Safe to just rebuild kernel after cvsuping src?

[Kevin D. Kinsey, DaleCo, S.P.]

Charles Ulrich wrote:

>Robert Huff said:
>  
>
>>Nikolas Britton writes:
>>    
>>
>>> Is it generally safe to just rebuild the kernel, and not make
>>> world, when your only tracking a release and not -STABLE,
>>> i.g. setting cvsup to track RELENG_5_3?
>>>      
>>>
>>	As a general rule, this is _never_ safe.  Unless you're
>>prepared to locate and understand all the changes - just bite the
>>bullet and make world.
>>	(This is not to say you can't do it and have it work - been
>>there, done that - but you're definitely increasing the odds of a
>>problem.)
>>    
>>
>
>I've been doing the opposite on some of my machines which run stable releases
>of FreeBSD. Is it relatively safe to build and install a slightly newer world
>without rebuilding the kernel?
>  
>

It's really a question for the programmers, and it's
likely that whatever they say will include a lot of
disclaimers and warnings.

I once got kernel and world a lil' out of sync, and
had things running apparently normally.  However,
top(1) broken; when I asked the list what was up,
someone else with experience was able to determine
the problem immediately.

In short, it's probably not a good idea unless one
can determine that relatively nothing has changed
in the source since the other part (kernel or world)
was built.  Looking at CVS, you don't see much
completely idle time.  And, unless you are aware
of every bit of code that might have changed, it's
impossible (for me, Joe Average) to know what might
break ....

Kevin Kinsey