Add the "url" launchable type#144
Add the "url" launchable type#144ximion merged 1 commit intoximion:masterfrom dchristidis:url_kind_web_app
Conversation
|
The fact that it has a web-app could be circumstancial. Sounds like it should be the value of the a See https://www.freedesktop.org/software/appstream/docs/chap-CollectionData.html
|
|
This needs some discussion first. Web applications have never been properly defined in AppStream, and it has been a long-time item on my todo list. |
|
Web apps don't really have a bundle, they're not bundled blobs, they really are just a URL and some metadata already in the AppStream XML. |
|
@hughsie Jup, that's my problem with using So, from that standpoint is see webapp-URLs being closer to bundle than to URL. In order for this to go into AppStream properly, we would also need to determine what a webapp is exactly, and how a software center should deal with it (just open the URL in a webbrowser? or something different?). |
|
Actually, the intended purpose here is closest in spirit to the As for how software centers should handle this, there are still open questions (should they use a dedicated browser? what is the difference from a web app compared to a simple hyperlink? which websites actually are apps?) I also wonder how we distinguish the current |
|
Launchable makes sense to me too. |
It's a different component type for sure. |
|
@hughsie At the moment, those are of the "service" component-type, because they essentially are services that run in the background. But let's not worry about that yet. |
This must be used for components of type "web-application".
|
Thank you for discussing this issue. I have submitted a new patch for which I’m hoping to receive feedback on. I’m not sure where to state that a |
Exactly. Web applications were not really supported in AppStream until now and were more of a hidden feature. It's likely time to change that, and I already started drafting the new section in the spec. Your patch looks fine to me, thanks! |
From my initial PR on appstream-glib:
I’m hoping that the justification is reasonable and the implementation correct.