Yes, <config-file> element would work fine for wp7+8 as well. I then see no other need for the <permission> element.
<config-file target="Properties/WMAppManifest.xml" parent="/Deployment/App/Capabilities"> <Capability Name="ID_CAP_IDENTITY_DEVICE" /> <Capability Name="ID_CAP_IDENTITY_USER" /> <Capability Name="ID_CAP_LOCATION" /> </config-file> @purplecabbage risingj.com On Wed, Apr 17, 2013 at 4:04 PM, Filip Maj <f...@adobe.com> wrote: > Nope nothing being forced on anyone here. I agree with you: developers > should be aware of the platforms they are targeting and should understand > how their app works, regardless if its written with a cross-platform > framework or not. > > We are talking about a spec helping standardize the way bits of > functionality can be exposed to cordova applications. If you have > something very app-specific, we (cordova) should not mandate that it be > written as a plugin. If you have a cross-platform plugin to access special > device functionality, then wrapping it up in a plugin with a manifest that > is standardized will help with exposure, discovery and usability of the > plugin. > > Back to the issue at hand, the contention was: do we need a special > section of the spec listing out platform permissions required by a plugin? > Originally my answer was yes, however, Anis brought up good points and now > I am not as convinced. We already have a section for generic XML munging > (the <config-file> element) which we can use in this case. The specific > use case that needed satisfying (installing multiple plugins with > shared/overlapping config changes, then uninstalling one of them, and > having the shared/overlapping config changes remain in the config) can be > satisfied with both approaches. However, adding a <permission> element is, > in a sense, redundant, as it doesn't satisfy any other use cases that we > deem necessary. > > If there are other use cases in plugin management that are not satisfied, > but would be by adding a <permission> element, then we should list them > out in this thread so we can keep pushing development of this plugin spec > and our tooling around it. > > On 4/17/13 3:55 PM, "Jesse" <purplecabb...@gmail.com> wrote: > > >The use case is any app that has any additional native functionality added > >by the developer. ( every app I have ever written ) > >Otherwise we force an all or nothing stance with cordova use, Android, > >iOS, wp7, wp8 all support using cordova as a view in an otherwise strictly > >native app, or adding a native view in an otherwise browser-control > >(typical cordova) based app. > >I am not suggesting we make drastic efforts to support every type of app > >that could exist, but if we can make a few simple changes to allow it, I > >think we should. > >Ideally everyone would write any additions to their app as a plugin > >according to our spec, but do we need to force it? > > > >@purplecabbage > >risingj.com > > > > > >On Wed, Apr 17, 2013 at 3:37 PM, Filip Maj <f...@adobe.com> wrote: > > > >> > >> >However, I do not think that removing permissions will end well. > >>Looking > >> >at the bigger picture, I can think of many cases where a permission is > >> >required by the app itself, and should remain regardless of any plugin. > >> > This needs to be decided by the app developer themselves imho. > >> > >> Can you provide one? I don't see how an application's permissions would > >> stem from the application itself and not the device capabilities the app > >> needs to access. > >> > >> I think for now leaving the permission stuff as children under > >> <config-file> is good enough, but tweaking our tooling implementation to > >> be smarter about handling it is the way forward. The idea behind the > >> <config-file> element is: all of its child elements will be appended > >>into > >> specific parts of the native manifests for the platform it is specified > >> under. So, as long as the tooling understands that permissions (and > >> possibly other areas such as "requirements" in windows phone land) are a > >> shared space between different plugins, we are good. > >> > >> > >