@xrevan86 @diogo One of users reported that #AndStatus app doesn't work with the new #GnuSocial server "v.2.0..x dev". I started investigation and see that /api/statusnet/config.json API point now returns 404. Looking in current Source code of the server I see that the API is documented now as /api/gnusocial/config.json - but that returns 404 also...
So the question is did Client API change radically or the server is simply broken, or... ? And where can we see any info on Client to Server API changes / current docs? @colegota
@diogo @xrevan86 Oops, I see that 404 error was due to the wrong domain :-). But my questions regarding API changes and their descriptions are still valid: now authentication fails with 401 error and I simply don't know, where to look at? Any reference client app/tool to lookup?! @colegota@quitter.es
There are no changes planned in V2 regarding existing APIs so, if it isn't working and it was in a prior version, that means we have either accidentally added a bug or some plugin in the faulty instance is somehow interfering :-(
Every GNU social installation has a doc concerning its APIs at instance.tld/doc/twitterapi, as seen, e.g., in: https://loadaverage.org/doc/api
If this problem is present in LoadAverage, please open an issue with instructions to reproduce this, so that we can further debug and get this fixed asap!; Otherwise, please ask the affected instance's sysadmin to open an issue requesting for help, so that we can assist them getting their instance back to normal :)
@diogo Thank you for your explanation. Now I really see a page on the loadaverage site that tells its server version. And #AndStatus still successfully connects to the server (I'm writing to you using it). This means that the authentication problem that I reported is caused by a configuration of a concrete server, and not by the latest version of #GnuSocial :-)
Using a chance to talk with you, I hope that you will find time for Client-to-Server #ActivityPub API also. I'm extending support of that API in AndStatus, and already see that not only #Pleroma, but even some #Mastodon sites begin to implement some parts of the API, allowing a Client app to request information directly from its sources, without a need to duplicate it. E.g. Notes by an Actor may be requested from the Actor's profile (from the Actor's outbox) at his site, and not from the server, at which the User of the Client app has an account...