W3C home > Mailing lists > Public > public-geolocation@w3.org > August 2011

Re: geolocation.setCurrentPosition({...}): a way to debug geolocation and give more power to users.

From: Wojciech Masłowski <wmaslowski@opera.com>
Date: Thu, 25 Aug 2011 09:20:43 +0200
Message-ID: <4E55F7CB.6080807@opera.com>
To: louis-rémi Babé <lrbabe@gmail.com>
CC: public-geolocation@w3.org
W dniu 2011-08-24 09:45, Charles McCathieNevile pisze:
> On Tue, 23 Aug 2011 19:33:29 +0200, Richard L. Barnes 
> <rbarnes@bbn.com> wrote:
>
>> +1
>
> Agreed. There are easy ways in Opera to control the information you 
> send websites. It would seem to suggest that this is something that 
> the user agent should handle in its UI, rather than being in the spec.
>
For debugging in Opera you might try Geolocation Provider unite app 
which implements alternative dummy geolocation provider service.

http://unite.opera.com/application/701/
> cheers
>
>> On Aug 23, 2011, at 1:21 PM, Doug Turner wrote:
>>
>>>
>>> I think this is out of scope of the working group.
>>>
>>> Regards,
>>> Doug
>>>
>>> On Aug 23, 2011, at 8:58 AM, louis-rémi Babé wrote:
>>>
>>>> Yes, this is what I want to do.
>>>> This kind of API shouldn't be able to control browser features such as
>>>> permission prompt.
>>>>
>>>> Looking at NetworkGeolocationProvider.js in Mozilla central, this
>>>> doesn't appear easy to do.
>>>> And it is impossible to do in all other browsers.
>>>>
>>>> I'm aware that allowing developers to debug an API is not usually part
>>>> of that API design. But most APIs are still easy to debug (it is
>>>> possible to generate custom events to debug event based APIs for
>>>> instance). Developers are able to create their own tools and do not
>>>> have to wait for browser vendors to implement ready to use tools.
>>>>
>>>> The geolocation API can currently be only debugged if you own the
>>>> source code of the application and insert "debugging hooks".
>>>>
>>>> I think this situation is not ideal. And I think giving users an easy
>>>> way to control the position info sent to websites is important too.
>>>>
>>>> Lr
>>>>
>>>> 2011/8/17 Doug Turner <doug.turner@gmail.com>:
>>>>> So, in general, you want a way to setup a mock position for 
>>>>> testing purposes?  Do you also want to be able to control the 
>>>>> permission prompt in this case?
>>>>>
>>>>> In FF's case, it is trivial to write an add-on that does this - 
>>>>> something that you can share with other users.  (In fact, our test 
>>>>> infrastructure basically has exactly this without the add-on 
>>>>> packaging).
>>>>>
>>>>> Doug
>>>>>
>>>>> On Aug 17, 2011, at 8:31 AM, louis-rémi Babé wrote:
>>>>>
>>>>>> Hi all,
>>>>>>
>>>>>> I'm currently frustrated by the lack of debuging tools regarding the
>>>>>> geolocation API, as opposed to other event based device APIs such as
>>>>>> deviceorientation.
>>>>>>
>>>>>> To debug deviceorientation, I can simply generate a custom event 
>>>>>> right
>>>>>> from my js console using create/init/dispatchEvent.
>>>>>> And I can do so on any website listening to deviceorientation events
>>>>>> (I don't need to be the author).
>>>>>>
>>>>>> It is impossible to do the same thing with the geolocation API. To
>>>>>> debug it I need the following kind of code:
>>>>>>
>>>>>>    // can't use inline handler
>>>>>>    function positionHandler( event ) {
>>>>>>        ...
>>>>>>    }
>>>>>>    navigator.geolocation.watchPosition( positionHandler );
>>>>>>    // need to create a global "debug object" to make the
>>>>>>    // positionHandler available from my js console
>>>>>>    window.updatePosition = positionHandler;
>>>>>>
>>>>>> It requires to add one extra line to my code for debugging purpose,
>>>>>> and I need to be the author of the website.
>>>>>> I can alternatively write a geolocation provider for my Web browser,
>>>>>> but that requires a far larger amount of work, a kind of knowledge
>>>>>> that Web developers do not necessarily have, and it isn't a cross
>>>>>> browser solution.
>>>>>>
>>>>>> geolocation.setCurrentPosition() would allow to create a fake 
>>>>>> position
>>>>>> in the flow of "real" position updates.
>>>>>> Users could simply ignore the "permission prompt" on a website 
>>>>>> and use
>>>>>> setCurrentPosition() to have complete control over the position
>>>>>> updates provided to a website.
>>>>>>
>>>>>> Thank you in advance for your feedback,
>>>>>> louisremi@mozilla.com
>>>>>>
>>>>>
>>>>>
>>>
>>>
>>
>
>


-- 
Wojciech Masłowski
Engeneering CORE Wrocław
Opera Software ASA
http://www.opera.com
Received on Thursday, 25 August 2011 07:21:34 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 19:51:02 UTC