Minetest logo

IRC log for #minetest-dev, 2015-06-14

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

All times shown according to UTC.

Time Nick Message
00:14 paramat now pushing 528 to game
00:21 paramat complete
01:11 paramat left #minetest-dev
01:17 Wayward_Tab It seems the Android build may have been broken by one of the latest commits
01:18 est31 Wayward_Tab, you do regular android builds?
01:18 est31 can you do a clean re-clone and re-build?
01:19 Wayward_Tab I do.
01:19 Wayward_Tab Sure
01:26 Wayward_Tab I'm getting an error now: /bin/sh: 5: svn: not found
01:26 Wayward_Tab I just built it successfully a few minutes ago
01:30 Wayward_One Here's the current error from the fresh directory: http://pastebin.com/HAurs124
01:34 est31 do you have svn installed?
01:34 est31 what does svn --version output?
01:35 est31 did you follow this guide: http://dev.minetest.net/Android#Bu​ilding_Android_from_master_branch
01:36 Wayward_One hmm, it seems it was not installed. i've been building successfully without it though
01:36 est31 heh
01:39 Wayward_One and yes, i followed the guide, and when i upgraded my os i copied everything (except subversion, it seems) over to the new install, and it continued to build correctly. the only time i got that error was from the fresh clone
01:39 est31 that makes sense
01:39 est31 because svn is only used for fresh clones
01:39 est31 not when you copy
01:39 Wayward_One ah, ok
01:40 Wayward_One anywho, it's installed now :P
01:40 Wayward_One re-building now
01:40 zat joined #minetest-dev
02:03 Wayward_One it finished with an error, here's the output: http://pastebin.com/ZWRnT9ri
02:04 est31 you need m4
02:04 est31 as described in the guide :)
02:04 Wayward_Tab Oh yeah, duh... new clone...
02:20 chchjesus joined #minetest-dev
02:30 Wayward_Tab Ok, finished and installed, and still getting the same error- ERROR: Couldn't resolve address: No address associated with hostname
02:30 Wayward_Tab Despite having an address
02:31 est31 ?
02:31 est31 same??
02:31 Wayward_Tab Yep
02:31 est31 why same?
02:31 Wayward_Tab No idea
02:31 est31 also when does it appear?
02:31 est31 while building?
02:31 est31 starting?
02:31 est31 playing?
02:32 Wayward_Tab When trying to join a server
02:32 Wayward_Tab Also, the input box for the password has a character in it already when opened
02:32 Wayward_Tab Unlike before
02:33 Wayward_Tab And the username has what seem to be spaces already input as well
02:33 Wayward_Tab Also unusual
02:33 est31 hrmm
02:33 est31 but only on android...
02:36 Wayward_Tab http://imgur.com/bMs3eu0
02:36 Wayward_Tab http://imgur.com/vqoMBqE
02:36 Wayward_Tab http://imgur.com/l1SCRKG
02:36 est31 yes I know
02:37 est31 I can reproduce it myself
02:37 Wayward_One ah, ok
02:58 Wayward_Tab Shall I file an issue...?
02:58 Wayward_Tab If it'll help any
03:08 paramat joined #minetest-dev
03:09 paramat currently fixing the settings of 'is ground content', why are trees and leaves 'false'? all other plantlife is 'true'
03:22 est31 Wayward_Tab, yes file one
03:34 paramat i feel like adding a newline at the end of files i work on, not sure if this is good or bad, or different code style for lua and c++
03:37 kahrl paramat: all text files should end in a newline
03:39 paramat thanks, not code files then
03:40 kahrl code is text :P
03:40 est31 hehe
03:41 paramat aha i see
03:41 kahrl basically I meant everything that file -i identifies as "text/..."
03:42 paramat i'm getting into the habit of adding them to everything i work on, they're missing in so many places
04:07 paramat Sokomine and all https://github.com/minetest/minetest_game/pull/539 -- Fix is_ground_content settings for nodes, remove line if the default of true
04:33 paramat left #minetest-dev
04:41 JeDa_ joined #minetest-dev
04:45 est31 Wayward_Tab, Wayward_One you can use this commit if you want, it fixes the issue, but its no proper fix https://github.com/est31/minetest/commit/​a456b3cf7d28e57ca80912e6aa07df59ffe3969c
04:48 Wayward_Tab Ok, I'll try it tomorrow
05:52 Elronnd joined #minetest-dev
05:52 srifqi joined #minetest-dev
05:58 hmmmm joined #minetest-dev
06:01 Hunterz joined #minetest-dev
06:07 RealBadAngel joined #minetest-dev
06:15 est31 somehow wide_to_utf8 is broken on android
06:15 est31 utf8_to_wide isn't
06:16 nore joined #minetest-dev
06:19 cvtsx joined #minetest-dev
06:20 cvtsx joined #minetest-dev
06:29 est31 gonna push in 5 minutes: https://github.com/est31/minetest/commit/​60f31ad52348c53209701d0112f9f809018f7188
06:30 sfan5 looks good
06:31 est31 pushing now then
06:31 sfan5 but why do we have wide_to_utf8 when wide_to_narrow does the same?
06:31 est31 does wide_to_narrow do the same?
06:31 sfan5 apparently
06:32 est31 it does the same for simple charsets like ascii
06:32 sfan5 otherwise you couldn't just forward the call to wide_to_narrow
06:32 sfan5 or does it not do exactly the same?
06:32 est31 but wchar_t (or wstring) isn't something you should send over the network
06:32 est31 so yes it does the same in most situations, but not all
06:32 sfan5 why not?
06:32 sfan5 oh ok
06:32 est31 e.g. when you have other system planes
06:33 sfan5 other system planes?
06:33 sfan5 so wide to narrow uses the native charset?
06:33 est31 yes
06:33 est31 I dont know what it does when the first byte isn't 0.
06:34 sfan5 iirc wide_to_narrow is a hack on android
06:34 sfan5 but it works so ¯\_(ツ)_/¯
06:48 RealBadAngel Taoki, minimap also is workin as a separate thread
06:48 RealBadAngel at least most heavy calculations
06:50 est31 btw, does it create a separate irrlicht context for that?
06:50 est31 or does it not access the irrlicht context?
06:52 RealBadAngel context?
06:52 RealBadAngel you mean things being displayed?
06:53 est31 the stuff thats accessed by the main thread already and cant be shared with other threads
06:53 RealBadAngel we are communicating
06:54 RealBadAngel mapblock mesh update thread makes also data for minimap
06:54 est31 so do you render it in the thread and pass a bitmap?
06:54 RealBadAngel send it to main
06:54 est31 or do you do something else
06:54 RealBadAngel when i got that data in block->minimap
06:55 RealBadAngel i make an image out of it and display it just before the HUD
06:56 RealBadAngel also, the way im displaying it is very important
06:56 RealBadAngel i display a mesh
06:56 est31 ah like the mesh thread then
06:57 RealBadAngel that means theres a simple way to display cubes in formspecs
06:57 RealBadAngel even animated
06:57 RealBadAngel like when hoovered over
06:58 RealBadAngel that will let us get rid of inventory images once and for all
06:58 est31 no why
06:58 est31 first tools
06:59 est31 second, sometimes you dont want the rendered block, but an abstraction
06:59 est31 like when one makes a sapling with a nodebox
06:59 RealBadAngel if we have ability to display mesh in a slot we wont need extra textures
06:59 est31 but still wants to have the known sapling symbol
06:59 RealBadAngel maybe we can allow forcing such image
07:00 RealBadAngel but in most cases we need a block or a mesh
07:00 RealBadAngel atm we render such blocks and save resulting image
07:00 RealBadAngel its SICK!
07:01 est31 so you want to render it every frame instead?
07:01 RealBadAngel fuck, why
07:01 est31 so whats the difference then between what was and what you want to change
07:01 RealBadAngel do you calll rendering a dirt node each frame, something that has to be done
07:02 RealBadAngel or ?
07:02 RealBadAngel and whats the difference where this node is/
07:02 RealBadAngel in the world or in inv
07:02 est31 a node in world can be seen from different angles
07:03 est31 and distances
07:03 est31 there can be other nodes before it
07:03 est31 and so on
07:03 RealBadAngel and its heavier
07:03 est31 it makes sense to render it every frame
07:03 est31 but a node in inventory only is viewed from one angle distance and light
07:03 RealBadAngel because you do render hundreds of nodes
07:03 RealBadAngel ok ok
07:03 est31 so why should we call the render loop on that over and over again
07:04 RealBadAngel now take 3500k of nodes
07:04 est31 the result is always the same
07:04 RealBadAngel each with 6 textures
07:04 RealBadAngel calculate ammount of ram
07:04 RealBadAngel to have those inv images
07:04 RealBadAngel and you will have an answer why mt needs 16gb
07:05 est31 so you say its about the inventory icons?
07:05 est31 they should be cached
07:05 RealBadAngel what cache means?
07:05 RealBadAngel once loaded takes mem
07:06 RealBadAngel its all this funny politics
07:06 est31 or another idea
07:06 RealBadAngel modders thinks that operators are simple and easy
07:06 est31 perhaps when somebody opens a formspec, the images are generated on the fly
07:07 RealBadAngel but simple operation on two textures creates 3rd one
07:07 est31 then they are only rendered once
07:07 RealBadAngel too slow
07:07 est31 too slow?
07:07 RealBadAngel ofc
07:07 est31 you want to render them every frame!
07:07 RealBadAngel preload item visuals
07:07 est31 thats even slower!
07:07 RealBadAngel NO
07:07 est31 so what are we discussing about
07:07 RealBadAngel GPU does that
07:08 est31 and for some people the GPU is shit
07:08 RealBadAngel compare inventory made of lets say 100  nodes
07:08 RealBadAngel to world made of  thousands (visible)
07:09 RealBadAngel do you get it?
07:09 est31 yes
07:09 est31 you still will have to cover comparable sizes of screen
07:09 RealBadAngel no need
07:09 est31 so 600x600 pixels filled with world
07:09 RealBadAngel you define a viewport
07:09 est31 or 600x600 pixels filled with formspec
07:10 RealBadAngel just a tiny small one for the image
07:10 est31 all full with items
07:10 est31 in the result, its the same
07:10 RealBadAngel not rly
07:10 est31 ok there is a small difference I admit
07:10 est31 or a medium one
07:10 RealBadAngel if you work on lets say 32x32 slot, how much computation power would you need?
07:11 RealBadAngel in pixels of course
07:11 RealBadAngel propably old timex watch could do some frames for us realtime ;)
07:11 est31 ok you convinced me
07:11 est31 what I want is perfection, but its ok if its only excellent :)
07:12 RealBadAngel theres no such thing as perfection
07:12 RealBadAngel theres only way to reach it
07:12 est31 still I want it :D
07:12 RealBadAngel so, youre on your way ;)
07:13 est31 I guess what you propose is much better than what currently is there
07:13 RealBadAngel less memory, faster fps
07:14 RealBadAngel look what i did with nodeboxes
07:14 RealBadAngel its fairly the same thing
07:14 est31 faster fps?
07:14 est31 by rendering each frame?
07:14 est31 what is exactly your plan?
07:14 RealBadAngel we dont have any nodeboxes anymore
07:14 RealBadAngel at least on c++ side
07:15 est31 ok
07:15 RealBadAngel my plan is to use native irrlicht ways of displaying things
07:16 est31 ok
07:16 RealBadAngel also i need the power that is atm spent on bullshits like those inv images
07:16 RealBadAngel and memory
07:17 RealBadAngel its an abomination that HDX texture pack occupies 12-13gb
07:18 * VanessaE hides
07:18 est31 RealBadAngel, still your plan is to render the nodes every frame, again and again, rendering the same thing just because its a "native" irrlicht way?
07:18 RealBadAngel est31, you dont get it
07:18 est31 yes or no
07:18 RealBadAngel i want to display MESHES
07:18 proller joined #minetest-dev
07:18 VanessaE est31: your question raises another:  what's the difference whether it eats a bit more GPU, if the player is merely in a formspec?
07:18 RealBadAngel not textures
07:19 RealBadAngel i want to display 3d objects
07:19 RealBadAngel the thing for we have hired 3d engine
07:20 RealBadAngel and which is letting GPU do the hard workd
07:20 RealBadAngel -d
07:20 est31 RealBadAngel, I agree that this will sink the memory footprint, but I still dont see how it will spare "power"
07:20 RealBadAngel let us higlight some words
07:21 RealBadAngel 4gb - 16gb with HD, swap
07:21 RealBadAngel swap is the most lovely here ;)
07:21 est31 what do you mean by swap?
07:22 RealBadAngel if game needs 16gb to run, but has physical 4gb, what happens?
07:22 est31 yes performance is shit, but thats still "only" a memory problem
07:22 RealBadAngel because of such things
07:23 RealBadAngel i can unroll thousands of loops for sake of performation but will hunt down each sign of such waste
07:24 est31 meh, I realize I dont even know how it does things right now (I recall that preloading item visuals is disabled by default)
07:24 RealBadAngel and done on request
07:24 RealBadAngel so at the very end you will end up with memory filled
07:25 RealBadAngel either you will pay with your time at the startup
07:26 RealBadAngel or face waiting for next page of inv to show up
07:26 RealBadAngel those images are generated no matter you love them or not
07:27 est31 and if its done by the gpu, its done every frame
07:27 est31 yes, the gpu is more efficient
07:27 RealBadAngel first of all, when youre in a formspec fps is limited to 20fps
07:27 est31 but even if it were 1000x more efficient, the advantage would be gone after about one minute of viewing
07:28 RealBadAngel it gives you 40 fps worth time to jump around
07:29 RealBadAngel est31. do you actually think that rendering a texture takes no time?
07:29 est31 RealBadAngel, you think that rendering a texture takes more time than a mesh?
07:29 RealBadAngel its being converted to a mesh plane
07:29 RealBadAngel then displayed
07:30 RealBadAngel maybe with differnt engine it could be faster (but theres still price of having the look prerendered)
07:31 RealBadAngel but not with irrlicht
07:31 est31 ok then
07:31 est31 didnt know that
07:31 est31 I agree then to you :)
07:32 RealBadAngel you will notice only speed up
07:32 RealBadAngel and less memory used
07:32 RealBadAngel that i can promise
07:35 Amaz joined #minetest-dev
08:00 Calinou joined #minetest-dev
08:06 Yepoleb joined #minetest-dev
08:10 kilbith joined #minetest-dev
08:11 rubenwardy joined #minetest-dev
08:40 est31 if we only measure by commit number, we have more commits since 0.4.12 release than between any two following releases from 0.4.12 to 0.4.8 (havent checked before that)
08:47 Puma_rc joined #minetest-dev
08:55 kilbith celeron55, there are 2 bugs so far on the forum : https://forum.minetest.net/vi​ewtopic.php?p=181773#p181773
09:01 LNJ joined #minetest-dev
09:19 kilbith sfan5: feedback on this please : https://github.com/minetest/minetest_game/pull/540
09:20 est31 ++1
09:20 est31 ah dammit +1 is no l-value
09:20 est31 then +1 :)
09:27 celeron55 sorry for the forum kind of hanging up occasionally, i'm working on some backup stuff and apparently it just locks up mariadb completely
09:29 celeron55 (i hate mysql, and myself for not using postgresql for this originally)
09:30 sfan5 can't you migrate to postgres?
09:30 celeron55 it's pretty much impossible
09:31 celeron55 there's a reason why postgresql is better
09:31 celeron55 the reason involves it not being compatible with mysql
09:31 celeron55 seriously, try taking a phpbb database and migrating it 8D you will find out what i mean
09:38 Megaf joined #minetest-dev
09:47 celeron55 i guess i could explain a bit: the first thing you will see is that a large amount of the types do not match
09:47 celeron55 and the second thing you will notice is that constraints work differently
09:47 celeron55 and you will also notice that they are impossible to automate and there is like a hundred tables to convert
09:48 celeron55 and indexes work differently
09:56 celeron55 it looks like people are now managing to register on the forum so the one broken blacklist was apparently the only broken one
09:56 MinetestForFun joined #minetest-dev
10:03 est31 pushing in 5 minutes https://github.com/est31/minetest/commit/​8b29699b0df0bd746437d4cbb3aba06900195b8f
10:03 sfan5 est31: testUTF -> testUTF8
10:04 est31 *shrug* changing
10:04 sfan5 also do you really need all the std::string constructor calls in testStartsWith
10:04 sfan5 ?
10:04 est31 yes
10:04 est31 unfortunately
10:04 sfan5 won't it automatically convert to std::string
10:05 est31 usually yes but not here, because I think its a template method
10:05 sfan5 hm
10:05 sfan5 est31: "the shovel digged a crumbly node!" -> "the shovel dug a crumbly node!"
10:06 est31 ah english tenses
10:06 est31 https://github.com/est31/minetest/commit/​c5de5c5f7a677f9c6ffc4ddf022089a485d78057
10:07 sfan5 looks good
10:08 est31 pushed
10:08 est31 (commit mostly needed for the utf8 tests)
10:25 proller joined #minetest-dev
11:10 H-H-H joined #minetest-dev
12:16 chchjesus joined #minetest-dev
12:31 nore joined #minetest-dev
12:40 proller joined #minetest-dev
12:52 Krock joined #minetest-dev
12:58 kilbith joined #minetest-dev
13:21 kilbith joined #minetest-dev
13:25 celeron55 bleh, mariadb filled /var with some kind of dumb binary logs and now everything is stuck
13:25 celeron55 hang on there, lol
13:29 celeron55 i can't even login to mysqld now, i guess i'll just remove the files 8D
13:32 celeron55 now my spoon fell into my trash can; i think this is exactly the moment when i should be working with databases
13:37 celeron55 (i actually knew i was running out of disk space, but due to this spoon-into-trashcan mode i thought it was /home instead of /var 8))
13:38 celeron55 (anyway it works now, except that not *right* now because it's again running a backup thing that locks up everything)
13:50 MinetestForFun joined #minetest-dev
14:33 RealBadAngel #2783 ive updated textures displacement PR here
14:33 ShadowBot https://github.com/minetest/minetest/issues/2783 -- Improve displacement mapping by RealBadAngel
14:34 RealBadAngel i believe its rdy to merge, anybody wanna test it a bit?
14:35 RealBadAngel nore?
14:53 ajmiles joined #minetest-dev
15:06 Amaz joined #minetest-dev
15:14 celeron55 fuck, how does one add a redirect in nginx from something like /viewtopic.php?id=<id> to /viewtopic.php?t=<id>
15:16 RealBadAngel celeron55, RedirectMatch ?
15:16 ElectronLibre joined #minetest-dev
15:17 kahrl that sounds apachy
15:19 RealBadAngel kahrl, can you check #2783?
15:19 ShadowBot https://github.com/minetest/minetest/issues/2783 -- Improve displacement mapping by RealBadAngel
15:21 kahrl RealBadAngel: I don't understand the changes in minetest.conf.example
15:22 kahrl why the newline above enable_shaders? why remove the comment about enable_bumpmapping?
15:23 sfan5 celeron55: rewrite "^/viewtopic.php?id=([0-9]+)(.*)$" "/viewtopic.php?t=$1$2" || and the forum is 502 currently
15:23 kahrl the convention is to have the description of an option above the option
15:23 celeron55 well 502 is better than serving php file sources like a moment ago
15:24 celeron55 oh... or now
15:24 celeron55 well this isn't too good
15:24 celeron55 sfan5: that doesn't work; nginx does not include get parameters in the uri like that
15:25 sfan5 it doesn't?
15:25 sfan5 thats stupid
15:25 celeron55 it's usually useful
15:25 sfan5 hm
15:25 sfan5 didn't xyz have some code snippet in viewtopic.php that used id if there was no t param?
15:26 celeron55 yes
15:26 celeron55 it had a patch in that code and i want to get rid of it
15:26 sfan5 // Replace naughty words such as farty pants
15:26 sfan5 $row['post_subject'] = censor_text($row['post_subject']);
15:26 sfan5 hm
15:26 sfan5 why does phpbb have such functions
15:26 sfan5 (the viewtopic.php i just downloaded line 1399)
15:28 celeron55 i don't read phpbb code
15:28 celeron55 i refuse to
15:30 kahrl RealBadAngel: in translateMesh/scaleMesh, I think u32 would be better than u16 for vertex_count and i
15:30 kahrl (not your fault, but change it anyway)
15:31 kahrl the type u16 of indices restricts the number of vertices to 65536, but (u16) 65536 is 0
15:33 kahrl what's the motivation behind making MeshCollector a class?
15:38 Dartmouth joined #minetest-dev
15:41 celeron55 i guess i have no other option than to add the extra PHP junk to viewtopic.php then
15:41 jbb joined #minetest-dev
15:41 celeron55 http://pastebin.ubuntu.com/7406797/
15:41 celeron55 this
15:41 jbb are you woking at the forums error?
15:41 celeron55 no
15:42 celeron55 i'm closing the forums down because phpbb sucks
15:42 jbb ok
15:42 jbb will it comeback soon?
15:42 celeron55 never
15:43 celeron55 this is the end
15:43 jbb oh
15:43 jbb left #minetest-dev
15:44 RealBadAngel kahrl, ooops, i made it a class when was coding templates (dropped idea)
15:44 RealBadAngel will change it back to struct
15:45 Krock HAHAH
15:45 Krock It's the end.
15:45 RealBadAngel jbb is on the other side already
15:46 RealBadAngel c55 you shouldnt be so cruel ;)
15:46 celeron55 i take no responsibility
15:46 celeron55 the one is responsible who caused the switch to phpbb
15:49 jbb joined #minetest-dev
15:50 Calinou "PunBB: No pun intended."
15:50 Calinou "phpBB: As long as it works! We share our motto with WordPress."
15:51 celeron55 i'm starting to get the hang of being a proper BOFH
15:52 celeron55 i cannot feel the pain of forum users enduring the 2 minute hang of my backup script during european afternoon times anymore
15:52 celeron55 it's nothing
15:58 hmmmm joined #minetest-dev
15:59 RealBadAngel kahrl, ok changed that
16:01 VanessaE celeron55: is there a better alternative than phpbb to which the forum data can be ported?  (and which is powerful enough and actively supported)
16:01 Calinou lots of them, the exporting is the hard part
16:01 Calinou anti-spam too but less known forum systems have naturally less spam
16:01 Calinou phpBB is a victim of its success
16:08 jbb at freeminers websites is a forum which copies all articels from minetests forum
16:08 jbb here is the link http://forum.freeminer.org/
16:09 VanessaE we don't talk about that program here.
16:12 celeron55 VanessaE: well all of them are kind of experimental
16:12 celeron55 and then there are some paid ones
16:12 celeron55 which aren't really worth their price
16:13 ajmiles "Almost infinite world and a fairly good map generator (limited to ±31000 blocks in all directions)."
16:13 celeron55 (and which don't really suit this community anyway)
16:13 ajmiles infinite or 31,000, which is it?
16:13 kilbith *almost* infinite
16:13 celeron55 ajmiles: nothing is infinite, ever
16:14 ajmiles not really what I was asking
16:15 celeron55 infinite limited to 31000 is 31000
16:16 ajmiles might start telling people I have infinite money in that case :)
16:16 ajmiles seems strange to brag about 31,000 being almost infinite when Minecraft itself has worlds many times larger than almost infinite
16:16 RealBadAngel spending money is easier than for example 100km trip on foot ;)
16:17 celeron55 ajmiles: your decision to compare with minecraft is your own, though
16:17 celeron55 we don't generally do that
16:17 ajmiles you're not Freeminer either though right
16:18 ajmiles was just referring to a quote on the project page for the link jbb posted
16:18 celeron55 well minetest's marketing material generally says the same
16:18 ajmiles does it have the same limitation or +- 2^15 ?
16:18 ajmiles *of
16:18 celeron55 yes
16:18 ajmiles right, I see where the confusion might have occurred
16:19 jbb joined #minetest-dev
16:20 celeron55 hmm yes 8)
16:20 celeron55 that was weird
16:22 ajmiles someone (I'm not sure if they're related to this project) left a comment on my blog (https://raytracedminecraft.wordpress.com/) pointing me in this direction
16:22 ajmiles not sure if "Akos" rings a bell
16:27 RealBadAngel kahrl, hmmmm can i merge it?
16:28 celeron55 never seen anyone called "Akos"
16:28 celeron55 anyway, what you're doing and how minetest's rendering is set up is.... ehm
16:28 celeron55 ...kind of far apart
16:28 ajmiles I don't doubt it
16:29 celeron55 feel free to look around and see if there's something you can do, though
16:33 ajmiles I'm more of a graphics guy, looks like you have it covered
16:33 celeron55 modernizing a system designed to efficiently and dynamically manage and render mixed voxel and custom geometry, built on a rather outdated 3D engine, is not really a small or an iterative task
16:34 celeron55 that's the issue
16:34 celeron55 it works but there are really no upgradeable parts
16:35 celeron55 "looks like you have it covered" :D i can't but laugh at this
16:36 celeron55 [X] Minetest has graphics
16:37 ajmiles assuming it's not more than a couple of years old, why is it built on an outdated 3D engine?
16:37 celeron55 well, MT is from 2010
16:37 celeron55 back then Ogre3D had an unsuitable license
16:38 celeron55 so that's some history for you 8)
16:38 est31 joined #minetest-dev
16:38 est31 Elronnd, yes you can punch holes in NAT "firewalls"
16:39 est31 but not always and in all situations
16:39 est31 e.g. with double NAT
16:39 Elronnd Elronnd?
16:39 est31 ?
16:39 Elronnd I think you're talking to the wrong person, I haven't said *anything* since you joined
16:39 est31 I've seen the log
16:39 Elronnd Ah
16:39 Elronnd "ports system" freebsd ports system
16:39 Sokomine hm, rendering (voxel) worlds with a raytracing program? isn't that similar to what jordach does? though that's not real-time
16:40 est31 ah that ports system
16:40 RealBadAngel mt is not blender game
16:40 ajmiles it's similar to what Chunky does, but real-time. Not sure who jordach is though
16:41 RealBadAngel ajmiles, Jordach plays a lot with blender, he makes very nice renders in it
16:42 ajmiles right
16:45 Sokomine ...and RealBadAngel knows a lot about graphics and shaders and the like in minetest and develops that part further
16:46 est31 yeah so much that I can't make a real deep review of his PR :(
16:46 ajmiles draw distance (or lack thereof) was what always bothered me most about MC
16:46 est31 I'll try my best to not just point out style issues
16:47 jbb sokomine?
16:47 Sokomine hm, yes, i'm afraid that's a problem of mt as well. probably of all those games. you may stand right in front of a huge mountain and never notice it because that part of the map which contains the mountain hasn't been sent to you yet
16:48 est31 also, we should have a poll about how to document inside minetest.conf.example.
16:49 est31 first the description then the option or other way round
16:49 est31 (unrelated from the pr which)
16:49 est31 should be merged independently
16:51 ajmiles Sokomine, that's where ray tracing really comes into its own. draw distance is only really limited by memory and not by CPU/GPU powe
16:51 ajmiles you only pay the cost of the extra draw distance in the pixels that actually need to travel that far
16:52 celeron55 and you need a $1000 GPU
16:52 ajmiles not at all
16:53 ajmiles my card is about 3 years old and was only mid-end at the time
16:53 ajmiles you're not going to catch the Intel GPU crowd, but anything even vaguely recent from AMD or NVIDIA will be fine
16:53 celeron55 anyway, the issue with rendering far is that it all has to be dynamic and procedurally generated well enough to be actually worth playing in and looking at... and it all has to be run on old hardware too
16:54 RealBadAngel est31, there are no varying bools in glsl
16:54 celeron55 and you still need non-voxel things too
16:54 Sokomine hm, may still be too much? the game also runs on phones and tablets and old weak hardware. my own machine only has the integrated gpu of a celeron g1610
16:54 celeron55 can't get away with just having blocks only
16:54 troller joined #minetest-dev
16:54 est31 RealBadAngel, are there varying ints?
16:54 ajmiles the plan was always to rasterize non-voxel content the way I would now
16:55 ajmiles torches, animals, etc
16:55 RealBadAngel est31, propably yes
16:55 ajmiles can just do that after the voxel stuff has been rendered
16:55 RealBadAngel but since everything is float in there i prefer using floats
16:55 RealBadAngel and ints only for counters
16:55 est31 ints are usually better than floats for boolean like usage, but its only a nitpick
16:56 RealBadAngel also shaders behaves very strange on different gpus
16:56 RealBadAngel if i do have 0.0 value, check == 0.0 doesnt work all the time
16:57 RealBadAngel like it was having some discontinues
16:57 RealBadAngel so i have to check > 0.0
16:57 RealBadAngel its pretty weird i know
16:57 est31 yea thats the reason why floats are bad, no good == checks
16:58 est31 but didnt know that > is safe
16:58 proller__ joined #minetest-dev
16:58 ajmiles there's nothing inherently wrong with doing floating point equality tests so long as you understand how floats work
16:59 RealBadAngel ajmiles, way it work has nothing to do with that 0.0 sometimes is not equal 0.0
17:00 ajmiles 0.0 always equals 0.0
17:00 RealBadAngel thats a fucking bug in it nothing more
17:00 RealBadAngel no it doesnt
17:00 ajmiles one of your numbers probably wasn't 0.0
17:00 * est31 is rather too scared than having to debug an issue with float equality
17:01 RealBadAngel ajmiles, there are not many  numbers. just a few variables that can be 0.0 or 1.0
17:01 RealBadAngel also reading colors is affected
17:02 RealBadAngel if you will sample texture for 0.0 sometimes you will get it
17:02 est31 ok RealBadAngel +1 to your PR if it works, the style issues should be fixed too.
17:02 RealBadAngel but sometimes dont
17:02 RealBadAngel generally shaders work differently on each setup
17:02 ajmiles sure it's not a lowp/mediump/highp issue?
17:03 RealBadAngel just see my code
17:03 RealBadAngel and replace >0.0 with ==0.0
17:03 ajmiles if you link it I'll take a look
17:03 RealBadAngel and you will see that shaders will have discontinues
17:03 RealBadAngel sometimes the check will simply not work
17:04 est31 >0.0 is the opposite of ==0.0 isnt it?
17:04 ajmiles > is more like the opposite of <=
17:04 est31 yes ofc
17:04 RealBadAngel enerate_heightmaps > 0.0 means logic 1
17:04 RealBadAngel +g
17:04 est31 only for values >= 0 now
17:05 VanessaE kinda surprised you can't pass a true bool there.
17:05 RealBadAngel theres no way to pass bools between vertex and pixel shader
17:05 est31 graphics cards are number crunchers
17:05 ajmiles I can't find enerate_heightmap, did you mean enumerate?
17:05 est31 crunch crunch crunch
17:06 RealBadAngel ajmiles, https://github.com/minetest/minetest/pull/2783
17:06 RealBadAngel that code is my new PR
17:07 VanessaE and it doesn't properly apply...
17:07 est31 " This pull request can be automatically merged. "
17:08 est31 or you mean whitespace errors?
17:08 VanessaE est31: no, http://pastebin.com/uaiqTbG7
17:08 est31 w3ird
17:10 est31 yup I have it too
17:13 celeron55 sfan5, VanessaE: have you seen any spam so far since yesterday?
17:13 VanessaE celeron55: not yet, but I haven't had a chance to look
17:13 sfan5 nope
17:15 celeron55 we have gotten new registered users since then, but none of them has posted anything
17:15 celeron55 i hope there are no problems there; too lazy to check 8)
17:15 ajmiles RealBadAngel, are there particular cards/vendors you see that issue on?
17:16 VanessaE celeron55: well there's certainly plenty of new posts since yesterday anyway
17:16 est31 also, celeron55 can you give me "developer" status on the forums?
17:16 VanessaE don't see any spam
17:16 est31 not that it would matter
17:16 RealBadAngel ajmiles, i can see such behiaviour on mine. nvidia gti 550
17:19 ajmiles I'm more of a D3D guy where there's only one left of precision for floats (32 bit), but I suspect that's what it is
17:19 est31 gtg 4 now
17:19 est31 bye
17:19 ajmiles an issue of precision, perhaps a different precision mode in the VS vs the PS
17:21 ajmiles if it's a bool equivalent you're after then a lowp float that's set to 0 or 1 with a test of > or < 0.5 would seem most likely to be robust
17:21 celeron55 https://forum.minetest.net/vi​ewtopic.php?p=181773#p181773
17:21 celeron55 i wish i would be able to solve this
17:21 proller joined #minetest-dev
17:22 sfan5 more .php file patching maybe?
17:22 VanessaE celeron55: I get the same error.
17:22 sfan5 (not saying it is a good idea)
17:23 celeron55 well everyone gets that error
17:23 celeron55 but i have no idea how to fix it
17:27 VanessaE celeron55: here's something to look into:  https://forum.minetest.net/viewtopic.ph​p?f=9&amp;t=12596&amp;p=181673#p181673
17:28 VanessaE (that broken image links to the attachment below it, which renders only partly when clicked on)
17:29 VanessaE I assume it's just a failed upload, but figured I should point it out in case it isn't.
17:30 celeron55 oh that's easy
17:31 VanessaE delete the attachment? ;)
17:31 celeron55 ...mayabe
17:32 celeron55 i fixed this for www.minetest.net already, but it looks like i can't immediately find how i did that
17:33 celeron55 ...i think it was due to imagemagick not getting used
17:35 celeron55 wait, why would phpbb even edit an attached image at all?
17:35 VanessaE maybe to produce a thumbnail?
17:35 VanessaE (not that we ever actually *see* such things on this forum)
17:36 celeron55 why would the image break from that
17:36 celeron55 this makes no sense at all
17:37 VanessaE *shrug*   why would a shader variable set to 0.0 not equal exactly 0.0? :)
17:37 celeron55 maybe the user uploaded a broken image?
17:37 VanessaE that's what I figured, since only part of it actually renders when clicked, but since you've been having other problems, idk for sure
17:37 sfan5 now the attachment is gone
17:38 sfan5 o.O
17:38 VanessaE ?  wasn't me
17:38 celeron55 now i cleared phpbb's cache and the image is nowhere to be seen
17:39 sfan5 uhhhh
17:39 troller joined #minetest-dev
17:39 sfan5 wat
17:40 celeron55 at least zip attachments seem to still work, eg. here https://forum.minetest.net/v​iewtopic.php?f=11&amp;t=9917
17:41 VanessaE image attachments here work:  https://forum.minetest.net/viewtopic.ph​p?f=3&amp;t=156&amp;start=3425#p181413   but one of the two here does not:  https://forum.minetest.net/viewtopic.ph​p?f=3&amp;t=156&amp;start=3425#p181316  (half-broken image, like the post I initially mentioned)
17:41 VanessaE this post has another example of one good, one broken:  https://forum.minetest.net/viewtopic.ph​p?f=3&amp;t=156&amp;start=3400#p180760
17:42 sfan5 both attachments are broken at https://forum.minetest.net/viewtopic.ph​p?f=3&amp;t=156&amp;start=3425#p181316 for me
17:42 VanessaE I know for a fact that those two on p181316 were good before, so it isn't a "partially uploaded" problem.
17:43 neoascetic joined #minetest-dev
17:44 neoascetic Hi all. Why there is dev-0.5 branch in the repo? It will be same as master until stable 0.5 release, am I right
17:44 neoascetic ?
17:44 VanessaE neoascetic: dead branch
17:44 VanessaE ignore it
17:45 neoascetic Ohhh, why not delete it then? :D
17:53 Sokomine hm, i get a partial image (only about the topmost quater is shown) at https://forum.minetest.net/view​topic.php?f=9&amp;t=12464&amp;
17:57 Tauro joined #minetest-dev
17:57 Taoki hmmmm: I might have discussed this before, long ago. But any chance that you could add a builtin function to read the size of a schematic? I need one for my (apparently now functioning properly) town generation mod... and although I have one which still works, it relies on some binary hacking Lua-side which I don't believe I can maintain in case anything changes (and which might be slow).
17:59 Sokomine use my handle_schematics mod. it reads schematics and returns information about them
18:15 est31 joined #minetest-dev
18:18 Hunterz joined #minetest-dev
18:22 Sokomine hm. i'm having trouble with downloading a binary attachment (.zip file) from the forum now as well
18:28 est31 further opinions about 0.4.13 release date in 2 months?
18:31 VanessaE est31: fine by me, depends how long the feature freeze is though
18:33 est31 VanessaE, as an ex-victim of a feature freeze, I recommend that we use a -dev or similar branch to continue feature development, perhaps the moment the freeze starts, perhaps after
18:33 est31 with after I mean that branching is done at the intended release date when it has to be moved because blocker bugs are found
18:35 VanessaE well, I'm okay with most any development model, but a hard feature freeze at least encourages people to ONLY fix bugs
18:35 VanessaE whereas a separate dev branch doesn't come with that incentive.
18:36 est31 thats true too
18:36 rubenwardy joined #minetest-dev
18:41 ElectronLibre left #minetest-dev
19:25 MinetestForFun joined #minetest-dev
19:29 RealBadAngel https://github.com/RealBadAngel/minetest/commit​/936cfe1bcf86d2d582ad2138e6ed4cd301c52a15.patch
19:30 RealBadAngel ok, we have solved some AMD related issues
19:30 RealBadAngel going to push it soon
20:13 CraigyDavi joined #minetest-dev
20:36 est31 joined #minetest-dev
20:36 est31 RealBadAngel, you want to push 2783 now or not?
20:36 est31 because its at master branch
20:36 est31 https://github.com/minetes​t/minetest/commits/master
20:36 VanessaE I think he already did, but not from the PR
20:37 est31 I wonder because before it was that way too, then he removed it again
20:37 est31 http://irc.minetest.ru/min​etest/2015-06-14#i_4287632
20:38 est31 http://irc.minetest.ru/min​etest/2015-06-14#i_4287584
20:40 est31 also I wonder why he didnt fix the whitespace issues. I mean its no big deal they are in there, but I still wonder...
20:40 est31 so that together with him not closing the PR looks strange
20:40 VanessaE the whitespace problems were keeping it from going in
20:40 VanessaE plus there are still some corner-case performance issues to sort out
20:41 est31 you say so?
20:41 est31 could have said earlier :)
20:41 VanessaE didn't have a chance :)
20:41 VanessaE RBA and I were discussing in /msg about it.
20:41 est31 aha
20:42 VanessaE short version:  it works fine, but aniso+mipmap has some conflict if bumpmapping is enabled, causes a huge performance drop then
20:42 VanessaE and apparently only on AMD devices.
20:43 est31 so three options to go now: 1. leave it as it is, with all the whitespace errors (I only checked the style whitespace errors, not the trailing ones)
20:43 est31 2. hard reset to "Add UTF and other utility unit tests"
20:43 est31 3. extra revert commit
20:44 VanessaE 4. leave it as-is and do a whitespace cleanup as an additional commit
20:44 est31 when was the last time you spoke to RBA?
20:44 VanessaE about an hour ago.
20:44 est31 perhaps he's still around then
20:44 VanessaE he's in bed now
20:44 est31 oh
20:44 VanessaE been there since an hour ago
20:48 est31 so this is no bug where the build is totally fucked up
20:48 VanessaE nah
20:48 est31 neither its one where you cant log in
20:48 VanessaE it built fine for me once he just outright pushed it
20:48 est31 I guess it can stay in there until RBA fixes it :)
20:48 VanessaE yeah :)_
20:49 VanessaE however I'd suggest an automated whitespace fixup
20:49 VanessaE he was having major troubles with it
20:49 est31 what I wonder is why the commit message isn't "Improve displacement mapping"
20:49 est31 perhaps he pushed an older revision
20:50 VanessaE well he had to make a couple of AMD-specific tweaks to it
20:50 VanessaE and he did reset + push -f when it didn't work right the first time (it was in the 10 min window), but idk why the message is different
20:50 Wuzzy joined #minetest-dev
20:51 Wuzzy Any thoughts about this?: https://github.com/minetest/minetest/pull/1439
20:51 est31 yes Wuzzy seen it
20:52 VanessaE that would be useful
20:52 Wuzzy it will see its 1st birthday in a few weeks, lol
20:52 VanessaE but the standard keys for that are Q and E
20:52 VanessaE (in other games)
20:52 Wuzzy for what?
20:52 est31 ah sorry that isnt the new error message PR
20:52 Wuzzy hotbar stuff?
20:52 Wuzzy too bad that E is already occupied in Minetset
20:52 VanessaE yeah
20:53 VanessaE in games like quake/OA, Q and E are very handy to switch weapons
20:53 Wuzzy In CounterStrike, Q switches to previously seleted thing
20:53 Wuzzy anyways
20:54 Wuzzy I don't care much about the default bindings, so you could easily convince me to change my proposed defaults
20:54 Wuzzy unless its some arcane key combination of course XD
20:54 Wuzzy but if you want Q and E, then Use needs to go ... where?
20:54 VanessaE well I'd just go and change then myself anyway :)
20:54 est31 hrmm diff <(git show 43fcfbfe05578d747) <(git show 936cfe1bcf86d)
20:54 VanessaE "use" could be moved to 'f'
20:55 est31 this is quite empty
20:55 VanessaE I think that key is unused right now isn't it/
20:55 est31 so I'll mark it as merged
20:55 VanessaE est31: yeah
20:55 VanessaE just do a whitespace cleanup I think.
20:59 est31 its great the PR got merged
20:59 est31 now we can get minimap in future :)
21:00 Wuzzy ?!?!?!
21:00 Wuzzy minimap???
21:00 Wuzzy wow
21:00 VanessaE yep, that's next
21:00 est31 yes :)
21:00 Wuzzy actually part of core? I am surprised
21:00 est31 yes :)
21:00 VanessaE yep
21:01 VanessaE part of core and super-fast
21:01 VanessaE client-side
21:01 Wuzzy screenshot, maybe?
21:01 Wuzzy :)
21:01 est31 bin extra thread
21:01 est31 -b
21:01 VanessaE http://digitalaudioconcepts.com/vanessa/hob​bies/minetest/screenshots/random/Screenshot​%20-%2005202015%20-%2003%3a53%3a03%20PM.png
21:01 est31 client/shaders/nodes_shade​r/opengl_vertex.glsl:114: trailing whitespace.
21:01 est31 +       eyeVec = (gl_ModelViewMatrix * gl_Vertex).xyz;^M
21:01 est31 dammit
21:01 VanessaE (that's just the upper corner of the screen)
21:01 est31 windows line ends
21:01 VanessaE http://digitalaudioconcepts.com/vanessa/hob​bies/minetest/screenshots/random/Screenshot​%20-%2005262015%20-%2002%3a06%3a34%20PM.png
21:01 VanessaE another view
21:02 VanessaE and this is without the visual enhancements he's done since then
21:02 Wuzzy cool shit :-)
21:02 Wuzzy thanks to whoever did this :)
21:02 VanessaE RBA.
21:03 Wuzzy maybe I need to check GitHub more often XD
21:03 est31 no PR yet
21:03 est31 it was only in chat and a forum site
21:04 VanessaE https://github.com/RealBadA​ngel/minetest/tree/minimap2
21:04 VanessaE it's way behind master though
21:04 Wuzzy oh, its not official yet
21:04 VanessaE (there's a 'minimap' and a 'minimap3' also)
21:06 VanessaE he needs that parallax code that just went in, in order to make the minimap work properly
21:07 VanessaE ("properly" as he defines it that is ;) )
21:12 celeron55 VanessaE, Sokomine, sfan5: report problems here: https://forum.minetest.net/viewtopic.php?t=12611
21:12 est31 so here it is https://github.com/est31/minetest/commit/​238eae0b41f2864759012d55c3fab0c785067e13
21:12 celeron55 i can't really keep track of these on this channel
21:12 est31 pushing in 5 minutes if nobody minds
21:12 est31 did a git diff HEAD~2 HEAD --check
21:12 est31 and was empty
21:12 VanessaE celeron55: ok, no prob
21:12 est31 OTOH git diff HEAD~2 HEAD~1 --check
21:12 est31 that one is quite long
21:13 VanessaE est31: this by you is "long"? :P
21:13 est31 ?
21:13 VanessaE oh the other one derp
21:13 VanessaE anyway a quick glance it looks good
21:14 est31 try running git diff 43fcfb^ 43fcfb --check
21:14 est31 I've found a command on how to do such a thing with a rebase
21:14 est31 so then I did a rebase, and saved the diff between the two commits into a file
21:14 est31 then I applied that diff
21:14 est31 and commited it :=)
21:15 VanessaE nah, that's not so bad :P
21:15 VanessaE interesting that it throws whitespace errors just because of windows EOL differences.
21:16 est31 yea
21:16 est31 git has some magic in the underground
21:16 est31 but I'm not sure how it works
21:20 Sokomine my git complains that "error: Your local changes to the following files would be overwritten by merge:". which is quite astonishing, as i'm not aware of any changes i might have done to src/shader.cpp   does anyone know a way around that? i'm surprised that such a message shows up as well. i thought i had a clean version
21:21 celeron55 i now have a working system for creating such backups of the forum that contain everything except private information, and i plan to start publishing these in an automated way
21:21 celeron55 could some people (that i might trust) maybe take it and check through for what they can find?
21:23 est31 Sokomine, whats your git status?
21:24 Sokomine est31: http://www.pastebin.com/4rAgDcKJ
21:26 Sokomine i'm not aware of any changes i might have applied to that file. it seems highly unlikely
21:26 est31 do a git diff to find out :)
21:26 est31 or git diff src/shader.cpp
21:27 Sokomine yes, a git diff seems obvious, but: it doesn't show any changes. which makes the whole affair even more mysterious
21:28 est31 no idea then
21:28 * est31 suggests #git
21:28 est31 I had issues similar to this too
21:29 Sokomine ah. well...maybe. i had a similar problem earlier...
21:38 celeron55 VanessaE: okay now the attachments work properly; it was a permission issue with a weird nginx temporary directory where it buffers or caches replies from fastcgi (just writing it here so that i can figure out a permanent fix if it occurs again)
21:42 VanessaE celeron55: ok.  now for an acid test.....
21:43 VanessaE acid test failed :)
21:43 VanessaE https://forum.minetest.net/v​iewtopic.php?f=11&amp;t=7422  -> edit (but don't change anything) -> preview -> http://pastebin.com/b0svDMTE
21:44 VanessaE (figured I'd mention it now while you're messing with perms)
21:44 VanessaE happens on a lot of Dan's old posts for some reason.
21:45 est31 about #2785 can I merge that?
21:45 ShadowBot https://github.com/minetest/minetest/issues/2785 -- Remove reference to deprecated privilege by CraigyDavi
21:46 kilbith don't need to ask for such thing
21:47 est31 kilbith, I dont know when the last usage of that priv was
21:47 kilbith maybe VE knows
21:47 est31 and you know that xkcd with the "fixed 100% CPU bug"
21:48 VanessaE I've never seen that priv in use before
21:48 VanessaE besides, if a mod needs it, it can define it on its own
21:52 Taoki RealBadAngel: Is the new paxallax mapping in master now? Or will it be tonight? Didn't see more GIT logs about that fix
21:52 est31 *shrug* gonna merge it then, I guess nobody will complain that their admins can now revoke interact_extra without the privs privilege
21:52 VanessaE it is in.
21:52 Taoki Just curious :P
21:53 Taoki Oh, awesome
21:53 VanessaE it needs tweaks and tuning still
21:53 VanessaE but it works.
22:05 Sokomine i was given the advice in #git to hard-reset my copy via "git reset --hard origin/master", and that did work out. i had no local changes anyway
22:06 paramat joined #minetest-dev
22:09 est31 #2405 is another trivial PR
22:09 ShadowBot https://github.com/minetest/minetest/issues/2405 -- Ignore punches with no damage in the actionstream by SmallJoker
22:09 est31 gonna merge it
22:13 VanessaE est31: as long as you're in the damage/no-damage code, .....
22:16 est31 heh
22:17 VanessaE ;)
22:29 Wayward_Tab joined #minetest-dev
22:29 paramat left #minetest-dev
22:41 crazyR joined #minetest-dev
23:32 Taoki RealBadAngel: Got latest master, confirming the new parallax effect works for me (AMD / MESA). Thank you... it looks so sweet and realistic!
23:36 Taoki Strange however... it only seems to work sometimes o.o I switched back to the window as I wrote the above line, and the effect was gone... I was getting the old parallax. Also when I quit to the main menu than back in. So something might be causing it to switch between the two randomly.
23:56 jordan4ibanez joined #minetest-dev

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