Minetest logo

IRC log for #minetest-dev, 2016-01-29

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

All times shown according to UTC.

Time Nick Message
00:27 paramat left #minetest-dev
00:51 kaeza joined #minetest-dev
00:56 paramat joined #minetest-dev
00:57 paramat will merge game 778 810 soon
01:10 twoelk|2 joined #minetest-dev
01:43 ssieb joined #minetest-dev
01:49 redblade joined #minetest-dev
01:49 redblade hi
01:50 redblade i removed irrlicht from my vps because now that i compile minetest myself, i (don't think i) need it
01:50 redblade but is irrlicht supposed to be required anyway?
01:50 redblade because i got an error message over lack of a header
01:50 redblade when i removed irrlicht and compiled without sound (as i normally do)
01:51 redblade i compile only minetestserver
01:51 redblade i just did a pull on the latest
01:53 redblade so i dont know if there's a #define missing somewhere or if minetest should be requiring irrlicht
01:54 sofar did you read the wiki page on compiling?
01:54 redblade yes
01:54 sofar http://dev.minetest.net/Compiling_Minetest
01:54 sofar it explains
01:54 redblade and have done it several times
01:54 sofar You still need the Irrlicht headers for this
01:54 redblade irrlicht libraries arent in ldd
01:54 redblade it does?
01:54 sofar it's in the wiki page
01:54 redblade - If you build a bare server, you don't need to have Irrlicht installed.
01:54 redblade In that case use -DIRRLICHT_SOURCE_DIR=/the/irrlicht/source
01:55 sofar installed != present
01:55 redblade ok
01:55 sofar read the sectin "Building without Irrlicht / X dependency"
01:57 redblade i dont see it here https://github.com/minetest/minetest
01:57 sofar http://dev.minetest.net/Compiling_Minetest
01:57 redblade oh ok
01:57 redblade i thought that was dated information and you're supposed to go by the github instructions
01:58 sofar - If you build a bare server, you don't need to have Irrlicht installed.
01:58 sofar In that case use -DIRRLICHT_SOURCE_DIR=/the/irrlicht/source
01:58 sofar it literally says the same, you still need the irrlicht source (for headers)
02:03 redblade ok
02:03 redblade i dont really know how irrlicht works, i just thought that if ldd doesn't show it you don't need it
02:04 sofar internally the server uses a lot of the 3d data types
02:07 redblade i've been having issues with lighting (more specifically, sign text being blank or half-visible in certain areas) and i recompiled to see if there was a fix
02:08 redblade it only recently started so that's my guess
02:08 redblade i dont think it was a change in any of the mods
02:08 redblade though i do update them a lot
02:08 redblade and it isn't signs_lib because it affects the normal wooden signs too
02:11 DogePony joined #minetest-dev
02:26 STHGOM hmm, I think text on the signs get messed up with mipmapping (the ones that dont put text on your hud)
02:26 STHGOM is it on?
02:33 redblade the server?
02:33 redblade hold on
02:33 redblade i just ran it
02:33 redblade other people were having that problem
02:33 redblade the terrarium
02:33 redblade i'll log on in a minute
02:33 redblade it wasn't just me
02:34 redblade ok just logged on
02:36 redblade wtf i logged on and it's 70 lag?
02:36 redblade i did a clearobjects but it's over
02:37 redblade wtf did i do it's 50 lag?
02:38 redblade STHGOM try logging on and see
02:38 redblade it's dropping 43 now
02:38 STHGOM where?
02:39 redblade hmm no slowness in playing
02:39 redblade the max lag value
02:39 redblade i did a clearobjects but it ended
02:39 redblade and when i logged on again it was 70 lag
02:40 redblade i had just recompiled
02:40 redblade but the blank signs are still black
02:40 redblade nk
02:40 redblade log on and type /teleport redblade7
02:42 redblade all signs on ground and some signs on blocks
02:44 leat joined #minetest-dev
02:45 STHGOM log on where?
02:46 kaadmy !server redblade
02:46 ShadowBot kaadmy: server [--{name,address,ip,players,ping,port} <value>]
02:47 kaadmy sry about the ping
02:47 redblade how do i set that
02:49 STHGOM set what? also pehaps it would be better to move to just #minetest
02:54 redblade no i mean set the !server redblade
02:55 redblade is that like a search? i'm redblade7
02:56 STHGOM yea, that was a search
02:56 STHGOM i think
02:56 STHGOM !server just test
02:56 ShadowBot STHGOM: server [--{name,address,ip,players,ping,port} <value>]
02:56 STHGOM hmm
02:57 STHGOM !server 87.110.8.195
02:57 ShadowBot STHGOM: server [--{name,address,ip,players,ping,port} <value>]
02:57 STHGOM idk, tbh :P
02:57 redblade !server 45.33.65.68
02:58 ShadowBot redblade: server [--{name,address,ip,players,ping,port} <value>]
02:58 redblade !server redblade7
02:58 ShadowBot redblade: server [--{name,address,ip,players,ping,port} <value>]
02:58 fireglow --address
02:58 redblade just join "THE TERRARIUM"
02:59 STHGOM k
03:02 red-001 joined #minetest-dev
03:04 kaadmy maybe !server only works in #minetest?
03:05 kaadmy yup
03:07 dfelinto joined #minetest-dev
03:17 kaeza joined #minetest-dev
03:27 fireglow !server --address 4.firc.de
03:27 ShadowBot fireglow: No results.
03:27 fireglow !server --address 4.firc.de --port 30000
03:27 ShadowBot fireglow: No results.
03:27 fireglow oke
03:46 paramat now merging game 778 810
03:54 paramat merged
04:24 paramat left #minetest-dev
05:15 `PenguinDad` joined #minetest-dev
05:49 dfelinto joined #minetest-dev
05:57 ShadowNinja Merging #3593.
05:57 ShadowBot https://github.com/minetest/minetest/issues/3593 -- New timer design. by sofar
06:02 sofar ShadowNinja: make sure to close #3443 #3592
06:02 ShadowBot https://github.com/minetest/minetest/issues/3443 -- Fix minetest.after sometimes calling timers too slowly by GunshipPenguin
06:02 ShadowBot https://github.com/minetest/minetest/issues/3592 -- minetest.after is unreliable
06:02 sofar well, I can close 3592 myself
06:18 Calinou joined #minetest-dev
06:22 PenguinDad joined #minetest-dev
06:32 Hunterz joined #minetest-dev
07:49 sofar ShadowNinja: just in time, needed to record something every 1.0sec interval, tried without my timer patch... 3 second interval steps occurred XO
08:01 red-001 game#794
08:01 ShadowBot https://github.com/minetes​t/minetest_game/issues/794 -- Add an on_burn callback. by red-001
08:37 Amaz joined #minetest-dev
08:38 nrzkt joined #minetest-dev
09:00 jin_xi joined #minetest-dev
09:43 Calinou joined #minetest-dev
10:47 Fixer joined #minetest-dev
10:49 nrzkt joined #minetest-dev
11:57 proller joined #minetest-dev
12:50 dfelinto joined #minetest-dev
13:00 red-001 joined #minetest-dev
14:11 Calinou joined #minetest-dev
14:20 Calinou_ joined #minetest-dev
14:37 turtleman joined #minetest-dev
15:15 neoascetic joined #minetest-dev
15:16 CraigyDavi joined #minetest-dev
15:17 hmmmm joined #minetest-dev
15:19 neoascetic #3628 please
15:19 ShadowBot https://github.com/minetest/minetest/issues/3628 -- Fix OSX building issue caused by ad884f2 by neoascetic
15:19 kaadmy joined #minetest-dev
15:24 est31 joined #minetest-dev
15:25 est31 wow somebody really likes minetest https://github.com/gravgun/Diggler/blo​b/master/doc/spec.md#goals-and-mindset
15:26 neoascetic Phhh
15:27 blaze joined #minetest-dev
15:28 sofar it's not a bad point to make, but tbh the amount of code needed to keep support for older versions isn't that large
15:28 est31 ^
15:30 fireglow lol nice
15:30 kaadmy wow
15:30 fireglow I like him already
15:34 sfan5 >Specific file headers (and corresponding footers) for easy recovery and verification in the case the Universe data gets lost/"deleted"
15:34 sfan5 hm
15:35 sfan5 i'm not sure how that's supposed to work
15:35 hmmmm looks like that guy coded up quite a bit in a very short amount of time
15:35 Taoki joined #minetest-dev
15:35 kaadmy i'm also pretty sure he doesn't know how to write code
15:35 hmmmm why?  it looks sane
15:36 kaadmy "i want a perfect game that works withoutbany bugs!"
15:36 sfan5 >Compression algorithm
15:36 sfan5 >proceeds to list a lot of compression algorithms
15:37 sfan5 why not decide on one that works fine
15:37 hmmmm because i'm guessing it autoselects the best compressor for a given chunk of data and uses that
15:37 sofar that's a common tactic
15:38 sofar albeit throws a ton of CPU against a few bits of gain
15:38 sfan5 i'm not sure whether compressing each chunk 8 times saves time
15:38 sofar it won't
15:38 sofar it will be slower
15:38 hmmmm slower on save
15:38 sofar you'll always have to wait for all the compressions to save
15:38 sofar and finish
15:38 est31 it can be easily parallelized though
15:38 hmmmm in any case, i've tried that before in a different non-related application
15:39 est31 still compression is as slow as the slowest compression alg
15:39 hmmmm it doesn't work that well.  lzma almost always comes out to be the optimal algorithm or your money back.
15:39 hmmmm but deflate is the best in terms of speed/compression ratio
15:39 sofar I would just pickle through his commits and take the good stuff
15:39 hmmmm that code is not a fork of minetest
15:39 est31 yeah
15:39 hmmmm that's the impressive part
15:40 hmmmm so far all of the "i hate minetest" forks have just been that - forks
15:40 est31 yup
15:40 hmmmm this guy can code though because he did.  the work there is clearly a WIP but it's still an impressive amount to get done in a month
15:40 sofar initial commit is now a year old
15:40 hmmmm but yea, the "legacy" stuff in minetest is not as bad as people are saying
15:40 hmmmm oh it is?
15:41 hmmmm didn't realize that lol
15:41 Obani joined #minetest-dev
15:41 sofar 20150219
15:41 hmmmm I think most players expect that new stuff can only be achieved through breaking compatibility
15:42 hmmmm it makes them feel like the experience is more exclusive or something
15:42 hmmmm total psychological bs
15:43 neoascetic #3628 please
15:43 ShadowBot https://github.com/minetest/minetest/issues/3628 -- Fix OSX building issue caused by ad884f2 by neoascetic
15:44 hmmmm "new timer design"??
15:44 sofar that just reimplements minetest.after
15:44 sofar it's the commit before that
15:45 hmmmm hmm
15:45 sofar sorry, spaced out on the lack of OSX machines in my house (0)
15:45 hmmmm I usually don't like refactors but the new timer design one seems nice, look at all the code savings
15:45 sofar I couldn't even make sense of the old code in there, honestly
15:45 hmmmm i was trying to get an osx vm that is able to build minetest a while back and it's actually really hard
15:45 sofar all that mintime stuff was ... weird
15:46 sfan5 est31: #3628 is a "trivial fix", you can merge it without waiting for other coredevs to approve it
15:46 ShadowBot https://github.com/minetest/minetest/issues/3628 -- Fix OSX building issue caused by ad884f2 by neoascetic
15:46 hmmmm yeah i agree, that is trivial
15:47 hmmmm but i also approve
15:47 nrzkt lz4 is faster than our gzip method, tested and approved, and used less CPU :)
15:47 sfan5 but lz4 is usually worse than zlib
15:47 sfan5 concerning the compression ratio
15:48 nrzkt for mapblocks i didn't notice a huge difference, maybe ~10%
15:48 hmmmm sofar, you're not supposed to prefix a function name with _
15:49 sofar I took that out, didn't I?
15:49 hmmmm _os_get_clock()?
15:49 Hunterz joined #minetest-dev
15:50 hmmmm in any case, could you put _os_get_clock inside of porting.cpp??
15:50 hmmmm the rule is that if a function is over 3 lines of code it should go into the .cpp file
15:50 sofar oh, hm, even there? It's not something that should be called by anyone directly, and inline too
15:50 neoascetic So, can anyone merge #3628 please? I wan't to make new PR for OSX but I need a fixed master
15:50 ShadowBot https://github.com/minetest/minetest/issues/3628 -- Fix OSX building issue caused by ad884f2 by neoascetic
15:50 hmmmm the intent of the "_os" prefix could be better signified by porting::
15:51 sofar I'll gladly fix this up, but right now I have to drive the kids to school ;^)
15:55 hmmmm lol neoasthetic is your name really "phuckin"?
15:55 hmmmm oh nvm i misread
16:01 neoascetic :(
16:25 STHGOM joined #minetest-dev
16:31 jin_xi joined #minetest-dev
16:32 neoascetic #3629
16:32 ShadowBot https://github.com/minetest/minetest/issues/3629 -- Implement OSX Travis builds by neoascetic
16:32 neoascetic please
16:34 sofar neoascetic: it's nice that it's already testing the osx addition on .... OSX ;)
16:35 neoascetic Yeah, Travis-CI added support recently
16:37 est31 thats the cool thing about PRs that modify travis.yml
16:37 est31 they are instantly tested on the new version
16:55 H-H-H joined #minetest-dev
16:56 est31 so, neoascetic you want to adjust the PR?
16:56 est31 then it only requires a second +1
16:57 sofar is there an :getparent() method to get a reference to the attached-to object?
16:57 neoascetic But why?
16:57 neoascetic est31
16:58 est31 why waste resources neoascetic
16:58 sofar clang on OSX isn't going to expose much more than on Linux
16:58 est31 I dont think that there will ever be a bug thats neither catched by osx + gcc nor catched by clang + linux
16:59 neoascetic Ok, no problem, gimme a sec
17:01 GhostDoge joined #minetest-dev
17:02 neoascetic There are such many excludes now in travis file, shouldn't we build clang && gcc within single build?
17:03 est31 having excludes is imo due to a shortcoming of travis' build matrix system
17:04 est31 its probably more useful for ruby or whatever travis has been designed for but for minetest it isnt really useful
17:06 est31 sfan5, can you look at #3629 ?
17:06 ShadowBot https://github.com/minetest/minetest/issues/3629 -- Implement OSX Travis builds by neoascetic
17:06 sfan5 soon
17:06 est31 (or anybody else)
17:10 Krock joined #minetest-dev
17:40 neoascetic joined #minetest-dev
17:40 neoascetic joined #minetest-dev
17:41 sfan5 est31: +1 from me except that some unittests are failing right now
17:50 neoascetic est31 sfan5 this is strange issue that appears (only?) on osx/gcc. In that test timings are checked. I guess because build run under VM it is pretty slow and because of this that specific test failing
17:50 neoascetic It is failing from time to time, not every build
17:50 est31 yeah
17:52 neoascetic So, merge?
17:54 neoascetic I have a weird idea to implement minetest server protocol in Erlang, just for fun. It is that weird as I think?
17:57 Calinou I would rather work on documenting the protocol or such
17:57 Calinou alternative implementations will never be 100% complete, and nobody will really use them
17:57 neoascetic I said just for fun. Docs are not fun
17:57 est31 lol
17:58 neoascetic But I want to see protocol docs definitely
17:58 Calinou there is a point to reimplementing proprietary software, like TrueCraft does for Minecraft
17:58 Calinou but not much for reimplementing existing libre software
17:58 est31 well idk it did happen
17:59 est31 look at gcc and clang
17:59 neoascetic https://github.com/rvirding/luerl :)
17:59 est31 or at firefox and webkit
17:59 Calinou Clang isn't really a reimplementation… it has tons of new ideas
17:59 est31 (or KHTML)
17:59 est31 clang tries to be compatible to gcc in many regards
17:59 est31 even implements unofficial behaviour
18:01 neoascetic was clang made just for fun? I don't think so :)
18:02 neoascetic What is wrong with #3143, btw?
18:02 ShadowBot https://github.com/minetest/minetest/issues/3143 -- Add CONTRIBUTING.md by davisonio
18:06 Calinou people just never agree in Minetest
18:06 Calinou might as well rename it to Arguetest
18:08 neoascetic What does @ in some labels on github means?
18:08 neoascetic there is @Build, but OSX, enhancement...
18:09 Amaz Calinou, +100
18:10 Gael-de-Sailly joined #minetest-dev
18:15 nrzkt joined #minetest-dev
19:23 Player_2 joined #minetest-dev
19:30 Gael-de-Sailly joined #minetest-dev
19:44 Taoki joined #minetest-dev
19:51 ssieb joined #minetest-dev
20:06 H-H-H joined #minetest-dev
20:26 Hijiri I'm having some issues when a magic missile (an entity) tries to punch an item entity. Does get_inventory (as an ObjectRef method) return nil when the entity has no inventory?
20:27 Hijiri actually I don't see any way to give non-player entities inventories, is there a way?
20:28 Hijiri or maybe they have them by default
20:41 sofar I think you can just add a "main" inventory to the entity
20:44 sofar actually looking at some forum posts people are saying there isn't a way
20:44 sofar you'd have to make a detached inventory for each entity :/
20:51 Taoki joined #minetest-dev
20:55 neoascetic joined #minetest-dev
20:56 Wayward1 joined #minetest-dev
21:01 est31 joined #minetest-dev
21:12 STHGOM joined #minetest-dev
21:18 Wayward1 not sure if it's just my build environment, but the android build keeps failing for me with this error: http://pastebin.com/dhi9fiRr
21:19 est31 there is a simple test for build environment
21:19 est31 just check an old revision
21:20 Wayward1 oh yeah, of course...
21:20 * Wayward1 is tired :P
21:41 Taoki joined #minetest-dev
21:45 Wayward1 that one failed too... and there is no "libiconv.so" on my computer. although, i *did* upgrade to mint 17.3 and kernel 4.2 recently... i'll try a vm of 17.2
21:58 STHGOM joined #minetest-dev
22:09 proller joined #minetest-dev
22:10 proller joined #minetest-dev
22:14 est31 in theory, it should build iconv while you run the build process
22:14 est31 can you put the full stdout of the build process somewhere?
22:14 est31 or /msg me a link
22:16 Hijiri sofar: the issue I'm having, I think, is that the on_punch callback of the builtin item entity gets the inventory of the puncher and uses it without checking if its nil
22:17 Hijiri so making a detached inventory wouldn't work I think, because :get_inventory() would still be nil
22:17 Hijiri wouldn't work to fix the crash, I mean
22:19 sofar you're making an entity hit a player, I assume?
22:20 Hijiri It's a homing projectile, it hits whatever it's aimed at
22:21 Hijiri there's a crash when people fire it at item entities, because it punches its target
22:21 Hijiri at least I think so
22:21 Hijiri I don't get a stack trace because it's in an entity on_step
22:21 Hijiri but I'm pretty sure that's the issue, since it only happens with item entities
22:22 Hijiri as far as I've seen at least
22:22 Hijiri I could just exclude immortal things from being punched
22:22 sofar why would the on_step() crash?
22:22 Hijiri the punching happens in the on_step
22:22 sofar the crash happens in the on_step of the puncher?
22:23 Hijiri the missile punches the item, the item entity tries to index the inventory of the missile, which is nil
22:23 Hijiri sofar: yes
22:23 sofar can I see the code?
22:23 Hijiri oh, I guess I could put it in an after for debugging
22:23 Hijiri ok
22:23 sofar I would narrow it down to the exact problem line first
22:24 Hijiri https://github.com/raymoo/magical_realm/blob​/master/combat_spells/magic_missile.lua#L269
22:24 Hijiri I could put the punch in an after for debugging I guess
22:24 Hijiri I want to wait to open minetest until I get home though
22:24 Hijiri since I'm in the library
22:24 sofar lol
22:41 paramat joined #minetest-dev
22:42 paramat hmmmm or anyone, this just needs a +1, the code is already thoroughly reviewed #3590
22:42 ShadowBot https://github.com/minetest/minetest/issues/3590 -- Add V5 caves to Valleys mapgen. by duane-r
22:58 est31 joined #minetest-dev
23:08 paramat est31 kahrl nore nrzkt sfan5 can anyone +1 3590 ? code is reviewed
23:23 STHGOM joined #minetest-dev
23:23 est31 I dont even know what the difference between valleys mapgen caves and mgv5 caves are
23:24 est31 its fixing the MSVC errors thats good
23:26 paramat yeah it's a bit complex. it replaces duane's 'simple caves' which are 3D noise tunnels with mgv7-type 3D noise tunnels plus our standard large pseudorandom caves, it also adds the giant 3D noise caverns from my subterrain mod
23:27 paramat calling them 'mgv5 caves' is a bit confusing, that's referring to the large pseudorandom caves
23:28 est31 I guess the PR is doing alot of other things except what the title says.
23:29 paramat yes, some simplification of parameters, using mymin/mymax/myround
23:30 paramat we've been working hard on it for 11 days, many updates
23:31 est31 and kwolekr doesnt want to review it?
23:31 paramat so mgvalleys will have the most kick-ass underground of any core mapgen
23:31 paramat seems busy
23:32 paramat on holiday :)
23:33 est31 Well, I'd say lets wait a day so that the core team can respond and look at the PR, and if nobody does -1 then you merge it
23:34 paramat ok cool
23:34 est31 its not really touching sensitive areas, the mapgen isnt stable no network compat issues or so
23:35 est31 I dont really think that it helps minetest if we dont merge PRs anymore bc we are understaffed
23:38 paramat luckily within the mapgen files themselves i roughly know what i'm doing
23:38 Hijiri Is it possible to specify with a texture string just a single color fill?
23:39 Hijiri right now I'm using a 1px png texture with the color I want
23:40 paramat 'colourise' with a white base texture?
23:40 paramat just guessing
23:40 Hijiri the base texture has transparency, so it doesn't work out that well
23:40 Hijiri that was what I tried originally
23:44 paramat ah
23:58 sofar https://mesamatrix.net/

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