Richard L. Hamilton wrote:
...
> OTOH, I'm not convinced that replacing patching with package replacement
> is a good idea; I have a feeling it would be less stable, although I don't
> know that I could prove it  (after all, a patch is nothing but one or
> more (usually partial) package updates that are applied together, along
> with whatever scripts and such are needed to handle any related loose ends).

Our data indicates that it's actually almost certain to be more 
reliable.  As Bart likes to say, this is the "dim sum" problem; the 
current patching methodology with sparse packages allows you to create 
combinations of binaries or libraries which have never been tested 
together and really can't be, given the combinatorics and the limits on 
the resources we can apply.  Providing fewer combinations and testing 
them more thoroughly is a proven technique for increasing the 
reliability of the customer experience.

Dave
_______________________________________________
indiana-discuss mailing list
[email protected]
http://mail.opensolaris.org/mailman/listinfo/indiana-discuss

Reply via email to