Expose all base URLs that can be used for invites/jump links/etc. for parsing purposes

[Jackenmen]

Current problem

Currently, there is no way to know from the data found at https://canary.fluxer.app/.well-known/fluxer what URLs can be used as invites/jump links/etc.

For example, the hosted (canary) Fluxer instance web app accepts all of the following as channel links:

https://web.canary.fluxer.app/channels/1471691126960477817/1474459104743985187
https://web.fluxer.app/channels/1471691126960477817/1474459104743985187
https://fluxer.app/channels/1471691126960477817/1474459104743985187

but there is no way to derive that fluxer.app and web.fluxer.app are valid from instance discovery object since they're all subdomains of canary.fluxer.app.

This is a problem because users can, depending on whether they used the stable or canary client, post both of these on what's essentially the same hosted instance, making it impossible for the bot to parse channel links as input, even though the canary web app has the ability to recognize both just fine (i.e. it renders all of them the same).

More generally, all links used by the web app can either be [web.]fluxer.app or [web.]canary.fluxer.app depending on the client used but they cannot be detected in a message.

One concrete use case where this may especially matter is moderation of invite links being posted - I should be able to know that people can post any of the following and have them render as invite links:

https://fluxer.gg/abcd1234
https://web.fluxer.app/invite/abcd1234
https://fluxer.app/invite/abcd1234
https://canary.fluxer.app/invite/abcd1234
https://canary.web.fluxer.app/invite/abcd1234

Proposed change

I think it would be good, if the endpoints in the instance discovery object included some notion of alt URLs. I'm not sure which endpoints would need to have those - at the very least it seems that it should include the alternative webapp base URLs (e.g. web.fluxer.app would be an alt listed on canary.fluxer.app/.well-known/fluxer) though that would still not make the bare fluxer.app known. The bare fluxer.app happens to be a marketing endpoint but I imagine that the marketing endpoint wouldn't necessarily have /invite, /channels, etc. paths available in other cases so I guess there might also need to be a notion of alt origin URLs? Though clearly the instance discovery objects returned by canary.fluxer.app and fluxer.app are different so perhaps alt URLs wouldn't be the best way to call these.

Additional information

No response

Acknowledgements

• I searched existing discussions.