| 3 Jan 2022 |
 Samantaz Fox | TheFrenchGhosty mind shortening the output next time? ^^ | 14:52:24 |
 TheFrenchGhosty | irc doesn't like codeblock? xD | 14:52:50 |
| Samantaz Fox | If the bot didn't post a link instead, that would be cahos, yeah | 14:53:39 |
| Samantaz Fox | (And my matrix client shows a complete copy of the message whenever somebody replies to it) | 14:55:54 |
 zehka | Do you assume the bug to be inside invidious or inside crystal? It seems TLS is handled by openssl but wrapped by crystal? | 15:10:25 |
| Samantaz Fox | I'm pretty sure it's NOT coming from invidious | 15:10:50 |
| Samantaz Fox | invidious is only a "user" of the Crystal library on that part. | 15:11:12 |
| Samantaz Fox | So the preblem lies either in the crystal implementation, the library bindings, or the version of openSSL shipped on this distro. | 15:11:49 |
| Samantaz Fox | I have something to do, brb | 15:13:39 |
| zehka | Well, it might be more complicated than we thought.
Good news first: I found a workaround that makes invidious work: ln -s /etc/pki/tls/cert.pem /etc/ssl/ | 15:30:59 |
| zehka | Strange news: If i run a small test program that opens a https connection to youtube and prints the status code, that works without any need for a workaround and opens /etc/pki/tls/cert.pem directly. | 15:32:58 |
| zehka | # test.cr
require "http/client"
client = HTTP::Client.new("www.youtube.com", tls: true)
response = client.get "/"
puts response.status_code
| 15:33:16 |
| zehka | So the problem seems to be induced further up than inside the HTTP library | 15:33:47 |
| Samantaz Fox | mmmh, interesting. Mind posting that on the issue? | 15:35:20 |
| zehka | I'll do that in a few minutes | 15:39:40 |
|  Guest26 joined the room. | 15:52:18 |
| Guest26 left the room. | 15:54:37 |
| zehka | In reply to @samantazfox:pussthecat.org
mmmh, interesting. Mind posting that on the issue? https://github.com/iv-org/invidious/issues/2753#issuecomment-1004193972 | 16:08:12 |
| Samantaz Fox | zehka nice | 16:08:50 |
|  Guest48 joined the room. | 20:14:45 |
| Guest48 left the room. | 20:15:03 |
| Samantaz Fox | zehka welp, I've been loonking for quite some time now and I'm unable to reproduce the issue on a small subset of the codebase.... | 23:50:46 |
| Samantaz Fox | I'd recommend to use docker instead | 23:50:54 |
| 4 Jan 2022 |
| Samantaz Fox | Ok, I think I found out.... | 00:33:08 |
| Samantaz Fox | That's what I suspected since the beginning, but because of a typo, couldn't be ruled out | 00:33:26 |
| Samantaz Fox | I really feel the cheapness of this VPS xD | 00:38:15 |
| Samantaz Fox | 5 minutes for a debug compilation | 00:38:26 |
| Samantaz Fox | Ok, that's lsquic | 00:42:17 |
 RSS (bot) |
New commit to Invidious: Flag to disable QUIC should be 'disable_quic'
| 00:47:14 |
| Samantaz Fox | zehka you can try again to compile without QUIC (using latest master) | 00:49:42 |
