[webkit-dev] lots of red in the tree.

[David Levin]

On Tue, Aug 30, 2011 at 12:54 PM, David Levin <levin at chromium.org> wrote:

>
>
> On Tue, Aug 30, 2011 at 10:55 AM, Jarred Nicholls <jarred at sencha.com>wrote:
>
>> On Tue, Aug 30, 2011 at 1:52 PM, David Levin <levin at chromium.org> wrote:
>>
>>> It means that I'm much more likely to cause regressions because I miss
>>> new test failures caused by my changes among the 12 to 62 failures
>>> already occurring on the OS X bots.1
>>>
>>
>> Yeah it was wigging me out this morning.
>>
>>
>>>
>>> I think there are a few solutions to the current situation:
>>>
>>> 1. Live with it. All of the red and green reminds us of the festive
>>> winter holidays.
>>>
>>
>> presents!
>>
>>
>>>     Downside: More regressions get in as nobody notices them much even if
>>> they try to be careful.
>>>     Upside: Requires no more extra work, so it is quick to do!
>>>
>>> 2. Get folks working on every red test.
>>>
>>
>> Was thinkin' about diving head first into this.
>>
>>
>>>     Downside: May not be able to get folks to drop what they are doing
>>> and work on them.
>>>     Upside: More stable code, easier to work with, etc.
>>>
>>> 3. Add them to skipped and file bugs.
>>>
>>
>> Good first step.  Some of the tests are flaky though - I'll get different
>> results on subsequent runs.
>>
>
> I'm ok with letting flaky ones run for now as they are re-run by the
> harness and won't hide new failures. I would like to take care of those that
> are always failing.
>
> I'm in favor of this one -- Adding them to skipped and filing bugs.
>

So far it sounds like there are no objections, so I plan to start doing this
soon.

I'll add new baselines (which may be incorrect) and file bugs to fix the
results (cc'ing folks whom I'll guess will be interested in that issue). If
it is crashing, I'll add to skipped and file bugs.

dave



>
> dave
>
>
>
>>
>>
>>>     Downside: Not having the tree red may lower the urgency and having
>>> them in skipped list may mean that folks just ignore them.
>>>     Upside: We'll catch regressions more quickly and perhaps stop the
>>> current decent which it seems like we've been proceeding on.
>>>
>>> 4. Your idea!
>>>
>>> What do other folks think?
>>>
>>> Dave
>>>
>>> _______________________________________________
>>> webkit-dev mailing list
>>> webkit-dev at lists.webkit.org
>>> http://lists.webkit.org/mailman/listinfo.cgi/webkit-dev
>>>
>>>
>>
>>
>> --
>> ................................................................
>>
>> *Sencha*
>> Jarred Nicholls, Senior Software Architect
>> @jarrednicholls
>> <http://twitter.com/jarrednicholls>
>>
>>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.webkit.org/pipermail/webkit-dev/attachments/20110831/63d73d3f/attachment.html>