Time Nick Message 00:19 paramat #4100 hmmmm zeno can anyone review this simple mapgen-related tweak? 00:19 ShadowBot https://github.com/minetest/minetest/issues/4100 -- Defaultsettings/Android: Increase 'max block generate distance' to 3 by paramat 00:19 hmmmm okay wait a minute 00:19 hmmmm who signed off on the android default settings commit 00:19 hmmmm all the people who voted for it 00:19 hmmmm i know i gave a -1 00:20 paramat c55 was possibly a +1 00:21 hmmmm and zeno 00:21 paramat i was going to +1 it with this change 00:21 hmmmm well, ask them 00:21 hmmmm you should get approval from c55 and zeno, not me, since you're basically overwriting the changes they approved 00:22 hmmmm (i don't think any of the settings should have been touched unless a developer actually had access to an android device and *tested* each setting's effect) 00:22 hmmmm gasp, how dare he want evidence for the snake oil! 00:23 paramat ok. but as it's mapgen related your input is useful 00:23 hmmmm you already know what my opinion is, but i won't give any approvals because it's not my place tbh 00:24 paramat ok, but i guess you would prefer 3 to 2 as it's less nerfed :) 00:24 hmmmm also same goes with you 00:25 hmmmm have you actually tested experimentally to see if there's a difference between 3 and 2 on android? 00:25 hmmmm perhaps the issue is that developers are attempting to offer support to a platform that they are not able to test on 00:25 paramat no, no android 00:26 paramat i mean i don't have an android device 00:26 hmmmm i think this could be addressed however 00:27 paramat so ok i'll ask for a test with 3 instead of 2 00:27 hmmmm the money gathered from the donations should go to purchase one of those raspberry pi clones that sell for $10 and given to a developer who has the time and motivation to perform android testing 00:27 paramat zeno has android 00:27 hmmmm mm 00:28 paramat we may have others too with that device 00:29 paramat btw wishing you health 00:29 hmmmm th..thanks 00:31 paramat i can't be bothered with a smartphone, too busy and i kinda wish we didn't bother with them 00:34 hmmmm i don't have a smartphone either 00:35 paramat your mapgen changes look good, i guess we merge those soon after release 00:35 hmmmm i am not even getting started 00:35 hmmmm i haven't had any time to work on them but there's still a lot more to go 00:46 paramat good. i'll merge #4092 soon 00:46 ShadowBot https://github.com/minetest/minetest/issues/4092 -- Keep scroll position constant in ChatBuffer::deleteOldest() by kahrl 00:51 Wayward_One speaking of mobile devices and different settings, FWIW, i have the settings on my $99 tablet similar to the settings on my pc, and i often hit 40 fps 00:57 paramat a difficult balance due to varying touchscreen device power 01:01 paramat those with more powerful tablets would be better off plugging in a keyboard and using the PC version 01:02 paramat android seems to cause many issues, personally i would love to drop it :} 01:14 hmmmm i suppose 4092 is good, it's just that i lack knowledge of the surrounding code to make a quick decision 01:15 hmmmm the code looks great if it does what it's supposed to do, but for example, i don't know what del_formatted is vs del_unformatted and which is the correct position to use 01:17 hmmmm i'd rather somebody else take a look 01:47 paramat i'm going to merge 4092 now because close to release it's important it gets wider testing 01:58 paramat merged 01:59 paramat uh over half of engine milestones are andoid 02:01 paramat so it seems sapier was our android dev, but is not around anymore. i think we need a new core dev specialising in android, otherwise drop it 02:02 paramat because i oversee and organise PRs and issues i seem to spend a lot of time on android. after this release i might go on droid-strike :) 02:05 paramat MT on android is simply crap, better to not do it at all than have a crappy version of MT which also takes up dev time 02:10 paramat it just doesn't scale to a small screen with touchscreen controls. smartphones need dedicated ultra-simplified voxel games 02:33 jhcole is there a way to get the un-aliased version of a node name? node.name is returning the alias 02:45 paramat not sure 03:20 Zeno` maybe for next release we should add all the android settings tweaks to a text file and load them from the java :/ 03:21 Zeno` or host them somewhere and use wget hahah 03:47 gregorycu tell Fixer no 03:47 gregorycu !tell Fixer no 03:47 ShadowBot gregorycu: O.K. 03:52 Wayward_One Tbh, I don't get what all this android-bashing is about. Perhaps it does depend on the device, but from the perspective of a player on a 6" screen I find the android app rather enjoyable. The only major downsides I can think of are the arrows instead of a virtual joystick and the double-tap to place nodes 03:52 Zeno` Wayward_One, I'm not bashing it (I use it as you know) 03:53 gregorycu Bloody hell 03:53 Zeno` I'm referring to that twice now I've committed that particular PR only for the author to add a comment 10 minutes later that he'd forgotten something in it lol 03:53 gregorycu Where are all these new issues coming from 03:53 gregorycu Bugfix? 03:54 Wayward_One Zeno`, I know, but I've just heard lots of negative things about it lately from various people 03:54 Zeno` gregorycu, someone decided that adding Bug label to a PR that fixes a bug didn't quite make sense (and I agree with them) 03:54 Wayward_One So I thought I'd give my two cents on the matter :3 03:54 Zeno` Wayward_One, they are strange yes 03:55 gregorycu Is there some way I can pair PR and bug on the issues screen 03:55 Zeno` Wayward_One, there are not really any coredevs who even test the Android builds. I do, sapier does (but he is not really around) and I think SN might 03:56 Zeno` gregorycu, not if nobody provides a cross-reference and even then I don't think you can pair them on the same screen 03:56 gregorycu Well, the lack of cross reference is a data issue 03:56 gregorycu (Which we should fix when we can) 03:56 Zeno` yeah 03:57 Wayward_One Ah, ok 03:57 gregorycu The sooner we get this release out, the sooner I may become core dev 03:57 Zeno` anyway, I agree that people who don't even use Android shouldn't really be providing non-stop feedback on something they've never experienced :) 03:58 Zeno` yeah, and then we can blame gcu for all Windows issues 03:59 Wayward_One Hehe 03:59 gregorycu I'm not the only windows dev! 03:59 gregorycu I think 03:59 * Zeno` starts adding Windows labels to bug reports 03:59 Zeno` sfan does a lot of Windows stuff 03:59 Zeno` and umm... 03:59 Zeno` hmm, he's not a core dev 04:00 Zeno` anyway, TheJoker does also 04:00 Wayward_One "Dang it gcu, the windows binaries don't work on my linux install" 04:00 Wayward_One :3 04:00 Zeno` Wayward_One, install WINE 04:00 Zeno` :3 04:00 Zeno` (TheJoker == Krock) 04:01 Zeno` Krockodile Dundee 04:01 Zeno` except he's not Australian 04:16 OldCoder WINE is FINE 04:16 Zeno` The Red Red wine goes to my head 04:16 gregorycu Try the Wine 04:16 OldCoder I've found that my WIN32 builds work well under Linux with WINE 04:17 gregorycu Are they performant? 04:17 Zeno` OldCoder, it does 04:17 OldCoder Performant? Do you mean fast? Yes... 04:17 OldCoder But no special Windows graphic card features are emulated AFAIK 04:18 Zeno` OldCoder, latest versions of WINE *may* have changed that 04:18 Zeno` But it runs fine without them anyway 04:18 OldCoder I'll try them, thank you. And, yes. 04:18 OldCoder On my system, speed differences seem minimal. 04:19 Zeno` gregorycu, why wouldn't it be "performant"? It's not emulated code -- it's a native API 04:19 gregorycu Isn't wine an emulator? 04:19 gregorycu (troll) 04:19 * Zeno` grumbles and sits with OldCoder 04:20 Zeno` us cranky old people can sit in this corner 04:21 OldCoder gregorycu, you are aware of what the acronym WINE stands for :-) And this seems more appropriate for the other channels 04:22 gregorycu *shrug* 04:22 * OldCoder likes recursive acronyms 04:22 gregorycu And yes I do 04:22 OldCoder For the benefit of non-Linux people, WINE stands for WINE Is Not an Emulator 04:25 gregorycu Zeno, when you ran your dodgy code checker 04:25 gregorycu Did anything around PlayerSAO pop up 04:25 Zeno` my what? 04:25 gregorycu Did you run like valgrind or soemthing? 04:26 gregorycu To detect race conditions or something 04:26 Zeno` oh race conditions... umm I wasn't checking around PlayerSAO so I'm not sure. I think I would have noticed though. Looking 04:27 Zeno` nope, nothing related to PlayerSAO 04:27 gregorycu Curses 04:27 Zeno` I'm not sure I'd call valgrind "dodgy" lol 04:27 Zeno` gregorycu, but that's the reason (sort of) I asked tenplus1 if he had logs 04:28 gregorycu Of course, you asked for as much info as you can 04:28 * Zeno` runs helgrind again just in case something has changed recently 04:29 gregorycu you ask * 04:40 Zeno` it's a bit hard to make make myself die when I can't move lol 05:09 jin_xi it's not exactly mt dev but i hope related enough to be asking this here: anyone can tell me where to start looking for a good way for particle systems to share and update a cache of meshes? i dont even know what to google for 05:10 gregorycu That sentence didn't really make sense to me 05:11 jin_xi sorry. not native english speaking person here. 05:12 jin_xi need to keep meshes of mapblocks for doing collisions, its expensive to create them per step per particle system 05:12 jin_xi would be nice to cache and only remesh on map change event but share updated mesh with all particle systems operating on that block 05:19 gregorycu All good 05:20 gregorycu If it's not exactly MT related, I'm not sure what advise we can give you 05:21 thePalindrome jin_xi: Is this to have the particles collide with a world? 05:21 gregorycu If you need to keep a cache of meshes for collision detection with particles... well... keep a copy of the meshes for collision detection 05:31 jin_xi thePalindrome: thats what i am trying to do 05:38 thePalindrome In that case just ensure that they have pointers (not copies, pointers only) to the world meshes, and that should be more than enough 05:39 thePalindrome If you need to do some calculations, have the particle_manager (or whatever is keeping track of the particles) do the calculations, then make sure that the particles can access the manager 05:43 gregorycu Having pointers is a great way to ruin performance 05:44 gregorycu Chances are that the meshes will be updated infrequently, and you'll constantly have to use them 05:44 thePalindrome er...? 05:44 thePalindrome I fail to see how pointers ruin performance 05:44 thePalindrome Maybe I"m missing something 05:44 gregorycu Because when you go to use the only copy of the meshes, you have to lock 05:44 thePalindrome Why? 05:45 thePalindrome It doesn't sound like he's writing the meshes 05:45 thePalindrome But I see what you mean now 05:45 thePalindrome that's only an issue if that step is multi-threaded 05:45 gregorycu it should be multithreaded 05:45 gregorycu If it's single threader, sure, use a pointer 05:45 gregorycu threaded 05:46 thePalindrome If you're just reading the pointer, you don't need a mutex 05:46 gregorycu wrong 05:46 thePalindrome heck, if you're multithreaded, you need a mutex anyway! 05:46 thePalindrome What if the mesh changes as you read it and you walk off the data structure? 05:46 gregorycu The particle you mean? 05:46 gregorycu This is a common misconception 05:47 jin_xi so the basic problem is that currently every single particle is a full scene node and uses same collision detection as a player. so i am trying to use meshes and irrlicht triangle selectors to do batch collision detection against block meshes 05:47 thePalindrome I mean while you copy the mesh 05:47 jin_xi these meshes currently are not mts mapblock meshes but simpler ones ignoring non walkable nodes and such 05:48 thePalindrome still, doubling the memory usage just to hold a seperate read-only copy of the meshes 05:48 thePalindrome It really comes down to what's multithreaded and what is bound to a single thread 05:48 gregorycu If you have shared data, and one of the users can change the data, you need some form of locking around every single use of the data 05:48 thePalindrome That's true even while you copy it 05:49 thePalindrome the only way to do an atomic copy is to use mutexes 05:49 gregorycu Unless the data isn't shared at that point 05:49 thePalindrome In which case just don't copy it 05:49 gregorycu The meshes are actually generated in a thread, and posted to the main thread 05:49 gregorycu They could be copied in that other thread 05:49 gregorycu Before the meshes are shared 05:49 thePalindrome oh, that makes sense then 05:49 thePalindrome I see what you're saying 05:50 thePalindrome I thought the meshes didn't need to be computed 05:50 thePalindrome Like you would deep copy the world :P 05:50 gregorycu We get a block data, we need to resolve that into something the GPU can draw quickly 05:51 thePalindrome For collision? 05:51 gregorycu No 05:51 gregorycu For drawing 05:51 thePalindrome I thought the question was just about collision? 05:51 gregorycu Yes, but I'm explaining what already happens 05:52 thePalindrome oh, continue 05:52 gregorycu We get block data from the server, we generate a mesh off it, and then feed it to the GUI thread 05:53 gregorycu If we were to make collision detection use the same copy of the GUI thread, that would mean that we would need locking in the GUI thread to make sure the mesh doesn't get updated while the collision detection system is using it 05:54 gregorycu It is often better to double memory usage and have two different data sets 05:54 gregorycu Unless it's unreasonable for those two data sets to be slightly different 05:54 gregorycu bb in 10 min 05:54 thePalindrome But since the collision and gpu are only reading the meshes, why not use an immutable pointer? 06:01 gregorycu Back 06:01 gregorycu You mean a pointer to immutable data? 06:02 gregorycu That's fine, provided that both threads have their own copy of the pointer 06:03 thePalindrome ? 06:04 thePalindrome oh 06:04 thePalindrome I meant more like an immutable pointer to immutable data 06:05 thePalindrome Hullo monsiour 06:05 gregorycu How do you update the data? 06:06 jin_xi currently i just regen every step in every particle system using new MapMakeData and generate my own mesh off it. still beats mt particles for speed 06:06 thePalindrome I'd presume that you lock the threads while you update the data, since in most cases I can think of you wouldn't do both at the same time 06:08 est31 I meant more like an immutable pointer to immutable data 06:08 est31 thats what const was invented for I think 06:09 thePalindrome aye 06:09 thePalindrome My inexperience shows, eh? :P 06:09 gregorycu I don't think we are talking about the same thing, and I have bugs to solve 06:09 thePalindrome Alrighty, sorry for keeping you :P 06:09 gregorycu jin_xi: So, is this minetest or what? 06:09 jin_xi yes 06:09 thePalindrome oh est31 would I talk to you or sfan about minetestmapper? 06:10 gregorycu thePalindrome: I kept myself, all good bro 06:10 thePalindrome gregorycu: just out of curiosity, what do you usually work on? 06:10 gregorycu You mean real life? 06:10 thePalindrome in minetest :P 06:10 gregorycu Performance related fixes 06:11 thePalindrome I see why you've given so much thought about this then :P 06:11 thePalindrome I guess I should thank you for the fact that minetest runs on a potato :P 06:11 est31 thePalindrome, I think it reached sfan5 I'd suggest waiting a day or so 06:12 est31 if not you can ping him on github inside the issue 06:12 thePalindrome Oh no, I didn't mean that issue 06:12 thePalindrome I can wait on that for a while 06:12 gregorycu It can be a lot better 06:12 thePalindrome Just having some weird issues with the install step for minetestmapper 06:12 gregorycu jin_xi: MeshUpdateThread::doUpdate 06:12 thePalindrome gregorycu: fine, it'll run on two potatoes and a carrot :P 06:13 gregorycu Take a look at this guy 06:13 jin_xi thanks, will do 06:13 gregorycu It generates MapBlockMesh and then adds it to a queue to be used by the GUI thread 06:13 thePalindrome Ah, now I see where I got confuzzled 06:14 jin_xi i wonder i never saw that. i looked at how inventory gets its meshes 06:14 gregorycu If you want, you could add some sort of hook in there to know when to update the mesh data for collision detection 06:14 jin_xi turns out for some cases (glass and such) it loops through 48^3 blocks to render a node 06:15 gregorycu Of you can clone the data there, and push to a collision detection thread 06:19 jin_xi that sounds great, i wonder if such a thing could also be used for mobs and entities to avoid issues with static collision boxes 06:21 gregorycu Is this for the server or client? 06:21 gregorycu I assumed this was for the client 06:23 jin_xi oh yes client 06:26 gregorycu Cool, I would guess the mobs would be handled server side 06:26 gregorycu Well, the currently are 06:26 gregorycu they 06:28 jin_xi yes, i was just rambling without thinking much, atm just looking at particles and client, sorry 06:30 gregorycu All good 07:03 gregorycu nerzhul 07:04 gregorycu hmm... 07:04 gregorycu I think I found the bug 07:05 gregorycu Zeno`: I'll get you to confirm, it's the one you asked for additional info on 07:05 Krock congrats, you may keep, roast and eat it, so it doesn't multiply 07:17 Zeno` what an I confirming? 07:17 Zeno` am* 07:19 gregorycu Server::SendPlayerHPOrDie 07:20 Zeno` oh god... not that function again 07:20 gregorycu It calls both SendPlayerHP and DiePlayer 07:20 gregorycu or rather, one of the two 07:20 Zeno` just after (or maybe just before) that function was looked at a lot 07:20 gregorycu Look how they both handle the result of getPlayerSAO(peer_id) when it returns null 07:20 Zeno` well not looked at but added for something else 07:21 gregorycu So, the "fix" is pretty clear 07:21 gregorycu I just wonder if https://github.com/minetest/minetest/commit/ce8a9ed94b7a3a4cbce04bee27bf8d4167ac6161 was actually wrong though 07:22 Zeno` well yeah that was my first thought as well 07:23 gregorycu If playersao can become null due to another thread, we probably have a race condition here 07:23 Zeno` I think possibly it's actually not correct for the assert to be there (that was my gut feeling anyway), but felt that just removing it might (just *might*) be covering a bug up 07:24 Zeno` tenplus1 must be using debug builds on a server 07:24 gregorycu Yeah, he must be 07:24 gregorycu Kudus to him 07:24 * Zeno` thinks 07:24 Zeno` I hate just removing it... I actually thought about adding a log warning instead 07:24 gregorycu I'll try and cause a crash locally 07:25 gregorycu try to * 07:25 Zeno` which is another reason I asked for how often it happens :) 07:25 Zeno` the problem with the warning is that if it *is* a bug nobody is ever going to report it if it doesn't crash 07:25 gregorycu What will the warning say? 07:25 Zeno` Warning peer_id %sdjfksj doesn't exist or something 07:26 Zeno` but I was uneasy about just doing that for the aforementioned reasons 07:26 gregorycu Apparently we know this can be true 07:26 Krock The rules from 150207 are still active, right? https://forum.minetest.net/viewtopic.php?f=3&t=11172 07:26 Krock If so, where's the feature freeze branch? 07:26 gregorycu We get a bug report about it, we look at the code and go: "Oh yeah, we expect that sometimes" 07:27 Zeno` so if it's expected sometimes then using assert() is not correct 07:27 gregorycu Yeah 07:27 Zeno` there should be a condition as you said 07:28 gregorycu My contention is that the assert (and even the early return) may be hiding a race condition 07:28 Zeno` yes, that was what I was worried about also 07:28 Zeno` and by doing what 0a41326 does is just hiding a bug rather than fixing it 07:28 gregorycu I'll try to force that race conditionally locally 07:28 gregorycu Yeah 07:28 Zeno` ok, thanks 07:28 Zeno` I did run helgrind and drd (data race detector) and killed myself btw 07:29 Zeno` nothing showed up related to PlayerSAO 07:42 Zeno` I'm inclined to trust the DRD and after looking at the code there probably isn't a race condition but testing is good 07:43 Zeno` if there was a data race it would have shown up 07:44 Zeno` like the million others that did 07:45 gregorycu Yeah, it's all done in the one thread 07:45 gregorycu So no data-race 07:45 gregorycu I'll do change the code in the same way 07:45 Zeno` ok 07:46 Zeno` I'll merge now as trivial since both of us have looked at it and come to the same conclusion 07:46 Zeno` I was about to do it anyway when you added the comment that you'd look at it :) 07:46 gregorycu Well, if you want to do the code change, you can 07:46 gregorycu If that's easier than me creating a PR 07:46 Zeno` Well you can.. I have to cook Mum dinner anyway 07:46 gregorycu ok 07:49 Zeno` thanks 07:49 Zeno` bbs 07:51 Zeno` if you're interested in the other contentions and stuff I can email you the analysis output but it's ugly 07:51 Zeno` most are log related anyway (and we know about that) 07:52 Zeno` might be a bit annoying to read the text by hand but all the data is there 07:52 Zeno` Actually I'll do it again after I cook dinner but disable sound 07:52 Zeno` that'll make the file about 1/4 the size 07:56 gregorycu Yeah, I've hit the log stuff a few times 07:57 gregorycu #4102 07:57 ShadowBot https://github.com/minetest/minetest/issues/4102 -- Fixes Dieplayer playersao failed error by gregorycu 08:29 gregorycu I'll be back later, family dinner time 08:50 Zeno` will merge #4102 shortly (see above) 08:50 ShadowBot https://github.com/minetest/minetest/issues/4102 -- Fixes Dieplayer playersao failed error by gregorycu 08:51 Zeno` est31, if you can quickly read the logs and see 4102 that would be appreciated 08:53 est31 yes would very appreciate it 08:53 * Zeno` looks around blankly 08:53 est31 but fix the hrammar 08:53 est31 damn 08:54 Zeno` the hammer? 08:54 Zeno` oh 08:54 Zeno` yeh sure 08:54 est31 I wanted to say fix teh grammar 08:54 est31 not the 08:54 est31 stupid autocorrect 08:54 est31 ruins my jokes 08:54 Zeno` lol 08:54 Zeno` https://patch-diff.githubusercontent.com/raw/minetest/minetest/pull/4102.patch 08:54 Zeno` what would you like changed? :P 08:55 Zeno` cases 08:55 Zeno` I think he just copy pasted from that other part of code 08:55 Zeno` if you're referring to the grammar in the comment 08:56 est31 yes 08:56 Zeno` I guess I can change it :P 08:56 est31 nrz is a nice guy but his grammer is not parfect 08:56 Zeno` but... what if you don't like my grammar! 08:56 Zeno` then we will end up in an almighty fight 08:56 Zeno` this is a dangerous game 08:57 est31 minetest is dangerous? 08:57 Zeno` eys! 08:57 Zeno` merging now ;) 08:57 Zeno` I'll try and fix the grandma 09:00 Zeno` I wish I could understand it 09:03 Zeno` pushed 09:03 Zeno` I dunno if you'll like my grammar 09:04 est31 its okay 09:05 Zeno` phew 09:05 est31 I prefer multi line git commit msgs though but in this case one line is fine 09:05 * Zeno` puts the whiskey away 09:05 est31 hehe 09:05 Zeno` oh man 09:05 est31 just sayin 09:05 Zeno` :) 09:06 Zeno` I should have changed that .. didn't notice tbh 09:41 gregorycu I read that comment, didn't even realise the bad grammar 09:42 gregorycu Only remaining milestone bugs are android things 09:44 gregorycu Which menas my work here is done 11:27 gregorycu Time to try to see if I can do something about #3602 11:28 ShadowBot https://github.com/minetest/minetest/issues/3602 -- Serious stuttering (large instantenious fps drop) (rendering) 11:31 gregorycu lol 11:31 gregorycu Speak of the devil 11:35 Zeno` you should not speak of the devil. The devil can hear and is drawn towards the speech 11:36 Zeno` speaking of the devil is equivalent to a summoning. Very bad. 11:49 Fixer i do have android tablet, but I need compiled builds to actually install them for some limited testing, i'm not playing on it much 11:50 Fixer just reading 12:25 gregorycu About programming? 12:28 Fixer sometimes yes 12:28 Fixer but not fast 12:28 Fixer don't hold your breath 12:40 gregorycu I'm not sure it's worth working on jitter, considering we will have farmap changes soon 12:41 gregorycu The sooner we get this release out the door 12:41 gregorycu The sooner we can merge that in 12:41 gregorycu The sooner I can look at these stutter issues 12:41 Fixer ugh 12:42 Fixer we have 3770 :) 12:43 gregorycu #3770 12:43 ShadowBot https://github.com/minetest/minetest/issues/3770 -- Fix superflous shader setting updates by ShadowNinja 12:43 gregorycu Ahh 12:43 gregorycu Your fav 12:43 Fixer it boosts fps a lot, not only for me 12:43 Fixer needs bugfixing 12:44 Fixer with it i can boost my view range to nice levels (close to 200) 13:23 gregorycu Alright, time to take a look 13:26 gregorycu This stuff is a bit of a nightmare by the way 15:42 gregorycu What idiot used the factory pattern here 15:50 PilzAdam gregorycu, sorry, I don't see any idiots around here 15:52 Calinou maybe Terasology dev infiltrated us 15:55 gregorycu What is Terasology? 15:56 Calinou #terasology, http://terasology.org 15:56 gregorycu yeah, main page gives no indication what it is 15:57 Calinou Minecraft-style game made in (enterprisey) Java 15:57 gregorycu lol, ok 15:57 Calinou we even wrote on our dev wiki that their code style is horrible 15:57 gregorycu Looks pretty though 15:57 Calinou it's also slow as hell 15:57 rubenwardy How I hate the phrase "enterprisey" 15:57 Fixer Calinou, it is just you too fast %) 15:58 gregorycu I wonder if they have shader issues 15:58 Calinou it looks mostly fine, it's just very slow 15:58 Calinou the defaults are rather poorly tweaked, so it's slow even on fast PCs 15:58 gregorycu Sounds like they need me to help them 15:58 Calinou I wouldn't be surprised if you couldn't play the game maxed out on the upcoming GTX 1080 15:59 Fixer Calinou, ghhh 15:59 Krock a yay for minetest :) 15:59 Calinou Minetest is not exemplary either :| but it's still much faster 15:59 gregorycu Hey, we are making minetest fast 15:59 Calinou if I had more time, I'd remake it in Godot ;) 15:59 Fixer on the other hand Minecraft performance is damn impressive (with optifine) 15:59 gregorycu I'm working on that right now 15:59 Calinou Fixer: yes, they did a lot of optimization 16:00 Fixer it is like 80-100 fps with view range similar to our v_range 240, and no fucking stutters, flawless 16:00 gregorycu We plan to fix those things 16:00 Fixer 64 bit java though 16:00 gregorycu I plan to fix those things 16:00 Fixer gregorycu, +1 16:00 gregorycu Just a matter of time 16:00 gregorycu And this shader bullshit 16:01 gregorycu I like your comment 16:01 gregorycu "Easy to reproduce after a few hours" 16:01 Fixer well, it depends on map 16:02 Fixer if you have fresh singleplayer one without builds, bug is very rare, but if you play on servers with lots of builds, or on that saved world, you can catch it faster, i catched it multiple times, even after restarts 16:02 gregorycu The problem is, a map with a lot going on, it's very difficult to debug a specific issue 16:02 gregorycu Maybe it's just the fact there is a lot going on that causes the issue 16:03 Fixer gregorycu, https://cloud.githubusercontent.com/assets/16494741/13375919/b4ea1dc8-ddb5-11e5-909d-c2e3d081fb35.png 16:03 Fixer this example... 16:04 Fixer it is somehow can be connected to doors, slabs, panels (nodeboxes?) 16:04 Fixer protectors are nodeboxes iirc 16:04 Fixer doors are too 16:05 gregorycu What is the specific issue 16:05 gregorycu I don't play minetest heaps 16:05 gregorycu It's the dark nodes above right? 16:05 Fixer no wait 16:05 Fixer so it is day on this screenshot, nodeboxes are ok, but everything else is dark 16:06 Fixer i can put a torch here, it will not fix anything 16:06 gregorycu It will not fix anything? 16:06 gregorycu Torch won't light it up? 16:06 Fixer yeah, it looks the same, torch will bright it up around itself, but nothing else 16:07 gregorycu Hang on a sec 16:07 gregorycu Torch will bright up around itself 16:07 Fixer yes 16:07 gregorycu And fix nodes in its radius, right? 16:07 gregorycu But only within the radius 16:07 Fixer i think it brightens nodes within usual radius, not fixes, if i remove it, it will go back to dark 16:07 Fixer within radius 16:08 gregorycu Would it be fair to say that it's like the world thinks it's nighttime but it isn't? 16:08 gregorycu Are the nodes being lit like it is nighttime? 16:08 Fixer gregorycu, could be, example: https://cloud.githubusercontent.com/assets/16494741/13375926/c648667e-ddb5-11e5-8eec-e563ba8270cd.png 16:09 gregorycu If it were nighttime in the second picture, would that look correct? 16:09 Fixer gregorycu, could be, it looks a bit dark :/ 16:09 Fixer gregorycu, another example, this time it is reverse: https://cloud.githubusercontent.com/assets/16494741/13375923/c642a4c8-ddb5-11e5-97f5-cdcaa761b6fd.png 16:10 Fixer gregorycu, some nodeboxes are dark, but it is day 16:10 Fixer look in shadow 16:10 Fixer unusual darkness 16:10 Fixer even for night that is too much 16:10 Fixer maybe second screenshot is very dark too 16:10 Fixer something wrong with lighting 16:11 asl97 #4091 doesn't seem to improve performance by much, not even on android 16:11 ShadowBot https://github.com/minetest/minetest/issues/4091 -- mesh tiling bug and possible fix 16:12 Fixer gregorycu, just look at this: https://cloud.githubusercontent.com/assets/16494741/13375920/b4ee864c-ddb5-11e5-9088-ee278da1130b.jpg 16:12 Fixer gregorycu, right up corner 16:12 Fixer gregorycu, you can see totally black protector in daylight (it is nodebox), and rock with correct lighting below 16:12 Fixer gregorycu, protectors are that things in air with yellow center 16:13 Fixer gregorycu, if you look closer to crosshair, you will notice same protectors, not completely dark now, also, doors are much more darker :/ but not completely 16:14 Fixer gregorycu, it does feel strange, bug can occur at day change, i noticed that 16:14 gregorycu What a mess 16:14 Fixer yeah, but it gives some hints 16:15 Fixer blocks, nodeboxes, day change, lighting, torch won't fix, etc 16:15 gregorycu We know it isn't related to lighting code 16:15 gregorycu As that wasn't touched 16:15 gregorycu So, we wouldn't expect the torch to fix it 16:15 gregorycu Like it does with other lighting bugs 16:16 gregorycu Though it's no obvious to non-dev, this lighting is different to the lighting that is done by torches 16:17 Fixer this one is interesting too: https://cloud.githubusercontent.com/assets/16494741/13375929/c660d5ce-ddb5-11e5-9094-e2cd4d883713.png 16:17 gregorycu It's almost as if the ambient light isn't set correctly for the nodes that are black 16:17 Fixer yellow protector inside building looks kinda ok, but on outside far, it looks megadark 16:18 gregorycu Does "upright_sprite" mean anything to anyone? 16:19 Fixer nope 16:19 jin_xi yes its an old drawtype 16:19 jin_xi was used for players and mobs way back 16:20 gregorycu Fixer: with the stuff that isn't drawn properly 16:20 gregorycu Is it ever regular blocks? 16:20 gregorycu Are regular blocks ever fucked up? 16:20 Fixer gregorycu, yes, regular blocks too 16:20 gregorycu Yeah, it looks like they are 16:20 gregorycu Fuck 16:21 gregorycu Thanks jin 16:21 Fixer gregorycu, prime example: https://cloud.githubusercontent.com/assets/16494741/13375926/c648667e-ddb5-11e5-8eec-e563ba8270cd.png 16:21 Fixer or this https://cloud.githubusercontent.com/assets/16494741/13398253/251176ae-df06-11e5-984b-1053bfa9a1d0.png 16:21 Fixer funny how sign is ok 16:22 Fixer and water 16:22 gregorycu Someone who probably knew what they were doing would have this solved by now 16:22 gregorycu I'm not a graphics guy 16:22 Fixer what is funny about that bug, is how it kicks in 16:23 Fixer you walk around everything looks normal, normal, and after sometime you changed location and notice it looks different now 16:24 jin_xi ever noticed some minimal distance you have to move or is that random? 16:24 Fixer random 16:24 Fixer changing height could help 16:24 Fixer changing height and daytime too 16:24 Fixer restarting game too 16:26 Fixer gregorycu, maybe it is shaders that ruin the lighting? 16:26 gregorycu Oh, yeah, it is 16:26 gregorycu I just don't know why 16:27 Fixer any way to gather more data? 16:27 gregorycu Not really, I'll have to have a play around 16:28 Fixer gregorycu, try my saved world, go to that building near spawn, just walk, leave minetest, go back, walk around to other location, go back to house, fly slowly to platform above, go back, walkaround, try multiple restarts near building, shaders must be on obviously 16:29 Fixer it kicks in after some time 16:29 gregorycu Yeah, but once I reproduce, well, I need a plan 16:29 gregorycu There can be hundreds of shader calls, not sure what a bad one would even look like 16:29 gregorycu But I'm not giving up 16:30 Fixer isn't minetest can show shader warnings? 16:30 Fixer how can I turn them on? 16:30 gregorycu What are shader warnings? 16:30 gregorycu I mean, is this a thing we used to have? 16:31 hmmmm i think it's a setting in irrlicht 16:32 gregorycu Do you have any ideas with regards to this hmmmm? 16:32 celeron55 try setting the verbose minetest log level 16:32 gregorycu Graphics isn't my strong suit :( 16:32 hmmmm no 16:33 hmmmm aghh you've piqued my interest now 16:33 hmmmm let me read up and see the issue 16:34 hmmmm [11:58 AM] I wouldn't be surprised if you couldn't play the game maxed out on the upcoming GTX 1080 16:34 hmmmm the GTX 1080 is a rebadge of the 980 16:35 celeron55 their marketing says it's twice as fast; i guess someone has to test it 8) 16:35 celeron55 (but obviously it's useless for MT) 16:36 gregorycu Because, as it was put before, minetest runs on a potato? 16:45 Zeno` like in the Terry Pratchett thing? 16:46 Zeno` "In The Long Earth, potatoes are elevated to a new level of significance when the humble tuber becomes the power source of a world-changing machine." 16:46 Calinou hmmmm: say that in ##hardware :p 16:46 Zeno` wait a sec... I could play minetest maxed out on a gtx-680 :/ 16:46 hmmmm so it's actually a big improvement? 16:47 Zeno` with something like 7000 fps 16:47 hmmmm shrug, *i* heard it was a rebrand 16:47 Calinou hmmmm: yes 16:47 Zeno` I dunno... I have a 970 now 16:47 Calinou 25% faster than the 980 Ti 16:47 hmmmm nice 16:47 Calinou up to 2x faster in VR applications, thanks to a special feature that allows rendering several viewports with the cost of only one 16:47 Calinou Zeno`: maybe next year I'll get 4K monitors and a 1080 :) 16:47 Zeno` Calinou, how much does it cost? 16:48 Calinou $600, $700 for the Founders Edition 16:48 Zeno` eek 16:48 Zeno` USD? 16:48 Calinou yes 16:48 Calinou the performance/price is quite better than the 980 Ti 16:48 Zeno` oh, well I think I'll wait a while :) 16:48 Calinou and maybe better than the 980 as well 16:48 hmmmm :d 16:48 hmmmm [ryan@titan] ~% grep nvidia /var/run/dmesg.boot 16:48 hmmmm nvidia0: on vgapci0 16:48 Calinou grep: /var/run/dmesg.boot: No such file or directory 16:49 gregorycu Zeno`: http://knowyourmeme.com/memes/recorded-with-a-potato 16:50 hmmmm if loonix is so good why doesn't it even have a dmesg.boot logfile!! 16:50 Zeno` doesn't need it! 16:50 hmmmm dmesg | grep nvidia should do okay as well assuming you haven't had 500 segfaults in the past boot 16:51 Zeno` $ lspci | grep NVIDIA | grep VGA 16:51 Zeno` 01:00.0 VGA compatible controller: NVIDIA Corporation GM204 [GeForce GTX 970] (rev a1) 16:52 Calinou [ 1.866916] nvidia: module license 'NVIDIA' taints kernel. 16:52 Calinou :-) 16:52 Zeno` lol yep :) 16:52 Zeno` tainted love 16:53 Calinou maybe next year I'll be able to use Nouveau 16:53 Zeno` *shudder* 16:53 Calinou ie. reclocking support, decent performance 16:53 Calinou someone has been able to drive 2 4K monitors using Nouveau 16:53 Zeno` that must be nice for them :D 16:54 Calinou http://www.phoronix.com/scan.php?page=article&item=nouveau-boost-perf 16:54 Zeno` I have no problem using the nvidia driver 16:54 Calinou performance is becoming acceptable 16:54 Zeno` I am not RMS 16:54 Calinou Zeno`: with nvidia driver, you don't have native-resolution TTYs, no flicker-free boot 16:54 Calinou no KMS 16:54 Zeno` no flicker free boot? 16:54 Calinou hybrid GPU is relatively cumbersome as well (it's a bit easier using Nouveau) 16:54 Calinou Zeno`: there's forced mode setting when you boot 16:54 Calinou at least one of them 16:54 Calinou with Nouveau, you can get rid of it, and speed up booting by ~2 seconds 16:54 Zeno` I dunno how I can stand not having flicker free boot hehe 16:55 Zeno` Calinou, I boot in text mode anyway 16:55 Zeno` so there is no flicker 16:55 Zeno` there might be when starting X... not sure 16:56 Calinou also nvidia-settings is quite bad 16:56 Calinou it looks like it was made in 2006 :P 16:56 Zeno` yeah, fuck you nvidia! 16:56 Calinou but sadly, for open source drivers, there's no GUI configuration tools yet 16:56 Zeno` wait, I like the driver 16:56 Calinou (apart of the unmaintained driconf) 16:57 Zeno` wait a sec... no native TTYs? 16:57 Calinou Zeno`: also NVIDIA's OpenGL implementation has lots of proprietary things, and isn't really standards-compliant 16:57 Calinou free drivers do it much better 16:57 Zeno` My ttys are native 16:57 Calinou better as in, more restrictive 16:57 Calinou Zeno`: 1920x1080 TTY? 16:57 Calinou not possible with nvidia 16:57 Zeno` they use only text 16:57 Zeno` oh, I have VESA tty 16:57 Zeno` i thought that's what you meant 16:57 Calinou no 16:58 Calinou you're limited to 1024x768 or something like that 16:58 Zeno` why would I want a 2560x1440 tty? 16:59 Zeno` It would be like having a 2560x1440 minetest console 16:59 Zeno` err 16:59 * Zeno` runs minetest to check something 17:00 Zeno` yeah, it'd be like minetest 17:00 Zeno` ;) 17:00 gregorycu Fixer: I may know what the bug is 17:04 gregorycu The CachedShaderSetting is not shared, there is one per material 17:04 gregorycu Or rather, there are many per material 17:04 gregorycu The class will not send shader information if it thinks it hasn't changed 17:04 gregorycu shader parameters 17:05 gregorycu I have to go to bed now, we can try this tomorrow 17:05 Zeno` Did your Mum enjoy her dinner, gregorycu? 17:05 gregorycu She did, Chinese 17:05 gregorycu Classic 17:05 gregorycu Night 17:05 Zeno` nice :) 17:05 Zeno` night 17:23 Zeno` does anyone else think that mese should be purple? 17:23 Zeno` yellowish is too natural 17:24 lisac Zeno`: I do. 17:24 lisac But dark purple. 17:24 Zeno` yeah really dark purple 17:28 Fixer Chinese classic food? 17:29 Zeno` I guess it is for some 17:29 Fixer no! yellow is nice 17:30 Fixer mese was nice glowy block in dark caves, don't rape it please :) 17:30 Zeno` maybe lime green would be better? 17:31 Void7 how about mese is much rarer and more powerful, but there's a purple mese-like substance? 17:31 Void7 more powerful as in the tools last ~10x longer and dig ~1.5x faster 17:36 Calinou purple looks like obsidian 17:36 Calinou lime green looks like slime/acid 17:36 Calinou yellow is fine 17:36 Calinou it's a bright yellow anyway 17:38 Zeno` a bit of green in it might be nice though 17:38 Zeno` alien stuff is normally greenish 17:41 Calinou is it really alien? 17:41 Calinou aliens need to go away 17:41 Calinou Doom did not have any aliens, same for Quake 17:41 Calinou and don't even tell me Minecraft has aliens 17:42 rubenwardy A suggestion from c55 17:42 rubenwardy And we're not allowed to add something if MC doesn't use it? 18:16 Zeno` 2 days, paramat 18:17 paramat hopefully 18:18 paramat i didn't test #4100 , no android, what do you think? 18:18 ShadowBot https://github.com/minetest/minetest/issues/4100 -- Defaultsettings/Android: Increase 'max block generate distance' to 3 by paramat 18:18 PilzAdam rubenwardy, I asked you this in the github issue, but you haven't responded (IIRC): what is the benefit of saying that mese is of alien origin? 18:21 Zeno` paramat, hopefully? Not hopefully, I've already ordered the catering! 18:22 paramat Zeno` the current setting is 7 so 3 will still help performance 18:22 asl97 does #4091 have any chance of making into release? 18:22 ShadowBot https://github.com/minetest/minetest/issues/4091 -- mesh tiling bug and possible fix 18:22 Zeno` paramat, it's fine. Merge the other kind of related PR with it 18:22 paramat ok 18:23 Zeno` asl97, no. It's a good PR but there is feature freeze after all 18:23 Zeno` asl97, but I imagine it will be one of the first after release 18:23 asl97 it is a bug fix 18:23 Zeno` it's not a bug 18:24 Zeno` how is it a bug? 18:24 PilzAdam Zeno`, it's a feature of continuous tiles to not be continuous? 18:24 paramat too risky to merge that now 18:24 PilzAdam paramat, I agree 18:24 Zeno` PilzAdam, it's not a bug 18:24 PilzAdam should be merged after the release 18:25 asl97 that's fine I suppose. It didn't affect performance as much as I expected 18:25 Zeno` I'm quite impressed by it 18:26 PilzAdam Zeno`, was it intentionally coded to only support stripes of up to 2 nodes? doesn't seem like it, but I don't know the code very well 18:26 paramat any thoughts on milestone #4086 ? est31 seems to have some issues with it 18:26 ShadowBot https://github.com/minetest/minetest/issues/4086 -- Fix locked hardware buttons on Android by MoNTE48 18:26 Zeno` PilzAdam, in that sense it's a bug, but the current code works 18:27 Zeno` PilzAdam, the PR is fine as far as I can see an I *would* merge it if not for the freeze 18:27 PilzAdam depends on the definition of "bug" then 18:27 Zeno` PilzAdam, yes 18:28 PilzAdam one possible definition is code that does not do what the author intended it to do 18:28 Zeno` it's probably less risky to merge than some others that were merged just before the proper freeze 18:28 PilzAdam as I said, I don't want this to be merged before release 18:30 Zeno` paramat, I'm not sure how 4086 "works" 18:30 paramat nor me 18:30 Zeno` it does, but I just don't understand 18:30 Zeno` I'm wary of it 18:30 paramat no prob i'll ask est to look at it 18:30 Zeno` ok 18:31 PilzAdam *o.k. 18:31 Zeno` aargh 18:31 Zeno` :P PilzAdam 18:32 Zeno` umm, if what est31 suggests is true it's an easy fix 18:33 Zeno` does that need to be in the milestone? 18:34 Zeno` since it's been around for 3 years anyway? 18:34 paramat i think yes because of volume control not working 18:34 Zeno` it's never worked 18:35 paramat heh 18:35 Zeno` seriously 18:36 paramat i mean, it's never worked because of this bug 18:37 paramat seems good to fix it 18:38 Zeno` *if* it fixes it and doesn't introduce some other bug 18:38 * Zeno` leaves it to est31 18:38 paramat indeed 18:40 Zeno` to merge that is more of a risk, IMO, than to merge 4091 18:41 Zeno` at least 4091 is understood 18:42 Zeno` whatever... just don't let it block the release :P 18:42 paramat agreed 18:49 paramat i've requested #4084 is renamed to 'resize' 18:49 ShadowBot https://github.com/minetest/minetest/issues/4084 -- Add [scale texture modifier by SmallJoker 18:50 paramat then it seems ready 18:51 Zeno` #4096 should be closed 18:51 ShadowBot https://github.com/minetest/minetest/issues/4096 -- Compilation warnings in Lua 18:51 paramat yes ok 18:51 Zeno` It's not something we should fix (although we could) 18:51 paramat done 18:51 Zeno` cool 18:59 Zeno` that *4 is crap 18:59 paramat heh yes 18:59 Zeno` even if it's better it's crap 19:00 Zeno` there is no way magic values like that should be in the source code even if it DID have a comment 19:01 paramat i won't allow that in this PR 19:02 paramat does android allow +-10 node increments of view range or is it fixed? 19:03 paramat (apart from 'view all loaded') 19:04 asl97 paramat: iirc, there is no button for that if that is what you mean 19:05 paramat it is, ok. that sucks 19:05 asl97 I am pretty sure the config file exist on android though 19:05 asl97 there are plenty of text editor on android 19:05 paramat ah 19:05 asl97 and there is the in game setting menu 19:06 Zeno` the config file does exist 19:06 Zeno` it's just not possible for the "normal" user to easily change 19:08 paramat so, all remaining problematic issues are android =/ 19:08 asl97 i doubt a "normal" user would even have a text editor on their android device 19:08 Zeno` :) 19:09 Fixer paramat, let me try running minetest from f-droid 19:11 Fixer paramat, ah, stable 0.4.13 has autorange finder, don't use newer one, sorry 19:11 paramat no don't bother, it's not important 19:12 * Zeno` watches lost in space 19:12 paramat we're fairly sure it's fixed in-game 19:12 Zeno` "Trip Through The Robot" episode 19:13 Zeno` lol 19:13 Zeno` Dr Smith blew the spaceship up 19:14 Zeno` it's just lucky they don't need it 'cause they're lost anyway 19:18 paramat heh mese wars 19:19 Fixer PilzAdam, it is either alien or human made or magic made (?) 19:19 sofar that's what you get when you include a trademark in your game of a company that's hostile to open source ;) 19:19 PilzAdam Fixer, are you talking about mese? 19:19 Fixer PilzAdam, yes 19:20 Fixer PilzAdam, alien as of nonnature origin? 19:20 Zeno` they won't charge the robot and he is gonna die! 19:20 PilzAdam it's celeron made 19:20 PilzAdam why do people insist on giving it a background? 19:21 Fixer PilzAdam, don't know, leave that block as it is %) just saying 19:22 Fixer PilzAdam, some people want for things to have some background I guess 19:22 Zeno` I think the robot is deaded! 19:22 Zeno` his core mese power supply ran out 19:23 Fixer and mt crashed 19:23 paramat c55 only vaguely suggested 'alien', it's more 'unknown/mysterious' 19:23 paramat that avoids a background 19:27 paramat essentially i'm suggesting a weird yellow block of unknown origin/formation is found with unknown symbols that look a little like ME SE, so it is called MESE 19:28 paramat it's good it is flexible, some may consider it ancient, some alien 19:28 PilzAdam avoiding a background is easily done: 1) do nothing 2) ... 3) profit! 19:28 paramat doing a websearch for mese mostly brings up place names 19:30 PilzAdam lol, "Cognate with Old High German mias, meas (German Mus, Gemüse)" 19:30 PilzAdam I vote for Mese as a vegetable 19:31 PilzAdam (https://en.wiktionary.org/wiki/mese) 19:31 sofar I vote for getting rid of the trademark reference in it 19:31 paramat yeah, 'dish, food, meal' .. 19:33 paramat well seeing as search results return many things except messenger it's ok 19:33 paramat i found a website mese.fi ! 19:34 Fixer erm, why you guys GOOGLE, minetest should have built in docs and manual, wtf 19:36 PilzAdam Fixer, what? 19:36 paramat page 3 of googling: mese in minetest 19:37 Fixer minetest has only hotkeys info, no data about crafting or anything else in mtg or wtwr 19:38 paramat mese.tv cute https://lh4.ggpht.com/5lR507vabiqP4Hdex8QAUx_gr79YzTan5lk1QejUBqxhXh5vNGG5zBY14X_PSeNVwQ=w300 19:38 PilzAdam Fixer, how is that relevant now? 19:39 Fixer you can write down info on blocks in gameplay.txt with info about mese I guess 19:40 PilzAdam I still don't get what you want from us; why shouldn't we google? 19:41 Fixer whatever... games used to contain PDF manual for many pages, here you just google :} 19:42 PilzAdam but we are talking about what would go into that PDF 19:42 Fixer general info about game, blocks, items, crafting, mods, settings, faq, things that wiki has in general 19:42 PilzAdam we can't write a PDF based on the content of that PDF to be written 19:45 PilzAdam Fixer, do you understand that we can't refer to any "official" documentation, because we are the ones who write it? 19:45 Fixer ugh, you can't imagine countless questions about crafting and other stuff ingame, no tutorials, no docs for players 19:50 paramat fixer yes i agree 20:04 paramat cool looks like we have a fix for the last mtgame milestone 20:30 paramat nore PilzAdam sfan5 this isn't a PR yet but does anyone approve of this fix for a divide-by-zero crash in tnt? https://github.com/HybridDog/minetest_game/commit/d785f05397ef35efd1e1fe985deecf8bd8475a92 20:30 PilzAdam paramat, seems good 20:30 sofar it's a good fix, I like the simplicity and it shows exactly where the problem is :) 20:31 paramat ok. i might do the PR myself later 21:05 paramat now merging #4100 21:05 ShadowBot https://github.com/minetest/minetest/issues/4100 -- Defaultsettings/Android: Increase 'max block generate distance' to 3 by paramat 21:11 paramat merged 21:31 paramat here's the PR #1084 21:31 ShadowBot https://github.com/minetest/minetest/issues/1084 -- Add InvRef::get/set_lists() by ShadowNinja 21:31 paramat game#1084 21:31 ShadowBot https://github.com/minetest/minetest_game/issues/1084 -- Tnt: Avoid divide-by-zero errors in calc_velocity() by paramat 21:51 paramat now merging 21:55 paramat merged. all mtgame milestones cleared