[08:26] alard. I think I broke the warriorhq [09:26] soultcer: the trackers are broked as well. [09:27] Oh, then it might not have been me [09:27] We had the problem before that the tracker used too much memory [09:28] The tracker uses redis as database which keeps everything in ram [09:28] If punchfork is done we could take that down [09:28] Anyone besides chronomex and alard have access to the tracker server? [15:33] soultcer: If I'm not mistaken, the "tornado" python library also needs to be compiled (regarding earlier conversation) [15:56] thx [21:17] alard: Would adding http basic auth to the warrior as a simple configuration option in the config.json file be ok? [21:18] soultcer: I don't have objections to http basic auth. But config.json? [21:18] Well, all config stuff for seesaw is stored in there [21:19] And I made the ec2 image read the "userdata" and put it into config.json [21:19] In that case it's probably a file that gets overwritten. [21:20] But looking at the code I think it's saving everything, including new config values, so that should be okay. [21:21] You could even add it as an option in the configuration panel: https://github.com/ArchiveTeam/seesaw-kit/blob/master/seesaw/warrior.py#L146-L172 [21:21] If I make two ConfigValues "http_auth_user" and "http_auth_password", it should work fine, right? [21:21] Yes. [21:22] And then in seesaw/web.py I simply do a check if there is http auth set, and enforce it if it is [21:22] Yes. Do you know how that works with the websocket connection? [21:23] I have no idea but I guess I will have to learn it :D