Time Nick Message 00:00 VanessaE_ that looks pretty good actually 00:00 VanessaE_ screenshot coming.. 00:00 VanessaE_ http://digitalaudioconcepts.com/vanessa/hobbies/minetest/screenshots/Screenshot%20-%2001222015%20-%2006%3a58%3a10%20PM.png 00:01 VanessaE_ left is HEAD, right is with that last patch (only) 00:02 VanessaE_ inventory looks good too 00:06 sofar Can you account for sun position? 00:07 VanessaE_ that's planned some time in the future 00:07 sofar excellent... little things like that really help with atmosphere 00:09 kahrl #2185 00:09 ShadowBot https://github.com/minetest/minetest/issues/2185 -- Fix NDT_GLASSLIKE normals by kahrl 00:23 VanessaE_ kahrl: there is one minor side effect: glass isn't smoothly-lit 00:24 VanessaE_ but that wasn't evident before I guess (or I never noticed it) 00:26 VanessaE_ btw, client is now doing 40 fps at 64m view range at VE-Creative spawn. best I've seen it in ages. 00:31 kahrl it isn't smooth in master, so no regression there at least :) 00:31 kahrl not that I want to complain, but my fix shouldn't improve performance :) 00:32 VanessaE_ actually 00:32 VanessaE_ I think it did 00:33 kahrl hmm, weird 00:33 VanessaE_ at least, there's no obvious reason for the higher fps/view range. maybe your patch combined with all the recent performance tuning just "worked">? 00:33 * VanessaE_ shrugs 00:36 VanessaE doesn't matter, I'm not one to complain either :) 01:50 sofar Assignment to undeclared global "srcstack" inside a function at .../minetest/bin/../games/minimal/mods/default/init.lua:1368. 01:50 sofar fresh git (10 minute back or so) 01:50 sofar placed an item in an oven (no mods) 03:24 sofar what is "BS"? some universal BlockSize factor? 03:25 VanessaE some of us interpret it more...literally ;) but yeah more or less 03:27 Zeno` heh 03:27 * sofar calls BS 03:27 Zeno` I see /* BS */ at the top of many source files (often next to an include) and begin to wonder 03:28 Zeno` How I interpret it depends on the mood I'm in :) 03:39 sofar I'm not seeing the player acceleration in the player object, is there a way to determine the player acceleration accurately somehow? 04:44 est31 Hello, after logging in to VE-S, I get ERROR[main]: generateImage(): Could not load image "door_steel.png" while building texture 04:44 est31 others report it too. 04:44 VanessaE ^^^ note that the textures exist and work fine, the problem he's having is client-side. 04:45 VanessaE I believe this is the same issue causing MichaelEh's client to crash (just a different texture being references) 04:45 VanessaE d* 04:45 est31 also accompanied by: ERROR[main]: generateImage(): Creating a dummy image for "door_steel.png" 04:51 VanessaE what's the story on https://github.com/minetest/minetest/issues/1425 > 04:51 VanessaE I believe this to be the cause of est31's issue 06:20 Gregorycu Is someone able to check out #2173 06:20 ShadowBot https://github.com/minetest/minetest/issues/2173 -- Speed up Profiler::avg by a factor of 10 by gregorycu 10:13 gregorycu What's the haps? 11:36 Zeno` looking at just the code, #2156 looks fine 11:36 ShadowBot https://github.com/minetest/minetest/issues/2156 -- Optimise MapBlockMesh functions by gregorycu 11:36 gregorycu Thank you Zeno 11:39 gregorycu Zeno, there is a settings callback system 13:38 gregorycu Why is the single player mode performance so terrible compared to client/server ? 14:58 T4im hm recently I noticed that the meshupdatethread noticeably spikes in cpu when moving through the world (but calms down fast again when stopping) 15:02 Nerz loading/creating area is intensive, yeah 15:02 T4im feels like it increased in recent commits though 15:05 T4im oh wait… that could simply be the optimization parameter changes 15:05 * T4im facepalms 15:07 Nerz :p 15:20 Nerz Zeno` ? 15:23 Nerz #2119 15:23 ShadowBot https://github.com/minetest/minetest/issues/2119 -- Rewrite client/server packet handling by nerzhul 15:26 Zeno` I've experienced no problems with #2119 on my server 15:26 ShadowBot https://github.com/minetest/minetest/issues/2119 -- Rewrite client/server packet handling by nerzhul 15:26 Zeno` I don't think VE has either 15:26 Zeno` hmmmm might merge it 15:29 kilbith why not you Zeno` ? 15:31 hmmmm first of all 15:31 hmmmm is it ready 15:31 hmmmm second of all, i've been hearing reports of network lag 15:32 hmmmm i'm not sure if adding a huge refactor is the best idea at this time, especially since we want to release 0.4.12 soon... this would push a release much further off 15:32 Zeno` network lag? 15:33 hmmmm in the meantime i doubt this pull request is as susceptible to rot since people don't modify the networking as often 15:33 Zeno` what network lag? 15:33 hmmmm the lag spikes wayword_one and others complained about 15:33 Zeno` Did you look at the profile he made? 15:34 hmmmm no, i did not 15:34 Zeno` That's not network lag, that's thread lag 15:34 Zeno` "lag" 15:34 hmmmm what's lagging in the thread?? 15:34 hmmmm anyway I sort of suspected that 15:34 Zeno` 20% of the CPU time is spent switching threads 15:35 Zeno` which is kind of crazy 15:35 hmmmm that doesn't sound right 15:35 hmmmm as in doesn't sound correct 15:35 Zeno` *shrug* 15:35 twoelk just for the records, my windows clients (krocks builds and related) still seem to crash on VE-C server. The amount of ram seems not relevant. All I need to to is fly around and load lots of blocks. The client will crash, disconnect but not load main menu and stay grey and unresponsive instead. And no I havn't much more info to share, am still testing local environmental possabilities 15:35 Zeno` the simple_single_player server does not belong in a thread (IMO) 15:35 Zeno` it should be a process 15:36 Zeno` but look at the valgrind output he gave us 15:37 hmmmm i'll have to check it out 15:37 hmmmm not irght now though i have work 15:37 Zeno` please do :) It's interesting 15:38 Zeno` and I've since confirmed it with my own tests 15:39 Zeno` twoelk, have you got mesh cache set to true? 15:41 * twoelk lets the virus scanner do it's reports and has a look at is minetest folders 15:41 twoelk +h 15:42 twoelk probably whatever default was 15:43 Zeno` ok; make sure it is false 15:43 Zeno` hmmmm, I will be merging it later today 15:44 VanessaE hmmmm: I'd like to recommend that ^^^^ cache be disabled by default and the variable name changed so that it can't possibly be turned back on by old configs. it costs 800 MB of RAM in dreambuilder for no tangible benefit. 15:44 Zeno` because AFAICS: a) the code is correct; b) it's faster; c) it's been tested for over a week; d) it does not seem to have anything to do with Wayward_One's problem 15:44 kilbith it happened to me (client crashes after 20 min) after 10-20 min spent on VE-C since the mesh nodes has been implemented 15:45 VanessaE kilbith, twoelk: minetest.conf, enable_mesh_cache = false 15:45 T4im the meshnodeupdate thread does seem to be quite cpu intensive indeed 15:45 VanessaE that'll at least extend your playing time anyway 15:45 kilbith anyways, this server is unplayable for the fairly old computers 15:45 Zeno` VanessaE, it should be disabled by default (it was at first, but maybe that's changed?) 15:46 VanessaE Zeno`: it didn't even *exist* at first, as far as I'm aware 15:46 Zeno` I know it didn't exist at first :P 15:46 Zeno` it was added during 0.4.10 dev 15:47 Zeno` VanessaE, have you had problems with #2119? 15:47 ShadowBot https://github.com/minetest/minetest/issues/2119 -- Rewrite client/server packet handling by nerzhul 15:47 Zeno` I haven't on MTZ 15:48 VanessaE nope, no problems that I can trace back to that 15:49 * twoelk notices he ripped apart his last installation a little too early 15:49 Nerz good new for #2119 15:49 ShadowBot https://github.com/minetest/minetest/issues/2119 -- Rewrite client/server packet handling by nerzhul 15:49 Nerz can this be merge before 23h00 GMT+1 ? :D 15:49 Zeno` What time is it now? 15:50 Nerz 16h55 15:50 Zeno` Personally I can't see why not 15:51 Brains If you were wanting to do a quick release, a rework of the network code would be contraindicated... 15:51 VanessaE I would tend to agree with Brains and hmmmm.. 15:51 kilbith has been extensively tested and it's just code design... 15:51 Nerz Brains, rework was tested since 1 week on not owned-by-me servers and no problem was reported 15:52 Nerz The problem are not in my code, the problem are on the protocol, and i haven't changed protocol 15:52 Brains Which is lovely... And I suppose similar things could be said about the last release. 15:53 Brains To be clear, pretty much everybody here uses a dev build so hiccups don't affect as much... But a release isn't for us. So there is a substantially higher bar. Do the release, then merge the code. "Everybody" is happy then. 15:54 Zeno` If *merge* it though there is a condition 15:54 Zeno` get rid of all those stupid exceptions in connection 15:54 Zeno` :) 15:54 Nerz personnaly i use my test server with 10 users and there are no problem 15:54 Zeno` not now, but in a soonish PR 15:54 Nerz i don't manage connection at this time zeno 15:54 Nerz but yes 15:54 Nerz i agree with you, and i'll discuss with sapier for this 15:55 Nerz next PR is focus on packet writing, because this PR is only rework the packet handling/reading 15:55 Zeno` k 15:55 Zeno` I'm on a mission to get rid of all the abused exception handling hehe 15:55 Nerz on third PR i think, i will rework the connection between packets and connection, because there are many buffer/unbuffer to convert data 15:56 Nerz but at this time i cannot remove these because old code use it :D 15:56 Nerz i need to change building, see you soon 15:58 hmmmm Zeno`: what?? 15:58 hmmmm merging what, the network rewrite, right before a release? 15:58 hmmmm exactly after I said no? 15:59 hmmmm code like that is going to cause problems 15:59 hmmmm after 0.4.12, I'm all for it 15:59 hmmmm but right now, please spare all of us.. 16:15 Nerz back 16:16 Nerz Zeno` , i agree to fight with exceptions :D 16:21 twoelk that setting did indeed help. I flew around for 20 minutes constantly loading new blocks until it crashed 16:22 twoelk so for normal playing it should be ok 16:23 twoelk btw, tacking screenshots seems to randomly trigger the same effect 16:25 twoelk 17:19:50: ERROR[main]: Some exception: "bad allocation" 16:25 twoelk In trans_func. 16:25 twoelk Access violation at 5C747365 write?=8 address=1551135589 16:25 twoelk In trans_func. 16:25 twoelk Access violation at 5C747365 write?=8 address=1551135589 16:25 twoelk 17:19:51: ERROR[main]: Some exception: "Access violation" 16:26 twoelk last message before crash 16:27 VanessaE obviously another memory leak 16:29 twoelk the message Acces violation is shown in chat, a few seconds later it seems blocks will not load and there is no interact. it takes a while until the client crashes after that. 16:30 twoelk the fovorite server file is not updated to the last server visited 16:31 twoelk and the config also seems to be not updated 16:32 twoelk the copy to local db seems ok though 16:41 twoelk according to irc messages it took about two minutes from the not loading of blocks anymore until the client finally crashed, no debug info is written after blocks failed to load. moving is normal during the dead time and allready loaded blocks are rendered correctly 16:44 Nerz erf 16:47 Zeno` hmmmm, ? 16:47 Zeno` what release 16:47 twoelk actually it is pretty similar to the lost-connection timeout except that the client hangs on it's way back to the main menu 16:47 Zeno` hmmmm, and why is you saying no stronger than me saying yes? 16:47 Zeno` that doesn't make sense 16:47 shadowzone Zeno`, we seem to have a problem. 16:48 hmmmm because one decision ruins the 0.4.12 release and the other does not 16:48 Zeno` hmmmm, what release? 16:48 hmmmm the 0.4.12 release 16:48 hmmmm in fact, weren't YOU the one to propose it? 16:48 Zeno` ages ago, yeah 16:48 hmmmm and we're going forward with it! 16:49 hmmmm are you saying you want to take it back? 16:49 Zeno` right! 16:49 Zeno` where is this mentioned? :( 16:49 Brains On the mailing list... 16:49 hmmmm i thought it reached a significant consensus 16:50 * Zeno` checks his pockets 16:50 Zeno` nope, nothing here 16:50 hmmmm alright whatever. *right now* do you agree with making an 0.4.12 release within a week?? 16:50 * Zeno` looks at the topic... nope, nothing there 16:50 Zeno` yes I would agree to that 16:50 hmmmm okay 16:50 hmmmm in fact I think I mentioned this last night 16:50 Zeno` so, in that case I won't merge 16:51 hmmmm no new features were added since then 16:51 Zeno` I wasn't here last night, I was in hospital 16:51 hmmmm Zeno`: a veto should override approval 16:51 Zeno` hmmmm, not if it's only one person 16:51 hmmmm what's the sample size though. 16:51 shadowzone I vote yes! 16:52 Zeno` anyway, I was in hospital after I got attacked by a crocodile so I had no idea there was a release planned for this wekk 16:52 Zeno` week* 16:52 hmmmm lol believable story 16:52 hmmmm what a crock :0 16:52 shadowzone O_O 16:52 Zeno` true fact 16:52 Zeno` look through the logs 16:52 shadowzone Did Krock get you? 16:53 shadowzone Zeno`, by the way, MTZ-Basic crashed. 16:53 Zeno` anyway I didn't know about an imminent release :( 16:54 Zeno` shadowzone, is deezl or CWz online? 16:54 Zeno` shadowzone, I can't ssh to it from windows 16:54 hmmmm 1 week sounds fair enough 16:54 shadowzone I could, and no. 16:54 shadowzone Yep 16:54 hmmmm note: after this release we switch to the new release model 16:54 hmmmm we NEED to for everyone's sanity 16:54 VanessaE what version number will it start at? 16:55 hmmmm I vote for 1.0.0 16:55 VanessaE I vote for 5.0 16:55 Zeno` yeah 1.0.0 is fine with me 16:55 shadowzone I vote for 1.0.0 16:55 shadowzone My inner 1st grader comes out saying "0.X, 1.X" 16:56 T4im vote for 5.0 16:56 Zeno` 5.0 or 5.0.0? 16:56 Zeno` there cannot be 5.0 16:57 Nerz no 16:57 Nerz you cannot improve it 16:57 T4im 5.0.0, 5.0 being the shortform omitting the "patch" field 16:57 Nerz the network protocol is purely shit 16:57 Nerz if you need to increment the version like this, wait the network protocol to be rewritten 16:57 T4im fair enough actually for any major increment ^^ 16:57 Nerz there are some ugly things, like a packet with U16 ID but the answer is U32 ID 16:57 * twoelk mumbles for 0.5.0 16:58 Zeno` Nerz, what we're trying to do is get a semantic versioning system in place (or something that at least resembles it) 16:58 Nerz or you have PASSWORD in a STATIC 28 CHAR instead of a dynamic packet sized string 16:58 Zeno` 5.0 just seems silly to me 16:58 hmmmm Nerz, that's because passwords are sent as hashes, are they not? 16:59 shadowzone Zeno`, just checked, deezl is not online, CWz is but away from keyboard. 16:59 Nerz username is a STATIC CHAR 20 16:59 Nerz it's not a hash :p 16:59 T4im zeno: many of us considered minetest at the version "4.11" at the moment… the 0 never was taken seriously 16:59 Zeno` it's actually malloc char 20 iirc 16:59 Nerz it's ugly 16:59 T4im with such a mental picture it stops looking silly :) 16:59 T4im and the other version starts to 16:59 Nerz std::string powa 17:00 Zeno` ok, 5 17:00 Zeno` whatever 17:00 Zeno` as long as it's consistent from then on :) 17:00 T4im don't you guys have someone in charge of public relations? I think it should be that persons call in the end 17:00 Nerz how can we announce feature freeze, there are so many PR and lags :s 5.0 will announce a major thing for users, and it's not 17:00 twoelk this will look stupid in history, people will think they missed something 17:00 Zeno` I am in charge of PR because I am fucking polite 17:00 T4im hehe 17:01 Zeno` i mean 17:01 Zeno` err 17:01 Zeno` nvm 17:02 Nerz increase major version mean major features, there are no major features here :) and sapier & me need to break protocol , a major progression is needed for that :p 17:02 Zeno` hmmmm, seriously... next time let me know if there is a planned release, please? 17:02 Zeno` or somebody at least 17:02 hmmmm it was planned yesterday 17:03 hmmmm you weren't around the entire day so... 17:03 Zeno` I wasn't on IRC yesterday, I was in hospital 17:03 hmmmm yes 17:03 T4im oh :( you're ok? 17:03 VanessaE 5.0(.0) because every idiot user (as opposed to intelligent ones) keeps calling 0.4.x just 4.x. this will avoid any chance of confusing it in the future. 17:03 Zeno` so, after this release maybe there needs to be a protocol for that as well 17:03 hmmmm yes 17:03 hmmmm let's get a mailing list and announce things like these on it 17:04 shadowzone Now I see the logic in 5.0, I'm slow. 17:04 hmmmm like every other open source project ever 17:04 hmmmm mailing lists are more for developers, forums are targeted toward users 17:05 Nerz the prefix 0. is pretty good, this tell that the engine is not ready, and we all think it's the case 17:05 * Brains sees the logic in 5.0, he just thinks it is a mistake. *shrug* 17:05 twoelk My unrelevant idea is still to make next release 0.5.0 as it is understandible and from there fast forward to 1.0.0 within months with more frequent releases. 17:05 Calinou “is not ready” BS 17:05 Calinou thank you 17:05 VanessaE twoelk: then users will be calling it 5.0, 1.0, 4.12..... which one's the latest? 17:06 VanessaE (of course we know, but they don't) 17:06 Nerz the course to the upper number is stupid, we are neither firefox nor google chrome 17:06 Nerz and there is no marketing to do :p 17:06 Calinou if you are going to start a new version scheme, I suggest starting over to 1.0 17:07 Nerz +1 17:07 Calinou so next release would be 1.0.0 17:07 Brains VanessaE: If you are a "modern" user and can't tell that 0.4.11 is before 0.5.0, then I'm not sure how much hand holding should be invested in you... On the other hand, projects that do a 5.0 release from 0.4.11 look just plain silly and obviously doing a PR thing. 17:07 Calinou this is what Warsow did 17:07 Nerz but after 0.4.12 17:07 Nerz 0.4.12 for fixes, 1.0.0 big changes 17:07 VanessaE Brains: the problem is the proposed 1.0 - users call 0.4.x just 4.x. now we will be giving them a 1.x. 17:08 hmmmm Nerz: why do toclientpacket.cpp and toserverpacket.cpp only contain a constructor? 17:08 Nerz they are removed 17:08 Nerz in the next PR 17:08 Brains VanessaE: If the messaging on the website and engine itself is "0.4.11" and "0.5.0" or "1.0.0", then it is about as plain as can be. You can only go so far with beating people over the head with obviousness. 17:08 hmmmm next PR?? 17:08 Calinou only really unknowledgeable users call it 4.x 17:08 Calinou and they should be proven wrong 17:08 Nerz ofc hmmmm 17:08 Zeno` we could make it 0.5.0.0 17:09 twoelk the next highest logical step is 0.5.0 in my oppinion. from there on the pacing can be changed to second place instead of now third and when 1.0.0 is reached things get perfect - version number wise 17:09 Nerz #2119 is a read rewrite 17:09 Calinou please not 4 numbers, Zeno` 17:09 ShadowBot https://github.com/minetest/minetest/issues/2119 -- Rewrite client/server packet handling by nerzhul 17:09 shadowzone Zeno`, too many 0's 17:09 VanessaE Brains: but matter how obvious that leading zero is and how many times we correct them, they keep omitting it. 17:09 Calinou it's Freeminer silliness 17:09 Zeno` :D 17:09 hmmmm well, okay... 17:09 Nerz hmmmm next PR will be rewriting write on client side 17:09 Zeno` the leading 0 can be implied 17:09 est31 VanessaE: you can add a ~version to the bot in your server's irc channel 17:09 Nerz hmmmm : https://github.com/nerzhul/minetest/commits/packet-handling_2 17:09 Calinou or we can make the next version be 0.5.0, that makes sense 17:09 Zeno` does anyone want me to revert my last merge since I didn't know that there was a release imminent? 17:10 * Brains has no vote but suggests 0.5.0. bbiab 17:10 VanessaE Zeno`: nah, too late now 17:10 Zeno` can't remember what it was... something SN did 17:10 hmmmm send real port to server list? 17:10 Zeno` yeah 17:10 hmmmm it's fine... besides, how could that break 17:10 Zeno` yep, ok, I'll leave it 17:11 Zeno` I vote for 1.0.0 or 5.0.0 17:11 Zeno` heh :) 17:12 shadowzone I vote for 5.0.0. REASON: My brothers and sister called it "Minetest 4.11" not "Minetest 0.4.11" only if I correct them. 17:12 Zeno` shadowzone's brothers have spoken! Case closed 17:12 shadowzone Lol 17:13 * Zeno` bangs his gavel 17:13 Zeno` next! 17:13 shadowzone Wait...What's next? 17:14 Zeno` nothing is next apart from a short wait 17:14 Zeno` next is the release 17:15 * twoelk thinks 0.5.0 would be nice version number for announcing officiall android support at a somewhat higher level than before 17:15 Zeno` twoelk, 0.5.0 is no good because it won't accommodate semantic versioning 17:16 Zeno` (nor backporting of bugfixes) 17:16 Zeno` so, really, it's 1.0.0 or 5.0.0 (IMO) 17:17 twoelk then there is no choice but 1.0.0 no matter what achievement qualifies such a version jump 17:17 Calinou we don't need to accomodate semver for now, perhaps 17:17 Zeno` Calinou, I think we do 17:17 Calinou then go for 1.0.0 17:17 Calinou but be sure to fix a lot of bugs and add at least one major feature, to not deceive users like Linux did :P 17:17 * twoelk is still not convinced though 17:18 Zeno` quite honestly 1.0.0 should really have been the version number before scripting was added 17:18 Zeno` but, can't change history 17:19 T4im git-filter-branch hmmm 17:19 twoelk a job for the timecops 17:19 twoelk where is the doctor when one needs him? 17:20 DoctorWho Gah! 17:21 * twoelk imagines a dessinterested stupid face to that sound 17:21 Zeno` I'll let you guys nut it out. I can see the benefits of both 1.0.0 and 5.0.0 17:21 Zeno` 5.0.0 (although I hate it) might be better for the public 17:22 * twoelk quietly wispers 0.5.0 next week and 1.0.0 next month? 17:23 Calinou 5.0.0 sounds like a gigantic bump, don't do it 17:23 Krock wut? 0.4.12? 17:24 Nerz twoelk, there are no major change for using 0.5.0, 0.4.12 is pretty good, and then, break the ice, change ugly things and 1.0 will be there :D 17:25 twoelk yeah but clearing the last digit befor jumping to doeble zero seems more logical to me 17:26 twoelk whatever is decided will have to be explained lots of times anyways as there seems to be no logical sollution 17:27 Nerz in fact, 3 digit is legitimate: 1 for major updates (new feature for users), 2 for minor changes (rewrites, performances...), 3 for minor update (memory leak...) 17:28 Calinou major update shouldn't be just new feature 17:28 Calinou it should be in case of a total overhaul 17:28 Nerz ofc 17:30 rubenwardy Just phrase is as a dropping of the 0. 17:30 jin_xi so i say 5.5.0 is nice as it echoes the projects histrory when it was minetest-c55 17:30 twoelk use android support to justify jump to 0.5.0 and network rewrite for 0.7.0 and break formspecs for 1.0.0 17:30 jin_xi but seriously im for 1.0.0 or so, jump to 5 is to big and arbitrary imo 17:33 Nerz twoelk, legitimate 17:33 twoelk lets see that could be adding client side lua to justify 2.0.0 and maybe mob-api for 3.0.0 :P 17:35 rubenwardy I suggest 4222.0 17:35 twoelk leave real implemented infinity for 10.0.0 17:36 twoelk and elevate to be gods with 42.0.0 17:37 Acerspyro How about we just jump straight to 666.666.6666 17:37 twoelk nah thats evil 17:37 rubenwardy (4222 is the number of commits so far) 17:37 Acerspyro or 6969.696969 17:38 Krock jump to (int)-1 pls 17:38 Krock *unsigned int 17:38 Acerspyro but seriously, majorversion.commit? 17:39 Acerspyro right now, we're at 0.4.11, idk how many releases have happened but it would be the number of releases.GITcommitNumber 17:39 Calinou commit number should only be used for dev build 17:39 Acerspyro Calinou: Yes 17:39 Acerspyro So it would be, for example if we had 52 releases 17:40 Acerspyro Minetest v.52 17:40 Acerspyro Why does that look insane to me, while it is clearly logic? 17:41 Nerz twoelk idea: use android support to justify jump to 0.5.0 and network rewrite for 0.7.0 and break formspecs for 1.0.0 seems pretty good 17:42 Krock Acerspyro, and the dev releases? 17:42 Acerspyro Minetest v.52.4222.0 17:43 Nerz beuark 17:43 Acerspyro Why does using big numbers as the major revision number look crazy, but makes sense? 17:44 PilzAdam why are people so obsessed with version numbers? 17:44 PilzAdam it doesn't matter at all 17:44 Acerspyro Minetest v.Whatever version you downloaded. 17:44 Krock yeah. just continue with 0.5.0 and everyone is happy 17:45 Acerspyro I wonder what's the major release number for if it's never used. 17:46 twoelk PilzAdam: in your how to report a bug you want a version information. so that better use a system users may understand 17:46 PilzAdam twoelk, I don't need to understand how version numbering works when I just have to copy the text from the top-left or --version output 17:48 twoelk If it is logical I don't have to copy it bit for bit and may notice more easily when I make a mistake 17:48 twoelk I can copy chinese signs but it is easier if I understand them 17:48 Krock Acerspyro, it's used just not now. it might take some years 17:49 Acerspyro What's gonna trigger its use? 17:49 Acerspyro (forward prediction much) 17:50 Krock it's there because we do it this way. 18:22 Nerz everybody died because of too huge minetest version... IRC become a desert 18:24 hmmmm this is such a trivial thing to decide 18:24 hmmmm it's so ridiculous so much effort is being expended on the stupid things 18:24 hmmmm command decision: next version is 1.0.0. DONE 18:24 T4im big bikeshedding is followed by lots of cycling around 18:24 hmmmm LET'S MOVE ON 18:25 Nerz version is useless, commits are useful (but i think commit to change version is useless :p) 18:27 * twoelk hands hmmmm an umbrella for the shitstorm brewing on the horizon 18:28 VanessaE hmmmm: ok, as long as we can be sure of one thing else: release early, release often... no more of this 2-4 months between releases bs. 18:28 VanessaE (but you already addressed that previously) 18:30 VanessaE Nerz: version numbers serve a purpose: to signal to the user what they have, and to tell those of us who have to field support requests/bug reports definitively whether the user has an up-to-date-enough version for whatever it is they're doing. 18:30 Nerz ofc 18:30 VanessaE users will never understand commit IDs 18:30 Nerz subversion commits are more readable than git, right :p 18:30 VanessaE and frankly, they're a little cumbersome to compare versus an incremental version number 18:32 T4im plus commit ids can sometimes create accidental words you might not want to publish… like beef or babe 18:32 VanessaE heh 18:32 T4im (seen those already) 18:35 Nerz :p 20:32 * est31 senses version number bikeshedding 20:45 est31 ~vers 20:45 est31 sorry 20:46 VanessaE 0.4.11.kiss_my_ass 20:46 VanessaE ;) 20:55 shadowzone o.o 21:03 rubenwardy How about 1.1.0.1.1.1.0.0.0.1.1.0.1.0.1.0.1.0.0? 21:07 shadowzone 5.3.2.5.5.3.2.4.6.2.3.5.6.2.4.6.4.5.3.6.5.4.6.5.4 21:13 Brains 8.6.7.5.3.0.9-jenny 23:08 sapier VanessaE: any issues with the new packet code? 23:08 VanessaE not so far. 23:08 VanessaE but, 23:08 VanessaE it's been decided not to merge until after the 0.4.12 release (on jan 29) 23:09 sapier I see, a release? now? :-) 23:09 sapier well I guess this has been discussed long enough as usual :) 23:10 sapier any opinions about the autohiding button bar? merge prior or after 0.4.12 23:10 sapier ? 23:11 VanessaE it's technically a feature and we're in freeze now, but as it's android and it's kinda needed for usability, I don't see a reason to wait 23:12 sapier did you have time to test the apk? 23:12 VanessaE no, I didn't 23:12 VanessaE but I've pointed others at it 23:12 VanessaE I'm unclear if they were able to successfully use it though 23:13 VanessaE wait. 23:13 VanessaE lemme see if anyone's using it :) 23:13 sapier :-) guess the file download is timed out by now 23:14 VanessaE nope, it still downloads. 23:15 sapier good 23:15 sapier I'm unsure about the exact positions 23:15 VanessaE I get no responses from anyone if it's being used or not :-/ 23:16 VanessaE wait 23:16 VanessaE [01-22 13:52] ok have them installed and running on tablet and phone as well 23:16 PilzAdam hmmmm, date format should conform to ISO 8601 23:16 VanessaE I think he's talking about your apk, sapier. 23:17 VanessaE [01-23 10:19] I like the improvements for android - has more controls approaching the PC versions 23:18 sapier hmmmm I guess android apk will be a little bit out of sync anyway so I'm gonna merge it immediately after 0.4.12 release and do the andorid 0.4.12 version right after merging it 23:18 VanessaE that's about all I could get from anyone, and as you can tell by those timestamps, that's a little old. 23:19 sapier sounds like ppl see it as improvement 23:19 sapier well one person 23:21 acerspyro Should fix scrollbars 23:23 sapier acerspyro: could you at least one time please provide a suggestion telling what you actually mean? 23:23 VanessaE sapier: there may be an issue with the network patch.. 23:23 sapier what issue? 23:23 VanessaE https://forum.minetest.net/viewtopic.php?f=10&t=4057&p=168651#p168552 23:23 acerspyro oh sorry 23:23 acerspyro The scroll bar is hard to grab 23:24 acerspyro Something usual on android like hold and moving your finger to scroll might be better 23:24 VanessaE at first I thought MichaelEh was having inadequate RAM problems, but now his description looks like a network timeout since I enforce 0.4.11-or-higher on that server 23:25 sapier acerspyro: what scrollbars are you talking about? 23:26 acerspyro sapier: Server list 23:26 sapier VanessaE: any way to reproduce it by now? 23:26 sapier acerspyro: on android or pc? 23:26 VanessaE sapier: I can't reproduce it, no, but I don't run Windows, either. 23:26 kilbith MichaelEh's quote : "In fact, 0.4.11 for windows either 32 or 64 crashes on my windows 7.1 Enterprise even in single player mode. It's either not render SOME media types or NONE at all." 23:26 kilbith ... 23:26 acerspyro sapier: On Android 23:27 clymancer wtf kilbith 23:27 VanessaE kilbith: by that he means either he plays with 0.4.10 and some media types (e.g. meshes) can't be rendered, or none at all work (because 0.4.11 or newer just crashes) 23:27 sapier acerspyro: what's the issue there, to small or something else? 23:27 acerspyro Yes, it's a bit too small and hard to browse, but that's not what I'm talking about 23:28 acerspyro You have to grab the tiny scrollbar to scroll down 23:28 sapier 0.4.10 crashes if you join a 0.4.11 server? are you kidding? 23:28 acerspyro and if you move your finger outside of its area, it stops scrolling 23:28 VanessaE sapier: nono, 0.4.11 crashes way more than 0.4.10 did, and I even tried pointing him at sfan5's latest builds, to no positive effect. 23:29 sapier acerspyro: sounds like an issue hard to fix, I'm gonna think about it but I doubt there's gonna be a quick solution 23:29 VanessaE his "greyed out" symptom reminds me of what happens when trying too-new of a client to connect to say redcrab 23:29 VanessaE that's why I wondered if it was a network glitch 23:29 acerspyro sapier: Seems like this lacks flexibility on the GUIs 23:30 sapier VanessaE: I remember seeing this when running mt in valgrind too ... did assume it was a valgrind issue 23:30 VanessaE plus a LOT of people are still complaining about crashes after playing for a while (5 to 20 mins), indicating a memory leak somewhere. worse if the game already uses a lot of RAM, e.g. dreambuilder 23:30 sapier acerspyro: our gui is extremely limited and no there ain't any chance for a quick fix 23:30 kilbith ... since mesh node implementation, yes 23:31 acerspyro Hmm, not willing to be too demanding, but how about using another backend for the GUIs? 23:31 sapier VanessaE: could be one of those issues we fixed about a week ago, I don't think a lot of ppl run clients that recent 23:31 acerspyro but this isn't a quick fix either, but will probably make fixing other GUI issues in the future way easier 23:31 VanessaE kilbith: enable_mesh_cache = false will help, but doesn't stop the leak. 23:31 sapier acerspyro: if you wanna implement it no problem 23:32 VanessaE sapier: michael tried a build that's now only about 2 days old 23:32 sapier acerspyro: usually switching gui frontend just changes old issues for new issues you don't know yet 23:32 sapier and for our gui we've got a lot of history to cope with 23:33 sapier yet writing a new gui including formspec replacement is quite a lot of work so if there's any student willing to spend about 3-6 months of time? 23:33 acerspyro not me 23:34 sapier VanessaE: memory loss in that new clients is strange... wonder if it's a leak or just stall memory 23:34 acerspyro I suck at developing stuff and I know it. 23:34 acerspyro sapier: I got that leak too 23:34 VanessaE sapier: there are bad leaks in 0.4.11-release, most of which have been fixed in subsequent commits. there's apparently a few left is all. 23:34 VanessaE ditto for crashes 23:35 sapier well I'd suggest finding those leaks prior release 23:35 sapier any ways to reproduce them ? 23:35 acerspyro play for a long time 23:35 VanessaE sapier: connect and play for a while 23:35 est31 oh interesting... currently my minetest uses 3.4 gib... gonna pull and recompile 23:35 acerspyro est31: lelz 23:35 sapier let me be more precise any valgrind capable ways to reproduce them ;) 23:35 VanessaE est31: 3.4 GB? what the hell are you running? 23:36 acerspyro VanessaE: nothing 23:36 VanessaE sapier: catch zeno, he's good at that :) 23:36 VanessaE est31: any texture pack? 23:36 est31 VanessaE: the instance thats connected to your survival server 23:36 acerspyro It can be doing plain fuckin nothing, and it can go up to 2 TB if left long enough. 23:36 est31 no nothing 23:36 VanessaE est31: oh, it's just sitting idle to test the leak? 23:36 est31 its commit g9b0d77a 23:36 VanessaE and you turned off that mesh cache? 23:36 est31 sitting idle? 23:37 est31 mesh cache let me see 23:37 VanessaE idle as in you're not actively playing 23:37 VanessaE just leaving it sitting there connected. 23:37 kilbith only 1GB for me with nothing... 23:37 VanessaE kilbith: that's consistent with what you should see right after connecting. 23:37 acerspyro kilbith: still too much for a game like MT 23:37 est31 I have a caverealms cave loaded... gonna retry and sit idle 23:37 VanessaE (before map has had time to load) 23:38 VanessaE est31: eh, that doesn't mean too much, map data should be a few dozen megs tops 23:38 kilbith oops, sorry, it's the total RAM usage ._. 23:38 kilbith 281 MB for minetest 23:38 acerspyro lol 23:38 acerspyro better 23:38 VanessaE kilbith: doing what? 23:38 VanessaE what server/game? 23:38 kilbith nothing, standing, new world 23:39 kilbith with homedecor and UI 23:39 est31 should I try enable_mesh_cache = false ? 23:39 VanessaE est31: you should, it'll cut your usage by ~800 MB 23:39 VanessaE but it won't stop the leak 23:39 VanessaE (doesn't anyone read the forums? :P ) 23:39 sapier nope ;-) 23:40 sapier we assume you're gonna tell us anything important VanessaE ;-) 23:40 VanessaE I did...in like 6 places :) 23:41 VanessaE and I even set it so dreambuilder's .conf :) 23:41 VanessaE (not that that helps server players) 23:41 est31 weird when I try to return to menu MT crashes. lets see whether I can reproduce... 23:42 est31 no 23:42 sapier est31 sounds like some sort of cleanup issue 23:42 est31 src/script/cpp_api/s_base.cpp:75: ScriptApiBase::ScriptApiBase(): Assertion 'm_luastack' failed. 23:42 sapier no lua stack available? 23:42 sapier very strange 23:43 est31 http://pastebin.com/5Q55aVt6 23:44 est31 I've stopped starting minetest with gdb perhaps I should re-introduce this 23:44 sapier is this everything? 23:44 est31 nothing unusual on stdout/err 23:44 sapier this looks like a race condition 23:45 sapier seems like something is done while client did already shutdown/delete required things 23:46 kahrl perhaps the system is simply out of memory and so luaL_newstate fails to allocate memory? 23:47 sapier possible too but prior starting new lua stack we did shutdown client I'd expect this to provide enough memory for a lua stack? 23:47 kahrl yeah, that's true 23:48 kahrl maybe the heap is too fragmented or something obscure like that 23:48 est31 memory was ~60% 23:48 VanessaE kahrl, sapier: for the sake of memory, can we get the mesh cache to be disabled by default, and its minetest.conf/internal variable changed so that it can't accidentally be turned back on? 23:48 sapier could be ... guess we should fix that damn texture fetching issue soon 23:48 sapier or at least make it a proper assert 23:48 VanessaE (e.g. by someone's old config who wasn't paying attention) 23:49 kahrl VanessaE: why? it's a good thing in most cases 23:49 VanessaE kahrl: it's a bad thing. 800MB is not worth the intangible benefit. 23:49 kahrl it's only 800MB with dreambuilder 23:49 sapier "only" ;-) 23:49 VanessaE yeah I know, but even with minetest_game I'm sure it takes a significant amount. 23:50 VanessaE and I can't fix clients, I can only fix singleplayer 23:50 kahrl sapier: "only" refers to the game not the memory :P 23:50 VanessaE (with the game's setting I mean) 23:50 kahrl I don't think it takes much at all in minetest_game 23:51 kahrl how many nodebox/mesh nodes are there? less than a dozen? 23:51 * VanessaE reaches into the aether and pokes at Zeno` 23:51 VanessaE in minetest_game? probably a couple dozen. 23:51 VanessaE no mesh nodes, but plenty of nodeboxes (stairs/slabs) 23:51 kahrl that number is small enough that the mesh cache won't be a problem 23:52 VanessaE but afaik it affects everything that is converted internally to a mesh and which can use facedir 23:52 kahrl disabling it would be bad for mesh making performance 23:52 VanessaE zeno and I tested that...it doesn't seem to have a tangible effect on mesh-building 23:52 VanessaE but okay. 23:53 kahrl link? 23:53 VanessaE it was private chat 23:53 * VanessaE reaches into the aether and pokes at Zeno` some more