Minetest logo

IRC log for #minetest-dev, 2016-03-31

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

All times shown according to UTC.

Time Nick Message
00:02 jin_xi joined #minetest-dev
00:15 paramat closed #1998
00:15 ShadowBot https://github.com/minetest/minetest/issues/1998 -- Failed unit tests in Windows build
00:27 yang2003 joined #minetest-dev
00:28 DFeniks_ joined #minetest-dev
00:46 ssieb joined #minetest-dev
01:35 paramat closed #2513
01:35 ShadowBot https://github.com/minetest/minetest/issues/2513 -- Registering Nodes while minetest is playing.
01:44 gregorycu joined #minetest-dev
01:44 gregorycu paramat: Hello there
01:45 gregorycu That was a FYI more than anything
01:45 gregorycu Couldn't sleep last night so I took the day off work
01:45 paramat hi
01:46 gregorycu Did you happen to see that ContentFeatures PR I did?
01:46 paramat yeah
01:46 gregorycu Some idiot was copying ContentFeatures all over the place
01:46 gregorycu I tease, I tease :)
01:46 paramat heh
01:47 gregorycu Do you get stuttery frames?
01:47 paramat yes
01:47 gregorycu I have an idea about that, gunna try it out
01:55 paramat closed #2557
01:55 ShadowBot https://github.com/minetest/minetest/issues/2557 -- Wielding items that use minetest.inventorycube causes an error
01:59 paramat closed #2572
01:59 ShadowBot https://github.com/minetest/minetest/issues/2572 -- ABMs are not happening fast enough.
02:06 est31 joined #minetest-dev
02:07 est31 paramat, you may want to close #1531
02:07 ShadowBot https://github.com/minetest/minetest/issues/1531 -- Be Able to Load Mods Without Restart
02:07 paramat ok
02:08 paramat closed #2645
02:08 ShadowBot https://github.com/minetest/minetest/issues/2645 -- Add CONTRIBUTING(.md)?
02:08 est31 #3476 should stay open
02:08 ShadowBot https://github.com/minetest/minetest/issues/3476 -- A clear mission statement and contribution guide for Minetest are missing.
02:09 paramat ok
02:12 paramat closed #1531
02:12 ShadowBot https://github.com/minetest/minetest/issues/1531 -- Be Able to Load Mods Without Restart
02:14 gregorycu Talk about kicking arse
02:15 gregorycu Why was 1531 closed?
02:17 est31 joined #minetest-dev
02:17 est31 gregorycu, its far too complicated and the advantage is far too small
02:17 est31 also, this was closed as well and would be required by that issue: #2513
02:18 ShadowBot https://github.com/minetest/minetest/issues/2513 -- Registering Nodes while minetest is playing.
02:18 gregorycu For mod writers though, being able to reload....
02:28 paramat closed #2698 hope others agree
02:28 ShadowBot https://github.com/minetest/minetest/issues/2698 -- Show default text-based header in main menu when header.png is not present
02:30 gregorycu Ahh  const correctness... We should do more of this
02:33 hmmmm what do you think about #3922
02:33 hmmmm merge?
02:33 ShadowBot https://github.com/minetest/minetest/issues/3922 -- Move down description.txt text by Rui914
02:34 hmmmm but i'm not too sure about that commit message :)
02:40 * paramat omnomnoms will bbl
02:40 paramat left #minetest-dev
03:04 ssieb #3899?  Any devs around that know that area?
03:04 ShadowBot https://github.com/minetest/minetest/issues/3899 -- Add an option to colorize to respect the destination alpha by ssieb
03:19 gregorycu We don't seem to have a read write lock
03:27 Lunatrius` joined #minetest-dev
03:28 gregorycu No biggie
03:30 thePalindrome_ joined #minetest-dev
03:35 VargaD joined #minetest-dev
04:02 paramat joined #minetest-dev
04:07 gregorycu So yeah, I think I can remove some of the stutter
04:09 gregorycu EVERYBODY CALM DOWN
04:09 gregorycu Never seen so much excitement
04:10 paramat \O/
04:10 gregorycu That's all I wanted
04:11 paramat how then?
04:12 sofar gregorycu: golfclap
04:12 gregorycu I've noticed that a lot of CPU time is spent copying mesh data, so that the mesh update thread can do its stuff
04:13 gregorycu If I can make the mesh update thread do the copy instead, that frees up the GUI thread
04:14 gregorycu It requires that for every read operation of block data on the update thread I need to take a lock, and every WRITE operation of block data on the GUI thread also requires a lock
04:14 gregorycu The mesh update thread never modifies the block data, so it doesn't need write locks
04:15 sofar many readers, one writer?
04:15 gregorycu Yes
04:15 gregorycu 2 readers, one writer
04:15 sofar good
04:15 gregorycu At the moment we don't have any shared data, instead we clone the data
04:15 gregorycu Which is a good pattern
04:16 gregorycu But if your data size is large, it can be less efficient
04:16 gregorycu Which is the case here
04:32 gregorycu How hard will this crash... let's find out!
04:43 gregorycu Pretty hard
04:44 yang2003 joined #minetest-dev
04:57 gregorycu :/
05:09 Hunterz joined #minetest-dev
05:13 gregorycu Who wants to see a pic of me
05:13 gregorycu http://i.kinja-img.com/gawker-media/i​mage/upload/japbcvpavbzau9dbuaxf.jpg
05:13 hmmmm what about https://avatars0.githubusercont​ent.com/u/1516163?v=3&s=400
05:14 hmmmm you look like a cross between an alien and a primate when that thumbnail is at a lower resolution btw
05:14 gregorycu lol
05:14 gregorycu Just what i was aiming for!
05:14 gregorycu There is a more... realistic pic here: http://codebrew.net.au/
05:14 hmmmm hmm how much % of cpu is being used by mesh copying exactly?
05:15 sofar when people on the internet say "Who wants to see a pic of me"...
05:15 sofar I know how fast to run
05:15 hmmmm waat
05:15 hmmmm that's you?
05:15 gregorycu Could be worse
05:15 gregorycu That's me
05:15 hmmmm you look much more...
05:15 hmmmm middle aged than I thought
05:15 gregorycu lol
05:16 gregorycu Yes
05:16 gregorycu I do
05:16 gregorycu I'm 29 but look older
05:16 gregorycu I'm at peace with it
05:16 hmmmm ya your github avatar does not reflect reality
05:16 gregorycu A hat can make a difference, also it's a little old
05:16 gregorycu I was very sleep depped when that photo was taken
05:17 hmmmm you should wear a beige safari hat and vest
05:17 gregorycu DA?
05:17 hmmmm DA?
05:18 gregorycu https://www.google.com.au/search?q=david+a​ttenborough&biw=1580&bih=838&s​ource=lnms&tbm=isch&sa=X&ved=0​ahUKEwip776QlurLAhXBipQKHb8LBDYQ_AUIBygC
05:18 hmmmm yeah kinda like him
05:19 hmmmm are you using pthread rwlock?
05:19 gregorycu Nah, just the mutex
05:19 gregorycu And I'm on windows
05:27 gregorycu grr... Visual studio let me take a reference to a temporary
05:29 hmmmm oh
05:29 hmmmm windows has its own rw locks but they're vista+ only
05:29 hmmmm we support XP
05:30 gregorycu Curses on us
05:31 gregorycu Not even microsoft supports xp anymore
05:31 hmmmm yea but we also support solaris
05:32 hmmmm that being said, xp isn't too much of a stretch
05:33 gregorycu It is 14 years old
05:34 paramat left #minetest-dev
05:34 gregorycu By the way, I'll check out the collision detection stuff
05:34 gregorycu I'll run in debug, I'm pretty sure I'll still need it
05:39 hmmmm thanks
05:45 gregorycu 70% of my GUI thread spends its time in collision detection
05:45 gregorycu 75% of that time is pushing back a bool into a vector of bools
05:46 gregorycu I'll obviously have to test the fix
05:49 gregorycu At the moment, I'm testing my jitter fix
05:57 gregorycu I think I've made it better
06:03 troller joined #minetest-dev
06:33 gregorycu Can i make it perfect...
06:37 sofar YESSSS
06:37 sofar may I now politely beg for applause?
06:37 gregorycu *claps*
06:37 sofar I just made nodeupdate non-recursive
06:38 sofar and made 5000 blocks of sand fall all at once without breaking the game
06:39 gregorycu Nice
06:39 sofar I just timed it
06:39 gregorycu Is this for the stack overflow?
06:39 sofar 3000 nodes falling took not even 2 seconds
06:39 sofar let me video it
06:40 sofar it's crazy fast
06:40 gregorycu Compared to what?
06:40 sofar crash
06:40 gregorycu lol
06:40 gregorycu I suppose that time is oo then
06:40 sofar I don't think it would survive even 100 nodes
06:40 sofar since the lua stack overflows at less than 50 calls I think
06:40 sofar luajit, in my case
06:40 gregorycu oh, this is lua?
06:40 gregorycu Lua handles falling nodes?
06:41 sofar yup
06:41 gregorycu Shouldn't this be a core thing?
06:41 sofar it is in core
06:41 sofar builtin
06:41 sofar yes, in c++ was going to be my next guess
06:41 sofar but now that it works in lua, I don't think it matters much
06:41 sofar let me video it
06:41 sofar sec
06:44 sofar 10 seconds for 7920 nodes falling
06:44 sofar uploading
06:45 sofar https://youtu.be/liKKgLefhFQ
06:47 gregorycu You sound funny!
06:48 gregorycu Soft dutch accent?
06:48 sofar 10 years in the US and it's still not gone
06:48 sofar no seriously, I had the exact same accent 10 years when I moved to the US
06:50 gregorycu Nobody is perfect
06:51 sofar lol
06:52 hmmmm [02:41 AM] <gregorycu> Shouldn't this be a core thing?
06:52 hmmmm yes it should and it was but it got changed for some reason
06:53 sofar well it works darn well in lua now
06:53 sofar as my video attests
06:53 hmmmm it could be even better
06:53 sofar sure
06:53 sofar maybe later port it to c++
06:53 hmmmm also the client really doesn't need to freeze like that
06:53 sofar it's receiving 16000 node changes
06:53 hmmmm so much dev needed so little time
06:54 hmmmm well what I'm saying is maybe it'd be a better idea to have the client side predict the node falling
06:54 sofar yes, that would be interesting
07:01 gregorycu hmmmm: I'm going to commit to a branch of mine a bit of a fix for the stuttering
07:01 gregorycu It's a WIP
07:01 gregorycu Did you want to take a look?
07:01 hmmmm sure.
07:03 sofar you know, ~1000 or nodes actually animate properly
07:04 sofar they all fall at the same time, lol
07:04 sofar geesh, that's smooth
07:04 sofar 625 nodes falling as one big cuboid
07:07 gregorycu https://github.com/gregorycu/minetest.git
07:07 gregorycu Branch called Jitter
07:09 gregorycu There are some graphics artifacts mainly because of half of the system doesn't use the locks lol
07:09 gregorycu (Which I intend to fix)
07:10 sofar bed
07:38 yang2003 joined #minetest-dev
07:58 yang2003 joined #minetest-dev
08:01 gregorycu hmmmm: Taken a look yet? :)
08:02 gregorycu I have a game of football to play, be back in 2 hours
08:02 hmmmm yeah I saw it
08:03 hmmmm the last time i looked you basically copied MeshMakeData::fill and pasted it into its own class with a mutex around it
08:03 gregorycu Any difference, or not really?
08:03 hmmmm and did some other adjustments
08:03 hmmmm simple but probably does the trick
08:03 hmmmm oh you mean like actually testing?
08:03 hmmmm no i did not look
08:03 gregorycu Ahh ok
08:03 gregorycu Yeah, I meant testing
08:04 gregorycu It's a WIP at the moment
08:04 hmmmm ofc
08:04 hmmmm yeah i'm sorry, it's just that testing is a big pain in the butt
08:04 gregorycu haha, no problem
08:04 hmmmm looking at a webpage is sooo much simpler
08:04 gregorycu I think Fixer will actually be very exited to try it anyway
08:04 gregorycu We can depend on him to chase these things up :)
08:05 hmmmm to be perfectly honest i noticed that the stuttering issue has decreased in severity over the past couple of versions
08:05 hmmmm although it is still there enough to be noticed it's not distracting or anything
08:05 hmmmm no idea what change made the difference
08:05 hmmmm definitely need a more concrete objective way to test for performance...
08:08 hmmmm oh god no please no operator overloading when it isn't even necessary
08:08 hmmmm hasn't operator overloading caused enough problems?
08:08 gregorycu No
08:08 gregorycu What operator overloading?
08:08 gregorycu oh
08:08 hmmmm what does that code even do??
08:09 gregorycu Hands out a mutex per map block
08:09 hmmmm oh wow interesting
08:09 hmmmm it's a mutex for mutexes
08:09 hmmmm mutex for map of mutexes
08:09 gregorycu Yeah
08:09 hmmmm can you do that?
08:09 hmmmm have a mutex as the value of a map
08:09 gregorycu Yeah, std::map doesn't copy its elements, so it's safe
08:10 hmmmm okay
08:10 gregorycu With a vector it wouldn't be safe
08:10 hmmmm no kidding
08:10 gregorycu The solution isn't good at the moment with regards to those mutexes
08:10 gregorycu As the map will just grow and grow and grow
08:10 hmmmm is the no-copy behavior defined by the C++ standard?
08:10 gregorycu Yes
08:11 hmmmm sounds like you're good then
08:11 hmmmm heh besides, I think I added DISABLE_CLASS_COPY to Mutex a while back iirc, so it'd give a huge glaring error if you tried it and it wasn't supposed to work
08:11 hmmmm :)
08:12 hmmmm i'll look at it more tomorrow.  it's 4 AM for me soo
08:13 hmmmm night
08:13 gregorycu Goodnight
08:15 hmmmm btw do you think people get pissed if i change their commit text before pushing to upstream?
08:15 troller joined #minetest-dev
08:15 hmmmm some of the commit texts for these PRs frankly suck
08:16 gregorycu I wouldn't
08:16 gregorycu If it sucks it's probably because they don't consider it important to begin with
08:17 hmmmm touche
08:19 hmmmm only 130 more to go
08:28 nrzkt joined #minetest-dev
08:43 Amaz joined #minetest-dev
09:10 AnotherBrick joined #minetest-dev
09:27 Darcidride joined #minetest-dev
09:29 jin_xi joined #minetest-dev
10:03 Lunatrius` joined #minetest-dev
10:26 Fixer joined #minetest-dev
10:43 Elinvention joined #minetest-dev
10:52 Calinou joined #minetest-dev
11:12 iangp joined #minetest-dev
11:13 gregorycu Back
11:16 rubenwardy joined #minetest-dev
11:21 Warr1024 joined #minetest-dev
11:23 thePalindrome joined #minetest-dev
11:23 sfan5 joined #minetest-dev
11:28 rubenwardy joined #minetest-dev
11:31 proller joined #minetest-dev
11:39 gregorycu Oh, Fixer is here
11:43 Fixer yes
11:43 Fixer hi
11:45 gregorycu So, the jitter issue
11:45 gregorycu I tried something
11:46 gregorycu https://github.com/gregorycu/minetest.git Branch called jitter
11:46 gregorycu It's pretty win, you should check it out when you get the chance
11:46 Fixer gregorycu, \o/ i will check it now, let me compile it
11:47 rubenwardy joined #minetest-dev
11:48 Fixer gregorycu, i will need at leaat 20 min to compile it
11:52 rubenwardy A regression was introduced by this commit which caused the order of drawing things to change https://github.com/minetest/minetest/commi​t/6cd2b3b445bf558fda1e5a7908adef8e3a45449a
11:52 rubenwardy Well, s/regression/change of behaviour/
11:52 rubenwardy before buttons and text were drawn over images, now images are drawn over buttons and text
11:53 rubenwardy Preparing a patch to discuss
11:55 STHGOM joined #minetest-dev
12:04 Elinvention joined #minetest-dev
12:09 iangp 2016-03-31 09:00:10: ERROR[Main]: ModError: core.registered_lbms was not a lua table, as expected.
12:09 iangp I can't load any world... I'm using the latest commit (c8ff11b)... compiled on Ubuntu...
12:10 gregorycu Sweet Jesus
12:11 rubenwardy are you definitely latest Minetest
12:11 Fixer update minetest_game?
12:12 rubenwardy Make sure you're running the version you compiled (ie: remember to do make -j3 && sudo make install if you're running using `minetest`)
12:14 iangp rubenwardy: I'm running ./bin/minetest
12:14 iangp Fixer: Now I'm not sure, I'll try pull the most recent minetest_game
12:15 rubenwardy https://github.com/minetest/minetest/bl​ob/master/builtin/game/register.lua#L18
12:15 rubenwardy should be ok
12:15 Fixer gregorycu, just a little more time, compile at 100%
12:22 gregorycu There are probably going to be some graphics glitches
12:22 iangp Everything fine now I just typed the sudo make install therefore /usr/local/bin/minetest
12:22 Fixer gregorycu, hmhmm, i must be doing smth wrong, i did clone your repo, git checkout Jitter, git status shows I'm at Jitter, but git log says no your commit
12:22 Fixer gregorycu, did i missed smth?
12:23 gregorycu What does git log -1 say?
12:24 Fixer Mar 29 kwolekrs commit
12:24 Fixer strange
12:24 gregorycu wrong one :)
12:24 gregorycu I'm not git expert
12:24 gregorycu Do a git pull
12:24 Fixer gregorycu, let me apply the diff, it should help
12:24 gregorycu Try git pull first
12:25 proller joined #minetest-dev
12:25 Fixer error :(, let me do it easier way
12:25 gregorycu Cool
12:25 gregorycu It's worth it
12:26 turtleman joined #minetest-dev
12:27 gregorycu I think
12:32 rubenwardy joined #minetest-dev
12:32 Fixer recompiling, another 20 min
12:33 Fixer sorry for f--- up with git
12:34 gregorycu Un-fucking-believable
12:34 gregorycu :)
12:37 Fixer gregorycu, https://xkcd.com/1597/
12:37 gregorycu I was talking to a bitbucket dev the other day, mentioned the same thing
12:41 Fixer gregorycu, hmmmmm, now worse, compilation error
12:41 gregorycu That's not good
12:42 gregorycu What is it?
12:43 Fixer i will save error, and will remove other PRs that were applied, and try again, shittons of errors, will try on clean one
12:55 Fixer gregorycu, still error with cavegen, hmm, let me copy error for you
12:55 gregorycu cavegen lol
13:01 Fixer gregorycu, error: https://i.imgur.com/YkVm7cK.png (lol)
13:02 gregorycu I see
13:03 Fixer i'm cross-compiling for windows in debian (to 64 bit) via mingw64
13:05 gregorycu ah
13:05 gregorycu I'm relying on a new compiler feature without even knowing
13:06 gregorycu Windows ey?
13:06 gregorycu Maybe this weekend I'll clean it up
13:07 gregorycu I have to fix up the thread safety
13:07 gregorycu Also, there is one more area I want to clean up slightly
13:09 Fixer gregorycu, if you have your compiled binaries for windows, just give them to me (that will be even faster) %)
13:12 gregorycu Building now
13:15 Player_2 joined #minetest-dev
13:20 gregorycu Urgh
13:20 gregorycu This is going to take ages
13:20 gregorycu Some other time
13:20 proller joined #minetest-dev
13:20 gregorycu Off to bed for me
13:26 Taoki[mobile] joined #minetest-dev
13:58 kaadmy joined #minetest-dev
14:09 iangp Someone changed the rendering order for the formspec?
14:09 iangp because my mod doesn't display a label widget above a image_button widget anymore
14:15 damiel joined #minetest-dev
14:28 iangp joined #minetest-dev
14:32 Taoki joined #minetest-dev
14:44 Samson1 joined #minetest-dev
15:00 rubenwardy joined #minetest-dev
15:06 hmmmm joined #minetest-dev
15:11 Obani joined #minetest-dev
15:11 rubenwardy #3942
15:11 ShadowBot https://github.com/minetest/minetest/issues/3942 -- Correct formspec draw order: Images should be drawn before buttons by rubenwardy
15:12 Fixer rubenwardy, can you look in inventory? try holding one item over another, looks ok with your patch?
15:13 rubenwardy yes, that works
15:13 rubenwardy the selected item which I'm dragging is drawn over
15:14 Fixer ok
15:25 Samson1 joined #minetest-dev
15:35 rubenwardy STHGOM, https://github.com/rubenwardy/ctf_pvp_engine/c​ommit/0db05b54ef75d686422cba64d1020f5e74b2f7c2
15:42 Elinvention joined #minetest-dev
15:46 STHGOM thanks again :D
15:57 nrzkt joined #minetest-dev
16:10 Calinou joined #minetest-dev
16:44 rubenwardy joined #minetest-dev
16:51 Krock joined #minetest-dev
16:51 kaadmy joined #minetest-dev
16:52 hmmmm rubenwardy, why is that an issue?
16:53 hmmmm that draw order behavior is intentional
17:05 rubenwardy it's a change in behaviour which breaks mods
17:24 Elinvention joined #minetest-dev
17:28 sofar items that should be interacted with should be drawn over items that should not be interacted with
17:29 sofar but the problem is buttons with images
17:30 ud1_ joined #minetest-dev
17:44 rubenwardy background[] can be used to draw images under buttons etc
17:49 nrzkt joined #minetest-dev
18:10 sofar nrzkt: you may want to look at #3938
18:10 ShadowBot https://github.com/minetest/minetest/issues/3938 -- Prevent nodeupdate from recursing. by sofar
18:24 rubenwardy joined #minetest-dev
18:24 Fixer this moretrees bug drives me nuts
18:37 nothing\at\\all joined #minetest-dev
18:47 Warr1024 joined #minetest-dev
19:09 Void7 joined #minetest-dev
19:17 rubenwardy joined #minetest-dev
19:29 Obani is xyz still administrator of minetest forum ?
19:34 rubenwardy no, afaik
19:34 rubenwardy celeron55 was always the admin, xyz hosted the forum for a while (which is why he was an admin) but now celeron55 hosts
19:34 Obani So is it a choice to keep as administrator ?
19:35 rubenwardy probably an oversite
19:39 celeron55 oh i guess i should take care of that
19:41 proller joined #minetest-dev
19:42 Obani celeron55, I don't know, maybe anothoer label to know that he was a dev at past, but admin,... :p
19:42 Obani Same for Blockmen, maybe there should be a label for the "oldies"
20:07 proller joined #minetest-dev
20:17 Amaz joined #minetest-dev
20:22 jin_xi joined #minetest-dev
20:24 Calinou I suggest "old farts", like on Red Eclipse :)
20:24 Calinou with a 1337 logo
20:46 paramat joined #minetest-dev
21:20 paramat thoughts on #3941 ?
21:20 ShadowBot https://github.com/minetest/minetest/issues/3941 -- Proposal for dealing with feature request issues
21:21 paramat Fixer what cavegen error?
21:26 Fixer paramat, nevermind, gregorycu tries to fix jitter and commit has some compiling problems, not in default anyway
21:27 paramat ok
21:44 iangp joined #minetest-dev
21:45 turtleman joined #minetest-dev
21:55 rubenwardy joined #minetest-dev
21:56 dfelinto joined #minetest-dev
21:57 luizrpgluiz joined #minetest-dev
21:58 luizrpgluiz there is forecast to exit the 0.4.14?
22:06 luizrpgluiz left #minetest-dev
22:10 paramat est31 ShadowNinja sfan5 what shall we do with #3894 ? is anyone able to solve the path search problem soon? otherwise we could perhaps ask sofar to use a workaround fix for doors
22:10 ShadowBot https://github.com/minetest/minetest/issues/3894 -- Add serverside noclip enforcement by est31
22:11 sfan5 i thought it wasn't needed for doors
22:11 paramat iirc there's a problem with meshdoors that needs this
22:12 paramat i think it's time we move towards full-freeze
22:16 nrzkt we are in feature freeze and this PR is a feature
22:16 paramat no we're semi-frozen
22:16 nrzkt wtf
22:17 paramat sorry channel message is wrong
22:17 nrzkt release is in 5 days and we are not in a feature freeze ?
22:17 paramat see forum thread, news section
22:17 paramat i should change the channel message
22:18 hmmmm i dunno if the release should come too soon
22:18 paramat (but i forgot how to)
22:18 hmmmm there are a ton of changes that really ought to make it in
22:18 paramat fine by me
22:18 hmmmm Make Minetest Great Again
22:18 paramat heh
22:19 hmmmm we're gonna build a great big... wall
22:19 hmmmm to keep out all the freeminer users
22:19 hmmmm and we're going to make them pay for it too in the form of tarrifs
22:20 hmmmm lol that joke probably went over 70% of this channel's head
22:21 paramat i was always planning to announce a freeze then at last moment when inevitably much is left to do, delay it
22:21 Fixer we don't even test it for good
22:21 thePalindrome But isn't it called mine-test? :P
22:22 Calinou Minestable: released every 2 years, with frozen mods
22:22 Fixer some servers updated to near last git, I play on one of them, but it is mod heavy
22:22 Fixer paramat, did you ever see papyrus and waterlilies in valleys?
22:23 nrzkt don't forget one thing: package maintainers needs a version sometimes, and we have many client features atm which are good for a release, others can wait a little bit time, and we could release a 0.4.15 in 3 months if needed, there are no schedule
22:24 hmmmm i'm conflicted on fixed schedule releases
22:24 hmmmm on one hand, it helps with predictability and frequency of releases, and it could also encourage developers to kind of hurry it up
22:25 paramat fixed just doesn't happen
22:25 hmmmm but on the other hand it could force us to miss a lot of great contributions that would've been great to have beforehand
22:25 nrzkt we don't miss, we delay
22:25 nrzkt releases also permit to communicate about the product
22:25 hmmmm isn't that the point of celeron's blog and the forum?
22:26 nrzkt without communication for a too long time, it's difficult to get new players
22:26 paramat anyway we should certainly be concentrating only on priority stuff for release, that's why we are 'semi-frozen'
22:26 nrzkt paramat, semi-frozen doesn't exist
22:26 paramat i created it
22:26 nrzkt you are frozen or not :p but not semi frozen
22:26 nrzkt if we add features, we are not
22:26 sofar I'm late in the discussion, and would love est31 path collision stuff in the release, but I think it's unrealistic time-wise
22:26 paramat we're in the fridge door next to the fruit juice
22:26 nrzkt sofar, agreed
22:27 nrzkt take the fruit juice in your head :p
22:27 sofar however, for the discussion, I see it as a feature, not a bugfix
22:27 sofar it's certainly not a regression
22:28 sofar are we perhaps considering a branch? would that help?
22:28 sofar branch for 0.4.14, then run a week of testing so that development can continue
22:28 nrzkt sofar, no, tried it but seems outside of MT scope
22:28 sofar np, fair enough
22:29 sofar a week isn't so bad, if that's what it is
22:29 nrzkt for 0.4.13 we tried this but doesn't work
22:29 sofar it takes.. practice and control
22:29 sofar anyway
22:30 sofar at this point in time I'd start going into milestones and kick items out that aren't going to make it
22:30 nrzkt i agree with you, but the MT devel model is currently not mature for this
22:31 paramat ok we're not semi-frozen, more chilled
22:32 sofar only 2 prs open in 0.4.14 for minetest
22:32 sofar so, we're really scraping the bottom already there
22:32 sofar 3 for minetest_game
22:33 paramat btw https://github.com/minetest/minetest​/issues/3266#issuecomment-199906559 who's the person for secure settings?
22:34 paramat we may have to delay nodetimer crops
22:34 sofar that's fine
22:34 sofar xpanes is out as well, there's more
22:35 paramat tnt would be good if the current problem can be fixed
22:35 sofar remind me, what are we talking about there?
22:36 paramat oh was nore's bug fixed?
22:37 paramat https://github.com/minetest/minetest_​game/pull/862#issuecomment-202301115
22:37 sofar https://github.com/minetest/minetest_game/pull/862​/commits/973b0bfadb0651105f3f0462faf5e8d1fd2f4ad1 fixes those
22:37 paramat removed nodetimer crops from milestone
22:38 paramat ok i think tnt can be in release, we have time to work on it
22:38 iangp joined #minetest-dev
22:39 paramat my biome tweaks are trivial
22:40 paramat game#978 can be closed? or is only partially fixed?
22:40 ShadowBot https://github.com/minetes​t/minetest_game/issues/978 -- furnaces dont work after rejoin
22:41 sofar not fixed, although the mods using them did merge workarounds for the actual problem
22:42 sofar hmm soon we can make tnt call nodeupdate
22:42 sofar that'll be awesome
22:50 sofar OH
22:50 sofar MY
22:50 sofar GAWD
22:50 sofar this is fantastic
22:50 sofar sec, recording vids
22:56 Void7 joined #minetest-dev
22:57 sofar https://youtu.be/4omndVZijLc
22:58 paramat wow
22:58 sofar if you pee your pants laughing watching that, I'm not responsible
22:59 sofar I've got a patch limiting entity speed, btw. It's badly needed :D
22:59 sofar I've seen falling nodes disappear for minutes without it
23:00 paramat if you can get that stable for release that would be good
23:00 sofar that has my nonrecursive nodeupdate code
23:00 sofar without that it would probably lag out
23:01 sofar it's not unstable. I haven't seen any bugs with any part of it. it certainly has been tested ;)
23:02 sofar of course, only sand and gravel does this
23:02 paramat i see. is the tnt PR ready for merge? i think now's the time
23:02 paramat looks like that last change is tested
23:03 sofar it doesn't have the speed limit in the PR yet
23:03 sofar that should really be added
23:03 paramat ok
23:03 sofar you can otherwise get crazy high speeds
23:03 sofar pushing
23:05 sofar the reason it's needed - without it you can explode items to the world edge
23:05 paramat eh
23:05 sofar this way they don't make it that far
23:06 sofar this is only a problem for multiple tnt explosions in the same spot, with 1 tnt it doesn't happen
23:06 paramat what happens if you explode tnt next to world edge, any big problems?
23:06 sofar no, engine just warns, and deletes the item
23:06 Fixer sofar, looks nice
23:08 kinokoio joined #minetest-dev
23:09 sofar I have sand blocks all, all, ALL over the place now
23:09 kaadmy joined #minetest-dev
23:20 Dragonop joined #minetest-dev
23:21 jin_xi nice
23:23 sofar 400 nodes away still sand
23:53 kinokoio Hi, can I send a pull request of just one line?
23:55 kinokoio I have found a problem in the line 47 of dungeongen.cpp. It must be "c_torch  = mg->ndef->getId("default:torch");" instead of c_torch  = ndef->getId("default:torch");
23:58 paramat yes a 1 line PR is fine
23:58 paramat those torches are disabled by default and only used for dev work

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