ubuntu-phone team mailing list archive

[Jamie Strandboge <jamie@xxxxxxxxxxxxx>]

On 10/21/2013 03:13 PM, Pat McGowan wrote:
> On 10/21/2013 04:06 PM, Michał Sawicz wrote:
>> On 21.10.2013 21:31, Michael Hall wrote:
>>> Is there a reason you are trying to use Python from QML instead of it's
>>> native Javascript?
>>>
>>> While PyOtherSide may work, it's not a supported language or platform
>>> for the Ubuntu SDK, so we can't guarantee that it will always work or
>>> work properly.  I would encourage you to use Javascript instead of at
>>> all possible.
>>
>> From what I can see - Python here is supposed to replace not the JS parts of a
>> "standard" approach to a QML app, but the C++ part. And while I do understand
>> some of the reasoning, I must agree that this is rather shoehorning it in. And
>> to make sure that your app is working, you'll have to ship a Python
>> interpreter itself, and all of the modules you want to use, with your
>> application - and that starts to become rather cumbersome.
>>
>> You should, indeed, reconsider using C++ and/or JS where you thought of using
>> PyOtherSide.
>>
>> HTH
> We would also need to address app confinement which would not work as shipping
> today.
> 
So, if the app shipped the entire python interpreter we should be able to get
away with not having to update the application confinement policy.

However, shipping the interpreter is cumbersome, and I think what is really
being asked is if app developers could reliably make use of system python
interpreter. This would require updates for app confinement (which is solvable),
but more importantly, click packages don't have a concept of dependencies beyond
the framework that is declared in their click manifest, so to be sure that a
python app worked on a given device, python needs to be an official part of the
ubuntu-sdk-xx.xx framework.

-- 
Jamie Strandboge                 http://www.ubuntu.com/

Attachment: signature.asc
Description: OpenPGP digital signature