Conversation

Notices

  1. Looks like RDN got their stuff fixed. I was able to follow nerthos on my gorf.club account.

    Sunday, 29-May-16 20:59:23 UTC from shitposter.club
    1. @normandy It did? Now that you mention it, I can see avatars properly again. Thanks @ceruleanspark and whoever else might have been working on it

      Sunday, 29-May-16 21:03:44 UTC from web
      1. @nerthos Just me, ultimately.

        Sunday, 29-May-16 21:05:01 UTC from web
        1. @ceruleanspark Well, thank you for it. I don't think it's said enough, considering you do it on top of your job and for free.

          Sunday, 29-May-16 21:07:10 UTC from web
      2. @nerthos meh, I think RDN is the only major GS instance that doesn't use TLS. As soon as I pass a notice from that instance in whole known network my SSL session breaks. Is setting up a free Let's Encrypt cert really that hard? D:  @ceruleanspark @normandy

        Sunday, 29-May-16 22:39:43 UTC from sealion.club
        1. @fl0wn "Is setting up a free Let's Encrypt cert really that hard?" I don't know, does switching to HTTPS still completely break existing federation relationships and require every other instance to update our URLs in their databases?

          Sunday, 29-May-16 22:44:22 UTC from web
          1. @ceruleanspark fair point. It's really something that should have been set up from the start but IDK if that was an option considering the age of your instance. AFAIK switching from http to https is a quick way to bork federation. Hopefully there will be a solution in the future because it is a source of major frustration for some of us and currently I think RDN is one of the only instances that breaks SSL on our end but I'm not willing to black hole the server for the sake of added privacy/ security :s

            Sunday, 29-May-16 22:49:08 UTC from sealion.club
            1. @fl0wn I actually already set up letsencrypt on this server to try and turn it on a few days ago, but even enabling it "Sometimes" to cover logins and account management seems to cause major problems with session handling.

              Sunday, 29-May-16 22:59:10 UTC from web
              1. @ceruleanspark as an admin on one of the larger instances maybe you have the weight to submit that as a bug report that would hopefully be taken seriously. Obviously, considering your unique UI you guys have a clue re: how the software works, but as the 'unofficial admin guide' referrers to the inner workings as 'voodoo magic' we still have a long way to go :s

                Sunday, 29-May-16 23:05:48 UTC from sealion.club
                1. @fl0wn @ceruleanspark The software needs better maintained IMO

                  Sunday, 29-May-16 23:14:02 UTC from shitposter.club
                  1. @maiya it's getting to the point where forking may be the better option @ceruleanspark

                    Sunday, 29-May-16 23:15:40 UTC from sealion.club
            2. @fl0wn Also, FYI, the instance was I think, already 2 years old when I inherited it from its old owner, so it was way too late for me to activate it then.

              Sunday, 29-May-16 22:59:45 UTC from web
              1. @ceruleanspark kudos on what you've managed to pull off. I'm just ven5ting a bit bc holiday weekend and something something very strong drinks ;P

                Sunday, 29-May-16 23:07:34 UTC from sealion.club
        2. @fl0wn @nerthos @ceruleanspark The thing about setting up SSL after the fact is that it can get pretty squirrely. I think their version of StatusNet (1.0.1) can have dual http/https configured, but recent versions of GNU Social had that ability removed.

          Sunday, 29-May-16 22:45:54 UTC from shitposter.club
          1. @normandy yeah, this software is a spider-web of WTF at times. It doesn't help that pull requests sit unnoticed for months at times. Don't mind me, I'm just being slightly mango bc beering on a Sunday afternoon due to 'murican holiday FTW =p  @nerthos @ceruleanspark

            Sunday, 29-May-16 22:52:49 UTC from sealion.club
            1. @fl0wn @nerthos @ceruleanspark There needs to be more active maintainers for GNU Social. Not particularly excited with development grinding to a halt just 'cause they went on a holiday.

              Sunday, 29-May-16 22:55:12 UTC from shitposter.club
              1. @normandy agreed! @nerthos @ceruleanspark

                Sunday, 29-May-16 22:55:55 UTC from sealion.club
          2. @normandy How do you do dual HTTP/HTTPS and have it work with remote subs?

            Sunday, 29-May-16 23:00:22 UTC from web
            1. @ceruleanspark I honestly have no clue how to do it. Perhaps someone in !gnusocial or !fediverse knows how to do it.

              Sunday, 29-May-16 23:01:30 UTC from shitposter.club
              1. @normandy @ceruleanspark The reason it got disabled in newer gnuSocial is it uh, basically doesn't work very well at all with federation.

                Sunday, 29-May-16 23:08:04 UTC from shitposter.club
                1. @maiya @ceruleanspark That's a bummer. I've already seen what happened to FragDev when the change happened; a bunch of people moved instances as a result.

                  Sunday, 29-May-16 23:11:48 UTC from shitposter.club
                  1. @normandy @ceruleanspark There really needs to be a good way to switch modes when neccesary.

                    gnuSocial in general isn't very fault-tolerant.

                    Sunday, 29-May-16 23:12:28 UTC from shitposter.club
                    1. @maiya @ceruleanspark I remember there was some work with getting the feed import/export working properly, but I dunno what happened to it.

                      Sunday, 29-May-16 23:13:58 UTC from shitposter.club
    2. @normandy I don't see him still.

      Sunday, 29-May-16 21:03:43 UTC from gs.smuglo.li
      1. @miki Unsub and resub him, see if that works

        Sunday, 29-May-16 21:04:12 UTC from shitposter.club