svn commit: r193734 - in head: lib/libusb sys/amd64/amd64 sys/i386/i386 sys/i386/xen sys/sys

Robert Watson rwatson at
Tue Jun 9 11:06:15 UTC 2009

On Mon, 8 Jun 2009, M. Warner Losh wrote:

> In message: <200906081823.n58INhX8038125 at>
>            Ed Schouten <ed at> writes:
> :   It turns out our compiler in stable/7 can't build this code anymore.
> :   Even though my opinion is that those people should just run `make
> :   kernel-toolchain' before building a kernel, I am willing to wait and
> :   commit this after we've branched stable/8.
> As a matter of policy, the project has always supported a RELENG_X -> 
> CURRENT upgrade with a simple buildworld.  The issue of building a kernel by 
> hand across releases has been targeted as one of those areas that can break, 
> but only if there's a really good reason to do so.
> Just FYI to everybody..

As compared to building 7.x kernels on 6.x, we got significantly farther 
through the release process before bumping into a problem, FWIW.  I certainly 
appreciate the ability to short-cut builds by doing them incrementally with 
existing toolchains, and while there are good reasons to break that (such as 
taking advantage of new syntax to improve the kernel -- for example, C99 
structure initialization) my hope is generally to avoid breaking it unless 
such a good reason turns up.  Usually, in the past, it's been share/mk 
incompatibility that's broken the build rather than C syntax, though.

Robert N M Watson
Computer Laboratory
University of Cambridge

More information about the svn-src-all mailing list