Api changes or changes to undocumented behavior, call it whatever you like, is fine. Never said anything against that, but for the sake of the users lets do the best out of it. But it's ok now. we agreed Im gonna drop xbmcswift2 from this addon because I need some features that's not possible with swift. I just need some time to do it. Lesson learned: be clearer about what's not supported; for instance that the original xbmc api must not be used in combination with xbmcswift, or use at your own risk.



On Fri, Feb 22, 2013 at 7:29 PM, Tristan Fischer <sphere@dersphere.de> wrote:
My 2 cents on this:
I helped Jonathan with checking the backwards compatibility of swift
using my ~10 swift2 add-ons and I gave my ok for the update.
You are using swift in a way which it was not designed for (and you
know that) - IMO you can't make the maintainer of a framework
responsible for checking every add-on - especially when you use it in
a way which it was not created for.

Also you are following the development of swift2 and making your
comments - the dev branch (which became the new master with this
update) includes the changes which break your add-on since more than
two weeks. Btw, this change is not a failure - it's a feature for all
other swift2 add-ons.

I agree that here is not the right place to discuss that, I only did
that because you did.

2013/2/22 takoi <takoi@substep.org>:
> That's what it looks like to me as well. I use the xbmc api too,
> intentionally, with a good reason. But let's not discuss that here. I hope
> Martijn and co. will agree with me if I say what needs to happen now is
> either:
> 1. Make it backwards compatible.
> 2. Contact everyone who is affected by this and give us some time to update
> before pulling xbmcswift (for me it currently looks like it means removing
> xbmcswift dependence completely)
>
>
>
> On Fri, Feb 22, 2013 at 3:01 AM, Jonathan Beluch <jbeluch2@gmail.com> wrote:
>>
>> You're right, not working. Just cleared everything out and tested again.
>> It's the returning None from a view thing.
>>
>>
>> On Thu, Feb 21, 2013 at 2:04 PM, takoi <takoi@substep.org> wrote:
>>>
>>> The nrk plugin? Care to test that again? I just double checked my setup
>>> and it doesn't seem to work at all here. Can't even enter the top plugin
>>> directory. I can't investigate further right now but it wouldn't surprise me
>>> if it's the succeed thing again.
>>>
>>>
>>>
>>>
>>> On Thu, Feb 21, 2013 at 6:59 PM, Jonathan Beluch <jbeluch2@gmail.com>
>>> wrote:
>>>>
>>>> What change doesn't work with your addon, the succeed=False change?
>>>>
>>>> Sorry about that, I tested your addon real quick and it seemed to work,
>>>> although I didn't test every single path.
>>>>
>>>>
>>>> On Thu, Feb 21, 2013 at 12:39 AM, takoi <takoi@substep.org> wrote:
>>>>>
>>>>> Hold please. This is not backwards compatible and will need some
>>>>> existing plugins to update.
>>>>>
>>>>>
>>>>>
>>>>> On Thu, Feb 21, 2013 at 6:40 AM, Jonathan Beluch <jbeluch2@gmail.com>
>>>>> wrote:
>>>>>>
>>>>>> *addon - script.module.xbmcswift2
>>>>>> *version - 2.4.0
>>>>>> *url - git://github.com/jbeluch/xbmcswift2-xbmc-dist.git
>>>>>> *tag - 2.4.0
>>>>>> *xbmc version - frodo
>>>>>>
>>>>>> CHANGES
>>>>>> - Allow ints to be passed to plugin.url_for() and str() them.
>>>>>> - plugin.set_resolved_url() now takes an item dict instead of a URL.
>>>>>> (The URL
>>>>>>   is still allowed for backwards compatibility but is decprecated).
>>>>>> - Ability to replace the context menu items by setting
>>>>>> 'replace_context_menu'
>>>>>>   to True in an item dict.
>>>>>> - pluign.get_setting() now requires an extra paramter, *converter*,
>>>>>> which
>>>>>>   converts the item from XML to a python type specified by converter.
>>>>>> - Major bug fix to accomodate handles > 0 in XBMC Frodo.
>>>>>> - Auto-call plugin.finish(succeeded=False) if a view returns None.
>>>>>> - XBMC now decides the default player when using plugin.play_video().
>>>>>> - Storages now call sync() when clear() is called.
>>>>>> - Added plugin.clear_function_cache() which clears the storage behind
>>>>>> the
>>>>>>   @cached and @cached_route decorators.
>>>>>> - Add ability to set stream info for list items. Also, devs can now
>>>>>> add
>>>>>>   stream_info to item dicts.
>>>>>> - Added the actions module. Contains actions.background() and
>>>>>>   actions.update_view() which are convenience wrappers for RunPlugin()
>>>>>> and
>>>>>>   Container.Update().
>>>>>> - Allow passing of functions to url_for.
>>>>>> - 'properties' key in an item dict should now be a dictionary. A list
>>>>>> of tuples
>>>>>>   is still allowed for backwards compatibility.
>>>>>> - Addon user is now prompted and storage automatically cleared if it
>>>>>> becomes
>>>>>>   corrupted.
>>>>>> - Added subtitles support. A 'subtitles' parameter was added to
>>>>>>   plugin.set_resolved_url().
>>>>>> - xbmcswift2 URLs and routing code is now indifferent to a trailing
>>>>>> slash.
>>>>>> - Bug fixes. See
>>>>>> https://github.com/jbeluch/xbmcswift2/issues?milestone=3.
>>>>>>
>>>>>>
>>>>>> ------------------------------------------------------------------------------
>>>>>> Everyone hates slow websites. So do we.
>>>>>> Make your web apps faster with AppDynamics
>>>>>> Download AppDynamics Lite for free today:
>>>>>> http://p.sf.net/sfu/appdyn_d2d_feb
>>>>>> _______________________________________________
>>>>>> Xbmc-addons mailing list
>>>>>> Xbmc-addons@lists.sourceforge.net
>>>>>> https://lists.sourceforge.net/lists/listinfo/xbmc-addons
>>>>>>
>>>>>
>>>>
>>>
>>
>
>
> ------------------------------------------------------------------------------
> Everyone hates slow websites. So do we.
> Make your web apps faster with AppDynamics
> Download AppDynamics Lite for free today:
> http://p.sf.net/sfu/appdyn_d2d_feb
> _______________________________________________
> Xbmc-addons mailing list
> Xbmc-addons@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/xbmc-addons
>