W3C home > Mailing lists > Public > public-test-infra@w3.org > July to September 2012

Re: [Bug 18468] New: Identical results should be merged

From: Richard Ishida <ishida@w3.org>
Date: Mon, 06 Aug 2012 08:11:38 +0100
Message-ID: <501F6E2A.6000500@w3.org>
To: Jeanne Spellman <jeanne@w3.org>
CC: public-test-infra@w3.org
hi Jeanne,

I don't mind if the multiple results show up on a page such as 
http://w3c-test.org/framework/details/i18n-html5/bidi-chrome-001

My main concern is that on a page such as 
http://w3c-test.org/framework/results/i18n-html5/section/10.6.4 the result

bidi-chrome-001	 2/./.	1/1/.	2/2/.	3/./.

is misleading.

Richard Ishida
Internationalization Activity Lead
W3C (World Wide Web Consortium)

http://www.w3.org/International/
http://rishida.net/

On 02/08/2012 19:54, Jeanne Spellman wrote:
> I am concerned that this could be problematic when I want to compare
> results from multiple testers.  Because my tests are all manual (and
> prone to human error) an important step in CR is resolving any
> discrepancies.  If we decide to merge identical results, would I still
> be able to tell that the test had been completely run by separate
> testers?  How do you envision these test results being displayed?
>
> jeanne
>
> On 8/2/2012 8:49 AM, bugzilla@jessica.w3.org wrote:
>> https://www.w3.org/Bugs/Public/show_bug.cgi?id=18468
>>
>>             Summary: Identical results should be merged
>>             Product: Testing
>>             Version: unspecified
>>            Platform: PC
>>          OS/Version: All
>>              Status: NEW
>>            Severity: normal
>>            Priority: P3
>>           Component: Test Framework
>>          AssignedTo: mike@w3.org
>>          ReportedBy: ishida@w3.org
>>           QAContact: dave.null@w3.org
>>                  CC: public-test-infra@w3.org
>>
>>
>> Currently, if someone submits test results for the same browser,
>> version and OS
>> with the same result multiple times (I have several cases in mind),
>> the results
>> summary tables increment the numbers displayed (ie. 1/./. can become
>> 5/./.
>> without any additional useful information being actually added.
>>
>> Please could we only report unique configurations in the results
>> summaries.
>>
>> This could be done by either checking when results are requested, or
>> by not
>> capturing identical configuration tests in the database (perhaps
>> showing only
>> the latest one, since date can sometimes be of interest).
>>
>
Received on Monday, 6 August 2012 07:12:01 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 20:20:33 UTC