[Qgis-psc] Motion: git write access for Etienne Tourigny

[Tim Sutton]

Hi

On Wed, Jul 11, 2012 at 1:45 PM, Etienne Tourigny
<etourigny.dev at gmail.com> wrote:
> On Wed, Jul 11, 2012 at 6:15 AM, Tim Sutton <tim at linfiniti.com> wrote:
>> Hi
>>
>> On Wed, Jul 11, 2012 at 4:07 AM, Etienne Tourigny
>> <etourigny.dev at gmail.com> wrote:
>>> Thanks all
>>>
>>> I will start by committing a slightly modified version of my histogram
>>> improvements (in a pull request).
>>>
>>> Speaking of tests - any idea how the histogram could be tested, and if
>>> it's worth it?
>>
>> Ideal would be to move histogram logic down into the gui lib as we
>> don't yet have a test system in place for app. I could envisage a nice
>> call like:
>>
>> QSize mySize(800,400);
>> QImage myImage = myRasterLayer ->histogramImage(mySize);
>>
>> I don't know if that will make sense though given that you are adding
>> interactivity to the widget....
>>
>>>
>>> It think the proper approach would be to generate a reference image
>>> and test that it matches, like in the render tests. But not sure this
>>> can work with qwt plots though.
>>>
>>
>> Yes there are a few approaches:
>>
>> - use the render checker approach if you expect there render to be
>> non-deterministic
>> - simply compute a md5 sum if you expect it to be derterministic
>
> I'd expect the images to be correct bit-wise so I would expect
> deterministic - but I'm not sure what you mean by deterministic.
>

I meant -'predictably render in exactly the same way on all platforms
and situations'.

>>
>>
>>
>>> Generally - how does one make tests that deal with widgets without
>>> actually displaying them (like the test environment requires)?
>>
>> The test environment can actually deal with gui things - on the test
>> server the tests run under a 'fake' X environment. So you could e.g.
>>
>> - add a new top level test dir for 'app'
>> - create an instance of the raster props dialog
>> - open it, set its tab to histogram
>> - grab the histogram as an image (I already did code to do that I think)
>> - use the render checker to validate the image
>> - close everything down again
>>
>> Adding app testing is certainly on the roadmap so feel free to forge
>> ahead in that direction if you have the time & inclination.
>
> While it would be a great thing, I'm sorry to say that I probably
> won't have the time to forge ahead, but I can certainly contribute and
> more easily replicate existing methodologies...
>

No problem!

Regards

Tim

> cheers
>
>>
>> Regards
>>
>> Tim
>>
>>>
>>> Etienne
>>>
>>> On Tue, Jul 10, 2012 at 4:28 PM, Sandro Santilli <strk at keybit.net> wrote:
>>>> On Tue, Jul 10, 2012 at 08:46:30PM +0200, Tim Sutton wrote:
>>>>> Hi
>>>>>
>>>>> Thanks all. Etienne you now have write access - please liase with the
>>>>> community / Marco before making any major changes, write lots of tests
>>>>> and make things stabler, better, faster and more awesome :-)
>>>>
>>>> Please let me stress that out:
>>>>
>>>>  _write_lots_of_tests_
>>>>
>>>> Thank you ! :)
>>>>
>>>> --strk;
>>
>>
>>
>> --
>> Tim Sutton - QGIS Project Steering Committee Member (Release  Manager)
>> ==============================================
>> Visit http://linfiniti.com to find out about:
>>  * QGIS programming services
>>  * Mapserver and PostGIS based hosting plans
>>  * FOSS Consulting Services
>> Skype: timlinux Irc: timlinux on #qgis at freenode.net
>> ==============================================



-- 
Tim Sutton - QGIS Project Steering Committee Member (Release  Manager)
==============================================
Visit http://linfiniti.com to find out about:
 * QGIS programming services
 * Mapserver and PostGIS based hosting plans
 * FOSS Consulting Services
Skype: timlinux Irc: timlinux on #qgis at freenode.net
==============================================