Time Nick Message 00:39 OldCoder Comments please on http://minetest.org/crash.txt 00:44 VanessaE so... critical regression here and no one's gonna look at it? 00:55 * OldCoder relates 00:57 OldCoder VanessaE, any thoughts on that log? Sokomine, RealBadAngel > 00:57 OldCoder ? bbs 04:32 VanessaE bump #2877 04:32 ShadowBot https://github.com/minetest/minetest/issues/2877 -- Player files corrupting with "blank" items 04:32 VanessaE for G*d sakes someone please take a look at this, it's fucking up player inventories in ways they can't fix 04:37 AnotherBrick looks like somehow items count is decreared below zero and since it's unsigned word you got something near 65535 04:37 VanessaE yeah but notice the blank spaces 04:37 VanessaE there's no item name 04:38 VanessaE afaik a mod can't do that 04:39 VanessaE and players are reporting items "disappearing" from inventories as well 04:39 VanessaE stuff that was there at one point, will be gone or reduced significantly the next time they login 04:39 AnotherBrick can it be caused by some bugged android client? 04:39 VanessaE I don't see how 04:40 VanessaE a client doesn't directly access/write a server's player file does it? 04:40 AnotherBrick it doesn't 04:43 AnotherBrick it might probably perhaps send inventory action, like "take 26 items for that cell and put there" and if nothing really in that cell and server perform that operation you'll got -26 = 65510 empty items 04:44 AnotherBrick i don't really know how that works, but it probably might perhaps be caused by similar scenario 04:44 VanessaE in which case the server ought to be denying the action I'd think 05:22 cheapie AnotherBrick: But if you have 1, say, cobble, and you take away 2, you should then have 65535 /cobble/ instead of 65535 nothings, right? 05:22 AnotherBrick yeah, probably 05:23 AnotherBrick but if you have zero of nothing and takes one out, u'll got 65535 nothings 05:33 cheapie OK, so I guess we have two problems here. 05:34 cheapie Items are turning into nothing, and there's no underflow detection. 06:49 VanessaE cheapie: I can reproduce my bug with vanilla minetest_game + gloopblocks (only used because it uses the replacements feature of the crafting grid) 06:49 VanessaE see the revised description 09:27 crazyR could someone add a minetest.is_dir method to the engine so that we dont need to use os.execute() to do the checks 09:36 TeTpaAka VanessaE, does #2878 fix the problem for you? 09:36 ShadowBot https://github.com/minetest/minetest/issues/2878 -- Fix missing check for 0 in craft replacements by TeTpaAka 09:37 VanessaE I have not tested it yet. 09:37 TeTpaAka Ok 09:38 VanessaE building it now. 09:41 VanessaE TeTpaAka: that seems to have fixed it. 09:41 TeTpaAka Good. 09:41 VanessaE from a quick test anyway 09:47 crazyR hmm, seems that in a sense is_dir already exists... i think someone needs to update the dev wiki methods section. minetest.get_dir_list() needs documenting 12:28 est31 pushing trivial fix #2878 in 15 minutes. hmmmm pinging you so that you know that I push 12:28 ShadowBot https://github.com/minetest/minetest/issues/2878 -- Fix missing check for 0 in craft replacements by TeTpaAka 12:29 est31 also after that I would like to go over all the for loops inside that file, and replace iterators over vectors with actual number counting 12:29 est31 look at that for loop, the iterator is even postincremented 12:29 est31 (the fix has nothing to do with that) 12:30 est31 I will push that second commit too in ~15 mins 12:46 est31 RealBadAngel, adding information to the nodedef whether its seamless or not is not the way to fix 12:46 est31 because there will be still a bias 12:51 est31 but RBA is right, thats how parallax mapping works 12:51 est31 look at this vid https://commons.wikimedia.org/w/index.php?title=File%3AParallax_mapping.ogg 12:52 est31 the texture is 12:52 est31 biased depending on the angle 12:55 est31 I wonder whether we can move it back without losing the effect 14:25 est31 ok last chance to stop the two commits https://github.com/est31/minetest/commit/286fb9ece884c72c56d98c6cc141f9e199fddb8f 14:25 est31 and https://github.com/est31/minetest/commit/87b9cdab07f70a7f2dd6531650242b0893eb5930 14:26 est31 5 minutes now 14:26 est31 (two hours later) 14:30 est31 pushed 14:34 est31 that one gets pushed in 15 minutes https://github.com/minetest/minetest/pull/2880 14:35 twoelk are you sure any of those it concerns can hear you? 14:36 est31 twoelk, should I ping hmmmm? 14:36 kilbith we don't really need him for everything 14:37 Calinou does Irrlicht support relief mapping, or better, tesselation? 14:37 kilbith this heavy stranglehold is kind of strangling 14:40 twoelk hehe, so there are more awake than I thought 14:55 RealBadAngel est31, relief mapping doesnt use bias, parallax mapping do 14:56 RealBadAngel Calinou, irrlicht supports material with parallax mapping, but 14:56 RealBadAngel if parallax then no shaders 14:56 est31 RealBadAngel, did I say something wrongß 14:56 est31 ? 14:56 est31 I've only said its paralax mapping 14:56 RealBadAngel we do have parallax mapping with slope information and relief mapping 14:57 RealBadAngel only the first use bias 14:57 est31 yes but why not reverse the bias, depending on angle 14:57 est31 that should be possible, no? 14:57 RealBadAngel im trying with no luck for now 14:58 RealBadAngel propably the solution could be inverted heightmap 14:58 RealBadAngel we will see 14:59 RealBadAngel btw, we do really need info on texture properties 14:59 est31 for what 14:59 RealBadAngel connected textures will also need that info 14:59 est31 well thats another discussion 14:59 RealBadAngel and i could use it to solve parallax problems 14:59 est31 no 15:00 est31 I mean what would you do? 15:00 RealBadAngel because now either non tiling textures do work or seamless 15:00 est31 disable parallax mapping for non tiling textures? 15:00 est31 but what about tiling textures? 15:00 RealBadAngel shaders do not know what kind of texture is given 15:00 est31 there its still broken 15:00 RealBadAngel no, not disable 15:00 RealBadAngel offset the effect 15:00 est31 why not always offset 15:01 RealBadAngel or as vanessae said, glue texture to non tiling edge 15:02 RealBadAngel whatever offset i tried theres some texture visible on opposite edge 15:02 RealBadAngel so that damn asphalt node with white border stripes has stripes on both sides 15:04 SopaXT Hi 15:04 SopaXT Where can I find the detailed specs on the game protocol 15:05 est31 SopaXT, in the code 15:05 SopaXT :< 15:05 est31 there are some pages on the wiki 15:05 est31 but they are mostly outdated 15:05 SopaXT Is it well documented? 15:05 est31 what do you want to find out 15:05 SopaXT I want to impoement a server 15:06 est31 ok 15:06 est31 you will have to either target one protocol version, or have compatibility with multiple ones 15:07 est31 but the protocol isnt documented at all at that detailed grade 15:07 est31 http://dev.minetest.net/Network_Protocol 15:16 SopaXT Ok, thank uou 15:37 Calinou custom servers often die early and have little support 15:42 AnotherBrick SopaXT : what's your reason for implementing new server instead of improving current one? 15:42 SopaXT My interediotism. 15:42 SopaXT Just for fun 15:44 AnotherBrick oh. good luck then :-) 15:45 est31 too bad hermatite uses the minecraft protocol, not the minetest protocol 15:45 est31 while we all know that its fundamentally flawed, because modding happens on the client, not the server 15:48 Calinou hermatite? 15:48 est31 man Calinou you know it 15:48 est31 http://hematite.piston.rs/ 15:48 Calinou oh right 15:48 Calinou that server thing is new 16:25 est31 the minecraft protocol isnt that bad after all 16:31 Calinou it's still TCP for a real time game 16:31 Calinou using UDP we can achieve lower pings, provided we do it right... I wish Minetest used enet from the beginning 16:32 est31 enet is no magic "super fast" formula 16:32 est31 it depends on the skills of the person who implements it 16:32 est31 uses* 16:44 est31 what i didnt understand is how they transmit y data for "chunks" 17:20 giacomo986 hello everyone, just downloaded minetest-master and compiled, it start, I create a new world but i can't enter in it, the client freeze when it is creating the server 17:20 giacomo986 version 0.4.12-dev 17:20 est31 giacomo986, there is a current regression, I'm working on it 17:21 giacomo986 ok, happy that you know it 17:21 est31 giacomo986, can you put your debug.txt to a pastebin? 17:21 giacomo986 sure 17:22 Sokomine giacomo986: yes, same problem here. debug.txt won't be particulary helpful i'm afraid... 17:22 Sokomine it's a problem of latest git head 17:22 est31 Sokomine, might or might not 17:23 est31 I have that gamedef bug 17:23 est31 and it does give a hang in "creating server" for singleplayer 17:24 giacomo986 http://pastebin.com/mmkR5Rwi 17:24 est31 ok 17:24 est31 there is an error, see? 17:24 est31 o sorry 17:24 est31 no error 17:25 giacomo986 i was trying some mods but it is the same without ods 17:25 giacomo986 the game freeze so it wont write anything on debug 17:26 Sokomine yes. the server does not fire up properly. calinou reported that as well 17:26 est31 its a regression since 8a85e5e5 17:26 Sokomine must have happened very recently, yes 17:27 giacomo986 I installed official build trough rmp and it work nicely 17:27 Sokomine yes. it's a problem with the very current version. slightly older builds will work fine 17:30 est31 seems to be my fuckup 17:30 est31 yes, its my fault 17:31 est31 fixing now 17:31 est31 (when I've found the cause) 17:31 giacomo986 wow, I did not expect this speed 17:38 est31 well I have the bug now 17:39 est31 pretty stupid from me 17:39 giacomo986 what was? 17:39 est31 dumb replacing 17:40 est31 what can possibly go wrong here https://github.com/minetest/minetest/commit/96989e0a6aa3ab069b5aeeab44a6280d6d51364a#diff-05fc0edba5ae83586559fef2fca70b47R244 17:40 est31 (and cause an endless loop) 17:40 hmmmm use size_t instead of std::vector::size_type 17:41 est31 why is that better 17:41 hmmmm well for one, it's shorter 17:41 est31 vector::size_type uses allocator::size_type 17:42 hmmmm this way you can fit all of it on the same line, reducing clutter and making a for loop and its semantics quickly noticable 17:42 hmmmm in practice everything is going to be an unsigned integral type of sufficient capacity. 17:43 hmmmm moreover, you're implicitly guaranteed through the standard that sizeof(size_t) >= sizeof(*::size_type) 17:43 est31 is it guaranteed size_t is always larger than allocator::size_type? 17:43 est31 lol 17:43 est31 ok 17:44 est31 pushing this now, ok hmmmm ?: https://github.com/est31/minetest/commit/64dc0532098ceeaa8a9de01984bd6a309578ec1e 17:44 hmmmm sure. 17:45 est31 pushed 17:45 hmmmm so 17:45 hmmmm on expressiveness 17:45 hmmmm more compact or expressive does not necessarily mean better 17:46 est31 then I'll use size_t in future because of shortness 17:46 hmmmm it is better when it enhances the ability of expressions to be quickly recognized in their form 17:46 hmmmm when you see for (int i = 0; i != blah; i++) { you instantaneously know what this form does 17:46 hmmmm when it's on several lines... you force the reader to actually look at the loop conditions and what not 17:46 est31 that has been my thought with using ::size_type instead of size_t 17:47 giacomo986 Can I try it now? 17:47 est31 giacomo986, bug fixed 17:48 giacomo986 nice, ty 17:48 hmmmm so, that commit you made that caused the regression 17:48 hmmmm who +1ed it? 17:50 est31 I have announced to do it in chat 17:50 est31 hours before 17:50 est31 even pinged you 17:50 hmmmm but... who +1ed it? 17:51 est31 no "I want to see it" from you 17:51 hmmmm that's a mistake that I would've caught 17:51 giacomo986 It works! 17:51 est31 nobody, it was triviality rule 17:52 hmmmm then maybe we should redefine "trivial" 17:52 hmmmm or remove it 17:52 hmmmm the "trivial" exception was put there to not totally piss off developers 17:53 est31 so you want to totally piss off devs? 17:53 est31 (using your words) 17:53 hmmmm yeah I think it's better than having that happen 17:53 hmmmm we need to accept the reality that the bleeding-edge devel version is what most people use 17:54 hmmmm and it needs a higher standard of stable 17:54 est31 I think even more people use the stable releases 17:54 hmmmm haven't we gathered statistics on this? 17:55 est31 but they complain less on irc 17:55 kilbith sfan5 has the stats 17:55 est31 ^ 17:55 hmmmm ahh good it's sfan 17:55 hmmmm i thought it was xyz 17:56 Sokomine it's ok if the bleeding edge version has trouble sometimes. irc is helpful to report such problems 17:57 paramat i'll push #2875 later 17:57 ShadowBot https://github.com/minetest/minetest/issues/2875 -- Mgv7: Lower base of mountain generation to -112 and define constant by paramat 17:57 hmmmm well 17:58 hmmmm paramat: if you find yourself making lots of commits that don't add functionality, but adjust parameters, maybe that should be configurable 17:59 paramat i did briefly think about making mountain base level configurable 17:59 paramat deoends on ocean depth 18:00 paramat i considered -112 deep enough to cover most custom noise parameters 18:00 VanessaE paramat: you neevr did add those acacia trees, did you? 18:00 VanessaE never* 18:00 paramat working on it 18:00 VanessaE ok. 18:00 VanessaE also: 18:01 VanessaE [off] check the pine trees code. it appears to be capable of griefing structures the way the default trees used to 18:01 paramat leaves texture needs redoing, i discovered the leaves are tiny 18:01 paramat you mean when a sapling grows? the lua code? 18:02 VanessaE [off] I found a place in one of my worlds where someone had planted one below a bridge, and the trunk grew through the bridge. 18:02 kilbith maybe using the L-system instead of a schematic could prevent that (?) 18:03 paramat l system can't reproduce the design of pine trees 18:03 kilbith VE uses for the moretrees's pine 18:03 kilbith uses the L-system* 18:03 VanessaE actually it can, but it's tricky to get it right and probably would not be too beneficial for that size of tree. 18:04 giacomo986 VanessaE, does your pipework mod works on 0.4.12? because I can't let it works 18:04 VanessaE giacomo986: that's not for this channel, but yes it should. 18:04 giacomo986 i'm sorry 18:05 Sokomine est31: thanks! the fix works fine. the worlds come up again 18:07 paramat yes i meant it can't reproduce the hardcoded design exactly 18:08 paramat just checked, lua pinetree growing doesn't check for air or ignore when placing trunk, will fix 18:09 sfan5 stats? 18:09 sfan5 hmmmm: stats about which versions are in use? 18:09 hmmmm yea 18:10 sfan5 http://meow.minetest.net/tmp/serverlist_stats_2015-03-14.txt 18:11 sfan5 the high amount of old version usage comes from the mobile stuff 18:11 hmmmm so we can't actually tell who's using a devel version or not 18:12 paramat so yes i'll keep mountain base level configurability in mind, i'm just considering the simplicity of setting a constant to a fairly deep value (-512 or whatever) being preferable to another parameter. for now i'll push this as is 18:12 sfan5 hmmmm: yes we do 18:12 sfan5 it's just not included in the stats 18:12 hmmmm why.. 18:13 sfan5 idk 18:13 hmmmm can we please start collecting devel versions 18:13 sfan5 189.144.161.188 - - [04/Jul/2015:18:13:20 +0000] "GET /list?proto_version_min=13&proto_version_max=24 HTTP/1.1" 200 147700 "-" "Minetest/0.4.11 (Linux/3.4.34-ecs-011 i686)" "-" 18:13 sfan5 thats devel, right? 18:13 hmmmm i guess not 18:14 sfan5 hm 18:14 sfan5 that can't be .11 18:14 sfan5 the proto args for the serverlist were added recently 18:14 sfan5 after the .12 release 18:15 sfan5 there is also "Minetest/0.4.12-dev (Linux/3.2.0-23-generic i686)" 18:16 sfan5 but i can only seem to find servers using that (POST /announce) 18:16 est31 the proto args were added before 0.4.12 18:16 sfan5 really? 18:17 est31 it is theoretically possible that they use a custom compiled version of a commit between the proto args commit and 0412 18:17 est31 they were my patch 18:17 est31 and now I think they can be removed 18:18 sfan5 either nobody is using .12-dev right now 18:18 sfan5 or something is broken 18:18 sfan5 but we *can* know which servers use -dev 18:18 sfan5 e.g. 18:18 sfan5 - - [04/Jul/2015:18:17:36 +0000] "POST /announce HTTP/1.1" 202 61 "-" "Minetest/0.4.12-dev-0b76e85 (Linux/3.2.0-4-amd64 x86_64)" "-" 18:19 Sokomine it took a while for .12 to show up even though it ought to have been .12 for quite a while now. perhaps that's it 18:32 est31 nno 4.12-dev at all? 18:32 est31 Ive thought this was fixed 18:36 est31 RealBadAngel, it seems only your last commit broke with opengl 2.1+? 18:36 est31 perhaps the state before that commit can be saved into the "simple shaders" directory? 18:36 sfan5 % tail -n10000 serverlist.log | grep "GET /list" | grep -- "-dev" | wc -l 18:36 sfan5 87 18:40 est31 hmmmm, I think we can lock up master by removing the triviality rule, and by requiring 3 core devs for patches a core dev submits, if we have enough devs around to do the checking 18:41 est31 sfan5, you have windows, no? 18:41 sfan5 no 18:41 sfan5 >by removing the triviality rule 18:41 sfan5 why 18:42 est31 hmmmm wants it, not me 18:46 est31 any core dev with windows? 18:46 kilbith est31, it's up to c55 to decide that 18:46 VanessaE blockmen, if you can contact him :P 18:48 hmmmm kilbith, what the hell is up with you 18:48 hmmmm do you have some kind of problem with me? 18:48 hmmmm who the fuck are you even? you're not a core developer, you're just a very opinionated guy who keeps causing trouble and PMing people about other people 18:48 kilbith yes, i have 18:48 hmmmm stop causing drama 18:48 est31 kilbith is pming alot yes 18:49 hmmmm and why is it up to c55 to decide anything 18:49 kilbith when you don't have any argument, you attack by ad hominem, it's usual 18:49 est31 agree with hmmmm here 18:49 hmmmm I am not having a debate with you at all. I'm calling you out on your shit 18:49 kilbith you're spreading disinformation about the rules whereas there no ambiguity, no grey zone 18:49 hmmmm ok chief. 18:49 kilbith the 1st time i recall you the rules, you attacked me in private 18:50 hmmmm is that so 18:50 kilbith for no reason 18:50 hmmmm I never PMed you once 18:50 kilbith remember when you asked if i was in love with nrz ? 18:50 hmmmm you always initiated PMs. I believe I could have responded once or twice, I try to ignore them though 18:50 hmmmm look 18:50 hmmmm if you're dissatisfied with minetest, you're totally free to fork your own 18:50 kilbith no wonder why you're not much appreciated here 18:51 sfan5 the fuck 18:51 sfan5 kilbith: [citation needed] 18:51 hmmmm kilbith, I think you're just looking for a villian 18:51 hmmmm this is software development for some video game, not a drama channel 18:51 kilbith you're insulting people, ashaming them, hating what they does, cheating the rules 18:52 est31 hmmmm is the most competent dev around the last time 18:52 kilbith i wonder how the rest of team can stand with such sour attitude 18:52 est31 (cant judge the others) 18:52 hmmmm :') thanks 18:52 hmmmm kilbith: because attitude frankly doesn't matter. code does 18:53 kilbith you have "your personal values and philosophy" but you don't even apply them to yourself at 1st instance 18:53 sfan5 kilbith: before you continue you hate train how about you give some evidence? 18:53 sfan5 your* 18:53 hmmmm maybe you're looking for a community like Rust's who is too scared of hurting peoples' feelings 18:53 kilbith sfan5, sure but i need time to gather all them properly 18:53 kilbith this is scattered on years scale 18:54 hmmmm this is stupid and doesn't belong in -dev 18:54 kilbith how many conflicts you brought there... 18:54 est31 hmmmm is a bit too much conservative the last days for me. 18:54 Sokomine hmmmm: sending pms is for most people the friendlier, less invasive way of contacting, especially if there is a problem. it is not intended to be an annoyance. it takes time for all of us to learn how to deal with each other, which preferences each person has 18:54 est31 there I agree 18:54 hmmmm est31: maybe it's a knee-jerk reaction to the last set of bugs 18:54 hmmmm minetest is currently a mess and I don't like seeing that 18:55 sfan5 kilbith: I'm not asking for you to write a book, it should take you about 5 minutes to find evidence for e.g. him cheating the rules 18:55 hmmmm Sokomine: kilbith does this all the time to talk about people behind their backs 18:55 kilbith sfan5: an example, did you already saw him trying to obtain another +1 when he merge his commits ? 18:55 hmmmm not quite the same as telling them about a bug, or giving a suggestion, or discussing how a feature should work 18:55 kilbith he want more restriction, except for him 18:56 sfan5 kilbith: no, I'd like a link to that 18:56 kilbith a link for what ? 18:56 sfan5 "him trying to obtain another +1 when he merge his commits" 18:56 hmmmm kilbith is so full of shit. he's the one propogating the debunked myth of the noise param configuration bug 18:56 Sokomine hmm. i've talked to kilbith a lot as well - about other things of mutual intrest, and havn't heared anything bad about other people 18:56 hmmmm again, stop adding noise to #-dev 18:57 hmmmm take it to #minetest or somewhere else 18:57 kilbith he merge directly after presenting the diff puched on his repo 18:57 Sokomine i don't think the discussion now will help any of the people involved. over here, it's far too hot to think 18:57 hmmmm kilbith: trivial change exception! 18:58 kilbith you have nastily broken the seed already 18:58 hmmmm see look! it's in the text added on some unprotected wiki page! therefore it must be legally binding 18:58 kilbith i can dig out many examples where people has complained about rudeness and insults here on IRC and the forum 18:58 sfan5 >i can dig out many examples 18:58 sfan5 ffs 18:58 hmmmm I don't care about what people think about my rudeness 18:58 sfan5 then do that 18:59 hmmmm there have been plenty of complaints before, they go directly to the virtual wastebin though. like I said, code flies, bullshit does not. 18:59 hmmmm /good/ code, at that. 18:59 kilbith and sfan5, you took his defense whenever hmmmm is attacked (last one : when rubenwardy has censured him for personal attack on the wiki) 19:00 hmmmm and this is the inappropriate venue for this sort of non-development discussion 19:00 hmmmm can it. 19:00 sfan5 how about you gather some examples instead of blindly accusing me 19:00 sfan5 all this makes me thing is that you aren't right at all 19:00 kilbith what about pestitidigator also ? 19:00 kilbith prestitigator* 19:01 kilbith yeah, ban me 19:02 hmmmm -> #minetest 19:02 kilbith i'd be interested to discuss of all that with c55 after a decent report about you 19:03 hmmmm > Chit-chat goes to #minetest. 19:03 kilbith i stop now 19:06 hmmmm a couple lines back and forth about a joke, meaningless banter, mod discussion, etc. is fine, continuously going on a tirade for a page or more about how I'm this horrible dictator is clearly not allowed here 19:15 paramat good grief O_o anyway now pushing 2875 19:23 paramat complete 19:29 est31 dammit internetz broke 19:29 est31 using mobile internet now 19:30 Calinou I have 20 GB of mobile data here :P 19:30 est31 its unlimited for me 19:30 est31 but after some limit, I get cut 19:30 est31 then I have to use 64 kbit 19:30 est31 works good for irc, and I can even push git commits 19:32 est31 hmmmm, you care about the code thats good. Also, ok that you are rude. but I have another opinion like you how contribution rules should be. 19:32 Calinou same here, it's throttled to 512 Kb/s after those 20 GB 19:33 hmmmm regardless of how many people are actually using dev vs. stable versions doesn't change my opinion much 19:33 hmmmm the fact is, if problems are allowed to exist to begin with, it usually takes a much longer amount of time 19:33 est31 512 kbs is quite decent 19:34 est31 I agree with you bugs must be fixed, and if they are serious enough, one even has to rollback. but I dont think we should close ourselves from development because of that 19:35 est31 devs have to admit bugs caused by them, and have to fix them 19:35 hmmmm maybe what if we have a technical debt 'credit card' 19:36 hmmmm you're allowed to rack up unfixed bugs, poor documentation, sloppy code style, etc. 19:36 hmmmm but as time goes on it accrues interest and becomes more important 19:36 hmmmm and there's a credit limit that you hit eventually 19:36 hmmmm at which point you can't commit more crap until the rest of your stuff is fixed 19:36 hmmmm ?? 19:37 * VanessaE imagines how many people will have exceeded said "credit" limit :P 19:37 est31 if you get neutral again if you fix your stuff, then yes 19:38 hmmmm yeah this is actually a decent sounding system 19:38 est31 also sloppy code style should weigh differently than bugs 19:38 hmmmm the frequent offenders would quickly learn, and those who produce great code to begin with wouldn't be penalized 19:38 hmmmm of course things have different weights 19:39 hmmmm we'll have to adopt that bug scoring system that CERT uses 19:39 est31 CERT? 19:39 hmmmm yea? 19:39 hmmmm or is it NIST 19:39 hmmmm i forget, but basically they have a scoring rubric for determining how critical a bug is 19:40 hmmmm or not, those are for security vulnerabilities only 19:40 hmmmm meh 19:41 est31 such a system would only work if somebody is maintaining it 19:41 hmmmm i could script it up in a web application 19:41 est31 also what to do if somebody else fixes your bug? 19:41 hmmmm i'd much rather work on minetest though 19:41 hmmmm hmmmm 19:41 est31 that sounds better 19:42 hmmmm good question 19:42 est31 if you give the score to both, you create a positive only system 19:43 hmmmm but it also encourages people to let their bug go and have somebody else do it 19:43 hmmmm and how does interest on technical debt work? improving some other part of the code as part of community service? or something? 19:43 est31 halve the effect? 19:44 VanessaE est31: give the person who broke it half the otherwise-negative value of the bug, and half the positive value to the person who fixed it. 19:44 VanessaE or however you want to rate the severity 19:44 est31 yes 19:44 VanessaE so that people who just let others fix their screwups will still decline and have to work harder to make up for it. 19:47 paramat meh waste of time, work on minetest instead 19:48 VanessaE that's the point, paramat. :) 20:19 paramat sfan5 here's the fix for VanessaE 's bug report game#554 20:20 ShadowBot https://github.com/minetest/minetest_game/issues/554 -- Default/trees: Add checks for air/ignore/needles to pinetree trunk by paramat 20:32 sfan5 paramat: if you place a pinetree node for y == yy would it be better to add a single data[a:index(x, y, z)] = c_pinetree call and let the loop start at y+1? 20:32 sfan5 s/if/since/ 20:32 est VanessaE, whats your creative spawn coords again? 20:32 sfan5 s/would/wouldn't/ 20:32 est also there is a minor street grief 20:33 est at -7.5,2,2.7 20:33 VanessaE est: -307,19,-426 20:33 est ahh there is it 20:33 est nice greeting "suspiciously large amount of objects detected" 20:34 paramat erm.. 20:34 sfan5 paramat: also a source code comment explaining why there needs to be a pinetree node at x, y, z always would be nice 20:35 VanessaE est: bingo, that's the start of the bug 20:35 kilbith bye 20:36 paramat perhaps better i'll work on it, and add comment 20:37 paramat if i do i will do that for appletree and jungletree too 20:38 paramat looks like your suggestion would be slightly faster, will work on it tomorrow 21:01 RealBadAngel est, 2.1 is not compliant with parallax code, my very last PR was about fixing bumpmapping issues for 2.1 (among other tweaks) 21:01 est RealBadAngel, then remove latest parallax pr 21:02 est errr take the code of the state before your parallax pr 21:06 RealBadAngel est, thats the idea for "lite" shaders 21:06 est ok 21:06 est well, either another dir, or an ifdef or whatever 21:06 RealBadAngel im think on shaders not just on/off, but off/lite/full 21:07 RealBadAngel another dir for sure 21:09 RealBadAngel thx to that i can achieve two things, have shaders for opengl2.1 safe, tested and not to be touched 21:10 RealBadAngel for full option i could drop compability with GLSL 1.2 and pick newer shader model version 21:10 est it should perhaps be called somehow else 21:10 est off/compat/full 21:11 RealBadAngel up to be decided 21:11 RealBadAngel just drawing a picture now 21:12 RealBadAngel keeping compability with nine yrs old shader model is pain in the ass, when developing modern effects 21:12 RealBadAngel this way i can definitely get rid of such problems 21:13 est we need compat for android too 21:14 RealBadAngel propably lite will do 21:15 RealBadAngel 2.1 is old enough for handheld to run imho 21:16 est right now its broken 21:16 RealBadAngel i will need to make some cleaning on my hdd and install x86 droid to test things 21:16 est even before your pr 21:16 RealBadAngel whats broken now? 21:16 est everything is red 21:16 RealBadAngel droid? 21:16 est yes 21:17 RealBadAngel never had a chance to test it or see logs on whats not workin 21:17 est #2502 21:17 ShadowBot https://github.com/minetest/minetest/issues/2502 -- Shaders on android produce wrong colors 21:18 RealBadAngel we shall change one thing about shaders 21:18 est using opengl es 1 21:18 RealBadAngel if failed to compile fall back to no shaders 21:18 est yea 21:18 RealBadAngel number of ways to report broken shaders (i mean red colors) is amazing ;) 21:19 est btw what happens when shaders fail to compile now? 21:20 RealBadAngel red eclipse ;) 21:20 RealBadAngel irrlicht materials are broken and theres no lighting 21:20 RealBadAngel mt expects lighting to be done by shaders 21:20 est does the shader compiler somehow output compilation errors? 21:21 est perhaps those should be logged 21:21 RealBadAngel shader.cpp 21:21 RealBadAngel there on error (of any kind we have to set a fallback) 21:22 RealBadAngel i think with a config setting for it 21:23 RealBadAngel otherwise testing shaders would be a pain in the ass 21:24 RealBadAngel enable_shaders_fallback = default true, and then enable_shaders should be set to false 21:25 RealBadAngel developing shaders is quiting to menu and launching again, so fallback in my case would be just annoying 21:26 est the problem is it doesnt even print a single error to errorstream 21:26 est how can you find out you had an error? 21:26 RealBadAngel it does 21:27 RealBadAngel it says the reason, and the line where it occures 21:27 RealBadAngel it does it even many times, since shaders are instanced for each drawtype 21:27 est so then shader compilation is successful 21:28 est because I see no reason 21:28 est or is it not in errorstream? 21:28 RealBadAngel its dumped to console and logged 21:28 est where? 21:29 RealBadAngel edit your shader and instert there some obvious fail 21:29 RealBadAngel you will see the effects 21:29 est where are shaders stored again 21:29 RealBadAngel client/shaders 21:30 RealBadAngel btw line reported to have error != line number in shader source file 21:31 RealBadAngel we do add shitload of lines in header runtime 21:44 VanessaE RealBadAngel: didn't I suggest multiple levels of shader capability some tiome ago? :P 21:45 VanessaE -o 21:45 VanessaE good idea to auto-fallback to no shaders though - but perhaps fall back to "lite" first, then to none if that fails to compile. 21:46 RealBadAngel first lets make a fallback at all, then we can modify it with levels 21:48 RealBadAngel such multilevel fallback would be tricky anyway 21:49 VanessaE tricky, but useful in some cases I think 21:49 RealBadAngel it would require user interaction - restarting game all over to get working one 21:49 VanessaE why? can't you just auto-switch? 21:49 RealBadAngel and even then it wont mean user will get useable shader 21:49 VanessaE "just" :P 21:50 RealBadAngel from the issue: "Nope, no wrong colors. Although I did only check the shaders button. Bumpmapping makes minetest unplayable.: 21:51 RealBadAngel so propably droid shaders should be just lighting code and maybe waving stuff (its not that expensive effect) 21:51 VanessaE right 21:52 VanessaE I think that's roughly what they had around 0.4.10 anyway 21:52 est RealBadAngel, I can give you actual data after I've worked through millions of Loaded cached media: 55085959931f72f6662290664e6fd00bae681a47 "default_tool_steelaxe.png" likes 21:52 est lines* 21:54 RealBadAngel est, http://pastie.org/10273334 21:54 RealBadAngel this is fail output 21:54 RealBadAngel with default log level 22:00 RealBadAngel lol, i was afraid for relief mapping being too expensive with loop and 24 iterations 22:01 RealBadAngel look at piece of code from SEUS shaders: 22:01 RealBadAngel for (int i = 0; sampleHeight < pCoord.z && i < 240; ++i) 22:02 est RealBadAngel, dunno if that "Irrlicht log:" still gets spit out by android 22:02 RealBadAngel est, but definitely you got error in the engine 22:03 RealBadAngel https://github.com/minetest/minetest/blob/master/src/shader.cpp#L803 22:03 RealBadAngel heres the place for fallback code 22:03 RealBadAngel lemme test it how it could work 22:07 est hrmm 22:08 RealBadAngel putting there g_settings->setBool("enable_shaders", false); is half the job 22:09 est cant you disable them inside the scene? 22:09 RealBadAngel when we get there we already decided to assign another material to tile 22:10 * est wonders whats still against merging Tesseract's logging pr 22:10 est well a rebase is needed 22:11 est and does android really fail to compile or is it the buildbot? 22:55 est31 ok what about this commit https://github.com/est31/minetest/commit/0f04f0e36a37527003e491c316f27a476974b08f 22:55 est31 hmmmm, ^ 22:55 est31 its part of SN's logging cleanup, and I think we can push it already now 22:55 est31 also I judge it as "trivial" 23:06 est31 second commit https://github.com/est31/minetest/commit/99699c6e8d3932796bee8b915f596ae4c191fee3 23:07 est31 so is anybody around to approve? 23:07 est31 hmmmm, ? 23:07 est31 PUSHING IN ONE HOUR 23:11 kahrl est31: why does a function that deals with utf-8 use CP_ACP instead of CP_UTF8? 23:11 kahrl even if it's correct, this at least needs a comment imo 23:11 est31 well, one of the formats is CP_ACP, the other one is unicode 23:11 est31 thats how I think these functions work 23:12 AnotherBrick https://msdn.microsoft.com/en-us/library/bb202786.aspx 23:14 kahrl the documentation is kinda unclear but I'm pretty confident that the codepage argument specifies the encoding of the multibyte string 23:15 kahrl since wide strings are always encoded in the windows variant of UTF-16/UCS-2, so there's nothing to specify about it 23:15 est31 MultiByteToWideChar: "Maps a character string to a UTF-16 (wide character) string. The character string is not necessarily from a multibyte character set." 23:15 est31 first sentence 23:16 est31 hrmm 23:16 kahrl also CP_ACP can have different meanings on different systems 23:16 est31 thats the point isnt it 23:16 kahrl depending on what language version of windows was installed 23:17 kahrl no, we always want to convert from/to utf8 23:17 est31 yes 23:18 est31 AnotherBrick, what is the output for ffЖff when setting CP_UTF8 23:18 kahrl I'm a bit out of the loop, what issue # is this commit fixing? 23:19 est31 #2847 23:19 ShadowBot https://github.com/minetest/minetest/issues/2847 -- Font characters broken 23:19 AnotherBrick it is 6666d0966666 23:19 VanessaE est31: I can confirm that btw 23:20 VanessaE someone was writing in an asian language on my server the other day. came through fine to the IRC bot, but in-game it was just blanks. 23:21 kahrl clearly some part of minetest is storing string in the local multibyte encoding and some other part is expecting it to be utf8 23:21 kahrl the string* 23:21 est31 chat is sent in wchar 23:21 kahrl wish I had access to a working build on windows so I could trace what happens to the string 23:22 est31 perhaps AnotherBrick can help 23:22 kahrl chat should be sent in utf8, but idk how to easily change that 23:23 est31 we will have to make a new protocol version, and make a new chat packet 23:23 est31 for legacy clients, the server converts to or from wchar 23:24 AnotherBrick idk how chat is send, but it's trying to be displayed as narrow string 23:24 kahrl does the problem only affect surrogates? 23:24 kahrl if so, it might be due to the differences between linux and windows in how wide strings are handled 23:25 AnotherBrick with CP_UTF8 it correctly converts to utf-8 and than display some obscure character 23:26 AnotherBrick sooo, though i see my cyrillic perfectly with CP_ACP i have no idea what other players will see 23:27 est31 AnotherBrick, can you try to make CP_ACP for one function and UTF8 for the other, to find out which function has the "wrong behaviour"?? 23:28 AnotherBrick you mean one utf8_to_wide and second utf8_to_wide? 23:28 est31 no, utf8_to_wide vs wide_to_utf8 23:29 est31 or has only utf8_to_wide the bug? 23:29 kahrl I don't think either has "the bug" 23:29 kahrl the bug is that something is not using them properly 23:29 AnotherBrick ok, i'll check it... 23:30 VanessaE bbl 23:31 est31 kahrl, so the bug is only on windows then? 23:31 est31 or perhaps its on linux too, but there the functions are wrong? 23:31 kahrl ah sorry, I didn't actually check it 23:32 kahrl I assumed so because the fix only modifies the windows implementation 23:32 est31 no the bug is only on windows 23:32 est31 works great here on linux 23:32 est31 (if not, I wouldnt have commited the pr) 23:36 AnotherBrick hmm.. utf8_to_wide isn't called at all for chat strings. or i'm sleeping right now? 23:36 est31 and wide_to_utf8? 23:37 est31 well its called at some places 23:37 est31 mostly inside guiformspecmenu.cpp 23:37 AnotherBrick and wide_to_utf8 is being called and produces results which i already provided 23:38 est31 so, can you test whether its the "fault" of wide_to_utf8 or utf8_to_wide? 23:39 est31 just change the codepages 23:39 est31 first one, then test, then the other one then test 23:39 est31 then change both 23:40 est31 if one of the first two tests succeeded, its only the fault of one function 23:40 kahrl have you tried changing the call in game.cpp:186 to utf8_to_wide? 23:41 AnotherBrick kahrl : that makes sense, i'll try 23:42 hmmmm est31: 0f04f0e looks good to me, things like 99699c6 where the code change is trivial but the effect isn't obvious needs to be verified through actual testing, which it seems like you did 23:44 kahrl hmmmm: about 99699c6, see above 23:44 est31 pushing this then: https://github.com/est31/minetest/commit/a93838707a9aab104c20b6a971d8c622184a508f 23:44 est31 the other commit will have to wait until we've found out what to fix 23:44 AnotherBrick kahrl i love you. that fuckin works!! 23:44 kahrl nice :) 23:45 est31 AnotherBrick, now I still wonder, where the problem occured 23:45 AnotherBrick so, CP_UTF8 is ok, everything in string.cpp is fine, problem is in game.cpp 23:46 AnotherBrick just wrong function for displaying chat strings on client side 23:46 est31 its not at all just displaying 23:46 est31 its also sending over the protocol 23:47 AnotherBrick yeah, but for sending only wide_to_utf8 is used 23:47 AnotherBrick and it's fine 23:47 est31 not yet 23:48 est31 also we havent figured out how to fix fgettext yet 23:48 est31 eg like here https://github.com/minetest/minetest/issues/2847#issuecomment-116172443 23:50 AnotherBrick mmm can't help here. i compiled without gettext 23:51 kahrl est31: I tried to summarize what happens: https://gist.github.com/kahrl/50669c6beaf57fe1796e 23:52 kahrl if my reading of the code is correct 23:53 est31 kahrl, makes sense, 23:54 kahrl I think we should strive to eventually get rid of most narrow strings, and use utf8 everywhere instead 23:54 est31 ^ 23:54 kahrl so that errors like this can't happen 23:54 est31 yea long term goal is to deprecate wide_to_narrow and narrow_* 23:57 est31 I'm working on sending chat message as utf8 now 23:58 est31 what do you think, should we use new opcodes, or make it only dependend on proto version? 23:58 est31 second would involve less "formality" around network code 23:59 est31 e.g. you have to add handler methods, add the opcode to various tables etc