| --- Log opened Fri Jan 17 00:00:36 2025 |
| 02:14 | | Vornicus [Vorn@Nightstar-8k6f0b.res.spectrum.com] has quit [Connection closed] |
| 02:23 | | Degi [Degi@Nightstar-dieriv.pool.telefonica.de] has quit [Ping timeout: 121 seconds] |
| 02:24 | | Degi [Degi@Nightstar-9k8dhn.pool.telefonica.de] has joined #code |
| 07:16 | | Alek [Alek@Nightstar-e1815p.il.comcast.net] has quit [[NS] Quit: ] |
| 07:27 | | Alek [Alek@Nightstar-e1815p.il.comcast.net] has joined #code |
| 15:00 | | Vornicus [Vorn@Nightstar-8k6f0b.res.spectrum.com] has joined #code |
| 15:00 | | mode/#code [+qo Vornicus Vornicus] by ChanServ |
| 17:43 | <&ToxicFrog> | That's...a good question. |
| 17:43 | <&ToxicFrog> | I've always had a domain, so it never occurred to me to see if it could be used local-only without one. |
| 17:43 | | Emmy [Emmy@Nightstar-qo29c7.fixed.kpn.net] has joined #code |
| 17:43 | <&ToxicFrog> | You'll definitely need one to federate but if you're just using it as a local component in the bridge, that doesn't matter. |
| 17:46 | < abudhabi> | I've been looking into conduit and synapse, and so far I've not been able to make either work. |
| 17:46 | < abudhabi> | This technology seems highly premature. |
| 17:47 | <&ToxicFrog> | Synapse is very "so you want to host a 10k+ user instance" software AFAIK, it has lots of moving parts so I kind of disregarded it |
| 17:48 | < abudhabi> | I've gotten farther with conduit, but all I got for my trouble is "connection refused" when trying to hit the "is it working?" endpoints. |
| 17:49 | <&ToxicFrog> | Connection refused implies it's not even listening on those ports, so that's a problem that happens before stuff like certs |
| 17:49 | <&ToxicFrog> | Like you should at least be able to netcat to localhost:6167 (or whatever port you have it configured for) and get an open socket. |
| 17:50 | < abudhabi> | "Hello from Conduit!". |
| 17:51 | <&ToxicFrog> | That's what you get when you curl it? Ok, so that means it is up and running. |
| 17:52 | < abudhabi> | I'm using caddy (because that allegedly autohandles certificates), and opening 127.0.0.1 shows that the web server is working (ie. I get the default caddy HTML). |
| 17:53 | <&ToxicFrog> | Ok, so you're using caddy as an SSL terminator, and then conduit behind that. |
| 17:53 | < abudhabi> | What's an SSL terminator? |
| 17:54 | <&ToxicFrog> | (or I guess a TLS terminator these days, strictly speaking) |
| 17:55 | <&ToxicFrog> | The service that handles establishing an encrypted connection to the client, then forwards unencrypted data to the actual server that generates responses (typically running on the same machine, or at least in the same datacenter) |
| 17:55 | < abudhabi> | OK. |
| 17:55 | <&ToxicFrog> | So you need to write HTTPS support into every server, you instead handle it at the ingress point with caddy (or nginx or what have you). |
| 17:56 | <&ToxicFrog> | Er, so you DON'T need to |
| 17:56 | < abudhabi> | Got it. |
| 17:58 | <&ToxicFrog> | Anyways. Before fiddling around with certs, I'd try to connect your matrix client of choice to the server just using the IP and port directly. Which I don't actually know how to do beyond "at least some clients allegedly support this". |
| 17:59 | < abudhabi> | I'll try that. |
| 17:59 | < abudhabi> | In the meantime, I think I fixed the caddy problem. The instructions were apparently wrong. |
| 18:00 | < abudhabi> | https://docs.conduit.rs/deploying/generic.html#caddy told me to create a new file, but when I instead added stuff to the existing Caddyfile, I don't get connection refused anymore. |
| 18:00 | < abudhabi> | I get "curl: (60) SSL certificate problem: unable to get local issuer certificate" instead. |
| 18:04 | <&ToxicFrog> | Can you like...skip the SSL or use a self-signed cert or something for now while you get the client setup verified working? |
| 18:04 | <&ToxicFrog> | Assuming this is all running on the same machine you may not even need caddy, just listen on localhost, connect to localhost -- the conduit docs assume you want it net-accessible and federable. |
| 18:05 | < abudhabi> | Right. I'm trying. |
| 18:06 | < abudhabi> | Going through matrix clients until I find something that works. |
| 18:08 | < abudhabi> | FluffyChat and Cinny and Element have not. |
| 18:09 | < abudhabi> | I may be doing something wrong. |
| 18:09 | < abudhabi> | I'm launching these and trying to enter "localhost" in the server box. |
| 18:23 | | Vornicus [Vorn@Nightstar-8k6f0b.res.spectrum.com] has quit [Connection closed] |
| 18:32 | < abudhabi> | On http://localhost:6167/_matrix/client/versions I get: {"versions":["r0.5.0","r0.6.0","v1.1","v1.2","v1.3","v1.4","v1.5"],"unstable_features":{"org.matrix.e2e_cross_signing":true,"org.matrix.msc3916.stable":true}} |
| 18:34 | <&ToxicFrog> | "localhost" without qualification is going to try to connect to port 80 or 443, I think. Which means it's going to try to talk to caddy. |
| 18:34 | <&ToxicFrog> | Try http://localhost:6167/ instead? |
| 18:36 | < abudhabi> | Cinny gives an error message: "Failed to connect. Either homeserver is unavailable at this moment or does not exist." |
| 18:36 | < abudhabi> | FluffyChat just greys out the input box. |
| 18:36 | < abudhabi> | Err, button. |
| 19:32 | | * abudhabi looks into alternatives. |
| 19:33 | | * abudhabi finds ferdium. |
| 19:34 | < abudhabi> | It seems to be a "use official UI within one program instead of having a million of them open" kind of deal. |
| 19:37 | < abudhabi> | At least for the things that already are, essentially, running a browser in a container. |
| 19:37 | < abudhabi> | (Which I dislike for reasons that aren't quite clear to me.) |
| 19:43 | < abudhabi> | (Well, aside from the gigabloat.) |
| 19:50 | <@TheWatcher> | ... you know, if I didn't know better, I'd think you were just making up half the names in this conversation. |
| 20:08 | < abudhabi> | It's better than nothing, I suppose. I can keep a few of the garbage webapps confined to a container that isn't my browser. |
| --- Log closed Sat Jan 18 00:00:38 2025 |