Time Nick Message 09:50 emptty Hello there 09:50 emptty It seems that minetest 4.7 crashes with a floating point exception if the openGL support is not sufficient : http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=720385 09:51 Calinou "Severity: important" 09:51 emptty do you guys know about this issue, or should I report it as a new issue? 09:51 Calinou it's not like there are too few drivers supporting it... 09:51 Calinou it wouldn't run anyway... 09:51 emptty it could be a clean abort() after detecting that openGL is not sufficient, couldn't it? 09:52 emptty I agree that the severity is a bit overstated, but it's not my fault ;) 09:52 Calinou then that's a minor bug, and nothing important 09:52 Calinou and that almost never actually happens, the open source drivers supporting 2.1-3.1 09:53 emptty I will downgrade the severity. 09:54 emptty could you guys add a cleaner exit(1) where it belongs? 10:31 BlockMen morning, any comments on that? -> https://github.com/BlockMen/minetest/compare/pause_menu 10:41 Exio4 https://github.com/BlockMen/minetest/compare/pause_menu#L1R1465 10:42 Exio4 why do you do that? 10:44 BlockMen because when the game is paused the cpu usage reduces from ~20% to 5% with that 10:44 BlockMen and the user wont notice a difference 10:44 Exio4 the gui is rendered "with the same" delay though 10:45 BlockMen yes, thats why the delay is 100 and not greater 10:46 BlockMen upps..200 10:46 BlockMen :D 10:46 Exio4 it is 200.. 10:46 Exio4 ok 10:46 Exio4 https://github.com/minetest/minetest/pull/237 10:46 Exio4 calinou's request 10:49 BlockMen hmm..could be done, but idk if it makes that much sense, because it should be cut at 200 IMO (else you would have these gui lags) 10:50 Exio4 200ms is 5fps 10:51 Exio4 the input latency is "up to" 200ms with that delay 10:53 Exio4 anyway, a coredev should talk 10:54 BlockMen well, im not fixed on the delay, i just thought it would be nice to have it in 11:02 troller https://github.com/proller/minetest/compare/heat https://github.com/proller/minetest/compare/liquid63 11:03 nore you should make a bot for that ;) 11:04 troller 8( 11:04 troller simpler on join script 15:35 proller https://github.com/proller/minetest/compare/heat https://github.com/proller/minetest/compare/liquid63 20:25 sapier1 iqualfragile I've got a problem mobf is too big for mmdb filesize limit ;-) is there a way to specify a external link? 20:26 PilzAdam sapier1, external links are too insecure 20:26 sapier1 ok then I request a larger file limit ;-) 20:26 PilzAdam we dont support bloated mods :-p 20:27 sapier1 I'm going to kick your as sooon ;-P 20:29 sapier1 why does every tool have it's own markup language ... github mmdb forum 20:31 sapier1 where to find a doc how "unflavored markdown" works ? 20:37 sapier1 PA external links can be secured by md5 sums ;-) 20:37 sapier1 ok sha512 is better 20:38 sfan5 > secure > md5 20:45 sapier1 I guess for files even md5 is safe enough sfan5 ... not for passwords of course 20:46 sfan5 there are tools available that make it easy to produce an md5 collision for given files 20:47 sapier1 I guess this will result in huge bloated files ;-) but I already said sha512 would be better ;-) 20:48 sapier1 but I prefere increase of filesize limit 21:25 kahrl gentoo checks sha256, sha512, whirlpool and size 21:26 sapier1 whirlpool? 21:27 kahrl https://en.wikipedia.org/wiki/Whirlpool_%28cryptography%29 21:27 sapier1 wasn't that the one mathematics aren't exactly sure if it's really that safe? 21:27 kahrl dunno 21:28 sapier1 something about a mathematic operation used guessed to be very expensive but not prooven ... but I'm not sure if I mix this up with some other algorithm 21:29 sapier1 but I don't think hash support will be added that soon ;-) unless something else wants to add it 21:29 kahrl mathematics aren't even sure if P=NP; if P=NP then true one-way functions don't exist afaik 21:31 sapier1 yes P=NP isn't prooven ... not sure what you mean with one way ... security isn't based uppon one way or not but one way beeing way more expensive than the other one 21:32 kahrl that's what is meant by one-way function 21:32 sapier1 ok ok 21:33 sapier1 but the problem isn't security of hash algorithm but unknown modsize limit of mmdb ;-) 21:38 kahrl I wonder if should remove the color support from textlist when I add the tables 21:38 kahrl if i* 21:38 sapier1 tables? 21:39 sapier1 I'm against removing color support there are many cases where colored list elements are very usefull 21:39 kahrl http://paste.dy.fi/BJ9 21:40 sapier1 are you really sure this is usefull at all? I mean it's going to work of course but who the hell will be able to write this? 21:40 kahrl yeah but the tables cover textlists as a very simple special case, and I don't like how you have to escape every textlist element starting with '#' 21:40 kahrl oh I am coding away 21:41 kahrl (and nobody cares to escape textlist elements starting with '#') 21:41 sapier1 imho you're doing work better invested in a more complete solution 21:42 sapier1 but as I'm sure you wont stop either ;-) 21:42 kahrl what would that encompass? 21:42 sapier1 I guess you already have read why I didn't do treeview so no need to discuss again 21:42 kahrl I might have but it's been long ago 21:43 kahrl if you find it in the logs link it, or re-tell 21:43 sapier1 ok short I didn't do it because this is additional work to a already overcomplicated by design flawed gui language 21:44 sapier1 formspec is nothing but another gui description language 21:44 kahrl I don't see any replacement coming soon 21:44 kahrl and we (imo) need a better looking serverlist to release 0.4.8 21:45 sapier1 maybe but adding even more features won't make the task easier and if someone is gonna do this all the work you did is wasted 21:45 Exio4 the thing kahrl is coding will be useful "for laterz" 21:45 sapier1 it's usefull for NOW 21:46 sapier1 if the clean solution is implemented it's useless 21:46 kahrl I guess if it seems useful it could be adapted to a new GUI thing 21:46 sapier1 but of course I don't know if and when that is gonna happen 21:46 Exio4 isn't kahrl implementing in-text colors for that? 21:47 Exio4 (like it was "just that") 21:47 kahrl Exio4, not really 21:47 Exio4 ah 21:47 kahrl you can set a color for each cell 21:47 sapier1 no it can't be adapted as most of the work you do is defining parsing and evaluating data structures which most likely will be different if someone implements support for another more common gui description language 21:48 sapier1 that doesn't make your work bad of course it's quite good ... I'm just afraid it's something as damage handling ... working but way to complex to be usefull 21:49 kahrl the event handling can be reused (IGUIListBox and others can't be used without the hacks that are there now) 21:50 sapier1 I'm not even sure if event handling (except lua part) can be reused if new gui isn't done with irrlicht gui elements for example 21:51 sapier1 I suggest adding textlist compatibility support I guess it's not too difficult to map it to your new mechanism 21:51 kahrl yep that's what I am doing 21:51 kahrl GUITable has a setTable() and a setTextList() method ;) 21:51 sapier1 kahrl how do I specify different table options for e.g. two tables in a formspec? 21:52 kahrl tableoptions[] table[] tableoptions[] table[] 21:52 sapier1 no no 21:52 kahrl what's wrong this time 21:52 sapier1 that's adding context sensitivity to formspec are you really really sure you want to open up this thing? 21:53 kahrl isn't it effectively context sensitive already, with size[] 21:53 sapier1 no 21:53 kahrl and what's the problem when it is context sensitive 21:54 sapier1 a second size overwrites first one 21:54 proller https://github.com/proller/minetest/compare/heat https://github.com/proller/minetest/compare/liquid63 21:54 sapier1 it's not "wrong" but adds a lot of complexity 21:56 kahrl an additional field in parserData is a lot of complexity? 21:56 sapier1 e.g. what to do if options is missing? don't draw table at all? use default value? 21:56 kahrl all options have default values 21:57 sapier1 what if between options and tables is another field? use last options? yes that's all a matter of definition but you need to add this to documentation 21:57 kahrl I think it already states that 21:57 sapier1 no it doesn't 21:58 sapier1 your examples show tables without table options so what options are used? 21:58 sapier1 I guess default 21:58 sapier1 but it'd be perfectly ok if last options set where used 21:58 kahrl guess you can interpret "last" in different ways 21:59 kahrl but yeah that is what is meant, and if there are none then the defaults are used 21:59 kahrl you can use "tableoptions[]" to reset all options 22:00 sapier1 so actualy you specify a single formspec isn't "sometext[]" anymore but can be a sequence of those elements too 22:00 kahrl (if you say tableoptions[background=#AABBCC] table[...] tableoptions[background_alpha=#00] table[...], the second table doesn't use the background from the first tableoptions) 22:01 sapier1 but you'd actually have to post an error in this case as the formspec definition is broken 22:01 kahrl what? 22:02 sapier1 sorry didn't read the table in there 22:02 kahrl formspec doesn't have a way to report errors 22:02 sapier1 error is reported to log only of course 22:03 kahrl I won't add parsing for that kind of stuff 22:04 kahrl if you say tableoptions[...] tableoptions[...] the first will be silently ignored 22:04 sapier1 never mind I don't want to stop you just want to make sure you are aware you increase complexity level of formspec language by a whole class by doing it that way 22:05 sapier1 no kharl thats wrong the second one needs to cause an error message as this is a invalid formspec tag 22:05 kahrl huh? 22:05 sapier1 there's no effect of that statement there's no use to do it that way 22:06 sapier1 chances for this to be an error is 99.99999999% 22:06 kahrl you can code a nop in many languages without causing an error 22:06 kahrl (in fact the halting problem shows you can't even decide if some code is a nop) 22:06 sapier1 "error" in formspec is comparable to a warning 22:07 sapier1 as your table is a very very complex design you really should warn about invalid parts 22:07 sapier1 you can't tell in general that's completely different from you can't tell 22:08 kahrl I can see someone adding the same tableoptions[...] at the top of all his formspecs, to have certain defaults, but have (additional) tableoptions in formspecs that want different ones 22:08 kahrl that shouldn't cause a warning 22:08 sapier1 I just don't want another "damage system" cause where docs and implementation have been that bad I needed to use gdb do find out how it works 22:09 kahrl besides trying to warn about such things would be the *only* thing that adds complexity to the parsing 22:09 sapier1 I thought table options are used for the very next table only? 22:09 kahrl no, I thought I said that 22:10 sapier1 ok :-) I guess this is already first case of missunderstanding 22:10 kahrl they stay until they are changed 22:10 kahrl they affect any tables that are created while they are active 22:11 sapier1 ok so the setup a context ... ok then the warning is bogus of course 22:12 sapier1 is this same for tablecolumns? 22:12 kahrl yeah 22:13 sapier1 ok ... hope you know what you're doing ;-) 22:15 kahrl no doubt this will be the most difficult formspec element to master, but there will be examples of it in the main menu 22:15 sapier1 I'm sure it's gonna work somehow ... but I'm not convinced this is right direction to go 22:17 kahrl basically what the formspec docs will be saying is: here are some simple tools for your formspec, use them, but if these aren't enough, here is an advanced tool that can do a lot more stuff 22:18 sapier1 I suggest creating a classlike interface for this one maybe this will reduce complexity to a managable level 22:18 kahrl classlike? 22:19 kahrl don't know how that would work with the formspec syntax 22:19 sapier1 e.g. local table = fs_table.create() 22:19 sapier1 table.add_column(some name) 22:19 sapier1 ... 22:19 kahrl stress will do that ;) 22:19 sapier1 and last table.get_formspec() 22:20 sapier1 at least you won't have to count your "," 22:21 kahrl it'd be weird to have this in the core, but only for table and not for other elements 22:21 sapier1 or maybe table.insert_text/image/(row,column,text/texture) 22:21 sapier1 of course ;-) 22:22 sapier1 but I'm almost sure if I'm gonna use that thing I'll write a interface like that because I usually forget some "," 22:23 sapier1 have a look at the filterlist implementation that's similar, I didn't want to implement a filtered list for each tab on its own ;-) 22:25 sapier1 btw are there any plans to release 0.4.8? 22:25 PilzAdam I think 0.4.8 waits for kahrl HTTP fetch thingy 22:25 sapier1 why is kahrl working on tables then? 22:26 kahrl I'd really like to have the tables in the 0.4.8 main menu 22:26 kahrl these text flags aren't gonna cut it 22:26 sapier1 this is gonna delay 0.4.8 for additional 2-4 weeks ... long enough to get another "must have" feature 22:27 sapier1 imho we should release more often ... at least as long as we have a stable version 22:28 sapier1 and as far as I see current version is most solid one for quite some time 22:28 kahrl the textlist bugs that I fixed recently would have been bad to be in a release 22:29 sapier1 of course there are some bugfixes we should add 22:29 sapier1 but no big feature enhancements 22:30 PilzAdam I dont like releases blocking too much development, and each release needs a feature freeze, so too frequent releases are not good 22:30 sapier1 last version is 3 months ... that's biggest gap for quite a long time 22:30 sapier1 to frequent? have you read what you wrote? ;-) 22:30 PilzAdam I read what you wrote 22:30 PilzAdam imho we should release more often 22:31 kahrl aside from the odd thing here and there the stable and git versions should be mostly network compatible 22:31 kahrl right now, unless I'm missing something 22:31 PilzAdam there is this water glitch when connecting to an old server, should that be fixed? 22:31 kahrl oh that 22:31 sapier1 yes still git version is way more advanced than stable one and adding everything you want to add kahrl will delay release for at least another month 22:32 sapier1 I expect it to be 2 22:32 sapier1 do we really want to exclude less experienced users from those features for that much additional time? 22:33 kahrl building up anticipation is a good thing in the gamedev community ;) 22:33 sapier1 you know trend is to do more ci? 22:33 proller your glitch fixed here - https://github.com/proller/minetest/compare/liquid63 22:34 PilzAdam proller, your "fix" is just commenting out my code, thats not good 22:34 sapier1 we've even passed kernel innovation cycles by now ;-) 22:35 sapier1 maybe we should use release branches? 22:35 sapier1 so release branch is feature frozen while development on git branch can continue 22:35 proller PilzAdam, your code strange, have errors and must no touch visual 22:36 PilzAdam proller, if it doesnt change the drawtype then it looks weird with range of 2 or 3 22:36 kahrl we already don't have enough manpower to deal with all the pull requests, adding more rebasing and merging work isn't going to help 22:36 proller PilzAdam, who will use your "feature" ? 22:37 sapier1 so we keep on doing our stuff for our selfs and discover community has gone due to boredome ? 22:37 PilzAdam proller, how is that relevant? 22:38 proller PilzAdam, my code allow to change number of water levels, its can be used in same way as your "range" 22:38 kahrl I'm not paid so I don't really care how many people use or don't use minetest 22:38 ShadowNinja +1 for sapier's release branch idea. Only merge new commits into it that areproven bugfixes. 22:39 PilzAdam proller, hm? does your code change anything in normal liquids? 22:39 kahrl when there are releases, I do want them to be of good quality 22:39 proller PilzAdam, no 22:39 PilzAdam so it cant be used instead of my code 22:40 proller dont call them "normal" 22:40 sapier1 kahrl I'm not paid too but I'm still interested in minetest to be a success ... it's disapointing to have none use it because this game is dead becaus it's not updated for half a year 22:40 sapier1 the release branch would minimize the merge effort 22:41 ShadowNinja It is startingto get a bit long... 22:41 kahrl what keeps people interested is mostly the mods, not the engine 22:41 PilzAdam sapier1, minimize merge effort? 22:41 kahrl and mods don't follow the engine release schedule 22:41 sapier1 exactly and current git version is a huge jump in mod support 22:42 ShadowNinja Mods depend on the engine, a lot of the nice mods depend on features like LuaVoxelManip. 22:42 sapier1 the online modstore is a big benefit compared to old version that's why I want to get this thing out into wild 22:43 sapier1 imho modstore is a "killer feature" delaying it will harm minetests spread 22:43 kahrl use the time until the release to get all the popular mods into the modstore? 22:44 PilzAdam to accomplish that you need to make the modstore actually useable 22:44 kahrl if it is a killer feature it has to "launch" well 22:44 sapier1 you know that this is a chicken egg problem 22:44 sapier1 why add it if noone can use it? 22:45 kahrl git users != noone 22:45 sapier1 git users == nokia n900 users ... others == iphone users ... guess whom are more 22:45 kahrl also it will ensure that people who only use stable will see their mod once the modstore launches in a release 22:47 sapier1 there won't be any noticable mod count differenc if we launch in 2 weeks or 2 months ... ppl are lazy if you tell em it's 2 month they'll think "enough time I'll add it later" 22:47 PilzAdam sapier1, mmdb needs more developement, its not in a state for a release yet 22:47 kahrl why do the authors have to be asked at all? most mods are WTFPL 22:47 ShadowNinja sapier1: How does the mod manager work? I only see options to download or install mods that are already installed. 22:48 PilzAdam ShadowNinja, click on "Download" 22:48 PilzAdam its very unintuitive 22:48 kahrl PilzAdam: I agree 22:48 sapier1 any better options? 22:49 ShadowNinja There, but it took about five seconds to load. 22:49 proller PilzAdam, i dont know how to combine this place, your code incompatible with my 22:49 sapier1 rename download to "browse modstore"? 22:49 kahrl an interface similar to the firefox addon manager maybe 22:49 ShadowNinja And the down arrow takes a few seconds too... 22:49 sapier1 kahrl if you want to write this thing go on 22:49 PilzAdam proller, make the drawtype depend on the finite_liquid setting 22:49 kahrl maybe it should download the mod repository once and not on every page load 22:50 sapier1 shadow it's not gonna get any faster it's downloading on the fly 22:50 kahrl update once a day 22:50 proller PilzAdam, new drawtype only for your feature? 22:50 ShadowNinja It also shows four pages, two of them are empty. 22:50 PilzAdam proller, no, not new drawtype, just an if(finite_liquid) 22:50 proller client doesnt know it 22:51 sapier1 yes those mods don't provide downloadable versions ... but this isn't told in overview 22:51 ShadowNinja Also: WorldEdit isn't detected, which prings me to annother bug, I cannot enable WorldEdit because I have it installed inside a modpack. 22:51 PilzAdam proller, oh; maybe it would be nice to have an addditionl liquidtype "finite" in the nodedef, since finite liquids dont need source and flowing 22:52 sapier1 ok ... I understand you ppl want the golden edge version ... so I'm gonna wait until pilzadam requires a feature for his mods 22:52 PilzAdam so its not a setting in minetest.conf anymore but up to mods to register their liquids as finite 22:52 kahrl sapier1, it's just not usable as is 22:52 ShadowNinja It adds WorldEdit to world.mt, not worldedit, worldedit_commands and worldedit_infinity. 22:52 sapier1 no kahrl thats plain wrong it doesn fit your feelings 22:52 ShadowNinja (Notice caps) 22:53 kahrl my feeling are fine :P I know the code so I can use it 22:53 sapier1 it's does everything it's designed for of course this can be done better ... there's always a thing that can be done better 22:53 PilzAdam sapier1, kahrl is right, its not very useable 22:53 kahrl but everyone who talks about it is stumped by it 22:53 ShadowNinja +1 for finite in nodedef, it would allow for things like oil. 22:54 sapier1 the only ppl talking about it are the ppl who wanted to do it better for about a year but didn't do it 22:54 proller PilzAdam, now source better for optimizing abm, 22:54 PilzAdam proller, can you please write English? 22:54 proller and "finite" - now incorrect, better "real" or.. 22:55 sapier1 so yes this is a first step and yes it needs rework ... but NO its not unusable 22:55 proller PilzAdam, no 22:55 sapier1 maybe "download" isn't best label for that button but switching to e.g. "browse online" isn't a big deal 22:56 PilzAdam sapier1, the text is not a big problem, but the whole GUI design 22:56 PilzAdam the button is under the "Installed mods" list, so people figure it has something to do with it 22:57 sapier1 pilzadam I guess I'm gonna ignore your jelous comments 22:57 PilzAdam the "Remove" and "rename" buttons, which actually modify the list, are somewhere else 22:57 proller mixing finite and old liquids is hard, now only one queue, its need to split 22:58 sapier1 and installing a mod doesn't modify the list of course 22:58 kahrl sapier1: what? what PA says about the buttons is right 22:58 sapier1 no he isn't 22:58 sapier1 did you mention the free room above the depends list? 22:58 kahrl it's a basic rule of UI design that things that belong together should be positioned together 22:59 sapier1 this is reserved for mod icon and description 22:59 * VanessaE peeks in 22:59 sapier1 not beeing present atm 22:59 VanessaE saw the previous chat about the mmdb... tried to add pipeworks just now and it didn't show in the web page's list after adding. 23:00 VanessaE (figured I'd get a jump on it before 0.4.8 goes out) 23:00 kahrl VanessaE: you'll have a little bit of time ;) 23:00 ShadowNinja VanessaE: I added a mod and it worked, although the screenshots don't. 23:00 VanessaE ok :) 23:00 sapier1 so the right half is for modifying the thing described above (of course htat information isn't complete) 23:00 PilzAdam sapier1, if the whole area is about local mods, how should the user figure out that the bottom left corner is about downloading them from an external repo? 23:01 sapier1 while the left one is list manipulation 23:01 VanessaE oh there it goes. 23:01 VanessaE it just took a bit for the list to update. 23:01 kahrl imo the mod manager should initially show the online modlist 23:01 sapier1 no 23:01 ShadowNinja Better to add mods before the release. And the mmdb also needs a redesign, the|xyz posted a really nice one. 23:02 sapier1 if you do so kahrl you kick out everyone without curl support 23:02 kahrl the users who need it will see an empty list which isn't useful at all 23:02 sapier1 it is usefull as you can install mods from local files 23:02 PilzAdam sapier1, downloading a mod is more common than removing installed ones 23:02 VanessaE ShadowNinja: actually, I think the way the web page works now is fine. Just that screenshot/title pic issue, and a minor layout tweak is all that are needed 23:02 sapier1 thanks for your oppinion o blessed one 23:02 kahrl the vast majority use the windows build, which has curl support 23:03 PilzAdam btw, "Configure" in the world selection should be named something like "Mods" so its more obvious to users what it actually does 23:03 ShadowNinja VanessaE: This is what he posted: http://i.imgur.com/MZyGxJN.png 23:03 sapier1 your argumentation is inconsistent "most users are git users" and now you tell vast majority use windows builds 23:03 VanessaE ew. 23:03 VanessaE that's just too spartan 23:03 sapier1 plz decide what you wanna tell me 23:03 PilzAdam sapier1, nobody said that most users are git users 23:04 kahrl please show me where I said that 23:04 ShadowNinja VanessaE: Notice that the download/no release button overlaps with everything under it? 23:04 sapier1 (00:47:45) kahrl: git users != noone --> impies you assume a at least big git userbase 23:04 kahrl no, that statement implies there is at least 1 git user 23:04 PilzAdam no, he implied that there is 1 or more 23:04 VanessaE ShadowNinja: well yeah that page needs some tuning 23:05 sapier1 you're cherrypicking and reeinterpreting to your current needs ... no use to continue this discussion 23:05 ShadowNinja I like that desgin, anyone else? 23:06 kahrl so I can work on the table code again? :P 23:06 proller PilzAdam, old liquid can be switched to getLevel/getMaxLevel and starts support change number of levels 23:06 sapier1 do what you want with main menu I'm out until you settled what you wanna do make minetest your privat playground kicking all different oppinions or make it a successfull project 23:06 PilzAdam proller, I dont like how you want to push the whole core into finite liquids 23:07 kahrl sapier1, I don't know why but you seem to get very protective of your code and ideas 23:07 kahrl in a "don't change that!" sense 23:08 sapier1 no I'm not protecitve I don't have any problems with adding your tables 23:08 sapier1 I'm just annoyed all things I did are endlessly delayed although they are way better than before 23:09 VanessaE hm, another issue - why are there only ever 12 entries shown on the mmdb page? 23:09 VanessaE OH 23:09 proller PilzAdam, "finite" now only in Map::transformLiquidsFinite, other changes about all liquids 23:09 VanessaE that's pages 1 and 2.. not "there are 12 pages" 23:09 VanessaE duh 23:09 sapier1 minetest needs a mod manager if someone comes up tomorrow with a way better one than the one I did I'll be the first to demand releasing that one 23:09 PilzAdam sapier1, all the annyoing things were fixed lately by kahrl (lists jumping arround etc.) which you have labeled at "wont fix" before 23:09 sapier1 but I don't think this is gonna happen 23:10 VanessaE put dashes between the numbers and move them up next to "Pages:". 23:10 PilzAdam sapier1, this gives me the impression that you dont have high quality standards for releases 23:10 sapier1 yes if you have a look at the amout of work required to fix it and the benefit they gave you know why I didn't want to do it right upfront 23:11 sapier1 you're right I'm for release early release often 23:11 PilzAdam you underestimate the benefit 23:11 sapier1 as adding all that work creates facts whil we don't even know if this is gonna be the design ppl want 23:12 kahrl if the UI behaves extremely annoyingly users are going to reject it 23:13 kahrl it has to work reasonably well to get people's real opinions 23:13 kahrl besides, deleting code is cheap 23:13 sapier1 ok and we need pretty server flags for a reasonable well working ui true? 23:13 PilzAdam if I would have known about this beahaviour of the Lua menu then I would have never agreed to merge it 23:14 sapier1 if you didn't know about it that's your fault, it's been there for about a month prior beeing added 23:14 VanessaE also, all of the mods that come as part of minetest_game (or just the game itself? something) need to be added to the deps/soft deps lists. 23:15 PilzAdam sapier1, just wanna give you an impression how big these bugs are 23:15 PilzAdam *were 23:15 sapier1 and that this wasn't 100% finished wasn't a secret it's been told prior merging 23:16 sapier1 I don't deny there where big bugs when beeing merged that wasn't avoidable as none of us did even know how old menu did behave 23:17 proller https://github.com/proller/minetest/compare/heat -- small temp-humidity improvement 23:17 sapier1 but the big ones are gone and kahrl didn't only fix bugs in mainmenu some of the bugs have been present before too 23:17 sapier1 e.g. doubleclick bug 23:18 PilzAdam sapier1, but these bugs were fixed after being upstream for rather long time 23:18 PilzAdam so frequent releases arent possible (at least in this case) 23:18 sapier1 of course we can't release a version not beeing usable but that's not the case atm 23:19 PilzAdam there are 3 people complaining about the modstore 23:19 PilzAdam so it might be not useable 23:19 kahrl at least in my case, it took me quite some time to get used to the new mainmenu code 23:20 kahrl that's why I wouldn't have been able to fix the bugs in july 23:20 sapier1 I don't blame anyone for having to learn how code works that's pretty normal 23:22 sapier1 my simple question is do we want to give ppl chance to try it out and use their feedback for a real good version or do you wanna add your personal golden edge until christmas and then tell ppl "its done that way, I believe it's perfect, get used to it" 23:23 PilzAdam sapier1, people are free to test the git version; there are even windows build of its so we get quite a lot of "testers" 23:23 PilzAdam its not like every not released commit is hidden to the public 23:23 sapier1 I'll gladly help rewrite it to adjust it to users needs but I'm not gonna rewrite it to match pilzadams/kahrls personal feelings just to rewrite it again if other ppl feel different 23:25 sapier1 I alread did rewrite parts to match pa s feelings just to rewrite it again because it wasn't right either 23:25 PilzAdam sapier1, any examples for that? 23:25 sapier1 background 23:25 PilzAdam can you explain that a bit more? 23:25 sapier1 there have been 3 wrong versions how background is set 23:26 sapier1 no one did know what's correct but noone did admit to not know 23:26 PilzAdam I never changed my mind regarding the background, and it fits what we have currently 23:27 sapier1 of course you're perfect pa 23:27 PilzAdam I always told you how it behaved before the Lua menu 23:28 sapier1 no you didn't you've been wrong about overlay too 23:28 kahrl in all this time spent talking somebody could implement a perfect modstore :P 23:29 sapier1 I guess you're gonna present a perfect bugfree modstore tomorrow along with tables and http fetcher? 23:29 sapier1 of course don't forget about the offline modmanager feature and add the screenshots and description to all existing mods 23:29 kahrl don't take it so serious man 23:29 PilzAdam sapier1, judging the code kahrl submited in the past that is not very unlikely ;-) 23:31 sapier1 kahrl I guess you already mentioned I'm not gonna start big tasks anymore for quite a while due to merge policy enforced ... it's almost unpredictable what's gonna be merged and what not 23:32 sapier1 to mee it looks like if a core developer is interested in a particular commit every crap is merged if something doesn't match his current feelings a patch can be not good enough to be merged 23:33 sapier1 and this is anyoing and frustrating 23:35 sapier1 and PA your comment is quite useless no matter how good kahrl is 23:35 kahrl I don't have a solution to that, merging pull requests is unrewarding work so it's rarely done 23:36 sapier1 this could be interpreted as "either be a core developer or fork it yourself" 23:36 ShadowNinja 64 now, remember when 40 was very high? 23:37 kahrl forking is always an option... for example I don't think minetest-delta was harmful, in fact it seemed like a very productive fork 23:38 kahrl and most of it got merged back into minetest-c55 eventually 23:38 sapier1 that's an issue because I couln't even do a cleanup to thise things I wrote because I'd need to rebase it 20 times untill it's gonna added again 23:38 sapier1 it's easy to you kahrl because you can add it as soon as it's finished 23:39 sapier1 if I'd rewrote modmanage I guess it'd be october until it'd be merged 23:42 kahrl sapier1, being a core developer makes some things easier, yes, but I still have to wait sometimes to get some things merged 23:42 kahrl either because no one is around, or it sucks, or it conflicts with something else going on at the moment 23:43 kahrl it just reduces the number of other people having to agree with it from 2 to 1 23:43 sapier1 I know you can't do everything 23:44 sapier1 no the big difference is what you already told "merge is not a rewarded" ... you can do it yourself no need for someone to have interest in merging 23:44 kahrl if I did that I'd get kicked out of the dev team almost instantly 23:44 sapier1 I guess I'd rewrite about 20-50% of mainmenu after current feedback ... but it's useless 23:44 proller https://github.com/proller/minetest/compare/heat -- 3 days waiting for approve 23:45 PilzAdam proller, maybe thats because several other core devs already told you to develop that in a fork? 23:45 sapier1 kahrl I don't mean you can add it without approval but you can add it instantly after approval 23:45 sapier1 to be more precise "YOU" can add it 23:46 proller PilzAdam, and make new game weatertest? 23:46 sapier1 to others it's not likely that something is even approoved but not added because noone cares enough to add 23:46 sapier1 unlikely of course 23:46 kahrl if some pull request gets 2 approves it always gets merged relatively quickly 23:46 kahrl I haven't seen otherwise 23:46 kahrl the problem is getting the approvals 23:47 sapier1 I don't have numbers for this but it doesn't feel to be so not only to things I suggested 23:48 kahrl if it happens feel free to mention it here, that's what chat is for 23:48 sapier1 you know how often I mentioned different pull requests allways getting reply of a single core dev but rarely 2 devs at same time 23:48 kahrl I don't see a reason it wouldn't get merged in such a case, except "either because no one is around, [...], or it conflicts with something else going on at the moment" 23:49 sapier1 https://github.com/minetest/minetest/pull/846 for example I don't know exactly how often I mentioned this one at least 3 times 23:49 kahrl the same commit can get approvals from 2 devs at different times 23:50 sapier1 https://github.com/minetest/minetest/pull/640 is another candidate too 23:51 kahrl sapier1: did anyone actually approve them? 23:52 VanessaE that one pull there would mean deleting about half of plants_lib :P 23:52 VanessaE (which is a good thing) 23:52 sapier1 as ppl don't actually say "I herby hnoroably approov commit xxx" I can't tell this for sure 23:52 VanessaE maybe what's needed is a signed-off-by tag? 23:53 kahrl well that's the problem: getting someone to actually review the code (which is hard work) 23:53 kahrl I mean, for #640, I looked at the code briefly and found some minor things I thought were problems, and mentioned them 23:53 kahrl but I never thoroughly reviewed the code 23:53 sapier1 the only issue mentioned to 640 is that there are similar but yet different functions present ant those could possibly be merged with a some time ... I guess this is what you mean 23:54 sapier1 -a+e 23:54 kahrl in cases like omnicleanup where no one is ever going to review the code, yes being a core dev makes things drastically easier 23:55 sapier1 actualy beeing a non core dev only minor changes have a real chance to get merged ... I'm still astonished how I managed to get scriptapi split as well as mainmenu in. 23:56 ShadowNinja You're lucky if your pull gets merged at all, regardless of it's size. One line changes can wait months before being merged. 23:59 kahrl remember that even for one line changes, code review means thinking of all the places that might be affected by the change 23:59 kahrl it's not simple at all 23:59 sapier1 maybe really enforcing merge policys would make merging at least more transparent 23:59 sapier1 e.g. 2 core devs have to comment a commit as "approved"