On Mon, Mar 23, 2009 at 01:07:20PM -0700, Peter Memishian wrote:
> 
>  > There would otherwise be so many rapidly-changing contracted private
>  > interfaces that we'd be buried in needless ARC work.
> 
> Instead engineers often change interfaces in ON without realizing they've
> broken some code buried off in a test gate.

I think we can and should distinguish testsuites that test primarily the
thing that exports some consolidation private interface from those that
primarily test something else and happen to use those consolidation
private interfaces.

>  > There's probably a good argument to be made for weaning the test code
>  > from private interfaces and treating it as a full-fledged product with
>  > all the normal consolidation boundary rules, but that's not quite
>  > where we are today.
> 
> ... or for the test code to live alongside the source, but that's another
> discussion ...

Yeah, definitely a different discussion!
_______________________________________________
networking-discuss mailing list
[email protected]

Reply via email to