Minetest logo

IRC log for #minetest-dev, 2015-02-21

| Channels | #minetest-dev index | Today | | Google Search | Plaintext

All times shown according to UTC.

Time Nick Message
00:07 Taoki https://github.com/minetest/minetest_game/pull/434 Made a few changes to the default textures, if anyone's interested to mention an opinion.
00:14 zat joined #minetest-dev
00:25 shadowzone joined #minetest-dev
00:34 chrisf joined #minetest-dev
01:19 Taoki https://github.com/minetest/minetest_game/pull/436 Footstep sounds for liquids (swimming), feel free to test and mention your opinion.
01:19 paramat now merging 2175 .. =/
01:33 Wuzzy joined #minetest-dev
01:34 Wuzzy Is it OK to call updatepo.sh first when working on a translation update?
01:54 paramat phew okay 2175 seems to be merged successfully
02:06 paramat MTGame team .. game#412 should now be merged to replace the blobs removed from mgv5
02:06 ShadowBot https://github.com/minetes​t/minetest_game/issues/412 -- mapgen.lua: Add dirt/sand/gravel blobs to mgv5/mgv7. Use blob ore for clay. Tune flowers by paramat
02:10 Wuzzy2 joined #minetest-dev
02:10 shadowzone joined #minetest-dev
02:11 shadowzone joined #minetest-dev
02:21 domtron joined #minetest-dev
02:46 paramat left #minetest-dev
02:48 VanessaE joined #minetest-dev
02:54 bit_shifter_ joined #minetest-dev
02:57 AnotherBrick joined #minetest-dev
02:58 bit_shifter_ I've compiled the ogl-es branch of libIrrlicht, and compiled the latest minetest (0.4.11-242-g109f0ff) with -DENABLE_GLES=1. When I run it with "video_driver = ogles1" (or ogles2) in the conf file, I get "Irrlicht log: Fatal error, could not get visual."  Here's my full output: http://sprunge.us/ATFU.  Thoughts?
02:58 bit_shifter_ This is on an odroid-c1 with a mali-450 MP GPU.
03:04 ShadowNinja bit_shifter_: Do you have X11 and OGLES set up?  Is the DISPLAY env variable set?
03:09 bit_shifter_ ShadowNinja: Yes to X11 and OGLES.  glmark2-es2 runs just fine, so I assume OGLES is fine.  It lists my GL_VERSION as OpenGL ES 2.0.  DISPLAY is set to :0 -- I'm running LXDE as we speak.  Minetest runs with "opengl" as the video_driver, but the framerate is horrible as I don't believe I have a full OpenGL stack.
03:11 bit_shifter_ OpenGL ES is fairly new to me, so forgive me if I'm mistaken on anything I've said.
03:14 ShadowNinja bit_shifter_: Hmmm, I don't know then, and all the devs I know that are experienced with OGLES haven't been around for a while.
03:18 bit_shifter_ Doh.  I might have to start digging around in the source then when I get some time.  Though it could just as well be an irrlicht issue, I suppose.
03:19 ShadowNinja bit_shifter_: It works in the Android builds.
03:22 bit_shifter_ Hmm... is it built with the "ogl-es" branch of irrlicht?  Info about the state of OGLES in any branch of irrlicht seems sparse.
03:24 ShadowNinja bit_shifter_: No idea.  Check build/android/ (aka the directory that broke my tab-complete for builtin :-P)
03:25 est31 ?
03:25 est31 ah
03:25 est31 lol:D
03:44 bit_shifter_ Well, for better or worse, it seems to be an issue lower than minetest.  I tried to compile and run an irrlicht "example" that comes with the source, and it gives me the same error.
03:46 Zeno` joined #minetest-dev
04:59 OldCoder joined #minetest-dev
06:35 Weedy joined #minetest-dev
06:35 Weedy joined #minetest-dev
07:46 JakubVanek joined #minetest-dev
07:48 decimalguy joined #minetest-dev
08:07 Amaz joined #minetest-dev
08:31 MattJ joined #minetest-dev
08:31 sfan5 bit_shifter_: did you try ogles2?
08:32 Zeno` joined #minetest-dev
08:52 kilbith joined #minetest-dev
09:02 Krock joined #minetest-dev
09:04 nore joined #minetest-dev
09:18 Hunterz joined #minetest-dev
09:24 Zeno` 300 more assert()s to check
09:24 * Zeno` looks at his desk and thinks it looks comfortable enough to lay his head on
09:26 Zeno` then again.. maybe I can leave BOGO_ASSERT there as an indicator of what needs checking and replacing
09:27 Zeno` no need for me to do it all... there are quite a few I'm unsure of anyway (i.e. are these proper use of assert or some kind of lazy error checking cases)
09:31 Zeno` I don't understand why so few people find the current usage of assert() revolting :(
09:32 Zeno` I hate the word "pattern" but it's an "anti-pattern" and goes against all software design/implementation/style recommendations ever invented
09:34 gravgun joined #minetest-dev
09:36 Fritigern joined #minetest-dev
09:38 ImQ009 joined #minetest-dev
09:41 nrzkt joined #minetest-dev
09:41 nrzkt hi all :)
09:41 Krock testing #2310 now
09:41 ShadowBot https://github.com/minetest/minetest/issues/2310 -- Update zlib support by SmallJoker
09:44 nrzkt Zeno`, if i understand assert code will be removed from binary when compiling in release mode, yes ?
09:44 Zeno` nrzkt, normally it should yes
09:45 Zeno` but there is currently a hack to make it remain in release builds
09:45 nrzkt okay, that could be good but may be difficult to track some crashs :s
09:45 Zeno` and that hack is what I'd like to get rid of
09:45 nrzkt you're right. The wind has past in australia ?
09:47 Zeno` assert() is a no-op if NDEBUG is defined. All the BOGO_ASSERT need to be checked to see if they're being used properly. If they are used as assert() is meant to be used then they can be changed to assert. If they're being used as some kind of error checking then FATAL_ERROR or FATAL_ERROR_IF should be used
09:47 Zeno` yeah, wind has abated :)
09:47 Zeno` and I didn't even lose power!
09:48 Zeno` if there is a crash by removing an assert then the assert is not being used properly :P
09:48 Zeno` brb, profiling something (again)
09:49 nrzkt okay, that's good
09:49 Zeno` BOGO_ASSERT can stay... I might add a #warning deprecated though
09:53 BlockMen joined #minetest-dev
09:58 proller joined #minetest-dev
09:59 Zeno` I was going to merge #2242
09:59 ShadowBot https://github.com/minetest/minetest/issues/2242 -- Optimise MapBlockMesh functions by gregorycu
09:59 Zeno` but... it seems slower on linux :/
10:00 Zeno` well, there 2 profiles each of with and without I just did indicate that 2242 makes things slower
10:00 Zeno` hmm
10:01 Zeno` getTileInfo() is faster though
10:02 nrzkt if optimize don't optimize, please don't merge :s
10:04 Zeno` it's tricky to tell
10:05 nrzkt :)
10:05 Zeno` optimising it not just a matter of "making this faster"
10:05 Zeno` because of all the inter-connections/relationships
10:05 nrzkt right :)
10:05 Zeno` it's certainly not 33% as the PR claims hehe
10:07 Zeno` I'll keep testing
10:07 kilbith what about if reverting is considered and you'd have to discuss with gregorycu here ?
10:07 Zeno` kilbith?
10:07 Zeno` oh
10:07 Zeno` I just wouldn't discuss it with him
10:07 Zeno` lol
10:07 kilbith you can't merge that then
10:08 Zeno` at the moment (if profiles are correct) it's hardly worth merging anyway
10:09 Zeno` after (if) it's merged nothing needs discussion with the original author
10:11 Zeno` even though I think he's an idiot it would be remiss of me to not merge something worthwhile just because of my personal opinion of him
10:11 Zeno` but, it doesn't seem all *that* worthwhile atm anyway
10:13 nrzkt +1
10:13 Calinou joined #minetest-dev
10:14 nrzkt #2340 seems good for merge
10:14 ShadowBot https://github.com/minetest/minetest/issues/2340 -- Update Japanese Translation by Rui914
10:15 nrzkt someonee agree ?
10:16 Zeno` I don't speak Japanese
10:16 Zeno` :D
10:16 Zeno` But rui has done all the other translations I think
10:16 Zeno` +1 for merge'
10:17 Zeno` I am not merging 2242 now. There is hardly any difference (micro-optimisation). Perhaps the compiler is optimising it anyway
10:21 SudoAptGetPlay joined #minetest-dev
10:24 nrzkt do we close #1140 without merge ?
10:24 ShadowBot https://github.com/minetest/minetest/issues/1140 -- Fix pointers casting by Selat
10:24 Zeno` yeah
10:24 Zeno` RTTI is not good for performance
10:25 Zeno` we must pick that kind of stuff up while programming, not relying on RTTI
10:27 Zeno` stuff like that is a performance killer
10:35 Zeno` ok, 2242 should be merged I think
10:35 Zeno` I'll change his bogus commit message though
10:35 Zeno` nrzkt?
10:35 Zeno` I am not reviewing any of his other commits though
10:35 Zeno` lol
10:36 nrzkt there is a real change ? i haven't tested it
10:36 Zeno` slight. Maybe 3%
10:36 nrzkt it could be good. The code looks good for me
10:36 Zeno` logically the code is faster... the profiler disagrees
10:36 Zeno` so there is probably optimisation already going on, but *shrug*
10:36 nrzkt it look strange
10:37 Zeno` If he wasn't such an idiot it probably would have been merged long ago
10:38 Zeno` actually, I'll decide tomorrow
10:38 nrzkt okay
10:45 Zeno` there is something I am uneasy about and I dunno why
10:49 Zeno` just added comment
10:49 Zeno` if he doesn't reply I'll close the PR
11:05 kilbith joined #minetest-dev
11:18 SudoAptGetPlay joined #minetest-dev
11:24 proller joined #minetest-dev
11:39 est31 joined #minetest-dev
11:40 Anchakor_ joined #minetest-dev
11:43 Fritigern joined #minetest-dev
11:54 Fritigern joined #minetest-dev
11:55 b joined #minetest-dev
11:58 SopaXorzTaker joined #minetest-dev
12:11 Krock joined #minetest-dev
12:21 BlockMen comments on #2368 and #2367 ?
12:21 ShadowBot https://github.com/minetest/minetest/issues/2368 -- Fix particles draworder by BlockMen
12:21 ShadowBot https://github.com/minetest/minetest/issues/2367 -- Use skin font for usernames (fixes #2363) by BlockMen
12:29 Krock both are good
12:29 nrzkt okay for #2367
12:29 ShadowBot https://github.com/minetest/minetest/issues/2367 -- Use skin font for usernames (fixes #2363) by BlockMen
12:33 BlockMen nrzkt, whats with #2368?
12:33 ShadowBot https://github.com/minetest/minetest/issues/2368 -- Fix particles draworder by BlockMen
12:40 nrzkt i need to test it, wait a minute
12:43 SudoAptGetPlay joined #minetest-dev
12:49 nrzkt i'll test 2368 now
12:53 nrzkt what are the torch you are using ?
12:55 BlockMen you dont need the torches mod, there are builtin ways to get particles
12:56 BlockMen e.g. digging -.-
12:56 nrzkt i test different particles, like TNT particles, digging, etc... it's okay for me
12:59 nrzkt BlockMen stop using merge function on github...
12:59 nrzkt please use a proper git -am patch like everybody. It's on the devel guidelines
13:00 BlockMen github just said "merged" when the commit hash is the same as in the PR's branch
13:00 BlockMen i NEVER used merge function of github here!
13:01 nrzkt you suppress one of my commits
13:01 MinetestForFun joined #minetest-dev
13:02 nrzkt https://github.com/nerzhul/minetest/commit​/e9231cf76b631bf287cb86a97adb54dc09164efd
13:02 nrzkt which was on master before your merge
13:04 nrzkt if push on minetest branch doesn't work don't do a force, this will suppress commits...
13:04 nrzkt revert your patches to a common commit, then update from minetest repository and re-apply your patches
13:05 Krock it has been commited now..
13:05 nrzkt yes
13:05 nrzkt i recommited it
13:06 nrzkt but this commit was done after CMakeList cleanup, and before BlockMen merge, 1 hour ago
13:06 nore nrzkt: are you sure you did push the changes?
13:06 nrzkt i'm sure
13:06 nore because MinetestBot didn't report it on the #minetest channel
13:06 nrzkt oO
13:06 nore anyway, it is fixed now :)
13:07 Krock BlockMen, could you test this change please? https://github.com/SmallJok​er/minetest/commit/61cadb7 I successfully compiled it but I'm not sure if it works on onther systems aswell
13:07 BlockMen nrzkt, no. i did NO force push
13:07 BlockMen stop telling shit!
13:07 nrzkt then the problem is fixed, no problem :)
13:07 nore calm down both of you :)
13:08 nrzkt i'm calm ^^ it's the week-end, oh yeah :p
13:09 BlockMen nrzkt, btw. TELL here before pushing...
13:09 BlockMen its in general stupid to just push
13:10 nrzkt if you want, no problem :)
13:10 BlockMen good :)
13:12 nrzkt what we do for #2357 ?
13:12 ShadowBot https://github.com/minetest/minetest/issues/2357 -- Make git version discovery work on master by est31
13:13 BlockMen im pushing #2368 now
13:13 ShadowBot https://github.com/minetest/minetest/issues/2368 -- Fix particles draworder by BlockMen
13:16 BlockMen Krock, will look at it later, ok?
13:16 Krock ok.
13:17 nrzkt i will do a commit to move hex.h to util/ folder
13:18 nrzkt hmmmm later in fact :)
13:19 BlockMen nrzkt, you dont have to say your plans, just right before you push. and not even what, just that ;)
13:23 BlockMen nore, opinions on game#429, game#433, game#435 plz
13:23 ShadowBot https://github.com/minetes​t/minetest_game/issues/429 -- Add reverse recipes for hoes v4 by MT-Modder
13:23 ShadowBot https://github.com/minetes​t/minetest_game/issues/433 -- Update bucket mods init.lua by tenplus1
13:23 ShadowBot https://github.com/minetes​t/minetest_game/issues/435 -- Add gunpowder trails by Wuzzy2
13:23 domtron joined #minetest-dev
13:23 BlockMen also i guess game#
13:24 BlockMen game#412 is ok for you (see your comment)?
13:24 ShadowBot https://github.com/minetes​t/minetest_game/issues/412 -- mapgen.lua: Add dirt/sand/gravel blobs to mgv5/mgv7. Use blob ore for clay. Tune flowers by paramat
13:28 nore yep
13:29 nore 429 looked good last time (I did not check the code again though)
13:29 nore about gunpowder: how does connect_to_raillike work?
13:32 nrzkt ok, pushing the hex.h move now
13:33 nrzkt good, PR number is decreasing
13:40 BlockMen nore, https://github.com/minetest/minetes​t/blob/master/doc/lua_api.txt#L1039
13:40 BlockMen needs to be removed from PR, indeed
13:41 BlockMen well, no
13:41 BlockMen its needed for burning gunpowder
13:42 nore BlockMen: then we need to add something that will connect those nodes only if the value for that group is the same
13:42 nore and we shouldn't merge that pr before then
13:45 DFeniks joined #minetest-dev
13:46 BlockMen nore, we can merge it since rails are not in that group so it does not connect to rails: https://github.com/minetest/minetest/blo​b/master/src/content_mapblock.cpp#L1475
13:47 BlockMen but in general, yes, it should be the a value check and not != 0
13:47 nore there are mods relying on rails being in that group
13:47 nore and btw: how simple is it to change that?
13:47 nore this looks a quite trivial change
13:47 BlockMen easy. but it would be a visual glitch on older clients ofc
13:48 nore that's not a problem
13:48 nore there are visual glitches for lots of things like that too
13:48 nore what do you think happens with meshnodes?
13:49 BlockMen i will label it with "needs engine-change" for now
13:49 nore ok
13:49 BlockMen maybe i will do it later
13:50 nore thanks
13:54 BlockMen is it ok if i add "Notify others on #minetest-dev before pushing upstream" to http://dev.minetest.net/inde​x.php?title=Git_Guidelines?
14:07 kilbith BlockMen, should i close game#415 ?
14:07 ShadowBot https://github.com/minetes​t/minetest_game/issues/415 -- Optimize Sam II + textures by kilbith2
14:17 BlockMen kilbith, leave it open for now. it has been open and closed enough already
14:18 BlockMen joined #minetest-dev
14:18 cheapie joined #minetest-dev
14:20 twoelk joined #minetest-dev
14:22 SopaXorzTaker joined #minetest-dev
14:24 nrzkt okay BlockMen for your modification
14:27 Zeno` joined #minetest-dev
14:29 Zeno` BlockMen, you don't think the font is too large now?
14:31 Zeno` does anyone have an issue if I close #2320?
14:31 ShadowBot https://github.com/minetest/minetest/issues/2320 -- Fix superflous shader setting updates by gregorycu
14:31 Zeno` author does not respond to questions and the silly thing doesn't build anyway
14:32 nrzkt don't close it
14:32 nrzkt the rule is 1 month without answer
14:32 Zeno` why not?
14:32 Zeno` nah
14:32 Zeno` If it doesn't build and they don't answer questions they can make another PR
14:33 Zeno` a week is long enough
14:33 Zeno` if it actually built then that would be another matter
14:33 nrzkt maybe but c55 has defined it
14:33 Zeno` defined what?
14:34 nrzkt If a pull request or an issue does not get a response from its author in one month (when requiring more details), it is closed.  => http://dev.minetest.net/Git_Guidelines
14:34 Zeno` that we have to keep PRs that do not build?
14:34 nrzkt in one month he can fix it :)
14:34 Zeno` It hasn't built from day one though
14:34 Zeno` fuck it... I'm closing it
14:35 SudoAptGetPlay We got a basass over here :D
14:35 nrzkt as you want, it's your responsability and i don't have opinion on this PR
14:36 Zeno` seriously... if someone makes a PR that *does not build* there is no point in waiting more than a few hours :/
14:36 BlockMen Zeno`, let wait for feedback. if many complain we can still revert it ;) but it general its good to use ttf aswell (if present)
14:37 Zeno` BlockMen, yep
14:37 SudoAptGetPlay Well I like the new font
14:37 nrzkt Zeno` we need to talk with celeron55 about it, maybe
14:37 SudoAptGetPlay No problems under ubuntu
14:37 Zeno` nrzkt, what for? :)
14:37 Zeno` it didn't build... 6 days to fix... no progress... end of story :p
14:38 BlockMen Zeno`, but if a patch is boosting the performnce a lot (and it would be not much work to fix) it woud be stupid to let it flow with time
14:38 BlockMen idk if this patch does ^
14:38 Zeno` BlockMen, I dunno if it boosts performance because it doesn't build
14:38 Zeno` cannot test
14:39 Zeno` ok, I'll open it
14:40 Zeno` 40% is bogus
14:40 Zeno` it's like his last PR that claimed 33% which was more like 1%
14:41 shadowzone joined #minetest-dev
14:41 nrzkt tested #2325 it's good for me
14:41 ShadowBot https://github.com/minetest/minetest/issues/2325 -- Add camera smoothing and cinematic mode (F8) by rubenwardy
14:42 nrzkt do you agree for merge ?
14:43 Krock he said, there are still problems
14:44 Zeno` nrzkt, you do know that celeron55 has retired from the project, don't you?
14:45 nrzkt right Zeno` yes, from development, but management no
14:46 Zeno` nope, he has said many times he is taking a "hands off" approach
14:47 nrzkt ofc, but for some things we need an impartial cut, and he can have this role :)
14:47 Zeno` maybe he's changed his mind?
14:47 Zeno` #2309
14:47 ShadowBot https://github.com/minetest/minetest/issues/2309 -- Fix superflous shader setting updates by gregorycu
14:47 Zeno` this is just a mess
14:47 Zeno` now it's #2309
14:47 ShadowBot https://github.com/minetest/minetest/issues/2309 -- Fix superflous shader setting updates by gregorycu
14:48 Zeno` and it STILL does not build
14:48 Zeno` sorry #2320
14:48 ShadowBot https://github.com/minetest/minetest/issues/2320 -- Fix superflous shader setting updates by gregorycu
14:49 nrzkt i think he codes without a compiler on an ipad :p
14:49 Zeno` this is the same PR but, of course, he ignores how to fix it
14:49 Zeno` it's just clutter that will not be fixed by him
14:50 Zeno` he thinks gcc and clang are broken (look at the logs)
14:50 nrzkt lol !
14:50 Zeno` I'm serious :p
14:50 Zeno` it's in the logs
14:50 SudoAptGetPlay hehe you defintely have something with that guy ^^
14:50 Krock nrzkt, you're "voiced", can push stuff to the minetest repo but aren't member of the minetest group?
14:50 nrzkt which minetest group ?
14:51 Zeno` Krock, he is. He has his profile set to private though
14:51 Krock https://github.com/orgs/minetest/people
14:51 Krock Zeno`, oh, that's why
14:51 Zeno` SudoAptGetPlay, I don't really have an issue with him apart from him being an idiot
14:51 nrzkt i'm nerzhul
14:51 nrzkt :p
14:52 SudoAptGetPlay Zeno`, glad someone said it
14:52 nrzkt i think i'm an idiot too, but you talk to me :p
14:53 Zeno` nrzkt, you're a sane idiot
14:53 Zeno` like me
14:53 Zeno` well, maybe not sane
14:53 nrzkt oh Krock: i'm private on the team member, right
14:53 SudoAptGetPlay omg this project is led by fools ^^
14:53 nrzkt Krock: fixed
14:53 Krock nrzkt, yeah, now there are 13 members
14:53 nrzkt shit, we must kill one
14:54 nrzkt 13 is a bad number :p
14:54 SudoAptGetPlay Add a random peep, that ill be a round 14
14:54 Zeno` 13 is lucky
14:54 Zeno` it's prime for a start
14:54 Krock yeah, and then also add me for 15, just to make sure
14:54 Krock <.<
14:55 Zeno` prime numbers are the only real numbers
14:55 nrzkt next week, on saturday or sunday, i'll merge PR 3/4. We need to review related PR like #2261 because it's better to merge all protocol things together, whereas we should modify some new things later
14:55 ShadowBot https://github.com/minetest/minetest/issues/2261 -- damage_per_second negative number implementation by JakubVanek
14:55 Zeno` i don't mean real as in real though
14:55 Zeno` I mean real as in ... err, real
14:55 shadowzone What team?
14:56 Krock shadowzone, the A one.
14:56 shadowzone Oh.
14:57 SudoAptGetPlay So master won't be compatible with the majority of servers ?
14:57 Krock SudoAptGetPlay, no. the the majority of servers won't be compatible with master.
14:58 nrzkt yes, in fact all servers. master is not a stable branch, nobody must follow them for updating their servers
14:58 Krock I would follow, just for the lulz
14:58 blaze joined #minetest-dev
14:59 nrzkt if you want and doesnn't case about players. 0.4.12 client support a feature for testing protocol versions announces to master server, which permit to detect incompat. Older client won't then they don't know they can't connect to server
14:59 SudoAptGetPlay Well I'm used to regularly build master to get the latest features while their 're hot :p
14:59 nrzkt s/announces/announced
14:59 nrzkt those features will be hot, but not for stable clients :)
15:01 Zeno` should we add a compatibility layer?
15:01 SudoAptGetPlay Do we have a strategy for 0.5.0 ? like making the new network stack work, then release so that everyone quickly uses the new one and that I can go back to master ? ^^
15:01 nrzkt no Zeno
15:01 Zeno` nrzkt, most servers are going to lose 99% of their users
15:02 Krock ah I'm glad I've 0 players, there's nothing to lose :)
15:02 Zeno` lol Krock
15:03 nrzkt ofc Zeno` it's why 0.5.0 will be release in ~5 months, because we must fix many and many things, it's a major version bump :)
15:03 Zeno` I didn't realise there was a roadmap :(
15:03 nrzkt and we should have a good communication strategy, and servers admins too, because it's server admin problem to prevent those problem to their users
15:03 nrzkt no roadmap
15:04 nrzkt but i think we should think about a roadmap, because this protocol bump permit to cleanup and modify other things
15:04 nrzkt maybe add it on dev.minetest.net ? :)
15:04 Krock btw, is UTF-8 support also a 0.5.0 milestone?
15:05 nrzkt i think we should
15:05 nrzkt and i also think about fixing keyboard issue
15:06 nrzkt i think we should think about objectives for this 0.5.0 release, because we could break some MT bricks (except map compat)
15:08 Krock an other thing would be to allow modifying the x and z rotation of an entity
15:08 nrzkt Zeno`, nore, BlockMen, where can we write a draft to minetest 0.5.0 objectives before freeze ?
15:09 nore you can make an issue for that
15:12 nrzkt this could be good, have you got some ideas ?
15:15 nrzkt #2370
15:15 ShadowBot https://github.com/minetest/minetest/issues/2370 -- [roadmap] Minetest 0.5.0 objectives
15:16 SopaXorzTaker joined #minetest-dev
15:21 ElectronLibre joined #minetest-dev
15:22 nrzkt ok, i posted it https://forum.minetest.net/viewtopic.ph​p?f=3&amp;t=11288&amp;p=170727#p170727
15:22 Zeno` It's a start at least ;)
15:23 nrzkt we need to communicate with our users :)
15:23 Krock nrzkt, s/make -j2/make -j<num cores + 1>/
15:23 Zeno` How do I create a new label?
15:24 Krock Zeno`, here? https://github.com/minetest/minetest/labels
15:24 nrzkt if you want :p
15:24 nrzkt choose a beautiful color :)
15:24 nrzkt but don't add label for <stupid-user>
15:24 nrzkt :p
15:24 Krock :(
15:24 Zeno` Krock, there is no "add"
15:25 Krock Zeno`, on my private repos,t here's a button "Add label"
15:25 Krock *"New label"
15:25 BlockMen nrzkt, i think an issue and the forum post ist good for the comunication, but a dev wiki article should be done aswell for "official reference" where only the "decided" things are written down
15:25 Zeno` I don't see a "new label" :(
15:26 Fritigern joined #minetest-dev
15:26 Krock maybe limited to super users :P
15:26 nrzkt forum is used by more users than dev wiki
15:26 nrzkt it's not for the roadmap, it's for communicate users that a breaking PR will be merged in 1 week
15:26 BlockMen Zeno`, top right, a big green button "New label"
15:26 Zeno` Krock, so I have to MITM it?
15:27 Zeno` oh, I see now
15:27 Zeno` thanks BlockMen
15:27 BlockMen yw :D
15:27 Zeno` MITM seemed more exciting
15:27 Krock *cough* yeah, it's hard to see with closed eyes
15:27 nrzkt no need if you buy a Lenovo :p
15:28 BlockMen nrzkt, ik. i ment as addition, to keep an overview. IMO its hard in discussion areas (only)
15:28 nrzkt ofc
15:28 * Zeno` added Idiot label
15:28 Zeno` lmao... j/k
15:29 DFeniks joined #minetest-dev
15:33 domtron joined #minetest-dev
15:33 Zeno` #2635 is a bit of a worry
15:33 ShadowBot Zeno`: Error: Delemiter not found in "HTTP Error 404: Not Found"
15:33 Zeno` err, #2365
15:33 ShadowBot https://github.com/minetest/minetest/issues/2365 -- minetest.(de)serialize no longer persists stacked tables properly
15:37 SopaXorzTaker joined #minetest-dev
15:47 T4im joined #minetest-dev
15:51 SopaXorzTaker joined #minetest-dev
15:54 SopaXorzTaker joined #minetest-dev
15:59 nrzkt joined #minetest-dev
15:59 nrzkt BlockMen: can you remove https://forum.minetest.net/v​iewtopic.php?f=3&amp;t=11288 i don't have moderation rights, please
16:01 BlockMen sfan5^
16:02 BlockMen nrzkt, i dont have either, i can just access the admin panel ;)
16:03 domtron joined #minetest-dev
16:04 sfan5 BlockMen: sfan5^ does not highlight me, use sfan5 ^
16:04 BlockMen ups
16:04 sfan5 i removed the topic
16:06 monty joined #minetest-dev
16:07 nrzkt thanks sfan5
16:10 shadowzone joined #minetest-dev
16:21 nrzkt i find some memory leaks on networking PR2/4, i'll test it and push it to master branch soon
16:24 monty joined #minetest-dev
16:25 nrzkt somes are stupid, and nobody see thems when review :(
16:26 Wayward_One joined #minetest-dev
16:27 SopaXorzTaker joined #minetest-dev
16:27 nrzkt i merge it now
16:32 decimalguy left #minetest-dev
16:33 nrzkt ShadowNinja: do we need #1062 ?
16:33 ShadowBot https://github.com/minetest/minetest/issues/1062 -- Better mod naming. by kaeza
16:38 nrzkt i updated for android build too... i think we should find a way to build android as the same way as principal builds :(
16:39 monty joined #minetest-dev
16:40 domtron joined #minetest-dev
16:55 ShadowNinja nrzkt: It's good, provided that the special behavior with dashes is removed.
16:56 nrzkt ShadowNinja: we can close the issue, then ?
16:57 ShadowNinja nrzkt: No, modify it and merge it instead.
17:00 nrzkt Then i remove the spec.name modification when find a dash and keep the readConfigFile for mod.conf, right ? and the info.exist("name") ?
17:00 ShadowNinja nrzkt: Send() shouldn't delete it's passed packet pointer.  You should pass a reference and allocate on the stack instead.  That's faster and it's not prone to memory leaks like you just fixed.
17:00 ShadowNinja nrzkt: Yes.
17:05 Calinou joined #minetest-dev
17:05 nrzkt https://github.com/nerzhul/minetest/commit​/b3292556e2f77ce709e621a596fcade508642d0d: ok for merge ?
17:06 nrzkt ShadowNinja: which Send function ? Connection::Send or ClientIface::Send or Client::Send ?
17:07 Krock nrzkt, >ovide a 'mod_info.txt' file,
17:07 Krock >IM+"mod.conf"
17:07 shadowzone joined #minetest-dev
17:07 nrzkt and you mean it's better if i remove all NetworkPacket* pkt to use NetworkPacket pkt and do Send(&pkt), right ?
17:07 nrzkt right Krock :)
17:08 nrzkt Krock: we use mod.conf or mod_info.txt ?
17:08 Krock if you use it for minetest/mod settings, then it's .conf
17:08 nrzkt at this moment it provide the modname
17:09 ShadowNinja nrzkt: Yes but Send(pkt) with a const NetworkPacket & param.
17:09 ShadowNinja nrzkt: Not sure which one, but I think there's a server one too...
17:09 domtron joined #minetest-dev
17:10 ShadowNinja Any one that does that actually/
17:10 Megaf joined #minetest-dev
17:10 ShadowNinja nrzkt: It should be mod.conf
17:10 nrzkt ok ShadowNinja. You are right it's a good approach. Then i will fix it before PR 3/4 merge, maybe tomorrow
17:11 nrzkt ShadowNinja: ok the commit message is fixed. If you agree, i merge
17:11 ShadowNinja nrzkt: Change the commit message to "Add mod.conf with name field" or similar.
17:11 VanessaE I'd like to ask that one last stable-0.4 release be made before 3/4 goes in.
17:12 VanessaE (considering that a few useful things have already gone in since 0.4.12)
17:14 nrzkt ShadowNinja: fixed
17:15 nrzkt [Mods] Add mod.conf file support  and comment: This new file only permits to specify a mod name for now
17:15 nrzkt ShadowNinja: okay ? :)
17:16 ShadowNinja nrzkt: No [Mods] part.
17:16 nrzkt ok i remove it, that's all ? :p
17:17 ShadowNinja nrzkt: And by comment you mean the description of the commit, right?
17:17 nrzkt ofc
17:17 ShadowNinja Tis O.K. then.
17:17 nrzkt perfect
17:19 nrzkt ShadowNinja: what about #2371 ?
17:19 ShadowBot https://github.com/minetest/minetest/issues/2371 -- Improve FindIrrlicht.cmake module by apoleon
17:20 ShadowNinja nrzkt: Seems good.
17:20 nrzkt i also think #2136 could be great too, if a rebase is done and the patch is okay, we merge ?
17:20 ShadowBot https://github.com/minetest/minetest/issues/2136 -- Add Colored chat for Freetype builds, ignored on Non-Freetype builds. by TriBlade9
17:22 ShadowNinja nrzkt: It should have a reset token, but if that's added it's O.K.
17:23 Krock I would love a 3-character long color definition, not just RRGGBB
17:23 Krock wait no. nvm
17:26 roniz joined #minetest-dev
17:32 Wuzzy joined #minetest-dev
17:32 Wuzzy Hey, is it OK to run updatepo.sh first before updating a certain translation?
17:34 monty joined #minetest-dev
17:37 SopaXorzTaker joined #minetest-dev
17:38 ElectronLibre joined #minetest-dev
17:39 shadowzone joined #minetest-dev
17:48 monty joined #minetest-dev
17:48 leat joined #minetest-dev
17:50 BlockMen_ joined #minetest-dev
17:53 paramat joined #minetest-dev
17:59 paramat sfan5 BlockMen_ please can game#412 be reviewed/merged?
17:59 ShadowBot https://github.com/minetes​t/minetest_game/issues/412 -- mapgen.lua: Add dirt/sand/gravel blobs to mgv5/mgv7. Use blob ore for clay. Tune flowers by paramat
17:59 BlockMen_ yes. nore and I agreed on
18:00 BlockMen_ feel free to push this commit yourself, else i will do later ;)
18:04 monty joined #minetest-dev
18:06 paramat okay will push it myself, thanks
18:06 ShadowNinja Wuzzy: Yes, as long as you only commit the changes to the translation that you update.
18:07 ShadowNinja Woah, someone screwed up the menu.
18:09 ShadowNinja http://i.imgur.com/IxSK7ke.png
18:10 kilbith ShadowNinja, https://github.com/minetest/minetest/commi​t/4d744cf87ad1cd108d125c58559cef4316e80a70
18:10 ShadowNinja Also, game icons are very low-res.
18:11 Wuzzy SVG support, maybe? :P
18:11 ShadowNinja celeron55: The menu's huge now, it was fine before...
18:12 ShadowNinja Wuzzy: Probably not.
18:12 VanessaE ShadowNinja: it's set to fixed-scaling now, but it needs shrunk to 84 percent of its current size.
18:13 paramat i merged a PR for the first time last night, but was uncomfortable just using 'git push' for automatic branch choosing, i assume i use 'git push origin master' instead?
18:15 est31 joined #minetest-dev
18:16 domtron joined #minetest-dev
18:28 rubenwardy joined #minetest-dev
18:28 BlockMen paramat, yes. but better rename your remote to upstream
18:30 paramat okay
18:37 monty joined #minetest-dev
18:43 domtron joined #minetest-dev
18:47 ShadowNinja paramat: Make sure you have push.default set to simple.  And origin should be your fork, pushing to upstream should be explicit.
18:48 T4im ShadowNinja: why not use %f instead?
18:48 rubenwardy2 joined #minetest-dev
18:48 T4im with regatd to #2365
18:48 ShadowBot https://github.com/minetest/minetest/issues/2365 -- minetest.serialize doesnt serialize floating-point numbers properly
18:48 T4im regard*
18:49 paramat okay
18:50 rubenwardy joined #minetest-dev
18:50 est31 T4im, looks like he has already pushed it to master. and its more than 10 mins ago, so per c55 official dev wiki page that can only be changed by separate commit.
18:50 Wayward_One joined #minetest-dev
18:50 T4im so what? :)
18:51 * est31 shrugs
18:51 est31 :)
18:52 shadowzone joined #minetest-dev
18:52 T4im tostring still has weirdness when it comes to large numbers.. %f might solve a few other corner cases we might not see yet, and worse, might change its implementation of number representation
18:53 est31 ah ok, then its actually having an effect, thought it was only to have a nicer format
18:53 T4im I mean tostring might change its impl*
18:53 T4im format adheres to some form of standard I think, while tostring implies not to
18:54 * T4im updates the unit test to check for the floating point in tables thoug
18:56 est31 shouldnt you rather add an unit test than modify existing ones?
18:56 est31 I mean this way you are losing testcases :D
18:56 T4im well yes :D
18:58 T4im well it does fail serializing 10001312123123113.1234
18:58 T4im still I mean
18:58 T4im not sure if %f is able to fix that one though
18:59 T4im oh sorry its not the one that fails
19:05 monty joined #minetest-dev
19:08 T4im alright.. tostring assumes a precision of 14 I think at max
19:09 T4im iirc then number with more than 14 digits where also scientifically notated, for decimal numbers it just rounds htem
19:09 T4im so the hybrid created a hybrid bug now :P
19:09 T4im %.19f should fix it (or is more precision possible?)
19:10 T4im it currently does stuff like: 1.0001312123123e+16
19:11 T4im so effectivly a regresseion of #2206 for floats now
19:11 ShadowBot https://github.com/minetest/minetest/issues/2206 -- minetest.serialize() is not handling large numeric indexes right leading to forceloading additional blocks
19:16 T4im then again, string format seems to not be able to handle that either... how I hate floats
19:19 monty joined #minetest-dev
19:22 ImQ009 joined #minetest-dev
19:29 SudoAptGetPlay joined #minetest-dev
19:38 monty joined #minetest-dev
19:38 domtron joined #minetest-dev
19:49 pitriss left #minetest-dev
19:52 rubenwardy_ joined #minetest-dev
19:55 Taoki So after many years since I implemented animated models (the essentials at least), I'm now investigating why animations are often not updated... for example a player still walks in place after having stopped.
19:55 Taoki What I'm finding is very bizarre. Apparently the client often receives animations ranges of 0 to 0, which the server never sends
19:55 Taoki Unless of course the player dies... but that's not the case here.
19:56 Taoki Somewhere lies a big bug of sorts...
19:57 Taoki It's either this or packet loss / bad backet order that leads to the problem though.
19:57 est31 nrzkt, 1133, 1106, 1638 are also unicode related issues, perhaps tag them, so that they dont get forgotten?
19:57 rubenwardy Would be interesting to restart server without disconnecting clients fully, you could do it by sending a "reconnect after 100ms" message to clients, which would freeze them then reconnect.
19:57 est31 or remove them, if they seem duplicate
19:58 rubenwardy Loads the game again, of course
19:58 est31 rubenwardy, great idea, like it :) in fact had it too
20:00 shadowzone joined #minetest-dev
20:06 Taoki Hmm, no... it is a different and wose issue; Packages triggered by some Lua functions seem to never get to clients at all sometimes. Including in singleplayer...
20:06 paramat left #minetest-dev
20:07 Calinou Taoki, is the animation packet sent as reliable or not?
20:08 Taoki Calinou: How do I know for sure? It's sent like any other package
20:08 Taoki From what I remember when I implemented it that is
20:12 Taoki But I thought all packages should be reliable...
20:13 est31 rubenwardy, best would be to trap it with its own signal, so that you could restart minetest server instance from console with one ctrl+something
20:17 rubenwardy That's a good idea
20:17 rubenwardy Needs to support /restart as well
20:17 rubenwardy Can you send signals to minetest server like ctrl+something from bash/python?
20:18 est31 yes
20:18 est31 killall -SIGINT for example
20:18 Taoki Anyone know how to make sure that a package is reliable?
20:20 ShadowNinja nrzkt: Have you considered using SCTP for the protocol?  It supports ordered and unordered delivery, reliable and unreliable transfer, and is datagram-based: http://en.m.wikipedia.org/wiki/Str​eam_Control_Transmission_Protocol
20:20 est31 I think its a boolean value inside {Client, Server}CommandFactory listing
20:20 est31 Taoki, ^
20:21 ShadowNinja Seems like the perfect merging of TCP and UDP for out use case.
20:21 ShadowNinja our*
20:21 est31 Taoki, but ask nrzkt for details :)
20:21 Taoki est31: Can it be set for something in genericobject.cpp? In this case std::string gob_cmd_update_animation
20:21 Taoki Or can these not be reliable?
20:21 est31 whats the message name?
20:22 Taoki est31: GENERIC_CMD_SET_ANIMATION
20:22 nrzkt i'll look at it later, but it seems interesting
20:22 Taoki I need to set a few others as well, like attachments
20:22 Taoki But if I know how to do it for one I can do it for all.
20:23 Taoki I need to write a patch either way to fix several things with the animation system, so I'll add this myself.
20:25 est31 Taoki, it is sent as reliable
20:25 est31 as TOCLIENT_ACTIVE_OBJECT_MESSAGES to be precise
20:25 Taoki est31: Why does it sometimes fail to get there then? Even in singleplayer where there is no lag
20:25 est31 dunno
20:25 est31 :)
20:26 Taoki ok
20:26 monty joined #minetest-dev
20:26 Taoki Something is wrong somewhere then
20:30 nore joined #minetest-dev
20:42 Taoki What is the best way to read / write a boolean? Like for example, for floats it's readF1000
20:42 est31 floats have no best way to read them
20:43 est31 there is always a better one:
20:43 est31 )
20:43 est31 as you can halve every epsilon
20:43 est31 and strict checks are bad for floats :)
20:43 est31 ok that was bullshit, pls ignore :)
21:07 ShadowNinja Taoki: Use a flags field with writeU{8,16,32,64}.
21:08 ShadowNinja Taoki: Put multiple booleans in one byte with bitops to not waste space.
21:11 BlockMen left #minetest-dev
21:14 Taoki Interesting. lua_isnil returns true even if the value really is NIL
21:25 T4im joined #minetest-dev
21:27 T4im Taoki: that might be the reason for a handful of current issues
21:28 Taoki yeah
21:28 T4im #2015 and #2264 for example
21:28 ShadowBot https://github.com/minetest/minetest/issues/2015 -- minetest.dig_node(test_position) does not fail when expected -- always returning true
21:28 ShadowBot https://github.com/minetest/minetest/issues/2264 -- Settings(path):remove(key) always returns true
21:29 ShadowNinja Taoki: Not in Lua, only in LUA.  And should it return something else?
21:30 Taoki ShadowNinja: I'm experimenting with a function. Even if I leave the field blank, that check fails and a NULL gets through instead of the default value
21:30 Taoki Also I don't think anyone cares about case sensitivity in a chat nowadays
21:35 Taoki In either case, it looks like a serious bug exists
21:41 MattJ joined #minetest-dev
21:54 shadowzone joined #minetest-dev
22:11 CraigyDavi joined #minetest-dev
22:12 Taoki https://github.com/minetest/minetest/pull/2372 Important changes to the animation system, please look into
22:16 Calinou the death anim could be re-added
22:16 Taoki Sure. Though I never really made one, just the model in a death pose (which would be applied immediately)
22:17 Taoki This is something I should have done when I implemented animated models to begin with, I'm a few years late :P
22:18 Calinou you did make one in 2012
22:18 Taoki I mean the feature to support non-looped animations
22:19 Taoki Ah... I forgot then
22:19 Taoki Yeah, maybe it can be added back from there after this is merged.
22:26 nrzkt Taoki: your PR failed
22:27 Taoki nrzkt: It compiles and works fine here. Don't know what system Github uses to check these
22:27 nrzkt it's travis, please look at travis logs to resolve the issue
22:27 nrzkt i doesn't work on every platform
22:28 Taoki That makes no sense. I only added a new property like all the existing ones. Nothing that could possibly be cross-platform
22:28 Taoki I'll look, but that is seriously odd
22:30 Taoki Aaaah... I see what it is. I think it's using an outdated version of Irrlicht, which doesn't support loop control in animations.
22:32 nrzkt maybe you must check the irrlicht version then
22:35 ElectronLibre left #minetest-dev
22:36 Taoki I have Irrlicht 1.8.1. Which is way over an year old
22:36 Taoki Travis is using a really ancient version...
22:37 nrzkt 1.7 ?
22:37 Taoki 1.7 probably doesn't have it in that case.
22:37 Taoki In either case it sounds really old. Given that 1.8.1 was released in December 2013
22:41 Taoki Also, it probably means that this is why I didn't initially implement one-time animations. Back when I added animated models, Irrlicht 1.7 was the latest version. This function did not exist.
22:41 VanessaE I see no reason not to require 1.8 at this point.
22:42 Taoki Same. No need to keep supporting such an old version.
22:42 VanessaE exactly, especially if it's holding minetest back like that
22:54 domtron joined #minetest-dev
23:17 ShadowNinja nrzkt: Do you think I should lock #2370 so only devs can comment?  Seems like it's become a place for everyone to put their favorite feature request rather than a good discussion of goals.
23:17 ShadowBot https://github.com/minetest/minetest/issues/2370 -- [roadmap] Minetest 0.5.0 objectives
23:20 nrzkt i think we could let it for know, and pickup some interesting ideas
23:25 shadowzone joined #minetest-dev
23:26 nrzkt and if too many people comment, we close it :)
23:30 domtron joined #minetest-dev
23:42 domtron joined #minetest-dev
23:57 Miner_48er joined #minetest-dev
23:58 Taoki Ugh... why do some people still care about compatibility with Irrlicht 1.7? Apparently I'm being suggested to ignore a valuable bug fix or new feature, only so Minetest can stay compatible with this old version which no one should still have by now.
23:59 Wayward_One joined #minetest-dev
23:59 Taoki I can see innovation stalling if we hang onto ancient versions for every library...

| Channels | #minetest-dev index | Today | | Google Search | Plaintext