Martin Langhoff wrote: > Generally agree that it is important... however. > > On Wed, Aug 12, 2009 at 9:55 AM, Richard A. Smith<[email protected]> wrote: >> You have to QA the whole system regardless of what change you make so really >> it doesn't increase the QA that much anyway. > > No, and that is an explicit goal: keep the changes small and low risk > so that we can do QA focused on the very limited areas of the system > we touch.
While I understand the thinking behind this I disagree with the notion that you can make a _any_ change to a system as complex as what we (or any other distro for that matter) have and consider it low-risk. In some cases the very act of attempting to repack the _exact_ same bin .rpm resulted in problems. > As per the original email "low low risk stuff". We have no QA team, > and I am not proposing that I will take on a huge task. I'm not proposing you take it on either. But "someone" has to take this point release and test it. Could be you, me, some adventurous users, the New Zealand team, special deployment testers, whoever. But someone has to test it or they are just plain denying the realities of software development. The effort level of that testing is essentially the same regardless of this change. -- Richard Smith <[email protected]> One Laptop Per Child _______________________________________________ Devel mailing list [email protected] http://lists.laptop.org/listinfo/devel
