Time Nick Message 01:27 ANAND rubenwardy: That would be great, thanks 01:28 ANAND I'm surprised this hasn't been reported before though... o.O 01:30 thoughtjigs #7445 01:30 ShadowBot https://github.com/minetest/minetest/issues/7445 -- Xorg/net wm pid by thoughtjigs 01:30 thoughtjigs @nerzhul do you need to aprove the above before it can be merged? 03:21 Edgy1 larrythecow, so i switched from Linux Mint to Debian on a server i host 03:21 Edgy1 also switched from 0.4.16 to 0.4.17.1 stable 03:21 Edgy1 i'm having problems with the server timing out a lot for no reason. also rollback doesnt work, the server completely times out and is useless until i restart if i try to use rollback 14:06 johnnyjoy Sorry to interrupt. I'm about to spend a great deal of time looking into SRP to see if it can use leverage existing accounts in LDAP, similar to Active Directory.. Research in the last 24hrs has turned next to nothing. Does anyone know if this is just not possible, or is someone has already looked into this? 14:08 sfan5 unless you store your passwords in an SRP format compatible to minetest (or in plaintext), there's not much chance for this 14:14 johnnyjoy What a shame. Seems like a crippling incompatibility in the software world. It seems so normal just to plug into any auth system there is. 14:15 johnnyjoy Maybe there is a possibility of having the client negotiate authentication mechanisms with the server, opening up the possibility of an actual authentication handler mod? 14:43 sfan5 the client does support auth negotiation on the protocol level, however the only mechanism supported is SRP 14:43 paramat rubenwardy any reason to keep #6538 open? 14:43 ShadowBot https://github.com/minetest/minetest/issues/6538 -- Formspec: Allow 'position[]' anywhere, just like backgrounds 14:51 paramat please can core devs look at issues marked 'possible close' and add their opinions and/or judge whether the issue is worth keeping open? 14:52 paramat we need to do this more to keep issue number under control 14:53 sfan5 johnnyjoy: since you're talking about a school situation on the forums: schools probably have the plaintext password anyway, so you could create a centralized database of srp hashes generated from the plaintext passwords and then make mt use that 14:55 johnnyjoy sfan5: That is a great idea, but the whole school district here is run by a very good Linux team using think clients on the desktops. The are using a LDAP and have never stored a single password in clear text. 14:56 sfan5 oh do students choose their passwords themselves? 14:56 johnnyjoy sfan5: I'm not sure. They might. 14:56 sfan5 ah that ruins this system 14:57 johnnyjoy sfan5: It doesn't really matter. Most places are going to a federated login system usig SAML or oauth. 14:58 johnnyjoy sfan5: Honestly most software has some kind of auth layer that allows for a large number of options for authentication. It seems like a huge flaw that MT doesn't meet that standard. 14:59 johnnyjoy sfan5: I have worked out the problem for my situation, but it helps almost nobody else sadly. 14:59 sfan5 "most software" also sends the password of the network in plain-text 15:00 johnnyjoy sfan5: I will have people sign up only via a web site, and when they do a SRP hash will be created at that time, along with the normal hash. 15:00 sfan5 mt would have supported all of this if it wasn't for SRP 15:00 johnnyjoy sfan5: most software does send the password in clear text over an encrypted connection. 15:01 johnnyjoy sfan5: I understand that without SRP MT wouldn't be secure and any sniffer on the net could take all the passwords. 15:02 johnnyjoy sfan5: SRP as a solution is great. But, as a singular solution is limiting. 15:02 sfan5 true 15:02 sfan5 perhaps sending the raw password can be re-implemented, but disabled by default 15:03 sfan5 then schools could deploy mt with this enabled and still have their LDAP integration 15:03 johnnyjoy sfan5: Imagine having an option to use TLS with plain text passwords using a self signed cert as an option. 15:03 johnnyjoy I see MT as one of the most amazing things children in poor areas of the world can experience. 15:04 johnnyjoy They might not have any actual toys, but MT gives them something to play with that can be even better. 15:05 johnnyjoy My own son was never social at all, but playing online changed all that and change his entire life for the better. He's a very different person. 15:05 johnnyjoy Never doubt that your work makes a difference. 15:10 johnnyjoy Thanks sfan5. I'm going to get coffee and create my interim solution, then look into serverpackethandler.cpp to see how server can offer up different auth options to the client. It might be more than I can time to get to, but I'm going to look. Take care. 15:49 thoughtjigs johnnyjoy, sfan5: auth was the majority of what I did at pokemon. I would be happy to take a look at it. Is there an issue number? 15:49 sfan5 no issue open for this afaik 15:53 thoughtjigs ok, I'll go take a look at the source and create an issue ... probably a discussuion issue first. 15:54 thoughtjigs sfan5: with my Xorg stuff ... what needs to happen next? 15:54 sfan5 some other devs needs to merge the pr 15:55 sfan5 s/devs/dev/ obviously 15:56 thoughtjigs should I ping someone in particular? 15:58 sfan5 no, you can however ping Krock or nerzhul since both are active with developing the C++ part 16:14 thoughtjigs Ok, thanks 16:23 celeron55 it seems fine to me but i don't know much about X11 16:26 celeron55 (and can't test at the moment) 16:45 thoughtjigs celeron55: Thanks for reviewing! 16:49 celeron55 that wasn't much of a review, lol 16:52 thoughtjigs lol, well 5 minutes from the one and only celeron55 should be worth several hours from any other dev ... right?!? :-D 17:34 paramat #7455 seems good now 17:34 ShadowBot https://github.com/minetest/minetest/issues/7455 -- Remove deprecated functions from menu_lua_api.txt by ClobberXD 19:25 paramat another excellent mapgen feature/optimisation #7456 19:25 ShadowBot https://github.com/minetest/minetest/issues/7456 -- Decoration API: Add lightweight ability to have complete coverage by paramat 19:28 paramat oh forgot docs .. 19:29 paramat =s 19:38 paramat done