Time Nick Message 00:24 paramat now merging games 543 1021 00:36 paramat merged 02:18 paramat hi hmmmm i've updated #3973 as requested and tested, please could you review sometime tonight? i also updated and tested #3974 and will merge this later 02:18 ShadowBot https://github.com/minetest/minetest/issues/3973 -- Mgvalleys: Don't let cavegen place biome nodes everywhere by paramat 02:18 ShadowBot https://github.com/minetest/minetest/issues/3974 -- Mgv5: Optimise tunnels, add biome material in entrances by paramat 02:52 paramat thanks 02:55 sofar paramat: game#999 updated 02:55 ShadowBot https://github.com/minetest/minetest_game/issues/999 -- Dirt should turn to dry grass in some occasions. by sofar 02:55 paramat ok 02:56 sofar the single ABM came out nicely, I think 02:59 paramat seems good i'll have fun testing this 02:59 sofar I noticed it's hard for dirt to grow back to grass under trees 02:59 sofar (and I kind of like it that way) 03:12 paramat now merging 3973 3974 03:15 sofar hmm donbatman makes a good observation 03:15 sofar I need to tweak game#999 one more way 03:15 ShadowBot https://github.com/minetest/minetest_game/issues/999 -- Dirt should turn to dry grass in some occasions. by sofar 03:22 paramat merges complete 03:46 sofar if (foo == NULL) { 03:46 sofar or 03:46 sofar if (!foo) { 03:46 sofar ? 03:46 Zeno` !foo is fine 03:46 ShadowBot Zeno`: Error: You must be registered to use this command. If you are already registered, you must either identify (using the identify command) or add a hostmask matching your current hostmask (using the "hostmask add" command). 03:46 Zeno` oops 03:46 sofar hahaha 03:47 sofar Zeno`: obviously, ShadowBot takes offense! ha! 03:47 Zeno` I was just testing the bot 03:51 Zeno` I am pleased to report that ShadowBot is functioning as expected 04:39 paramat #3991 04:39 ShadowBot https://github.com/minetest/minetest/issues/3991 -- Falling nodes not updating as before 05:55 paramat now merging #3990 trivial code optimisation 05:55 ShadowBot https://github.com/minetest/minetest/issues/3990 -- Mgv7, mgflat, mgfractal: Tunnel generation code optimisation by paramat 06:01 paramat merged 06:23 Zeno` well that was fun 06:23 Zeno` I finally coerced everything to work without UEFI heh 06:25 Zeno` made the fedora people angry because I said "wtf" at one point and that abbreviation is not acceptable apparently 06:27 * sofar clasp 06:27 sofar intel applauds you 06:27 sofar DISCLAIMER: they pay my salary 06:28 Zeno` :) 10:55 rubenwardy Ok, I'm currently fixing the inventory hotbar bug. Changing to pos + offset * scale * density works. I used a text element to check that the scaling is working currently, but it turns out that the text does not even take hud_scaling into account 10:55 rubenwardy Why are statbar offsets scaled, but text offsets aren't? 10:57 Zeno` Probably an oversight 10:59 rubenwardy it doesn't appear that images are, either 11:01 rubenwardy It appears that statbars and the size of hotbars are the only things that are scaled 12:54 Zeno` umm: http://imgur.com/bxZVONQ 13:01 Zeno` testing #3933 now to see if it fixes the issue. If it does I will merge because it's critical (the game is unplayable) 13:01 ShadowBot https://github.com/minetest/minetest/issues/3933 -- Allow mods to control chat position 13:01 Zeno` err #3993 13:01 ShadowBot https://github.com/minetest/minetest/issues/3993 -- [WIP] Fix inventory hud scaling by rubenwardy 13:02 Zeno` if the PR doesn't work I'll revert the commit that caused the issue 13:03 rubenwardy It's only WIP because I'm not sure about the inconsistency between images, text and statbars 13:03 rubenwardy As in, statbar's offsets are scaled, but images and text's offsets are not scaled 13:03 Zeno` well, it's better than not being able to play the game 13:03 Zeno` did you see my screenshot? 13:03 Zeno` (and your PR seems to fix it) 13:03 rubenwardy the wielditem on? 13:03 Zeno` that's not a wielditem 13:04 rubenwardy oh 13:04 rubenwardy wow 13:04 Zeno` that's a part of the bar 13:05 Zeno` I'll commit in 10 minutes because it's a disaster 13:06 Zeno` I don't think it was your commit rubenwardy 13:06 rubenwardy yeah 13:06 Zeno` it seems more like PA's 13:06 rubenwardy PA fixed a small bug which hid a bigger bug 13:06 rubenwardy both my fault 13:06 Zeno` 'tis cool :) 13:07 Zeno` this is why we have a development version 13:07 rubenwardy Yeah, at least it was obvious 13:09 srifqi How about the bug that makes player with enable damage turned on die on first play (after register)? 13:09 rubenwardy hmmmm ^ 13:09 rubenwardy oh, not here 13:11 Zeno` just testing again now 13:13 rubenwardy If images and texts are not hud_scaled, then they will become out of alignment with hotbars and statbars. I don't see why this was done. Unless I'm misreading 13:14 Zeno` I forget how to embed images on github 13:14 Zeno` oh well 13:16 Zeno` rubenwardy, after I merge I'll close it but if you can keep looking at it I think we'd all appreciate it 13:16 rubenwardy Ok 13:17 rubenwardy It would be cool to have a way to automatically generate a table of screenshots of MT at different densities and screensizes 13:21 srifqi still no news? 13:23 Zeno` merged 13:24 Zeno` I know it wasn't a trivial fix, but it was pretty desperately needed 13:34 Fixer one and final fix? :trollface: 13:35 Zeno` probably not the final fix 13:35 Zeno` but it's better that the game is playable than not (which it was not because the HUD took up the entire screen) 13:35 Fixer i will try it ofc 13:35 Fixer need 30 min 13:36 Zeno` thanks 13:37 Zeno` my first commit since I've been back and it's likely to be controversial lol. But I made the decision and think it was the right one 13:42 Fixer there are no bug free people 13:44 Zeno` sure there are 13:44 Zeno` those who don't do dev 14:03 Fixer testing 14:07 Fixer hud issue fixed for me, looked in SP and MP, resized window few times 14:08 Zeno` two birds with one stone :) 16:04 Zeno` morning hmmmm 16:38 hmmmm hello 16:39 Obani ollé 16:39 Zeno` hola 16:40 hmmmm did something explode since last night? 16:41 rubenwardy just #3989 16:41 ShadowBot https://github.com/minetest/minetest/issues/3989 -- hotbar appearing at top of screen 16:41 rubenwardy and 3993 16:41 rubenwardy and #3993 16:41 ShadowBot https://github.com/minetest/minetest/issues/3993 -- [WIP] Fix inventory hud scaling by rubenwardy 16:42 Zeno` well, it was more 3933 that caused the meltdown 16:42 Zeno` well, not that PR 16:42 Zeno` but it fixed the explosion 16:43 Zeno` so I applied without getting the proper approvals because I thought that something that made the game unplayable was serious enough to warrant it 16:43 Zeno` (after reviewing the code and testing of course) 16:43 rubenwardy #3933 16:43 ShadowBot https://github.com/minetest/minetest/issues/3933 -- Allow mods to control chat position 16:44 Zeno` a single hotbar element filling the entire window was not great 16:45 hmmmm geh 16:46 hmmmm this is what hungarian notation was actually invented for btw 17:46 Zeno` What do other devs think about stopping asking people to squash their PRs in most cases? 17:47 Zeno` It seems to cause a lot of confusion and maybe people who are not "git experts" shouldn't have to do that 17:48 Zeno` I'm just trying to think of ways to make it easier for contributers and, well, squashing might not be so easy for many 17:49 Zeno` Not saying we should encourage PRs with 100000 commits, but... you know 17:50 sofar yes, if it makes sense people should squash themselves 17:50 rubenwardy commits should still be squashed when merging (if it makes sense) 17:50 sofar no, the merger shouldn't automatically squash 17:50 Obani Zeno`, if there's a way to keep the initial author by squashing the commits on the main repo 17:50 Zeno` yeah I think we should as well 17:50 Obani Flooding the history is not a good thing 17:50 sofar that's not the point 17:50 sofar you can't make a general rule 17:50 sofar sometimes it's better to make more commits 17:50 Zeno` I mean squashing when committing is not a big deal 17:50 sofar sometimes it isn't 17:51 sofar don't make a hard rule 17:51 Zeno` and then it leaves the decision up to us 17:51 Obani Well sometimes it's better to make more commits 17:51 sofar that's just gonna make people do the wrong thing 17:51 Obani When it's better, it's asked 17:51 Obani Two commits for one change are not a good thing 17:51 Zeno` Obani, but a lot of people have trouble with it. It takes me less time to squash myself than asking the person to do it for me 17:51 rubenwardy Obani, git rebase keeps the author 17:52 Zeno` It doesn't make sense because I spend more time doing other stuff before committing before pushing 17:52 rubenwardy thats why I said (if it makes sense) 17:52 Zeno` a final squash would be like 1% of the time 17:52 rubenwardy Only squash if it needs squashing 17:52 Zeno` yeah of course, rubenwardy 17:53 rubenwardy like if the history is "Add feature", "fix type", "a", "s" 17:53 Zeno` rubenwardy, well I'd encourage them to squash those if they can but it would not stop me applying the patch and squashing myself if it makes sense 17:54 Zeno` It just seems that sometimes asking "please squash" throws up a barrier that some people will have trouble with and that a core dev should not have 17:55 Zeno` I'd do git am, test, squash if necessary and commit 17:55 Zeno` I don't see the big deal 17:56 Zeno` takes less time than asking them to do it in a comment heh 17:57 Zeno` Obani, btw you're missing the point. We can squash ourselves before we push 17:58 Zeno` with less stress for us and the PR author (IMO) 18:01 hmmmm yeah probably a good idea 18:01 hmmmm i'm going to adopt that policy 18:02 Zeno` me too 18:02 hmmmm i think the reason why it previously wasn't done that way is because devs are afraid to touch other peoples' commits 18:02 hmmmm let's face it though, it's usually faster to just do things ourselves 18:02 Zeno` for sure 18:03 Zeno` and we can make sure the commit message remains "correct" 18:03 hmmmm i began fixing commit messages before merging PRs 18:03 hmmmm a lot of commit messages really suck 18:53 est31 I myself have done a bit of changing the author's patches 18:53 est31 when merging et 18:53 est31 c 18:53 est31 fun starts ofc when the merging changes add a bug 18:53 est31 then the guy who originally proposed the patch gets blamed 18:54 est31 I had been through both, both as the guy who got a bug added to his commit and to the one who added a bug to another one's commit 18:58 est31 with this new github feature squashing + merging is now very easy 18:58 est31 no more rebase -i etc 18:59 est31 well I guess it really depends on the pr, some prs have better history preserved, because the changes are in multiple focussed commits 18:59 est31 rather than one big "make everything great" commit 19:00 PilzAdam I, too, think we shouldn't bother PR authors to squash their commits 19:01 PilzAdam if it has to be done, then the core dev who merges it can easily do it 19:01 PilzAdam I think it's mostly paramat who is telling people to squash their commits, so someone should tell him to stop that 19:02 est31 no 19:02 est31 we should still ask people 19:02 Zeno` no, we shouldn't. 19:02 est31 just if they dont know what it means, we just say its okay 19:03 Zeno` There is no point to it 19:03 est31 well the thing is if you do the work for them, they will never learn 19:03 PilzAdam a comment "Please squash" in the PR always sounds like a "barrier" to the author 19:03 Zeno` they don't care 19:03 est31 they contribute more and more and you have to do it now 19:03 Zeno` and there is no reason why they should care 19:03 PilzAdam est31, why should they learn to squash commits? 19:03 est31 I mean its good that they contribute 19:04 PilzAdam just because we have this weird rules about our git history doesn't mean everyone has to obey them 19:04 est31 its not weird to have a clean history 19:04 PilzAdam the core devs already rebase all commits so we don't have merge commits 19:04 PilzAdam we don't bother PR authors to do that 19:05 est31 well the system does get more intransparent 19:05 est31 you dont know whether the change was from the core dev or from the contributir 19:05 est31 contributor* 19:06 est31 I did change the prs locally myself in the past, i do admit 19:06 est31 but well it does need to be documented 19:06 PilzAdam I'm fine with writing about squashin in contributing.md, but we should not tell authors to do so in the comments to a PR 19:06 est31 if you merge it say something like "I fixed some wording here or cleaned up style or sth" 19:06 PilzAdam at least not in a way that seems like they are required to do so 19:07 est31 Well its easiest if the contributors know the rules 19:07 est31 but of course, its an extra service to them to squash the commits for them 19:07 est31 if you want to do it, fine 19:08 Zeno` est31, the contributers don't know how (most of the time). And I don't think we should force them to learn how 19:08 Zeno` Because there is no advantage to either party 19:09 paramat ok, i won't request rebase :) 19:10 est31 rebase != squash 19:10 paramat er, squash 19:10 PilzAdam paramat, you can request, but not in a way that seems like the author is required to do so 19:11 paramat agreed 19:11 paramat it's good practice so i think we should gently encourage it 19:11 paramat many are happy to do so 19:11 PilzAdam and, at least from my experience, squashing while rebasing is so easy for the core dev who merges the PR, that it's mostly not worth bothering the author 19:11 paramat no pressure if they're not 19:11 est31 I honestly dont really care who squashes, just that in the end we should have a nice history with focused and explanatory commits 19:11 Zeno` PilzAdam, exactly 19:12 paramat i'll get in the habit of fixing commit messages too 19:13 paramat so this is all done after git am, before git push, using git rebase -i HEAD~* ? 19:14 PilzAdam you use git am to rebase? 19:15 est31 yes? 19:16 est31 its the recommended way 19:16 est31 on the dev wiki 19:16 paramat i do 19:16 PilzAdam I personally always checkout the branch and then do a git rebase master 19:16 est31 http://dev.minetest.net/Git_Guidelines 19:16 PilzAdam and adding an -i to the git rebase that I'm already doing is fairly trivial 19:17 est31 PilzAdam, so you add the person's repo as remote? 19:17 est31 that would be too burdensome for me 19:17 est31 especially as it clogs up the remote list 19:17 est31 and when i do git remote update, github becomes veeeery slow 19:17 PilzAdam it's really handy; after some time you have the usual contributors set up 19:17 est31 probably some QOS thing or rate limiting 19:18 PilzAdam I always only update a single remote 19:18 paramat i do: clone, wget , git am , push 19:18 est31 ^ my strategy too 19:18 est31 well everyone is free to decide how to merge prs 19:19 est31 I'm going again now. 19:19 est31 bye all! 19:20 paramat so after git am i'll use interactive rebase to reword message, squash 19:20 PilzAdam paramat, maybe try using remotes for all contributors; you don't have to copy anything around after the initial setup 19:21 PilzAdam although, I don't want to mess with your setup; just consider that there may be faster ways to do stuff 19:35 paramat ok 19:37 paramat is anyone working on #3894 ? seems the last big milestone 19:37 ShadowBot https://github.com/minetest/minetest/issues/3894 -- Add serverside noclip enforcement by est31 19:38 paramat i think est31 may be too busy to work on it 19:39 hmmmm i think that should wait until after the release... 19:40 paramat well, sofar says it's essential 19:40 sofar well no 19:40 sofar the thing was we didn't want people to add dig prediction to prevent this abuse 19:40 sofar to all sorts of nodes 19:41 PilzAdam this has the potential to ruin gameplay on servers, if not done properly 19:41 PilzAdam so we should do it after the release so we have enough time to test it on the servers that use the git version 19:41 sofar yes, it's quite a high risk patch 19:41 sofar ship without it 19:41 paramat ok great i'll remove milestone 19:41 sofar and skip the node dig prediction too, that can wait as well 19:42 sofar that's just a cool feature, really 19:43 paramat how about game#978 ? 19:43 ShadowBot https://github.com/minetest/minetest_game/issues/978 -- furnaces dont work after rejoin 19:47 paramat i'll review and hopefully approve tnt changes 20:01 paramat i'll mention game 978 to shadowninj 20:10 paramat +1 for tnt rework 20:11 sofar paramat: did you blow up enough TNT? 20:11 paramat sfan5 you may want to review it again due to extra commits 20:11 paramat not tested yet 20:13 sfan5 no time atm 21:18 paramat VanessaE concerning game#1011 please can we remove the check for 'group:flora_block'? as this adds a costly second 'find nodes in area' search to a frequently run ABM, and only for a mod that is most likely not used. i think it should be down to the poisonivy mod to do the extra work 21:18 ShadowBot https://github.com/minetest/minetest_game/issues/1011 -- Tweaked flower and mushroom spreading by tenplus1 22:16 VanessaE maybe, but what good is it if that ABM only runs every so often? 22:16 paramat yeah, it's unfortunate, but this check should not have been added in the first place, because it's for an optional mod 22:16 VanessaE honestly, I think that falls under the "micro-optimization" category 22:17 paramat the ABM runs very often, every long grass and flower is flora, and grasslands are dense now iin mgv7 etc 22:17 paramat nah that search is costly 22:17 VanessaE that's where the problem is then. 22:17 paramat nope 22:18 * VanessaE sighs 22:18 Calinou Minetest server still needs optimization, even with anticheat disabled and not a lot of mods :/ 22:18 paramat or we could stop long grasses being flora, but they wouldn't replenish 22:18 Calinou CTF server with doubled tick rate (dedicated_server_step = 0.05) and 20 players constantly uses 100% of one CPU core 22:18 Calinou I'd expect to be able to have like 28-30 players 22:18 VanessaE trust me, having grasses replenish over time tends to irritate people 22:19 Calinou the server uses dummy back-end, so there's literally no disk IO too 22:19 Calinou and people never generate the map 22:19 paramat yeah i'm not sure if grasses really need to spread 22:19 PilzAdam paramat, just leave the check there 22:20 PilzAdam you can optimize ABMs, but no at the cost of features 22:21 paramat i''d like to leave it, but the high cost is not justifiable for a mod that is mostly not used 22:21 paramat (more likely not used) 22:21 nore paramat: I agree with PilzAdam 22:21 VanessaE from what I gather, lots of folks use poison ivy via plantlife modpack, though not as many as mt_game 22:22 VanessaE that more mods don't use it isn't a good enough reason to remove the feature. 22:22 paramat well looks like i'm over-ruled :) 22:22 Warr1024 how much of an impact is the change, anyway? has someone measured? are there actual numbers? 22:22 VanessaE sorry paramat :) 22:23 paramat that's ok 22:23 paramat well it's a search over 512 nodes 22:23 VanessaE *furrows eyebrows* 22:23 VanessaE 512 nodes? 22:24 Warr1024 so you're talking about reducing 2 iterations over those nodes down to 1? 22:24 VanessaE I don't recall the check radius being THAT big when I maintained that code originally. 22:24 paramat radius 4 i think 22:24 Warr1024 seems almost as if you could optimize down the 2 group checks by just creating a combined group of flora_block_or_flora or something. 22:25 Warr1024 maybe there's some optimization possible in how group checks are done to begin with... 22:25 paramat yeah 4. used in 'find nodes in area' 22:26 PilzAdam can't find_nodes_in_area be used with multiple groups at once? 22:26 paramat it might be, but then how to distinguish the results? 22:26 PilzAdam and then search the resulting array for the groups we care about at the moment 22:26 VanessaE that would be slow if done in lua.. 22:27 PilzAdam searching a (short) array twice should be faster than searching 512 nodes twice 22:27 paramat i'm not sure 22:27 VanessaE you would think so, but biome_lib does just exactly that, and it's not very fast. 22:27 PilzAdam VanessaE, have you benchmarked it yet? consider that the array that is returned is very short 22:27 VanessaE I did, long ago. don't have the results anymore though 22:28 Warr1024 groups don't really change at runtime, and there are only so many registered nodes... so you could pre-build a lookup for whether a node is in the target set and do the per-node check in constant time regardless of how many groups you're searching for... 22:28 VanessaE but I dunno, maybe it'll be ok if the list to search is sufficiently small 22:28 PilzAdam we are talking about flora here, which usually is above ground 22:28 PilzAdam so we have an array with 16 elements in the worst case 22:31 VanessaE hm 22:31 VanessaE fair enough 22:31 VanessaE still, an API call ought to be added to perform a search like that 22:31 VanessaE for the sake of speed. 22:31 VanessaE (it could be useful for other things, of course) 22:31 VanessaE (biome_lib could use it ;) ) 22:31 paramat btw should long grass replenish? it's only used for seeds 22:32 paramat removing that from flora would greatly reduce the ABM load 22:32 VanessaE definitely not 22:32 VanessaE well.. 22:32 VanessaE hm 22:32 VanessaE maybe it should replenish, if slowly 22:33 VanessaE problem is, you clear an area for building on a busy server, get started, come back a week later to find random grass everywhere 22:33 VanessaE I used to do just that with jungle gras 22:33 VanessaE +s 22:33 VanessaE people didn't like it. 22:34 paramat also, i now use noise to shape the differing grass heights in grassland, spread would eventually result in all lengths everywhere which is boring 22:34 Calinou there should be a way to regrow grass then 22:34 Calinou something like bone meal, but Minetesty 22:34 Calinou Mese Meal :) 22:34 VanessaE G*d no 22:35 VanessaE just call it "fertilizer" and make it with something meaningful 22:36 VanessaE a block of dirt, grass node, some leaves, and whatever else you can think to put into the mix to make a big organic mass, plus one or two mese crystal fragments to "charge it" 22:36 PilzAdam didn't kilbith suggest adding bullshit in some PR? 22:36 VanessaE actually I think he did :) 22:36 PilzAdam that would make a good fertilizer 22:36 paramat heh 22:36 VanessaE and now you've done it 22:36 VanessaE you went and invoked him 22:36 VanessaE now he'll show up ranting :) 22:36 paramat junglegrass is flora, dry shrub is not 22:37 VanessaE paramat: fwiw, my junglegrass mod would grow grass from one size to another, including dieing, and if it spawned in the desert, it would turn to dry shrub when it dies. 22:38 paramat i used to think only flowers spread, because they are vauable, i was surprised to discover long grasses do 22:38 paramat well i'll open an issue, we can consider 22:44 VanessaE in any case, 23:05 paramat yeah 23:06 paramat i think #3836 can be closed 23:06 ShadowBot https://github.com/minetest/minetest/issues/3836 -- Player gets detached from object when leaving server