ok, that makes sense! it isn't even a Context. ya, bad. kill. with. fire. (and a deprecation notice)
maybe we leave it deprecated for a farther future date. I know it doesn't conform to semantic versioning but I think it might be nicer if all the plugins did work for 2.0 maybe, the policy should be not fixed to version number but rather a rough date. if we deprecate something its gone in, lets arbitrarily say, 6 months? On Mon, Jun 18, 2012 at 5:55 PM, Joe Bowser <[email protected]> wrote: > Back when we first started working on plugins, a ctx was a context because > that's what we needed. Along the way, when > we removed PhoneGapActivity and changed it to a CordovaInterface for an > earlier implementation of CordovaWebView, we changed ctx to be an > Interface. The problem is that a CordovaInterface may not be an activity > and this looks stupid: > > ctx.getContext() > > I tried in an earlier version of CordovaWebView to change this back to > Context, but we decided that it should be an interface for some reason > (although I don't remember the reason, something about breaking plugins I > think), so since we can't make ctx a Context like what the convention is, > we should conform to convention and call the CordovaInterface something > descriptive like cordova since that will be less disruptive. > > So, yes, we've been kicking this can around the parking lot for a while. > > On Mon, Jun 18, 2012 at 2:45 PM, Brian LeRoux <[email protected]> wrote: > >> cool w/ that, and of course I trust you, but can you explain the >> problem with ctx, a familiar convention since the earliest days of >> phonegap/android, so I understand the benefit of the proposed >> solution? >> >> (breaking plugins will cause some backlash and, as I mentioned, >> creating a more abstract interface does increase ramp up for new >> native devs) >> >> >> On Mon, Jun 18, 2012 at 4:35 PM, Filip Maj <[email protected]> wrote: >> > Brian, we're doing Android devs (potential plugin authors) a favor here, >> > trust me. >> > >> > 2.0 is our chance to break interfaces. >> > >> > Also, +1 to Bryce's comment re: get this change in for 1.9, deprecate the >> > .ctx member in 1.9 as well, and axe it in 2.0. >> > >> > On 6/18/12 12:15 PM, "Brian LeRoux" <[email protected]> wrote: >> > >> >>I'm of the opinion that native impl should *not* abstract the >> >>platforms at the plugin level. It breaks old plugins, which is fine, >> >>but for what benefit? Conceptual purity at that level will make it >> >>harder to recruit plugin authors from their respective navtive >> >>platforms. >> >> >> >>On Mon, Jun 18, 2012 at 3:06 PM, Michael Brooks >> >><[email protected]> wrote: >> >>> If we are planning to rename the Cordova interface object, then we >> >>>should >> >>> do it for each platform in a consistent manner. There should be a >> parent >> >>> JIRA issue with sub-tasks for each Cordova platform. >> >>> >> >>> On Mon, Jun 18, 2012 at 11:50 AM, Filip Maj <[email protected]> wrote: >> >>> >> >>>> Yeh "ctx" implies Context, especially for Android peoples, so +1 to >> >>>> renaming to something less Android-ey. >> >>>> >> >>>> On 6/18/12 11:45 AM, "Joe Bowser" <[email protected]> wrote: >> >>>> >> >>>> >Hey >> >>>> > >> >>>> >Since we're approaching 2.0 and since part of the goals of 2.0 is to >> >>>> >improve the plugin architecture, I'm wondering if we should take the >> >>>> >opportunity to give the CordovaInterface variable on Plugin.java a >> >>>>name >> >>>> >other than ctx, which on Android usually refers to a context. The >> >>>>reason >> >>>> >for this is the fact that there's a use case where the >> >>>>CordovaInterface >> >>>> >may >> >>>> >not be a Context. I propose that we change the name to cordova. >> >>>> > >> >>>> >I'm not sure if this needs a JIRA ticket or not. >> >>>> > >> >>>> >Any thoughts? >> >>>> > >> >>>> >Joe >> >>>> >> >>>> >> > >>
