FreeBSD.org