David Powell wrote: > Richard Lowe wrote: >> Tom Haynes <Thomas.Haynes at Sun.COM> writes: >> >>> Eric Schrock wrote: >>>> Pushing batched changes from >>>> ON requires temporarily disabling sanity checks, as well as locking >>>> and >>>> unlocking the gate. >>> I've hit the issue with disabling sanity checks when I merge the >>> nfs41 gate with >>> the ON gate every two weeks. >>> >>> I have to remember to go into the $GATE/.hg/hgrc file and comment >>> out this: >>> pretxnchangegroup.1 = python:hook.sanity.sanity >> >> That's primarily going to be because the sanity check for the ON gate >> prevents you integrating merge turds though, right? >>
Yes. >> So they're not precisely what you actually want, since in that case >> you *will* be merging, and you want the merges to hit your (project) >> gate? > > It depends. You want the project gate's merges with the upstream > gate to remain intact so its children aren't invalidated, but > contributors to a project gate usually don't push merges -- they are > just as useless in a project gate as they are in the ON gate. > > That said, since a project will be collapsing its changesets before > pushing to the upstream gate, enforcement at the project level is > probably overkill. > So, are you saying to leave sanity off at all times? > Dave >
