Heikki Toivonen wrote:
And I believe it has happened partly because the rules haven't been
followed. I've seen all of the rules you mentioned broken in the last
two weeks.
I'm sure I've broken all of them; there's been no other way to make
progress, and there's been no community pressure not to.
This is not my understanding. The last time we discussed these publicly
(winter?), I believe we recognized it a problem that several tests were
disabled and you could not rely on your code being correct because so
many tests were disabled. Because of that we agreed that it would be a
priority to re-enable all disabled tests, and no longer disable tests
but fix the issues instead, quickly (by backing out if nothing else).
That process is not working; if it was, I wouldn't have written the
message I did!
I believe we have followed that more or less, but when a code freeze is
getting nearer it seems the pressure to get fixes in is the driving factor.
I agree it is frustrating that Tbox has been orange so much lately. I
think the solution is to follow the checkin rules as written.
Unfortunately, this won't solve our current problem.
My initial reponse here was "OK, I hereby promise to follow all the
rules", but that means I won't get anything else done for alpha4, and
(assuming everyone followed the rules to the letter) you won't see any
changes until the tree goes green, when you'll suddenly get a whole
bunch of pent-up changes that haven't been tested together... and odds
are, the tree will go yellow again. If people follow the rule that says
"do many small checkins, not one big one", you'll find that backing out
the breaking change will only cause later changes (that depend on the
earlier backed-out stuff) to break.
This will snowball us back into the situation we're in now. Therefore,
I'm proposing that a change to the policy for situations like this is
necessary, and further, that the policy should be what I said in my last
message: when the tinderbox is yellow for an extended period of time
(and I think the current situation qualifies: it hasn't been green
across the board in the eight days!), we *close the tree* and declare a
switch to this policy:
> - P1 bugs get filed for each failure
> - The failing tests get disabled on the platforms on which they fail
> (with bug numbers in the comments)
> - People are assigned to fix those bugs and reenable the tests.
...Bryan
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
Open Source Applications Foundation "chandler-dev" mailing list
http://lists.osafoundation.org/mailman/listinfo/chandler-dev
