W3C home > Mailing lists > Public > public-geolocation@w3.org > March 2012

Re: MOSQUITO project to contribute to HTML5 tests

From: Lars Erik Bolstad <lbolstad@opera.com>
Date: Thu, 08 Mar 2012 10:08:07 +0100
Message-ID: <4F5876F7.1000807@opera.com>
To: "Clari, Fabrice" <f.clari@inno-group.com>
CC: public-geolocation@w3.org
Hi Fabrice,

We had a closer look at your DeviceOrientation test suite at our recent 
Geolocation group meeting. It looks very promising and we'd like to base 
our implementation report on it.

Would you have time to make a few more tests and some minor 
modifications to the existing tests?

We're lacking tests for the compassneedscalibration Event. Are you able 
to contribute those?

Also, we have the following feedback on the current tests:

1. The timeout values are too low
Please increase the values to 1 or 2 seconds at least

2. Please split tests 002 and 022 into two separate tests
We'd like to have one test for each assertion in the spec

3. Test 028
Would be good to test more than just the first two events

Finally, we think tests 012, 030, 031 and 032 are out of scope for our 
implementation report.

Thanks and best regards,
Lars Erik Bolstad
Chair Geolocation WG


On 05.12.2011 15:41, Clari, Fabrice wrote:
> Hi Steve, all,
>
> Sorry for my late answer. Indeed yes, some of the tests require human
> interaction.
>
> The tests developed so far are available on a dedicated page of the
> MOSQUITO website [1]. As this is the first set of tests we develop, any
> feedback will be appreciated. We still have time to contribute, so if
> you know specific area of the API that need tests, please let us know.
>
> Thanks,
> Fabrice.
>
> [1] http://html5.mosquito-fp7.eu
>
>
> Le 23 nov. 2011 à 23:12, timeless a écrit :
>
>> On Tue, Nov 22, 2011 at 7:04 AM, Steve Block <steveblock@google.com
>> <mailto:steveblock@google.com>> wrote:
>>> Hi Fabrice,
>>>
>>> Thanks for your email. You're right that currently there are no tests
>>> for DeviceOrientation.
>>>
>>> It's tricky to reliably test a sensor-based API like this using
>>> automated tests because you can't make guarantees about the accuracy
>>> of a given implementation. Our current approach is to provide a simple
>>> test application which would be used to test the API manually [1].
>>
>>> I'd certainly be interested to see the tests that you would propose
>>> for DeviceOrientation. Did you have a strategy in mind at this point?
>>
>> fwiw, Mozilla has plans to test such things by using a system-qemu and
>> adjusting the sensor values "Marionette" [1].
>>
>> [1] https://wiki.mozilla.org/Auto-tools/Projects/Marionette
>>
>
> __________________________________
> *Fabrice Clari - inno TSD*
> mobile +33 (0)6 19 64 46 97
> office +33 (0)4 92 38 84 19
> fax +33 (0)4 93 65 41 35
> mail F.Clari@inno-group.com <mailto:F.Clari@inno-group.com>
> webhttp://www.inno-group.com
>
>
>
>
>
Received on Thursday, 8 March 2012 09:08:44 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Thursday, 22 March 2012 18:13:56 GMT