Alec Flett wrote:
I think a key component to avoid regressions is going to be getting
historical performance graphs - as they say, pictures speak louder than
words. (kidding) - but in all seriousness: I think we're not going to be
able to spot performance regressions if we're just looking at a graph of
today's performance numbers (which is what we have now) - I think we
need to look back at least 30 days and see if there has been a gradual
and/or sudden regression in numbers and the fastest way to do this is
with a graph.
Yup, we have a bug on that. However, I also think that regressions that
are large enough to notice immediately after checkin can and should be
noticed and addressed in some manner that same day.
When we're close to a release, definitely. When we are at the beginning of a
new development cycle, most likely not.
Andi..
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
Open Source Applications Foundation "Dev" mailing list
http://lists.osafoundation.org/mailman/listinfo/dev
