[PEAK] Trellis: undetected circularity via optional rules

Sergey Schetinin maluke at gmail.com
Thu Apr 16 19:48:51 EDT 2009


On Fri, Apr 17, 2009 at 02:13, P.J. Eby <pje at telecommunity.com> wrote:
> At 12:37 AM 4/17/2009 +0300, Sergey Schetinin wrote:
>>
>> On Thu, Apr 16, 2009 at 18:29, Grant Baillie <grant at osafoundation.org>
>> wrote:
>> >
>> > On 15 Apr, 2009, at 11:31, Sergey Schetinin wrote:
>> >
>> >> Now I wonder if this will break any trellis properties, like circular
>> >> maintain rules that don't set conflicting values (the rules might
>> >> initialize each other even if they are not optional, when the
>> >> component is being initialized). The tests fail, but they always hang
>> >> up for me anyway, so I can't really use that as a reference.
>> >>
>> >>
>> >> So I have a question: for how many people in this list trellis tests
>> >> do run successfully?
>> >
>> > They hang for me, too (on Mac OS X 10.5). After I disable the tests
>> >
>> > test_trellis.TestReactorEventLoop.testSequentialCalls
>> > test_trellis.TestTasks.testNoTodoRollbackIntoTask
>> >
>> > I still get a bunch of failures in test_sets.TestUpdateOps (although
>> > these
>> > seem to pass when run by themselves).
>>
>> Um.. disabling those helps tests to finish but I still get 37 failures
>> on fresh checkout, 39 failures on my working copy (apparently I broke
>> something that makes discrete lazy cells become constant before
>> resetting or not notifying about the reset). Plus a warning from
>> SQLAlchemy on use of deprecated apis.
>>
>> Using WinXP, Py2.5.
>
> Just as a datapoint, I don't get any test failures or hangs on my own
> machine.  I'm particularly confused by the set-related failures.
>
>

I wanted to attach a test log but the test suite output somehow routes
around >, >> and |. I might be inept with console but when I tried the
stupid approach: setting console lines limit to 1000 and just
copy-paste the output, I discovered that running the test messes with
it and resets the console size so it's not big enough to capture the
output. Nice.

Anyway, I managed to capture the output with SciTE:
http://pastebin.com/pastebin.php?dl=m7ca82e2f

That's vanilla checkout with some tests disabled
(test_trellis.TestReactorEventLoop.testSequentialCalls,
test_trellis.TestWxEventLoop.testSequentialCalls,
test_trellis.TestTasks.testNoTodoRollbackIntoTask).


More information about the PEAK mailing list