Wisdom of automating upgrades

[Peter Risdon]

Robert Huff wrote:

>Peter Risdon writes:
>
>  
>
>> I suppose what I'm driving at is whether the RELENG_4 branch sees
>> many commits that are likely to be problematic.
>>    
>>
>
>	In general, no.
>	On the other hand ... think of this as a Murphy's Law scenario:
>if you automate, it _will_ break horribly two days before some
>absolutely critical deadline.
>  
>

QED

Peter.