flex-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Alex Harui <aha...@adobe.com>
Subject Re: [mustella] all recent SDK tests failed
Date Mon, 03 Nov 2014 19:12:58 GMT


On 11/3/14, 10:55 AM, "Erik de Bruin" <erik@ixsoftware.nl> wrote:

>>
>> >I didn't run these locally, as I didn't expect changes to
>>SystemManager to
>> >affect anything but the very core of the SDK.
>>
>> I would be surprised if your changes made a difference as well but you
>> never know.  I thought the “unofficial” policy was that if your checkins
>> are tied to a run that fails that you have to investigate, by at
>>minimum,
>> reverting or showing that the tests pass locally for you.
>>
>
>I "wrote" that policy ;-) But there weren't any tests that consistently
>failed, the failures have been all over the place. I could revert, but my
>assessment was that not my changes, but something on the VM was the root
>cause.

I’m wondering if there is some easy way to show which tests are failing
more often than not.  My records show that these two tests:

gumbo/components/DataGrid/Properties/DataGrid_Properties_editable
Editable_twoWayBinding_test Failed AssertMethodValue (method cannot be
shown)(body:step 13)  method returned , expected test1234

gumbo/components/RichEditableText/Properties/RichEditableText_layout_test3
RichEditableText_Property_maxChars_1 Failed DispatchKeyEvent(body:step 2)
Timeout waiting for change from retEditable1

failed in runs:


1137
1136
1135
1134
1133
1132
1129
1128
1126
1125
1124
1123

Yes, there were other intermittent failures, and run #1130 was full of bad
stuff.  But #1121 passed, then #1123 with your changes, fails, and those
two tests are the repeat offenders since.  I have no idea how they can be
broken by your changes either, but right or wrong, I think you have try a
revert or see if the tests pass locally.

-Alex

Mime
View raw message