Where is 4.9-STABLE?

Chris racerx at makeworld.com
Sat Mar 6 14:37:25 PST 2004


On Saturday 06 March 2004 01:43 pm, Kirk Strauser wrote:
> At 2004-03-06T17:29:13Z, Chris <racerx at makeworld.com> writes:
> > Doing a make world is perfectly acceptable. It's considered the
> > "traditional" way of doing things, and accomplishes the same results.
>
> No, it doesn't.  Specifically, it skips the reboot and mergemaster between
> the installkernel and installworld steps, which means that you'll end up
> running a new userspace against an old kernel and /etc for a little while.
>
> > If your going to inform users NOT to do one way opposed to another, at
> > least give specifics as to why you feel that way.
>
> There you have it.  "make world" is *not* the recommended upgrade process
> anymore.  The new method is detailed in UPDATING.

It seems I needed to be very explicit in my meaning. While in 4.9-RELEASE to 
STABLE, mergmaster is needed as is a rebuild of the kernel.

Once those have been accomplished, a simple make world does the trick. As I 
stated, and does the handbook, make world IS the traditional way of compiling 
and installing your src tree. When done in the correct steps, works very 
well.

I for one refuse to adopt the new way untill I moved to 5.1 - then it was 
simply a matter if having too. 

I have always done make world up untill 4.9-STABLE without issues, as many 
other users. So perhaps I didn't clarify my points (as I'm doing now).

I believe in the users case of 4.9, make world (with the proper steps 
involved) is and will continue to be ba a viable way of compiling your src 
tre and upgrading your system. 

Otherwise, the handbook would have removed that all togather.
You must remember, there are many of us that sometime refuse to adopt new ways 
of doing things unless need requires it - now that I run 5.2.1, I use the new 
style and actually like it.

make world has worked for me since 2.2.6 and I hated to see it go. Oh well.
-- 
Best regards,
Chris


More information about the freebsd-questions mailing list