Post B4sCJP31kOZemEBz04 by kirbius@clubcyberia.co
 (DIR) More posts by kirbius@clubcyberia.co
 (DIR) Post #B4sCJP31kOZemEBz04 by kirbius@clubcyberia.co
       1 likes, 0 repeats
       
       correct me if I'm wrong but fedibbs just can't gather all the replies to a post, you can view a post, the amount of replies using the "," command but thats about it, you gotta rely on the notification stack to actually see everything going on in a thread. I guess this is why it says not to use it lol. I don't know go, but i might just try and hack it.
       
 (DIR) Post #B4sCThzw96miRgf2Om by kirbius@clubcyberia.co
       1 likes, 0 repeats
       
       @p
       
 (DIR) Post #B4sJ4RjWBFQMRNsl9s by p@fsebugoutzone.org
       0 likes, 0 repeats
       
       @kirbius Yeah, that's on the list.  And there's some other stuff I have some vague ideas about, like, "all unread notifications" or "oldest unread page" and doing some of the notification-paging client-side because `u` works right but `d` doesn't (req is fine, it's that the paging server-side is not tuned to do this kind of thing, it shows the newest set of notifications that have a min-ID but that just gives you the newest notifications, so you'd have to keep track of the max-ID client-side; easier to just keep everything in the cache).Using `,` to go up a thread makes it easier to find out stuff like "What the hell was that a reply to?" but it doesn't explode the thread, which would be useful.  Like, I use bloat's quickreply feature a lot in the notifications (which is why the FSE version moved it out of the hover menu and under the post).Part of the hold-up on those is coming up with a way that doesn't potentially create disasters when a single action turns into something unbounded in terms of the number of outbound requests it creates.  "Oldest unread page of notifications" is like that, you know, it is potentially a large number of requests for something simple.There's some other stuff that's missing but it's stuff I almost never use, like updating bio/avatar (neither of which is really *hard*, just it's a thing to do) or polls (displaying and creating) and a few things that I do use that aren't implemented (thread-mute/unmute).  It should probably display the URLs for the new-ish quote-posts.  (My suggestion to feld a while back about just recognizing object URLs in the body and showing a preview for all of them instead of having a quote-post thing, that's kind of a non-issue in FediBBS, since you can just `, $object_url` and see whatever.  Obviously, `, $object_url` also makes it easy to go back and forth between bloat and fedibbs or to do stuff cross-instance like `p Postin' this on one instance` followed by `use other_account@other_domain` and then `R $url_from_previous_post Replying from other account.` or `! $url_from_previous_post`.  It really is a very fast client once you get used to it, but I cheated by being the author, so some features are non-obvious like `R $url` followed by `interject You can interject as a reply because it consumes the reply field from dot.`.  I kind of imagined more multi-user setups so the `rumors` would matter more.)You can kinda see where I was going with stuff if you look at the changes I made to the old BBS interface ( https://git.freespeechextremist.com/gitweb/?p=fse;a=blob;f=lib/pleroma/bbs/handler.ex;h=a649dbc34e9430d892274a493bacfd478a902481;hb=HEAD ) but Elixir's kind of a pain.The next thing I am probably doing is improving the inline help because parsing Go is doable in the stdlib and so instead of having the ugly block of special-cased stuff like `help .` or `help R` (the double entendre in `help p` is intentional), it could just read comments from the source and then if you add a command you just put a comment and then `help $cmd` works, and the usage message is right next to the code that implements it, so there's better locality when you're editing.  Or maybe some other stuff.> I guess this is why it says not to use it lol.This is mainly because if the README is hostile enough, I don't have to answer questions unless I want to, and nobody will ask me about making it easier to use or implementing stupid goddamn line-editing server-side.I might have to do more(1)-style paging server-side, I don't think there's a great way to do that otherwise.  Like, line-editing should be done client-side so you can use rlwrap if you aren't using it in acme/rio windows, but paging is harder to do externally.  This usually wouldn't matter except PicoCalc->ssh->fedibbs or TRS-80 Model 100->/dev/ttyS0->fedibbs kind of needs paging.  Another thing would be, like, pseudo-IDs:  it's annoying to type out something like `B4pZGmyivNTyyxYiqO` so assigning short IDs would be nice, like a 256-entry ring-buffer in bigmoney.go and just as posts are fetched, they get added there, too, and pick a piece of punctuation like `&` maybe or `~` so you could do `R &7f Replying with a pseudo-ID requires less typing` without introducing ambiguity.  (Like, `R 1 This is ambiguous because it is possible to find notification #1 or the post with the ID 1.` for example.  The thing that prevents it is that notifications are dumped every time you fetch a new page; that's something that'd have to stop for the client-side notification-paging above.)> I don't know go, but i might just try and hack it. :cereal: HACK THE PLANET :zerocool: