Time Nick Message 00:02 rubenwardy they never have the messages 00:02 rubenwardy they have an encrypted form 00:02 rubenwardy it's like how ISPs don't save encrypted web views 00:02 rubenwardy well 00:02 rubenwardy hopefully 00:03 rubenwardy this is a matter of law anyway, I personally think that E2E encryption is a necessary tool for privacy, and I prefer to use platforms which use it 00:03 IhrFussel So you basically can do crimes via WhatsApp and as long as the receiver doesn't report it nothing will happen? 00:03 rubenwardy like on your server? 00:04 IhrFussel Not exactly the same... I could see the messages but if my server would make it unreadable for me then it would be the same 00:06 rubenwardy but anyway, there's a massive different between a server DM and WhatApp PMs 00:06 rubenwardy server DM is like "come here" 00:07 rubenwardy "follow me" 00:07 rubenwardy WhatsApp PM is basically a person 2 person conversation. Requiring companies to unencrypt communication is equally bad as requiring citizens to have spying TVs in their walls blasting propaganda 24/7 00:09 IhrFussel Wasn't there a CSM that encrypts the PM and decrypts it for the other user again? Or was it books? Something like that 00:10 rubenwardy it was ineffectual 00:11 IhrFussel It wasn't actual encryption correct? Something custom I guess? 00:12 luk3yx Caesar ciphers are easy to implement, but you could just view all 26 messages side-by-side. 00:14 rubenwardy lool 00:15 rubenwardy "actual encryption" is meaningless 00:15 rubenwardy it used XOR to encrypt, which is not secure at all for that purpose 00:16 Jordach using XOR is like saying you enjoy having someone saw through your bones 00:20 IhrFussel Well, goodnight guys 01:50 Jordach sofar: changes aren't conflicting, ignore git's dumbassery 01:51 sfan5 git is being a git 01:51 Jordach it'll be squashed to a single on merge anyway 01:51 sfan5 haHAAA word play 04:23 BillyS sfan5: Quick question, if you have the time (and knowledge :P). In the MT authentication, when using AUTH_MECHANISM_SRP, when does the server send the client the salt? Is that in the TOCLIENT_HELLO? 04:39 sofar left some comments on jordach's PR. All in all looks very good. 12:16 sfan5 BillyS: there are one (two?) packets that are used especially for SRP iirc 12:32 Krock !tell Jordach The issue with your PR is that you (probably) merged master into your branch instead of rebasing 12:32 MinetestBot Krock: I'll pass that on when Jordach is around 12:33 Krock but "git am" will apply the patches one-by-one, resulting in conflicts which were solved in a later merge commit which becomes useless as soon you resolved the conflicts 12:34 Krock *in a later merge commit. that merge commit becomes 14:52 Jordach Krock: one moment 14:52 Krock how about two? 14:56 Jordach Krock, on my Linux box, the problem that sofar reported isn't there 14:57 Jordach ah, now i see what's going on 15:03 Jordach sofar, Krock fixed it 15:03 Krock how did I fix it? 15:04 Jordach i forgot the comma 15:05 Krock COMMAS ARE IMPORTANT PEOPLE 15:05 Jordach yes 15:05 Jordach turns out, y previously had a 0.5x multiplier hiding it under the cover faces 15:05 Jordach setting it back to previous value 0.22x was visible, 0.44x, hides it for good 15:07 ANAND lol 15:07 Jordach 0.22x was minetest's previous default setting 15:08 Jordach but interestingly enough, that multiplier picks how far it sits at the bottom, not top 15:09 Jordach yep, 0.44x is perfect 15:14 Jordach Krock, pushed the fix commit 16:18 * rubenwardy gets forgetting to type -j5 instead of -j3 16:18 rubenwardy although, with hyperthreading maybe even -j9 16:21 Krock why not $(nproc) ? 16:21 rubenwardy bit of a mouthful 16:21 rubenwardy alias time, maybe 16:22 rubenwardy nproc=8, so with +1 it is 9 16:28 IhrFussel Now I had the spammer on my server...the one who floods the chat with some MT hack link 16:29 rubenwardy I can run minetest at 400 view range with 60FPS 16:30 Krock how about with valgrind enabled? 16:40 Jordach rubenwardy: how's the hardware feeling 16:41 rubenwardy it's very good! 16:41 rubenwardy there's slight issues with the touchpad 16:41 rubenwardy I'm trying to work out whether it's hardware or software 16:41 Jordach but that's a linux problem however 16:41 Jordach afaik, windows precision is a bit pants on Linux 16:41 rubenwardy I *think* it happened in Windows too 16:41 Jordach didn't on thingys 15 16:41 Jordach i didn't sense it so to speak 16:42 rubenwardy if I can't fix it with linux settings, I'll make a copy of my disk for safe keeping then reinstall windows 16:42 rubenwardy to see if it happens there too 16:42 rubenwardy if so, dell repair people time! 16:42 rubenwardy they come to your house to fix it, which is nice 16:43 Jordach there's a few QC issues, but not many major ones 16:43 Jordach rubenwardy: feel free to demo my recentest pull on it 16:44 Jordach (it has appeared to fix the square shaped look of the skybox completely) 16:44 rubenwardy nice 16:45 Jordach at least with 8 threads, MT compiles in a snap 16:46 Jordach also, that M.2 drive probs boots linux in about 3 seconds 16:46 Jordach if not faster 16:50 Calinou rubenwardy: THREADS="$(($(nproc) * 3/2))" 16:51 Calinou Bash doesn't have built-in floating-point arithmetic, so you can do that for using 1.5× the number of CPU threads :P 16:51 Calinou the value is rounded down IIRC 16:56 Fixer rubenwardy: specs? 16:57 rubenwardy i7-8550U with 16GB RAM and a 512 SSD 16:57 Krock beast 16:57 rubenwardy https://www.dell.com/en-uk/shop/laptops-notebooks-and-2-in-1-laptops/new-xps-13/spd/xps-13-9370-laptop/cnx37016 17:00 Fixer rubenwardy: no proper arrow keys 17:01 Fixer rubenwardy: GPU? 17:01 rubenwardy integrated 17:01 rubenwardy for £200 I could've got one, but >nvidia 17:02 Fixer rubenwardy: do you have beard? 17:02 Fixer rubenwardy: what's wrong with nvidia? 17:02 rubenwardy They don't support Linux very well 17:02 rubenwardy ie: GPU switching doesn't work 17:03 Fixer ah, it is bad on laptps you mean 17:03 rubenwardy and desktops 17:03 rubenwardy well 17:03 rubenwardy GPU switching is way more important for laptops 17:03 rubenwardy but anyway, i7s have way better integrated graphics than other classes 17:04 Fixer linux people say nvidia just works and amd still needs one more tweak (and over 9000 later) 17:04 Fixer maybe a meme 17:04 Fixer rubenwardy: 4 core? 17:04 Fixer rubenwardy: 4 cores?* 17:05 rubenwardy there's 8 virtual CPUs 17:05 rubenwardy but 4 physical ones 17:06 Fixer nice 17:06 Fixer rubenwardy: beard ready? 17:07 rubenwardy I have no idea what you mean 17:07 Fixer rubenwardy: https://pbs.twimg.com/media/A4y2092CIAA4kKj.jpg 17:10 * rubenwardy misses AUR 17:11 Calinou yeah, I miss AUR on Fedora too 17:11 Calinou integrating tarball-originating software by hand is boring :P 17:12 Fixer Calinou: reason the arch is soooo popular 17:12 Fixer but 17:12 Fixer arch people suck too 17:12 Fixer reason? 17:12 rubenwardy manjaro kept breaking and I cba 17:12 Calinou I was on Manjaro last year but couldn't install it on my desktop (I couldn't boot the live ISO last December when I needed it) 17:12 Fixer where is full install manual on the install disc? 17:12 rubenwardy ie: latest thing was Android USB support 17:13 Calinou also, Fedora is more robust in my experience (and still offers recent enough packages for the majority of use cases) 17:13 Fixer rubenwardy: probably kernel problem 17:14 Fixer Calinou: do arch users still go to wiki before installing or there is now docs/manual included on disc itse;f? 17:15 rubenwardy the arch wiki is pretty amazing 17:15 Fixer yes 17:15 Fixer arch community is very dedicated 17:15 Fixer they just need to include TFM on disc itself 17:18 Fixer feels like Arch is the new Gentoo 17:18 Fixer remember Gentoo days 17:18 Fixer ? 17:19 Fixer that mindless all day everyday compiling to have +5% speed up? 17:22 Calinou heh 17:22 Calinou Gentoo really fell out of fashion these days 17:27 Fixer people finally realised they wasted 1xxxxxxxxxx hours of their and cpu time on literally useless task 17:28 Fixer new and hip https://distrowatch.com/table.php?distribution=mx 17:31 Krock Calinou: right! Slackware is the new to-go distro 17:32 Calinou to this day, I have more experience with PonyOS than Slackware 17:43 Calinou nope :( but it has Doom and Quake! 17:43 Calinou and arbitrary window rotation 17:44 Fixer arbitrary what? 17:44 Fixer Calinou: compiz? 17:45 Calinou nah, it uses its own window manager 17:45 Calinou "arbitrary" means you can rotate each window by its own angle 17:45 Calinou (not limited to 90 degree increments) 17:45 Fixer but why? 17:45 Calinou it lets you play Doom upside down! 17:54 Krock mirrored upside-down please 17:55 Fixer with wobble effect 17:55 Fixer https://twitter.com/UnusualVideos/status/1061365238467969024 19:21 Krock guys we missed the opportunity 19:21 Krock minetest 4.20 19:22 Krock or 0.4.20 20:20 Jordach Minetest 6.9 release soon:tm: 20:21 luk3yx s/:tm:/™./ 20:22 Fixer Jordach: in 2069 20:22 luk3yx s/20/22/ 20:23 luk3yx Server admins should get control over what CSMs are allowed, but then some of them would just disallow all CSMs. 20:24 T4im aren't they? i thought the csm's get deployed from the servers 20:24 T4im csms* 20:24 luk3yx CSM = Client Side Mod. 20:24 T4im i am aware 20:25 luk3yx Servers have no control over them. 20:25 T4im i thought they get deployed from servers 20:25 luk3yx No, they're installed and run from clients. 20:25 luk3yx I think servers should be allowed (sandboxed) CSMs, however that'd be hard to implement. 20:26 luk3yx Is there a way to switch a node in a CSM (until the server says otherwise)? Client-side door opening would be nice. 20:28 T4im "Transferring client-sided mods from the server to the client is planned, but not implemented yet." 20:28 T4im there we go 20:28 T4im a lot of the ideas wouldn't make sense otherwise 20:30 Fixer Jordach: To see Minetest 6.9 and die (c) 20:30 T4im which only leaves the "lets only run those that the server opts in to" feature for after that 20:30 luk3yx But then server admins would blatantly deny all CSMs. 20:31 T4im nah, the mods installed could opt-in; like the mtg doors mod could offer a CSM for door opening prediction 20:31 rubenwardy server-sent CSMs is the reason it was ever added 20:31 rubenwardy what will happen is servers will be able to stop clients loading their own CSMs 20:31 luk3yx That is... rude. 20:31 T4im reasonable 20:31 rubenwardy then servers can send CSMs 20:32 rubenwardy there's no good way of whitelisting specific CSMs 20:32 luk3yx It's reasonable for oredetect and other hacky CSMs, but for legitimately useful CSMs... 20:32 T4im why not, rubenwardy? 20:32 T4im idtech3 managed so decades ago :D 20:32 rubenwardy instead you can disable certain features 20:32 rubenwardy how would you do it? 20:32 rubenwardy hashes? 20:32 rubenwardy ew 20:32 T4im pretty much, yea 20:33 luk3yx Modified clients will be able to bypass CSM restrictions anyway, so that's not a valid argument. 20:33 T4im no matter what you do, indeed 20:33 rubenwardy the point is to prevent it when they don't modify 20:33 T4im hashes over the packages give you security for unmodified clients though; and allows the server to opt-in to already installed and known versions 20:34 luk3yx But what about CSM updates? Every server owner would need to update their hash. 20:34 rubenwardy the issue with hashes is it essentially makes it impossible to legitimately fork or whatever 20:34 rubenwardy it has issues 20:34 rubenwardy much better to have measures in the API for it 20:34 rubenwardy ie: make the client lie to CSMs, and say default:nyan is default:stone 20:35 rubenwardy and in the long-term, have actually secure mechanisms 20:35 T4im could be solved by merkle tree concepts or assymetric crypto signatures instead of hashes 20:35 rubenwardy like: send blocks with nyans as default:stone 20:35 rubenwardy then only send the actual nyan when the player has LOS 20:35 rubenwardy line-of-sight 20:35 * luk3yx prepares for "that's a bad idea" 20:35 rubenwardy so even modified clients can't ore detect 20:35 luk3yx What about GPG signing? 20:36 T4im well yes, that would be one way 20:36 rubenwardy the disadvantage is with performance 20:36 T4im during startup? 20:36 T4im who cares :D 20:36 rubenwardy agreed 20:36 rubenwardy it probably won't be that much of an issue if time is spent optimisating 20:36 rubenwardy and it would be off by default anyone 20:36 rubenwardy *anyway 20:36 rubenwardy would just be nice to have time spent optimisating to begin with 20:37 rubenwardy hey T4im, do you like optimising? 20:37 T4im hehe 20:37 T4im depends what and when 20:37 T4im i like optimising useful things; but not the things that don't make a difference anyway 20:38 rubenwardy Sure 20:39 rubenwardy Well, map loading and mesh gen would be awesome 20:39 T4im :D 20:39 rubenwardy There's an issue somewhere 20:42 T4im the thing you mentioned with selectively sending is quite common indeed; limiting the effect of wallhacks and such; but what i've seen so far had the benefits of static maps (i.e. precompiled LOS info) and some fancy spatial trees to know what was visible 20:42 T4im minetest might have to get creative there 20:45 * T4im would love to hear about any concepts; if anyone already came up with some 20:49 rubenwardy It's done in Minecraft 20:49 rubenwardy Well, on custom servers 20:49 rubenwardy Oh cool 20:54 T4im some server side collision detection as a side effect of something like that would probably not hurt either :D 20:56 rubenwardy yeah 20:56 rubenwardy there have been some PRs for that 21:00 Jordach >paramat complaining about visual changes 21:00 Jordach how about fix mapgen for once 21:36 Jordach nerzhul: by same project manager you mean mismanager