Time Nick Message 04:08 paramat merging #6361 04:08 ShadowBot https://github.com/minetest/minetest/issues/6361 -- World start time: Add as setting, default unchanged by paramat 04:16 paramat done, no more getting up early for me :D 11:52 nerzhul hello :) can someone can finish the #6369 review to make it merged ? 11:52 ShadowBot https://github.com/minetest/minetest/issues/6369 -- Implement minetest.register_can_bypass_userlimit by nerzhul 12:24 Fixer had crash on newest minetest 12:25 Fixer during connection to server 12:26 nerzhul please give us a backtrace 12:26 nerzhul and what is MT version of cleint and server 12:29 Fixer unfortunately i was running nondebug one, client: newest (a383...), server: 0.5.0-dev, "dcbl" on serverlist 12:30 nerzhul it's possible this server is not up-to-date no ? 12:30 Fixer there are two 0.5.0-dev servers available, i was on time travel, but server itself crashes like crazy, so I connected to dcbl one and just right before having entered - got crash 12:30 nerzhul it's not recommended to use 0.5.0-dev in production 12:30 Fixer nerzhul: possible 12:30 nerzhul if you can connect on dcbl with 0.4.16 client it's a not up to date server 12:32 Fixer dcbl just went down :/ can't check 12:33 Fixer nerzhul: not possible to connect via 0.4.16 to dcbl 12:34 Fixer I guess I need to compile debug version 12:42 nerzhul y 12:52 jas_ dcbl is running a383eb0e at the moment 12:53 Fixer oh wait 12:53 Fixer it crashed ^_^ 12:56 Fixer nerzhul: https://pastebin.com/raw/bCTf6RRR 12:58 nerzhul Fixer, i see another SharedBuffer copy problem 13:00 nerzhul when i will get much more time i need to enhance another time the network code to have a nicer packet buffer and cleaner reading, the packet movement inside the code is a little bit crazy to follow, it works, but it's not optimal 13:09 Fixer issue is opened here https://github.com/minetest/minetest/issues/6373 13:13 nerzhul ok 13:14 nerzhul i'm polishing mod channel PR and i will look at it after doing some work at work 13:21 nerzhul sfan5, i enhance mod channel PR (#6351). I prepared work for readonly state on channels (handled but not settable by API atm), i added sender in the callback (missed mandatory field) 13:21 ShadowBot https://github.com/minetest/minetest/issues/6351 -- Implement mod communication channels by nerzhul 14:19 Calinou I'm trying to compile Minetest using MSYS2 (just for fun), how do you solve the missing OpenGL libraries? I installed Mesa using pacman already 14:19 Calinou (I successfully compiled Sauerbraten using the same MSYS2 setup, so it should be possible) 14:22 sfan5 shouldn't your toolchain have the gl headers? 14:23 sfan5 my mingw one does 14:45 nerzhul I push another concurrency problem fix in connection.h, reported by Fixer, due to network refactor. Those copies are useful, sometimes... but we really should have a better network buffer with concurrency fixes, later :) 14:46 Fixer Calinou: i just use modified buildbot with libs provided by sfan 14:46 Fixer for msys2 14:57 Calinou sfan5: I have an OpenGL DLL, but no headers, it seems 15:11 jas_ nerzhul: now i'm crashing left and right, i'll let you know 15:12 nerzhul what does that mean ? 15:12 jas_ (i'm doing make clean now, maybe that will change it) 15:13 nerzhul do you have a backtrace i can see ? 15:13 jas_ i got new crashes and non-connections on current head, but let me verify and i'll let you know. : ) 15:14 Fixer nerzhul: still crashing, does server need to update as well? 15:14 nerzhul no it's purely binary problem, Fixer is this the same backtrace ? i need to be sure 15:15 nerzhul maybe there is another part i should fix, but i need to be sure 15:19 jas_ ye i gonna have to take my time and compile a debug version finally, lmao 15:21 nerzhul sorry i crashed, you can now send me a backtrace 15:22 jas_ ok iwill, i'll run DEBUG builds from now on. i just connected OK so it's possible i forgot to ctrl+c and restart the server (i'm tried) 15:22 jas_ gn? 15:23 jas_ no this was a client crash: https://gist.github.com/jastevenson303/e73ef5151fcdc8cf618cc4db08c48f90 -- i'll send the backtrack soon as i have it 15:24 nerzhul this assertion is client side 15:26 Fixer i need to compile debug one 15:28 nerzhul reverting the last fix, it's not good 15:28 nerzhul i will take more and more time to find the real issue 15:29 jas_ okie doke. 15:29 nerzhul it's a concurrency problem but it seems it's not here, i should take more time to find the deletion issue, i don't understand why data is removed while used 15:29 nerzhul i think i need to graph the code 15:30 jas_ quick question: should i be doing make clean every time? 15:30 Calinou you don't need to perform a "make clean" unless you're getting build errors 15:30 nerzhul jas_, it's not needed 15:30 jas_ ah very good, thx! 15:30 nerzhul except if you change compilation options 15:34 nerzhul oh t0ny2 found a typo 15:34 nerzhul it's why it bugz 15:35 jas_ *\:D/* 15:35 nerzhul okay, reapplying the fix with the typo fix 15:35 nerzhul jas_ = t0ny2 ? 15:35 nerzhul sorry for the history... 15:35 jas_ nope, jas = jas 15:36 jas_ (one day i'll be able to read c++ heh) 15:37 nerzhul okay it's pushed 15:37 nerzhul you can test it now Fixer jas_ 15:37 nerzhul withtout the typo fix it's better 15:38 nerzhul i prefer to see that typo error than searching deeper if there were an error elsewhere, whereas i found this should be the concurrency cause (2 threads using this object) 15:41 jas_ recompiling. sounds deep, nerzhul 15:42 nerzhul concurrency is not easy 15:42 jas_ fewer rewards if it were, perhaps 15:43 nerzhul network is clearer now, but some optimization side effets appeared, i need to finish my current opened PR to re-read another time the network code to cut that network lag problem we got hen someone is slow in queue 15:45 jas_ good luck, i salute you 15:46 nerzhul ? 15:46 jas_ lmao " 15:46 jas_ a gesture of respect, homage, or polite recognition or acknowledgment, especially one made to or by a person when arriving or departing. 15:47 nerzhul yea :) 15:47 nerzhul i just hope your server will be running smooth 15:47 nerzhul 0.5.0-dev is for crazy man :p 15:48 jas_ got plenty of time, let's hope it's well spent 15:49 jas_ my server is live. i thought you had a testing server too, no nerzhul? 15:51 jas_ off to play for a bit b4 bed, gn o> 15:52 nerzhul i run tests on my machine, i have APpletree live server but it's quite old with old MT code base and many changes, it's not easily upgradeable 15:53 nerzhul hopefully i backported may changes to MT core but i need to convert map 17:55 Fixer new crash on newest git 17:55 Fixer will run under gdb, crash mid game 17:55 nerzhul ty 17:55 nerzhul i keep an eye on this network problem, but i need to fix various other apps too 18:39 paramat celeron55_ Krock nerzhul nore rubenwardy sfan5 sofar , lhofhansl is ok with being a core dev https://github.com/minetest/minetest/pull/4959#issuecomment-326868784 any support / objections? 18:40 celeron55_ the core dev question? 18:42 Krock No objections from my side 18:42 paramat +1 for me 18:51 Fixer if someone tests newest git, use debug one, for backtraces, there are crashes 19:02 nore +1 from me as well 19:03 sfan5 no objections from me 19:05 paramat radical suggestion: are the official MT mobile apps free in the app stores? if so how about charging a low price (less than other apps) to bring a small amount of money in to cover c55's costs? these clone apps cause problems, this way we can turn the app market to our advantage 19:07 sfan5 you just turned a small amount of android users into zero android users, congrats 19:07 sfan5 also how does that turn the app market to our advantage? 19:07 Raven262 Add a donate option, maybe? 19:07 Fixer nerzhul: sharedbuffer related crash ;) 19:10 red-001 I agree with sfan5 19:11 Fixer nerzhul: https://pastebin.com/raw/qFGptxV3 / c05228f 19:11 red-001 even when it has no ads and is completely free the official app has a far smaller marketshare 19:12 paramat my thought is, very many players are willing to pay money for unofficial apps, so they might be fine with spending less for the official one 19:12 DS-minetest what about a premium version for mobile, with different set default values for settings, enable everything, that looks special and good, even if it costs performance, just add a different initial minetest.conf 19:12 rubenwardy paramat, they're not though 19:12 rubenwardy they use a bunch of free apps with *ads* 19:13 Fixer and ads are cancer 19:13 paramat hmm, my impression was most clones charge 19:14 paramat ok 19:14 paramat odd that most players choose unofficial apps 19:15 paramat must have poor taste 19:15 nerzhul paramat, first MT app on play store is free, second a free app cannot be non free after first publish, last, no 19:15 Raven262 Or the apps have fancy names. 19:15 nerzhul Fixer, in destructor i see 19:16 paramat can't be non-free due to licence? 19:16 paramat lol yes it's the fancy names 19:16 nerzhul no, when you register app you choose: free or paid, paid can become free, free cannot become paid 19:16 paramat ok 19:17 paramat could open a new app account 19:17 nerzhul gg travis, they break clang-4.0 builds 19:18 nerzhul where are the fucking packages 19:18 nerzhul https://travis-ci.org/minetest/minetest/jobs/271782487 19:18 paramat anyway i realise this probably won't happen 19:18 paramat and our app is in poor state also 19:19 Raven262 Somebody made lots of spam accounts on the forums. Check out the "Congratulations to" section. 19:19 Raven262 All randomly generated names. 19:19 Raven262 plus completely random bio on each one. 19:19 Raven262 and all 36 today 20:01 nerzhul merging #6369 in ~ 5 mins 20:01 ShadowBot https://github.com/minetest/minetest/issues/6369 -- Implement minetest.register_can_bypass_userlimit by nerzhul 21:01 nerzhul Fixer, can you test the following patch please ? i restore the prvious old behaviour in the critical network part which cause problem 21:01 nerzhul https://hastebin.com/efolayocun.hs 21:02 nerzhul i think this should do the trick, i don't see any other problem in the refactor code, this is the only difference on this critical part 21:02 nerzhul it seems the refcount cut is mandatory 21:04 Fixer nerzhul: will look, but it does not crash that fast 21:05 nerzhul take your time, test it for one day if possible, if it's okay tomorrow i will push it 21:06 Fixer nerzhul: https://pastebin.com/raw/DuT53LNc 21:07 nerzhul hu ? 21:07 nerzhul can you apply it manually ? it's tiny changes 22:02 nerzhul see you tomorrow, hope you can confirm it's okay :) 22:49 paramat trivial #6376 22:49 ShadowBot https://github.com/minetest/minetest/issues/6376 -- Network protocol documentation: Add 'can_zoom' to version 36 by paramat 22:52 ShadowNinja Megaf, nerzhul, celeron55_: Regarding auto-formating: I wrote an Astyle PR since Astyle is less nitpicky and doesn't require everything to perfectly match its "one true style". It ignores some minor stylistic differences which don't really matter that clang-format will "fix", resulting in an error. IIRC it's also a lot faster since it actually ignores whitelisted files instead of formatting them regardless and then ignoring the results 22:52 ShadowNinja if it's whitelisted. 22:57 paramat ShadowNinja any comment on lhofhansl possibly becoming a new core dev? 22:57 paramat sofar too ^ 22:58 ShadowNinja paramat: I don't know enough about lhofhansl to have an opinion on him. sofar sounds good. 23:00 ShadowNinja ... I've beed away too long, I thought you were asking about sofar too. I thought he was a core dev too but I couldn't remember for sure :-P 23:00 ShadowNinja been* 23:01 sofar lhofhansl is cool, +1 23:01 ShadowNinja Sorry sofar :-D. I'd better get back to other sutff now. 23:02 sofar no worries 23:05 paramat hehe 23:06 paramat conf.example needs updating by auto-generation so will do that now and merge with 6376 23:23 paramat #6377 will merge these 2 now 23:23 ShadowBot https://github.com/minetest/minetest/issues/6377 -- Conf.example: Update using auto-generation by paramat 23:33 paramat done 23:46 Megaf +1 for sofar as core dev 23:46 sfan5 wot 23:53 Megaf [23:57:32] ShadowNinja any comment on lhofhansl possibly becoming a new core dev? 23:53 Megaf [23:57:49] sofar too ^ 23:53 Megaf xP 23:54 Megaf [23:58:56] paramat: I don't know enough about lhofhansl to have an opinion on him. sofar sounds good.