Time Nick Message 00:01 hmmmm paramat: 4174 looks good to me if you include a note that you don't create the heightmap in generateBaseTerrain anymore 00:09 paramat ok thanks 00:23 hmmmm liquids are typically floodable, right? 00:25 hmmmm hmm nope, air is floodable, liquids are not floodable 00:25 paramat correct, i hadn't considered that 00:27 paramat not sure it's wise 00:28 paramat floodable sortof means vacuum or gas 00:29 hmmmm yeah I'm asking because I'm doing a review on that updateLiquids fix/optimization 00:29 hmmmm https://github.com/minetest/minetest/pull/4065#discussion_r65467058 00:32 hmmmm ahhh 00:32 hmmmm very subtle 00:33 hmmmm notice how the node being checked for floodable is the current node, not the node above the current node 00:33 hmmmm i and j kinda look alike 00:33 hmmmm okay, the logic for this function is much more logical now 00:56 paramat looking at 4065 now and will look more after food 01:08 paramat there is talk of adding an 'on flood' callback so that certain nodes can trigger an action when flooded (drop node, extinguish torch etc). someone suggested allowing water to flood lava to trigger a lavacooling action. so i guess a floodable liquid *could* happen 01:11 hmmmm yeah 01:11 hmmmm talking about floods 01:12 hmmmm in minecraft there's a mechanism of liquid mechanics where if enough liquid nodes surround a floodable area, it'll rapidly fill up the gap 01:18 hmmmm I think this is what I'm talking about: https://youtu.be/bi80MBXIAu4?list=PLvh8CGg6LWwqcebmuPL36z7UFp_aRFHlH&t=971 01:20 VanessaE aside from MC filling the water in faster, it looks more or less the same as MT. 05:56 paramat please can anyone review #4163 ? 05:56 ShadowBot https://github.com/minetest/minetest/issues/4163 -- Sky: Darker, bluer sky and improved horizon haze at night by paramat 06:18 paramat will merge #4174 once checks are done 06:18 ShadowBot https://github.com/minetest/minetest/issues/4174 -- Mgv7: Always carve river channels in mountain terrain by paramat 06:25 paramat merging 06:38 paramat done 06:48 paramat hmmmm, some explanation from t4im was added to #3176 06:48 ShadowBot https://github.com/minetest/minetest/issues/3176 -- Don't have Schematic sources assign schematics to globals 07:05 hmmmm I get what t4im means, it's just that i don't get what minetest needs to have in order to do that... 07:05 hmmmm he wants a .lua source code file to effectively work like a schematic file 07:19 paramat #1908 could perhaps be merged 07:19 ShadowBot https://github.com/minetest/minetest/issues/1908 -- Make maximal view bobbing speed faster by Calinou 07:36 hmmmm where to start with that one... 07:37 hmmmm first of all the PR title says "make maximal view bobbing speed faster" but that's not true 07:37 hmmmm that makes the minimum view bobbing speed faster 07:38 paramat hmmmm will your upcoming mapgen settings changes make #2561 non-applicable? (sorry for topic change) 07:38 ShadowBot https://github.com/minetest/minetest/issues/2561 -- Add mapgen settings to create world dialog by srifqi 07:38 hmmmm rather, nevermind 07:39 hmmmm i misunderstood 07:39 hmmmm i'll approve of 1908 07:39 paramat ok 07:39 hmmmm but wait though 07:39 hmmmm does the player's speed typically get above 40? 07:40 paramat well i think it's requested for sprint mods or potions etc 07:40 hmmmm i just want to check first 07:41 hmmmm if this is going to affect normal gameplay, then it becomes a matter of preference 07:42 paramat yes i'm unsure if it's just a rarely hit upper limit to stop too-fast motion during glitches 07:42 paramat guess i shouldn't have +1 07:44 hmmmm ahh okay 07:44 hmmmm so i just tested it, and walking normally is 40 07:45 hmmmm sprinting is 200 07:45 hmmmm so, yes, it does make a very noticable change to the default gameplay 07:46 Calinou paramat: the value might be too high 07:46 Calinou 60 should be plenty as max view bobbing speed 07:50 paramat is it in nodes per second x 10? 07:51 hmmmm i think so. 07:51 hmmmm i'm retracting my +1... 07:53 Calinou paramat: yes because it's in Irrlicht units 07:53 Calinou hmmmm: yeah that PR can be closed, I can make a new one 07:53 hmmmm why..? 07:53 hmmmm do PRs get moldy? 07:53 paramat a normal speed bob seems wrong if somehow travelling faster, i would tune it to a reasonable fastest bob 07:53 Calinou I could redo the commit if I find how 07:54 paramat PR doesn't need rebasing 07:54 Calinou hmmmm: by the way, you said that Signed-off-by commits are forbidden in Minetest Git, but does that disallow all GPG-signed commits? I've tried signing a commit and it has no Signed-off-by line in commit message 07:54 hmmmm wait, I didn't say that 07:54 hmmmm I'm pretty sure est said that 07:54 Calinou someone else did in Minetest 07:54 Calinou maybe not sure 07:54 Calinou maybe not you* 07:55 hmmmm I don't know anything about those to be honest 07:55 paramat oops it does need rebase 07:55 Calinou GitHub offers fancy "Verified" badges for GPG-signed commits, so people are using these more now 07:55 Calinou like https://github.com/godotengine/godot/commits/master 07:55 hmmmm why would it matter for adjusting a number... 07:55 Calinou it looks good :3 07:55 hmmmm well i'd hold off and see what everybody else says 07:56 hmmmm i don't get why you'd only want some commits to be verified and not all of them 08:03 Zeno` est31 had a good reason for disallowing them, I just can't remember what it was atm :) 08:04 Zeno` but at the time it was being discussed I do recall asking why they'd be a problem 08:18 paramat any more disapprovals for #3898 ? 08:18 ShadowBot https://github.com/minetest/minetest/issues/3898 -- Make item groups support floating point values by ShadowNinja 12:22 est31 There are generally two reasons for disallowing signed git commits, the second one being a half-reason: 12:23 est31 1. what is inside contributing.mc 12:23 est31 md* 12:23 est31 https://github.com/minetest/minetest/blob/master/CONTRIBUTING.md 12:23 est31 and 2. 12:23 est31 It of course breaks with how we merge prs 12:23 est31 because right now we just take the patch file and am it 12:23 est31 and that destroys the signature anyway 12:24 est31 so even if we *allowed* signed commits, we would not get them anyway 12:24 est31 except for commits from core devs which get pushed directly 12:25 est31 so essentially its this bug report which is the cause for the strict ban: https://bugs.launchpad.net/bzr-git/+bug/1084403 12:51 est31 Calinou, see my answer above 12:51 est31 hmmm is right it was me who forbid it :) 12:52 Calinou right, launchpad :p 12:53 est31 I have nothing against signed-off-by btw 12:53 est31 no problem with that, just it should be used consistently 13:36 est31 but generally Calinou in the future if that bug is resolved I might use signed commits myself 13:38 Zeno` It's a bit redundant 13:38 Zeno` I.e. if a dev commits something then... well, they "signed it off" 14:34 Calinou Zeno`: it's easy to impersonate Git commits IIRC 14:34 Calinou it's added security and verifiability, nice to have 14:40 est31 https://github.com/jayphelps/git-blame-someone-else 14:40 est31 https://github.com/jayphelps/git-blame-someone-else/commit/e5cfe4bb2190a2ae406d5f0b8f49c32ac0f01cd7 14:41 Zeno` that's just amend author isn't it? 14:41 est31 yeah 14:42 Zeno` so to actually push it'd still be me doing the pushing 14:42 est31 yeah 14:42 est31 well with this signed commit stuff github has added a feature where you can verify a change was really made by a certain person 14:45 GnomeKris I got over 25 emails last night for every little update to the github. lmao 14:45 Zeno` I don't get it to be honest. To push to minetest master it uses RSA anyway 14:45 est31 well you have to log in to github in order to do it 14:45 Zeno` If I decided to start amending authors I'm sure somebody would notice hehe 14:46 Zeno` to push to minetest origin? 14:46 est31 yeah 14:46 Zeno` How do you push without logging in? 14:46 est31 you can use github password based auth as well 14:46 est31 no ssh keys needed 14:46 Zeno` well, that's the problem not git (surely) 14:47 Zeno` I mean... how would anyone get my private key? 14:47 Zeno` even if they had my github password 14:47 est31 shrug, if minetest wasn't a stupid game but idk a crypto library or something I would support a requirement to sign all commits 14:47 Zeno` github should change that heh 14:48 est31 Zeno`, for example by opening a PR which contains malicious code and you download and compile it without looking at the changes 14:48 Zeno` so you can push without RSA? 14:48 est31 and then it copies the ~/.ssh/ to some hacker owned place 14:48 est31 then they have your key 14:49 Zeno` yeah ok 14:49 Zeno` and how does signing get around that? 14:50 est31 not at all :) 14:50 Zeno` haha :D 14:50 est31 you can of course get a hardware token that signs the commits 14:50 Zeno` I dunno if we need hardware tokens heh 14:50 est31 yubikey I think 14:50 est31 the kernel requires this from its maintainers 14:51 Zeno` If we did I'd lose mine within days... I lose everything 14:51 est31 you need a hw token to sign commits 14:51 est31 but we are a voxel game 14:52 est31 not security relevant software 14:53 Zeno` I actaully don't have a ~/.ssh/ directory 14:53 Zeno` mine is called ~/.secretstuffhere/ 14:53 Zeno` doh! 14:53 Zeno` now I need to change :( 14:53 est31 lol 14:54 est31 well of course the ultimate virus hacker reads your /etc/sshd.conf and then determines the dir based on that 14:55 est31 maybe in the future with wayland and this xdg-app and snappy stuff we'll see more sandboxing 15:57 est31 OMG 15:57 est31 I shouldnt have looked into github notifications 16:05 sofar haha, yeah 16:06 sofar the other day mine was longer than a whole screen full 16:21 Zeno` est31, how is writing to an out-of-bounds address going to get you the value that the address(es) contain(s)? 16:21 est31 idk 16:22 est31 but if that address contains the pointer to the next elem in a linked list 16:22 est31 and that address is currently 0, so it means the list is empty 16:22 est31 and now you write a new address into it 16:22 est31 to memory controlled by you 16:22 est31 and that linked list was the list of privs of a player 16:23 est31 or idk 16:23 est31 something else you can abuse to get shell access 16:23 nore est31: well, it won't happen with the lua api, will it? 16:24 nore I mean, even voxelmanips are not dangerous 16:24 nore since they convert their data into a lua table 16:24 nore and you can write anywhere you want in a lua table 16:26 Zeno` correct 16:27 est31 well VoxelManips dont put their data into a table 16:27 est31 they offer normal index based access, no? 16:28 Zeno` not that I know of but I'd have to check the code 16:28 est31 set_node_at 16:28 est31 http://rubenwardy.com/minetest_modding_book/lua_api.html#set_node_atposnode 16:30 est31 Hrmmm setNodeNoEmerge does do a check 16:31 Zeno` MapNode readnode(lua_State *L, int index, INodeDefManager *ndef) 16:31 Zeno` that's a copy 16:31 Zeno` and yes 16:32 est31 readnode is not the issue here 16:32 est31 reading nodes is fine 16:32 est31 just writing them outside of the boundaries of the manip 16:37 est31 tbh I don't really know what is going wrong here. 16:37 Zeno` well, walking through the code briefly I can't see a real issue but if someone wants to do an audit then *shrug* :) 16:38 est31 from what I see set_node_at does make a boundary check 16:38 Zeno` yeah it checks m_area.contains(p) 16:38 Zeno` (or i) 16:40 Zeno` It can't hurt to check 16:41 Zeno` I personally think the "issue" as reported doesn't really exist though heh 16:41 Zeno` I'm happy to be proved wrong 16:43 est31 See my comment on #2685 16:43 ShadowBot https://github.com/minetest/minetest/issues/2685 -- Writing outside the VoxelManip boundaries modifies weird locations on the map. 16:43 est31 my check was not too thorough, so an additional pair of eyes could help 16:44 Zeno` my check was not thorough either... if someone can provide a test case to demonstrate the "issue" that'd be better 16:45 est31 yeah 16:52 Zeno` the is in bounds lua function could be added though 16:52 Zeno` since the calculation of the index is meant to be opaque 16:52 Zeno` at least then mods could check in a couple of ways 16:54 Zeno` not entirely sure what it would achieve, but it's better than a big <= >= x, y, z conditional maybe 16:56 Zeno` then again if I was writing a mod I'd just make sure that the code kept things within range 16:56 Zeno` so *shrug* probably useless 17:23 hmmmm phew you scared me into thinking VoxelManip was a sec vuln 17:26 sofar someone write a minetest API fuzzer :D 17:27 hmmmm don't mean to sound cocky, but i'm pretty confident that i wouldn't have created such an obvious bug like that 17:27 hmmmm i'm not that careless 21:42 Fixer https://github.com/minetest/minetest/pull/4170 23:46 Fixer !tell paramat tried newest nether: https://i.imgur.com/g6xj0nN.png exit teleport was without one wall for some reason 23:46 ShadowBot Fixer: O.K. 23:48 Fixer !tell paramat entry point teleport position: https://i.imgur.com/5PCOgMm.png 23:48 ShadowBot Fixer: O.K. 23:54 Fixer https://i.imgur.com/GLQcc0l.png :/