Time Nick Message
00:07 Fixer !seen millersman
00:07 ShadowBot Fixer: I saw millersman in #minetest-dev 5 days, 1 hour, 36 minutes, and 42 seconds ago saying "not sure when i find the time to fix it, though"
00:31 paramat ShadowNinja i'm planning to merge #4086 later, MoNTE48 says he is 'not a programmer' so i doubt this would be reworked by his team soon, so it's either merge as-is, rework it ourselves or not merge at all. shall i go ahead?
00:31 ShadowBot https://github.com/minetest/minetest/issues/4086 -- Fix locked hardware buttons on Android by MoNTE48
00:31 hmmmm no don't merge it
00:31 hmmmm jesus christ
00:31 hmmmm where are the approvals
00:31 asl97 that can wait till after release
00:32 rubenwardy Could always do an Android patch release
00:33 hmmmm paramat: good enough is the enemy of perfect, but that commit is not good enough
00:33 hmmmm you risk becoming freeminer by defaulting to this attitude too often
00:34 hmmmm just understand that if it gets merged as-is, NOBODY will ever fix it, or maybe they well 3 years later when it gets rediscovered for one reason or another
00:34 hmmmm if you have nobody to review the code right now you probably won't have anybody to maintain the code
00:35 hmmmm if you don't have anybody to maintain it you might as well leave it out
00:36 paramat ok no problem. est said 'yeah its mostly harmless' and it is now tested, so i was seeing that as a possible approval
00:37 paramat thanks for feedback, i was feelng unsure about this
00:39 paramat it's a relief to be able to refuse it. we're not letting this delay release
00:45 hmmmm paramat, ridge_heightmap is the same as heightmap ever since you added updateHeightmap after generateRidgeTerrain
00:45 hmmmm except for areas with overhangs
00:45 hmmmm that happen to be above rivers
00:45 paramat ok
00:45 hmmmm i'm actually removing ridge_heightmap in a followup commit
00:46 paramat nice
00:46 hmmmm in any case there's a bug with the commit you commented on
00:46 hmmmm i need to change it
00:47 hmmmm i defined heightmap to be optional for cave generation
00:47 hmmmm so if it's NULL then it's not used and we'll just default to water_level
00:47 hmmmm but CavesRandomWalk was lacking the check for NULL
00:52 paramat overhangs over rivers, and river tunnels, are common, but i get the feeling the difference will be barely noticeable and possibly worth it for simplification
01:36 ShadowNinja celeron55: It looks like there's a good chunk of code in map.cpp for handling the old sectors format, that likely doesn't even work. Is this correct? Is there a point to any of this code or can large chunks of it just be deleted?
01:52 ShadowNinja I haven't found it in my logs (it was a long time ago) but IIRC we agreed to switch to C++11 as soon as Debian stable included a version of GCC in the official repositories that supported it. But that happened a long time ago. Now we're waiting on Ubuntu I guess, but you can just install a newer GCC or Clang from a PPA anyways. C++11 is already 5 years old, I think waiting any longer is kind of absurd.
01:53 paramat merging #4106
01:53 ShadowBot https://github.com/minetest/minetest/issues/4106 -- Update Fedora dependencies list by lonniganseaweed
01:54 hmmmm paramat, did something happen to mapgen v6? i can't find any trees
01:54 paramat !
01:55 paramat i'll check soon, maybe it's the larger biomes and new snowbiomes
01:56 hmmmm we can't release a game that cannot be played
01:57 paramat mgv6 by default uses my new biome system that adds tundra and taiga, the biome noises are larger to compensate
01:58 paramat biome size is a little larger but not by much
01:59 paramat mg flags were changed too, 'trees' and 'flat' were moved into mgv6 spflags
02:00 hmmmm welp
02:01 paramat but i haven't touched mgv6 for a long time
02:01 hmmmm could you fix this?
02:01 hmmmm it appears that trees are not a part of the default MapgenV6Params
02:02 hmmmm nvm i'll do it
02:02 hmmmm why bother somebody else for a 5 second fix
02:07 paramat merged. all milestones cleared (apart from your fix)
02:08 hmmmm https://github.com/kwolekr/minetest/commit/9b17dcda6c55057e9e5d8b38aa4574bb12061f73
02:08 hmmmm will push on approval
02:09 paramat looking
02:15 paramat looks like spflags were used as modifiers for the builtin defaults, this may need checking in other mapgens too then
02:17 paramat what about biomeblend, decided to disable it by default?
02:20 paramat won't this break old maps though, that have a particular mgv6_spflags modifying the builtin flags?
02:22 hmmmm didn't you want biomeblend disabled?
02:22 hmmmm no it'll never break reverse compatibility, because writeFlagStr is designed to write out all flags and their options explicitly
02:23 hmmmm this is how we're able to make modifications like this to begin with
02:23 paramat erm i don't think biomeblend is disabled currently
02:24 hmmmm well nevermind if it is or isn't, was your intention to disable biomeblend?
02:24 paramat nope
02:24 hmmmm remember this is the ugly dithering blend
02:25 paramat if that was my intention biomeblend would be disabled in mgv6 for the last few months
02:25 paramat yeah
02:25 hmmmm OK then, i'll re-add biomeblend
02:26 paramat i'm not too keen on it but i seem to remember others insisted on it
02:26 paramat yes go ahead
02:27 paramat i have to go eat and can't think straight, i'll try getting my head around this later
02:27 hmmmm maybe this is my fault for not making the flag string system more intuitive
02:27 hmmmm there are two kinds of options
02:27 hmmmm implicit and explicit
02:27 hmmmm a flag string specifies explicit flags to enable/disable from the baseline implicit set
02:28 hmmmm so with this modification, the implicit set would be JUNGLES, SNOWBIOMES, TREES, BIOMEBLEND, MUDFLOW
02:28 hmmmm so if a user writes "mgv6_spflags =" with no flags specified it won't mess anything up
02:28 hmmmm it's just that nothing would be modified from the implicit baseline
02:29 paramat aha
02:29 hmmmm if mgv6_spflags is set equal to the text descriptions of each of the implicitly set flags, then no change will happen either
02:29 hmmmm if you want to *disable* a flag implicitly set, such as MUDFLOW, you have to write "nomudflow"
02:29 hmmmm the explicit flag setting overrides the default implicit
02:30 paramat ok i think there has been confusion over this then
02:30 paramat as in what overrides what
02:30 hmmmm likewise, the intention was that to actually not have trees, you need to explicitly write "notrees" instead of forgetting to write "trees"
02:31 hmmmm but this got screwed up when you didn't synchronize the implicit defaults
02:32 paramat i think many people, and devs, have misunderstood this, for years
02:32 paramat i hope the other mapgens are ok
02:32 paramat +1 then
02:32 paramat but please check other mapgens
02:34 hmmmm mgv6_spflags is the only mapgen-specific setting in defaultsettings.cpp
02:35 hmmmm imho it would be best to remove that line from defaultsettings.cpp completely
02:35 paramat i was about to suggest that
02:38 paramat i'm not sure 'jungles' should be in there. if 'snowbiomes' is set then 'jungles' is ignored
02:38 paramat https://github.com/minetest/minetest/blob/master/src/mapgen_v6.cpp#L437
02:38 hmmmm que
02:38 hmmmm if (spflags & MGV6_SNOWBIOMES) {
02:38 hmmmm float blend = (spflags & MGV6_BIOMEBLEND) ? noise2d(p.X, p.Y, seed) / 40 : 0;
02:39 paramat 'snowbiomes' essentially means 'new biome system which includes snowbiomes and jungles'
02:39 hmmmm repetitive much?
02:39 hmmmm oh nevermind
02:40 hmmmm SNOWBIOMES looks like BIOMEBLEND when they're ALLCAPS
02:40 paramat oh i see :] was confused
02:42 paramat 'jungles' being in there may have been from my misunderstanding of flags
02:42 hmmmm it certainly doesn't hurt to include jungles, even if it's implicit via snowbiomes
02:43 paramat indeed. just wondering why i left that in there
02:44 hmmmm s/implicit/implied/
02:47 paramat bbl to check your eventual commit
03:10 hmmmm btw paramat i added a new commit in the biomegen branch: https://github.com/kwolekr/minetest/commit/33249cf54d7027c1592bc162931a17e501e0678d
03:51 paramat nice, more flexibility
03:52 paramat hmmmm, now what about mg flags? can 'dungeons' be moved into mapgen.h?
03:52 hmmmm i'd rather not touch it
03:54 paramat ok. just seems the same situation as mgv6 spflags
03:55 paramat ie if someone has 'mg_flags = ' in conf, then no dungeons
03:56 hmmmm well i suppose you could fix that one too if you'd like, but it's nowhere near as important as the trees one
03:57 paramat well just before release perhaps the less we touch the better
03:57 hmmmm yeah that was my rationale too
03:58 hmmmm just keep in mind, stable != "bug free"
04:06 paramat good you merged the fix +1
04:23 paramat hmmmm, when i attempted to move the 'dungeons' flag from mtgame's .conf into mapgen.h you warned against it (this might be how i got confused) so it ended up in 'mg_flags' in defaultsettings. discussion http://irc.minetest.ru/minetest-dev/2015-04-13#i_4222219
04:59 hmmmm i can't see the code so i don't get the full context of this conversation, but reading what i said back then it seems totally correct and basically what i said now
04:59 hmmmm okay, this was in regards to adding a new default
05:00 hmmmm if you have a new default implicit setting in the mapgen, the explicit setting doesn't exist in the map_meta.txt
05:00 hmmmm so when the map_meta.txt gets loaded it doesn't see a "nodungeons" flag, so it doesn't modify the dungeons flag in the implicit defaults, which happens to be on
05:01 hmmmm therefore it will begin generating dungeons in a world that did not previously have dungeons
05:01 hmmmm while that particular example is benign, imagine the effect on some other mapgen flag...
05:03 hmmmm why this same approach doesn't work for trees, however, is because you moved it
05:05 hmmmm you can't "move" a flag like that without breaking compatibility
05:05 hmmmm people would be required to change their config files to accomodate the change
06:00 hmmmm does anybody have an idea of how many people play minetest compiled with clang as opposed to gcc and msvc?
06:04 Zeno` most of the time I don't even have any idea of what OS people are using
06:04 Zeno` e.g. #4107
06:04 ShadowBot https://github.com/minetest/minetest/issues/4107 -- Clicking shader checkbox makes sub-entries “jump aroundâ€
06:04 Zeno` :)
06:05 Zeno` I cannot make them jump around :(
06:05 hmmmm well clang being the minority might be the safe assumption
06:05 hmmmm i've been working with the cavegen a bit recently and i noticed a lot of code like this:
06:06 hmmmm v3f foobar(pseudorandom.next(), pseudorandom.next(), pseudorandom.next());
06:06 hmmmm that's not portable.
06:12 Zeno` what do you mean?
06:12 Zeno` oh yeah
06:13 hmmmm so if the majority of minetest builds are done with msvc and gcc, it's safer to explicitly evaluate the order from right to left
06:13 hmmmm since caves will be compatible with most builds
06:13 hmmmm is what i'm thinking
06:13 Zeno` yes
06:14 hmmmm but i don't actually have any statistics
06:14 Zeno` well we should really avoid implementation defined stuff in any case shouldn't we?
06:15 hmmmm yes, my solution is to assign to temporary variables first instead of evaluating right in the argument list
06:15 hmmmm it's just a matter of knowing which order i should be doing to minimize the impact
06:15 Zeno` oh I see
06:15 Zeno` does clang evaluate them in a different order?
06:16 hmmmm yes clang is left to right
06:16 hmmmm and i think MSVC used to be left to right as well but they keep changing it
06:16 Zeno` great
06:16 hmmmm and of course with some optimizations enabled it could shuffle the evaluation order around
06:16 Zeno` of course
06:17 Zeno` compiler can do whatever TF it likes in cases like these
06:21 celeron55 ShadowNinja: it works
06:21 celeron55 ShadowNinja: but it's kind of weird that it's a global fallback like that
06:23 celeron55 i don't know if there's any reason for it to exist; util/ contains a converter script which hopefully works
06:45 Zeno` For review please: http://dpaste.com/16AEY6Y
06:45 Zeno` Reason: to be able to ask people for minetest --version and at least have an idea of their OS
06:46 Zeno` (that's in main.cpp btw)
06:46 hmmmm isn't that information in the logfiles already?
06:48 Zeno` probably... easier to ask people to type --version though
06:49 Zeno` actually I'm not sure it is in the log files
06:50 Zeno` the function doesn't seem to be used anywhere except exporting it to Lua
06:50 Zeno` getPlatformName() I mean
06:50 hmmmm irrlicht prints it on startup
06:51 Zeno` so should we then when the user asks for --version :P
06:51 Zeno` what log level causes it to be in the logs? I can't see it
06:52 hmmmm hmmm maybe it's been changed since ShadowNinja's logging refactor
06:54 Zeno` you could add compiler as well, but that would no longer be a trivial change
06:55 hmmmm costs outweigh the benefits
06:55 Zeno` yes, but I think mine makes sense
06:55 hmmmm you could also polish all the pennies in the world by hand
06:55 Zeno` I'd rather get all the build info than asking people for their OS and build info
06:56 hmmmm er.. if you're into polishing pennies, or just bored...
06:56 Zeno` grr, I'm trying to make getting information from the user easier. And I like polishing pennies
06:57 Zeno` after release I'll probably add to logs as well just to polish some more :D
06:58 * Zeno` polishes a penny while he waits
06:59 hmmmm have you ever seen this http://i.imgur.com/9fgiGxM.jpg
07:01 Zeno` geez
07:01 Zeno` well, it makes polishing them easier
07:01 hmmmm it's under laminate
07:01 Zeno` I thought pennies were silver
07:01 hmmmm murrika pennies
07:02 Zeno` they're all different colours!
07:02 Zeno` they should have polished before laminating
07:02 hmmmm they weren't zenos
07:02 Zeno` :)
07:02 hmmmm sounds like a breakfast cereal. have you had your zenos today?
07:02 Zeno` lol
07:03 Zeno` well they *are* full of vitamins
07:08 * Zeno` is confused
07:08 Zeno` are you objecting to it hmmmm or just saying it's penny polishing?
07:09 hmmmm huh? no, I'm just saying I doubt it'll help very much to know the compiler, but it'd be a lot of work to add. it's ultimately up to you what information you want to add.
07:09 Zeno` oh I don't want to add compiler
07:09 Zeno` just the platform name
07:09 hmmmm ok then :)
07:09 Zeno` :)
07:09 hmmmm you mentioned you wanted to add the compiler above, no?
07:09 hmmmm that's non-trivial
07:12 Zeno` oh, I just mentioned it offhandedly as something that might answer your compiler question earlier
07:12 hmmmm ahh
07:12 Zeno` if I was to add compiler info I'd probably do it in cmake anyway
07:13 Zeno` and since I don't like cmake that means i'd never get around to it
07:14 hmmmm i hate build systems in general
07:15 hmmmm cmake is probably better than most, i mean it's a necessary evil but it still doesn't make me enjoy it
07:16 Zeno` yeah
08:49 thePalindrome Is there any way to prevent an on_cheat event from going through?
08:51 thePalindrome The "obvious" way of registering an on_cheat and returning false should conditions be met
08:51 thePalindrome *doesn't work
08:53 thePalindrome In this instance I'm modifying a sprint mod to keep the server from rubberbanding people that are sprinting
09:26 Zeno` what could cause an active object message to be invalid?
09:27 Zeno` I can't reproduce (Client::handleCommand_ActiveObjectMessages: caught SerializationError: deSerializeString: couldn't read all chars)
09:38 celeron55 i have seen that but never really figured out why it could happen
09:40 celeron55 i'd imagine some kind of length overflow
09:40 celeron55 maybe on the serialization side
09:40 thePalindrome Is it possible that the null termination is being cut off?
09:41 thePalindrome Or is it all length-prefixed?
09:41 celeron55 everything in the protocol is length prefixed
09:41 thePalindrome oh yeah, that'd make sense :P
09:44 thePalindrome Bleh, searching for potentially non-existant code is difficult
09:46 celeron55 but really it makes no sense; on a quick glance the the serialization and deserialization code are dead simple and the same serialization primitives are widely used in the codebase
09:46 Zeno` celeron55, hmm. yeah
09:46 thePalindrome I'd presume that the serializationerror throwing code is simple enough to avoid bugs?
09:46 Zeno` I've added a breakpoint and trying to trigger it again
09:47 celeron55 there's maybe one weird part in it
09:47 celeron55 the number of messages isn't contained in the packet but it relies on the total length matching the packet length
09:47 Zeno` oh...
09:47 celeron55 which realistically speaking should work fine
09:47 celeron55 but there could be a bug
09:48 thePalindrome Do we have anything in common with the reports?
09:51 celeron55 thePalindrome: there's no way to prevent an on_cheat reaction from the server
09:52 thePalindrome Is there an issue I can watch?
09:52 celeron55 thePalindrome: if you look in network/serverpackethandler.cpp, the server code just calls on_cheat and then always proceeds to revert whatever the player did
09:52 celeron55 unconditionally
09:52 thePalindrome hmm, that would prevent most vehicle mods from doing too much
09:53 thePalindrome Would there happen to be an issue number I can track, or should I put one together?
09:53 celeron55 dunno, but github's search should serve you well; you could make an issue if there isn't one
10:06 thePalindrome celeron55: https://github.com/minetest/minetest/issues/4108
10:06 thePalindrome Is there anything immediately wrong with it?
10:07 celeron55 it's a valid issue
10:07 thePalindrome alright, just wanted to make sure I didn't make a convincing argument
10:09 thePalindrome *didn't fail to make
12:39 naltamur hey
12:39 naltamur any news?
12:45 Fixer don't forget to test the damn game before release ._.
12:47 Zeno` sounds a bit time consuming
13:10 sfan5 Fixer: why dont you
13:10 Fixer i'm, but I will not catch everything
13:12 Fixer compiled todays build, lets see
13:13 Zeno` Fixer, minetest --version
13:13 Fixer one sec
13:14 Zeno` I wonder how I can get people to attach the output of minetest --version to bug reports
13:14 Zeno` kind of hard :(
13:14 Fixer Zeno`, http://pastebin.com/raw/xdu0HZmR
13:15 Zeno` hmm... maybe I should have put a space before the ( of (Windows)
13:15 Zeno` nvm
13:16 Fixer you can fix it, it is trivial bug
13:16 Zeno` so trivial that I cannot motivate myself :)
13:16 Zeno` maybe I should
13:18 Zeno` pushing in 105 seconds
13:26 Zeno` Fixer, I nominate you to ensure that every bug report has that info attached
13:26 Zeno` thanks
13:26 * Zeno` runs away now
13:26 Fixer Zeno`, i always post git hash and os anyway (people also know I'm on windows anyway)
13:27 Zeno` I know that YOU do :)
13:27 Fixer i will post --version
13:27 Zeno` just make sure every other bug report from other people have it as well
13:27 Zeno` easy task
13:28 Fixer yeah, that will be nice, can't make sure though
13:31 Fixer Zeno`, there is contribute.md file that probably has info about posting issues
13:40 est31 wth
13:40 est31 c++ doesnt specify in which order to evaluate multiple params in a list of params passed to a function?
13:40 est31 thats bad
14:01 Fixer heh funny
14:02 Fixer if i light something small on fire i can hear that slight fire sound 1000 nodes away
15:01 ShadowNinja The compression API should be changed from a ostream-based method. It causes inefficient code like this: http://sprunge.us/fgUP
15:02 ShadowNinja This is how I did it: https://github.com/ShadowNinja/NBT-CPP/blob/master/src/compression.cpp
16:31 Krock Is it going to be released today?
16:31 est31 ~topic
16:31 ShadowBot est31: Minetest core development and maintenance. Feature freeze since 2016-05-01, release on 11th. Last release: 0.4.13, Aug 20 2015. Chit-chat goes to #minetest. Consider this instead of /msg celeron55. http://irc.minetest.net/minetest-dev/ http://dev.minetest.net/
16:31 hmmmm :)
16:49 Krock \o/
16:56 Krock Wasn't sure if it's really the definitive date
16:57 est31 lets see what paramat will do
16:58 Zeno` august 20?!
16:58 Zeno` lol nvm
17:00 Zeno` I think paramat may be on vacation
17:00 est31 one release per year lol
17:00 Zeno` haven't seen him today
17:00 Zeno` maybe he's due back Aug 20
17:00 Calinou est31: still more active than Sauerbraten
17:00 Calinou its latst release was in 2013-01
17:00 Calinou and the last release before that was in 2010-07
17:01 Calinou you have no idea how inactive most open source games are :(
17:01 Zeno` also more frequent than A Song of Ice and Fire books
17:04 Krock Calinou, how inactive are they?
17:05 Calinou very?
17:05 Calinou https://sourceforge.net/p/sauerbraten/code/5309/log/?path=
17:07 Zeno` sourceforge? :-o
17:07 Krock Calinou, I see. It's good to have this active commiunty here :)
17:07 Calinou Zeno`: yep
17:07 Calinou eihrul doesn't like Git
17:07 Calinou only enet uses GitHub, the rest use SVN
17:07 Calinou (Tesseract/Sauerbraten)
17:07 Zeno` oh, yuck :(
17:11 Fixer sourceforge gui makes me :S :S :S
17:11 est31 yeah
17:11 est31 github really meant progress
17:11 Calinou $ourc€Forg€
17:16 Fixer maybe this release cycle is not cool for you?
17:16 Fixer once a year
17:23 Calinou nah, regular releases are important for an active community
17:24 Calinou ideally we'd make 4 releases a year
17:25 est31 but each release requires feature freeze
17:25 est31 and feature freeze means less fun
17:25 est31 (for the devs)
17:25 sofar well today is release day, right?
17:26 est31 sofar, thats what i've heard too
17:26 Calinou est31: feature freeze doesn't need to be very long
17:26 est31 be patient, the guy doing the release will show up
17:26 Calinou something went wrong? release a patch
17:26 Calinou this is why we should increment minor version rather than patch versions on the releases we're doing right now
17:27 sofar I'm all for setting a new release date soon
17:27 Calinou to not end up like Kernel Adiutor
17:27 est31 that would require us to leave the 0.4.13 like versioning scheme
17:27 est31 we cant have 0.4.13.1
17:27 est31 thats ugly
17:27 Calinou https://f-droid.org/repository/browse/?fdfilter=adiutor&fdid=com.grarak.kerneladiutor
17:27 Calinou 0.9.9.2.3
17:28 est31 you dont have to look far from minetest: https://f-droid.org/repository/browse/?fdid=org.freeminer.freeminer
17:28 est31 0.4.13.7.90
17:29 nore 0.rewrite.major.minor.patch?
17:29 nore (and mt is 0.rewrite.major :D)
17:29 est31 something like that yeah
17:29 est31 we should get rid of the leading 0
17:30 est31 I think I became a proponent of the 4.13 crew
17:30 nore I suggest adding a patch/minor number at the end then
17:30 nore 4.13 is fine with me, as is 0.4.13
17:30 Calinou why can't we have 0.5 already
17:30 Calinou I don't think Minetest is 1.0-ready, but still
17:30 Calinou (it's not far, it just needs some optimization)
17:30 est31 Calinou, because 0.5 means "we break network compat"
17:30 est31 and every compat
17:31 Calinou we'd break it on 0.6 instead
17:31 Calinou what's wrong with that?
17:31 Calinou we can go over 0.9 just fine, version numbers are not decimals
17:31 est31 I dont have anything against 0.5 its fine with me, as long as its only just a name
17:31 est31 and doesnt mean "we break compat"
17:32 nore same with me
17:33 est31 going now
17:34 Krock semver rules must not be followed
17:38 Zeno` maybe we *should* break compatibility
17:38 Zeno` it will be necessary at some point so why not next release?
17:38 Zeno` Forever putting it off makes no sense
17:39 Calinou good things break compatibility often.
17:39 Calinou it's how innovation is possible
17:39 Calinou it's a damn game, by the way.
17:39 nore Zeno`: I'd break at least network compatibility, and add a way to load old mapss
17:39 nore s/ss/s/
17:39 Calinou Sauerbraten breaks compat at every release, so does Red Eclipse
17:39 Calinou Xonotic chooses not to, it's how they stay with their shit engine taking all of the CPU :D
17:39 Calinou (but eventually they will switch engine)
17:39 Zeno` yeah, adding a way to load old maps would have to be done
17:39 Zeno` or at least convert old maps
17:40 nore yes, that's important
17:40 Calinou old map support is important, network compatibility is not
17:40 nore but two-way network compatibility should be dropped
17:41 Calinou network protocol seriously needs fixes IMO
17:41 Calinou it should be possible to host 50 players on my VPS without any lag
17:41 Calinou remember, we have only 10 ticks per second instead of 20 by default…
17:41 nore Calinou: network protocol needs fixes, map protocol needs too (I'd modify light to start with)
17:46 Zeno` Look at the mess Microsoft ended up in by trying to maintain compatibility with old DOS apps
17:46 Zeno` plus as Calinou said, this is a game
17:46 Zeno` not a mission critical life saving application
17:47 Zeno` There have been so many PRs and bug fixes that have been ignored because they break compatibility I've lost count
17:47 Zeno` It will have to happen one day if things are going to progress...
17:48 Zeno` The biggest risk is, really, that we break unofficial forks
17:49 Zeno` Every server operator that I know of, and every mod writer keeps up with HEAD anyway
17:49 nore Zeno`: we've already broken a lot
17:49 Zeno` nore, so let's smash it properly :)
17:50 Zeno` if there is a year between releases it doesn't matter so much anyway heheh
17:50 nore easy: strict_protocol_checking=true :p
17:51 Calinou if we could, we'd be rolling release \o/
17:51 Calinou too bad some people just hate that versioning scheme
17:51 Calinou it makes sense for games
17:54 Krock Real Life 1.0.0-beta
18:01 Fixer i like system when there is 4 releases a year with merges that don't break stuff and separate brunch for all new compatibility breaking features that will be in a new major version (maybe this is too complex)
18:06 Zeno` That didn't work very well last time we tried
18:11 Zeno` better would be having the separate branch being the one that doesn't break and master doing the breaking
18:11 Zeno` but even then you'd need a person fulltime working on the non-breaking branch
18:11 Zeno` and... why?
18:12 Fixer nevermind
18:12 Zeno` nah, it's worth voicing if you have an idea
18:15 Fixer Zeno`, yes, that what i was thinking main branch for breaking, and separate one without breakers for point releases
18:15 Fixer i guess it is too complex
18:17 Zeno` maybe
18:17 * nore wonders if paramat is around
18:17 Zeno` it seemed to be last time it was attempted anyway (although it was the other way around with master keeping compatibility)
18:18 nore he is supposed to release :)
18:18 nore Zeno`: the dev-0.5 branch was dropped
18:18 nore Too difficult to maintain
18:18 Zeno` nore, that's what I'm saying
18:18 Fixer postpone release for few days and let tidyman install newest git on Xanadu
18:19 Fixer and test it with lots of players :p
18:19 nore Fixer: well, if there are any problems, we can do a patch release in a few days
18:20 Zeno` we can't really test the client extensively without releasing anyway
18:40 paramat hi, just woke up. was exhausted. sorry
18:41 paramat let me look at github then i think we can release
18:41 nore paramat: no problem :)
18:41 nore yay, finally!
18:42 paramat hmmmm, did you test mgv6 flags after last night's changes?
18:46 paramat #4107 #4109 seem minor, not blockers
18:46 ShadowBot https://github.com/minetest/minetest/issues/4107 -- Clicking shader checkbox makes sub-entries “jump aroundâ€
18:46 ShadowBot https://github.com/minetest/minetest/issues/4109 -- Ubuntu devs should be installing git, not git-core
18:46 Zeno` I can't reproduce 4107
18:47 Zeno` even if I could it's not a blocker
18:47 paramat there's this but others seem to be able to build the apk https://github.com/minetest/minetest/pull/4086#issuecomment-218513511
18:48 paramat hang on, that may be with this patch applied
18:48 Zeno` doesn't matter does it. For release only we have to be able to build it don't we?
18:50 paramat pinged DonBatman to confirm details
18:51 paramat i assume his problem is with this patch applied, so good that we didn't merge it
18:52 paramat see project channel for discussion
18:53 paramat Wayward_One, any problems building apk?
18:53 paramat WO tested it recently so it should be ok
18:57 paramat in game there's game#1087 but is possibly minor
18:57 ShadowBot https://github.com/minetest/minetest_game/issues/1087 -- Page counter behaves weirdly after creative search
18:59 paramat ok so the apk build thing seems not an issue
19:01 paramat Wayward_One tested the apk 19hrs ago and it was ok
19:01 asl97 I can build it (after modifying the buildtools version), just some minor deprecated warning, i am using buildtools 23.0.2 though
19:02 paramat now i'll quickly check mgv6 flags
19:08 paramat the game issue is very minor and not a blocker (tested it)
19:10 paramat mgv6 flags are fine
19:13 paramat +1 for release then
19:20 paramat i'm looking at the 'releasing minetest' page (and can't do most of this myself) so first, 'define a new version number' please can anyone do this?
19:25 celeron55 really the first thing before that would be to check that builds work on all major platforms
19:25 celeron55 i guess it's fine but it's literally the only thing that can go wrong
19:26 paramat i've just built on linux and is fine
19:27 Calinou we need an OS X tester btw
19:27 Calinou I barely hear about Minetest on OS X…
19:29 nore well, we've got travis checking that mt builds on OS X
19:29 nore (and even failing thread unittests :/)
19:35 paramat linux clean build is fine, just the usual warning 'iso c forbids conversion of object pointer to function pointer type [-Wpedantic]'
19:37 Calinou have you tried with GCC 6?
19:37 * Calinou can
19:37 Calinou 6.1 to be accurate
19:44 sofar got some warnings
19:44 sofar with gcc-6.1.0
19:46 sofar compiles, though
19:47 sofar runs fine, too
19:48 paramat to be clear although i am 'release manager' all the technical steps of releasing are beyond me, sorry
19:48 sofar src/rollback.cpp:41:31: warning: throw will always call terminate() [-Wterminate]
19:48 sofar src/ltablib.c:137:3: warning: this 'if' clause does not guard... [-Wmisleading-indentation]
19:49 sofar src/loadlib.c:76:21: warning: ISO C forbids conversion of object pointer to function pointer type [-Wpedantic]
19:50 paramat the 3rd is a non-issue
19:51 paramat #4096
19:51 ShadowBot https://github.com/minetest/minetest/issues/4096 -- Compilation warnings in Lua
19:52 sofar yeah, it's fine to build with warnings
19:56 sofar the other two we can probably address soon anyway
19:58 * MillersMan can't wait for release to be done so he can bother the devs with code style questions
19:59 Fixer MillersMan, any progress on that liquids PRs? I did play with them and only minor issue is that liquid gup with rereflow on block load
19:59 Fixer gap*
20:00 MillersMan yes, I know the reason for the gaps and i think i can also improve the performance a little bit
20:00 MillersMan just a little short on time at the moment
20:30 sofar paramat: at least you can tag 0.4.14 in git ;)
20:34 Wayward_One paramat: testing now
20:34 paramat maybe, but perhaps best to wait until builds on all platforms are tested
20:34 paramat ok
20:36 Wayward_One paramat: the build was successful
20:36 paramat thanks
20:37 Wayward_One Np
20:37 paramat can anyone build on windows?
20:43 Fixer paramat, for testing?
20:43 paramat just to check it builds ok
20:56 hmmmm paramat: there were no other commits after mine. what do you mean?
20:59 paramat just to be careful i thought mgv6 flags should be tested, i did and they're fine
20:59 hmmmm ?? it was tested when i fixed it
20:59 hmmmm thing doesn't work
20:59 hmmmm i make change
20:59 hmmmm try it
20:59 hmmmm thing works
20:59 hmmmm is this not what everybody does?
21:00 paramat ok thought you had, just checking to be extra careful
21:00 hmmmm paramat, this is more relevant to you than others: https://github.com/kwolekr/minetest/commit/937ac688f659f6a414b3d002427613f7fce8902b
21:01 hmmmm if you ever find yourself creating a random vector, be aware of evaluation ordering problems
21:02 paramat i see, this could make pseudorandom non-repeatable
21:02 paramat deterministic
21:03 paramat ok will keep in mind
22:24 est31 hrmm I have a patch to speed up the android build
22:24 est31 idk whether it should go into 0.4.14
22:25 paramat not sure either
22:25 est31 lets add it later
22:25 est31 paramat, I want to test whether the f-droid build will work okay?
22:26 est31 before you do the release
22:28 paramat no problem go ahead
22:30 paramat i won't be organising the release myself, i'm completely unqualified, it's now up to others with the experience and knowledge
22:31 paramat but of course i can help out with simple stuff
22:31 est31 you can do the util/bump_version.sh
22:31 est31 and tag mtgame
22:33 est31 (dont do util/bump_version.sh until you really are sure thats the git commit you want to have for the release)
22:33 est31 same for tagging mtgame
22:35 celeron55 it's a bit of an annoying part really
22:37 celeron55 i see that we still don't have validation on whether the windows build works
22:39 celeron55 travis confirms 32-bit and 64-bit mingw though, maybe it's enough
22:40 Fixer cross-compiled on debian with mingw64 for windows also works for me
22:41 paramat best i don't attempt util/bump_version myself, no idea what i would be doing. tagging is probably the only thing i am capable of. sorry, i have very little ability outside mapgen
22:43 est31 no problem
22:43 sofar eh, just look at previous release commits and repeat?
22:44 est31 thats what util/bump_version.sh is about
22:44 est31 sofar, you can try running it locally
22:44 est31 the only side effects it has is your local git repo
22:44 sofar New version: 0.4.14
22:44 sofar New android version code: 14
22:45 sofar wow that's too easy
22:45 sofar paramat: run it, hit enter 4x, push it
22:47 sofar paramat: at least try it in a separate clone or something :)
22:49 est31 oh oh
22:49 est31 we need to update openssl
22:49 est31 1.0.2h is out
22:49 sofar on the server?
22:50 est31 openssl for android
22:50 est31 single liner
22:50 sofar quick PR?
22:51 sofar minetest_game also needs tagging, I assume
22:52 paramat well i can attempt that
22:52 sofar ah I see, build/android/Makefile, I assume
22:56 est31 https://github.com/est31/minetest/commit/a9b5c2662fca2e38ccafe2bd8be675ef0a98ffdc
22:56 est31 PTAL
22:56 sofar +1, trivial, merge ;)
22:57 Fixer https://github.com/minetest/minetest_game/issues/1087
22:58 est31 sofar, I have to check first whether it builds
22:59 est31 I wont test it before pushing to master, as its very unlikely for stuff to fail
22:59 est31 and i will test later anyway
22:59 est31 with the f-droid build
22:59 sofar doesn't the integration test that part?
22:59 est31 ?
23:00 sofar doesn't travis build it with the new openssl version?
23:01 est31 ah
23:01 est31 yeah i could make a pr to test it
23:01 est31 but I test it locally already
23:13 est31 okay pushing it, 10 minutes after I proposed it
23:14 est31 > 10 mins
23:14 Fixer to f-droid?
23:14 est31 no to master
23:21 est31 okay the next f-droid build is now running
23:21 est31 the first failed and i dont really want to find out why
23:21 est31 if there is one build that produces a working apk i am happy
23:24 est31 still building
23:40 est31 aha
23:40 est31 seems its some server being slow
23:42 est31 hrmm its curl
23:55 est31 its hopefully not what i think it is
23:56 est31 if you download curl sources using wget, they have metered it?