Minetest logo

IRC log for #minetest-dev, 2013-03-31

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

All times shown according to UTC.

Time Nick Message
00:02 RealBadAngel hmmmm, have you seen the effect?
00:11 hmmmm woah
00:11 hmmmm i haven't seen it, but wow, that really looks nice
00:19 RealBadAngel hmmm, apply the changes to main.cpp
00:19 RealBadAngel and see it live
00:20 hmmmm okay
00:21 RealBadAngel got the source pictures?
00:22 hmmmm no
00:22 hmmmm that's alright though
00:22 RealBadAngel hold
00:22 RealBadAngel http://cisoun.dyndns.org/minetest.png
00:22 hmmmm what do i name it as?
00:22 RealBadAngel save it in /textures/all/menusplash.png
00:23 hmmmm ahh
00:23 hmmmm that white effect in the forground isn't a cloud
00:23 VanessaE nono menusplash_clouds.png
00:24 RealBadAngel http://i.imgur.com/Ub0L0Ky.png
00:24 hmmmm doesn't seem to matter to me
00:24 RealBadAngel and this into /textures/all/menuheader.png
00:24 hmmmm i just renamed menuheader to something else and it did the trick just fine
00:25 RealBadAngel so, got it working?
00:26 hmmmm yeah
00:26 RealBadAngel and?
00:26 hmmmm for some reason it's choppy
00:26 hmmmm the angle is right though
00:26 RealBadAngel ahhh
00:26 RealBadAngel i may know why
00:26 hmmmm nevermind that
00:27 RealBadAngel see khonkhortisan's pull request
00:27 hmmmm now that i'm thinking about it, i don't know if we should have this as the default background
00:27 hmmmm it gets screwed up if you don't have the same aspect ratio as the image
00:27 hmmmm (i don't)
00:28 RealBadAngel so fix the scaling just
00:28 hmmmm how do we do that?
00:28 RealBadAngel dunno yet :) but shouldnt be that hard to keep correct ratio
00:28 hmmmm :/
00:29 hmmmm make sure it doesn't look screwed up before it becomes a default part of minetest
00:29 hmmmm also this is not going in for 0.4.6
00:29 RealBadAngel i saw how its done
00:29 hmmmm feature freeze, seriously
00:29 hmmmm just cleaning up some code right now and fixing things
00:29 VanessaE feature freeze sounds fine to me also, btw
00:29 RealBadAngel it takes picture of any resolution, and fits the screen of given siz
00:30 RealBadAngel size
00:30 VanessaE (as if you give two shits ;) )
00:30 RealBadAngel without lookin at original size
00:30 hmmmm right now i'm going to fix the cave division by zero and whatever
00:30 hmmmm ahh
00:30 hmmmm i have an idea of how you do it
00:30 hmmmm you crop the edges
00:30 VanessaE cropping the sides is fine as long as you can avoid cropping too much
00:31 RealBadAngel well
00:32 RealBadAngel i got another idea
00:32 hmmmm (pic_cy/pic_cx) * screen_cy will be the width
00:32 RealBadAngel why forcing it to work with one image
00:32 RealBadAngel when we can split it as before?
00:32 hmmmm so screen_cx - ((pic_cy/pic_cx)*screen_cy) / 2 will be the proper offset
00:32 hmmmm horizontal offset
00:34 RealBadAngel that should keep proper ratio
00:34 RealBadAngel but also "minetest" logo can go for header and landscape for splash
00:34 hmmmm yeah
00:35 RealBadAngel this way folks can replace one or both
00:35 RealBadAngel so same logo different landscape
00:35 RealBadAngel maybe even default with time
00:36 RealBadAngel like slideshow
00:37 hmmmm hmm
00:37 hmmmm if you don't have that background, the clouds look weird at that angle
00:38 hmmmm if you have just clouds by itself, it needs to have the original angle
00:38 hmmmm how are we going to do this
00:38 VanessaE the config file idea maybe
00:38 RealBadAngel yeah i know, if theres no file specified they shall be reverted to ones before
00:38 hmmmm i don't want to overcomplicate adding in a texture pack
00:38 hmmmm we don't want the user to have to add a configuration line just to use a texture pack....
00:38 VanessaE something the tex pack maker supplies with the pack
00:39 RealBadAngel yup
00:39 hmmmm there needs to be some better way
00:39 hmmmm oh god, texture_meta.txt
00:39 hmmmm :/
00:39 RealBadAngel lol no
00:39 VanessaE think of how minetoon works, except a level below that
00:39 hmmmm map_meta, env_meta, texture_meta, meta_meta, blah blah
00:39 VanessaE nonono the tex pack maker will supply the config file as part of the pack, the engine shall look for it and read it from the same place the texture images come from
00:39 RealBadAngel lets just assume two possibilities
00:40 RealBadAngel or 3 even
00:40 VanessaE that way it's transparent to the user
00:40 hmmmm so you want to have a minetest.conf in the texture directory
00:40 hmmmm mmmmmm
00:40 RealBadAngel menu=clean, clouds, animated
00:40 VanessaE naw
00:40 VanessaE texturepack.conf
00:40 hmmmm that's basically the same thing as texture_meta.txt
00:40 VanessaE included right next to e.g. default_stone.png and so on
00:41 hmmmm all these .txt and conf files, i'm going mad
00:41 hmmmm game.conf
00:41 hmmmm conf.conf
00:41 VanessaE lol
00:41 RealBadAngel just one setting
00:41 hmmmm conf_meta.txt
00:41 VanessaE wait, you forgot to think of the future:  conf.conf.d
00:41 VanessaE ;)
00:42 hmmmm we should use systemd to make sure our configuration for the menu background is correct
00:42 hmmmm but seriously
00:42 hmmmm is it just me, or is there too much focus being put on a goddamn menu
00:42 VanessaE too much focus?  no.  too much on stuff that isn't really all that important in the grand scheme?  maybe a tad so.
00:42 RealBadAngel cmon, why do you paint walls at your house? :)
00:43 VanessaE for example, if I were making changes to the menu, the first thing I'd do is rip out that stupid 'tape and glue' and other stuff on the left edge of the menu
00:43 hmmmm i did that though
00:43 VanessaE nope.avi
00:43 hmmmm oh, you mean the lettering all together
00:43 VanessaE yeah
00:43 hmmmm i changed it from tape and glue to singleplayer
00:44 VanessaE is that what it says? it's so badly spaced on my screen I can't see the whole thing
00:44 VanessaE the tab already says singleplayer, why does the side need to also?
00:44 hmmmm yeah, i noticed that with a couple peoples' screenshots
00:44 hmmmm i don't know..
00:44 hmmmm celeron put it there, he thought it would look good
00:44 VanessaE that's one thing
00:44 VanessaE another is menu focus.  When I click over to, say, the credits tab, my focus is not on the scroll box,
00:45 VanessaE er text box/list
00:45 VanessaE so I can't scroll with my mouse wheel until I click on it.
00:45 hmmmm that's easily fixable
00:45 VanessaE it's the little trivial things that need polished on the menu
00:45 VanessaE the big stuff, not so much
00:45 hmmmm i can't test it though because my mouse wheel doesn't work anymore
00:46 * Exio hands an usb mouse to hmmmm
00:46 hmmmm i have one, it's just so worn out
00:47 VanessaE multiplayer public list suffers the same issue btw
00:47 VanessaE heh
00:47 VanessaE time for a new mouse
00:47 hmmmm too cheap
00:47 VanessaE I like my little wireless one.  13 bucks at wally world I think
00:47 VanessaE or 15, something like that
00:48 Exio my actual mouse was 35 ARS like a month ago
00:48 VanessaE hmmmm: that has GOT to irritate the fuck out of you when you browse the web though
00:48 Exio (if you wonder, ~6.8 USD)
00:48 VanessaE I feel completely helpless without a mouse wheel
00:48 VanessaE :)
00:49 RealBadAngel shall we donate for a mouse for hmm?
00:49 hmmmm not really, i adapted to it
00:49 VanessaE playing MT would become a challenge too
00:49 hmmmm i click on the middle mouse wheel and go down with my mouse in one motion
00:49 Exio for what is the mouse wheel needed in MT?
00:49 hmmmm and then let go
00:49 hmmmm scrolling down listboxes without having to click on the scrollbars
00:49 RealBadAngel i like A4tech mices
00:50 VanessaE Exio: it cycles through your hotbar, wield
00:50 RealBadAngel theyre rock solid
00:50 Exio oh
00:50 Exio am i the only who got used to 1-9? :P
00:50 VanessaE I guess so :)
00:51 hmmmm no
00:51 hmmmm i have to use 1-9 too now
00:51 hmmmm same with any FPS
00:51 hmmmm actually playing games is quite painful for my setup
00:51 hmmmm keeps me from wanting to waste my time on them
00:51 Exio that is true
00:52 Exio when i only had my netbook i was more "useful" in terms of code
00:52 Exio and tried to optimize things to the max
00:52 Exio now i just play 50% of the time, the other 50% is between irc, irc, irc, irc, vim, irc, irc, and irc
00:53 hmmmm i think i'm being more productive but in reality i'm just wasting more time on irc
00:53 RealBadAngel solution for better productivity: ban developers in irc
00:53 hmmmm mailing lists
00:53 RealBadAngel ban
00:53 VanessaE I wouldn't say time on IRC is wasted
00:53 VanessaE think about how much more you can get done when you can discuss things live
00:54 Exio suddenly engine reworked and 95% of the code rewrote in 10 days
00:54 RealBadAngel ofc not, just laughing
00:54 VanessaE after all, it's easy enough to ctrl-q if you need to avoid the distraction for a while
00:54 VanessaE (though no one discourages that of course)
00:54 RealBadAngel sharing ideas is also important
00:54 Exio VanessaE: it depends, every second you use in IRC "would be" a second in your IDE, or editor
00:54 RealBadAngel and discussing them
00:54 VanessaE Exio: the same is true for reading/writing emails on a mailing list.
00:55 VanessaE the only difference is how much focus you dedicate to an IRC chat versus an email
00:55 Exio the chat "eats" more seconds because realtime
00:56 Exio outside that, we wasting the time! i should be coding the net-lib for my project and watching videos for the course
00:56 RealBadAngel hehehe
00:56 Exio we are wasting our time*
00:56 RealBadAngel no, we are socializing :P
00:56 RealBadAngel use to it :)
00:57 Exio a cute and nice way to say "wasting"? :P
00:57 VanessaE joined #minetest-dev
00:57 Exio welcome back VanessaE :P
00:57 VanessaE oops
00:57 VanessaE speaking of ctrl-q..
00:57 Exio xD
00:57 RealBadAngel lol
00:57 VanessaE and rba
00:58 VanessaE and rba's right, socializing.
00:58 VanessaE anyone who doesn't like to socialize even a little probably needs to rethink their life.
00:58 VanessaE (or go get laid, one or the other ;) )
00:59 Exio I just "joke" about that, but the thing is, at some point it is useful
00:59 Exio but normally "it gets" just a waste of time
00:59 Exio when you were able to be coding, and i know what i'm saying
00:59 RealBadAngel i think prestidigator paid the price today for not being socialized
00:59 RealBadAngel dont you think hmmmm ?
00:59 Exio hehe
00:59 Exio that is an extreme
00:59 hmmmm not sure
01:00 VanessaE no, prestidigitator just didn
01:00 hmmmm with his personality, i bet he would've continue on anyway
01:00 VanessaE didn't communicate effectively.
01:00 RealBadAngel if he could start such massive change with some questions
01:01 VanessaE one more thing btw,
01:01 VanessaE offtopic:
01:01 hmmmm frog in the pot
01:01 hmmmm seriously
01:01 VanessaE well, no.  skip the offtopic part.
01:01 RealBadAngel he would end with good code
01:01 hmmmm he couldn't start out making small changes first?
01:01 VanessaE hmmmm: I agree.
01:02 hmmmm anyway what were you going to say
01:04 RealBadAngel hmmmm, propably he wanted to finish the task as he was seein it. his error was not to ask other if he is not breakin other things.
01:04 VanessaE me?
01:04 hmmmm yeah
01:04 hmmmm realbadangel, in his eyes he was fixing everything
01:04 VanessaE I was going to suggest a new default font and some minor text position tweaks to be inserted after 0.4.6 goes out
01:04 VanessaE http://forum.minetest.net/viewtopic.php?id=4436
01:04 VanessaE this.
01:05 hmmmm he believed that noise.cpp was horribly broken because it didn't follow his extreme OOP guidelines
01:05 Exio full OOP sucks unless you are insane
01:05 Exio it is ok at some point
01:05 VanessaE you know what?  fuck oop.  it's not the be-all, end-all of programming
01:05 VanessaE we got by just fine without it for what, 40 years?
01:06 Exio the OOP-way with objects and "the basic OOP rules" is good for me for very big projects
01:06 hmmmm vanessa, what do others think about it?
01:06 VanessaE hmmmm: other who have used it do seem to like it
01:06 sapier oop is (at least by now) best way to manage complexity ... maybe by some time someone will find a better paradigm
01:06 Exio it is not
01:06 sapier but as always to much of good is bad again
01:06 Exio it is just *other way* to manage the stuff, not the best
01:07 VanessaE hmmmm: as you can see on that thread, at least a couple of people saying it belongs as the default.  But some UI and formspecs will need text position tweaks, and the engine's default kerning needs adjusted slightly.
01:07 hmmmm changing it?  especially right before 0.4.6?  ehh..
01:07 sapier ok I forgot the imho  ;-)
01:07 hmmmm sapier, what do you think
01:07 Exio ah
01:07 hmmmm exio, you?
01:07 VanessaE nono, put it in right *after* 0.4.6
01:07 hmmmm oh
01:07 hmmmm i didn't read that
01:07 sapier think about what?
01:08 hmmmm a new default font
01:08 hmmmm dejavu sans with a black border for better visibility
01:08 Exio i don't bigger fonts really, but i can always change it
01:08 Exio with the same size?
01:08 hmmmm it looks slightly larger
01:08 hmmmm i don't know really
01:08 VanessaE Exio: it's slightly bigger than the old default, by maybe two points I guess
01:08 RealBadAngel you know what? i think coders that use to code on their own have to pass one test. stop thinkin their ideas are best ever and just ask others
01:09 hmmmm yeah, 1px border on each side
01:09 hmmmm 2 points larger
01:09 Exio hm, i should try it
01:09 hmmmm i've seen people use it in screenshots
01:09 hmmmm i'm not sure if it's better, but it certainly looks different
01:09 Exio any screenshot of it?
01:09 VanessaE there are some minor kerning issues in the engine - the font itself is perfet :)
01:09 VanessaE perfect* too
01:09 hmmmm and it would definitely help visibility
01:09 VanessaE Exio: um, first post of the thread?
01:10 VanessaE Exio:  http://forum.minetest.net/viewtopic.php?id=4436
01:10 Exio grr
01:10 VanessaE hmmmm: best part, it's free :)
01:10 VanessaE (as in beer and freedom)
01:10 hmmmm there is no screenshot of it
01:11 Exio in the game
01:11 hmmmm that's not a screenshot, that's the actual font file
01:11 VanessaE you mean like a screenshot of it in-game, sure
01:11 VanessaE hold a sec
01:11 VanessaE http://digitalaudioconcepts.com/vanessa/hobbies/minetest/screenshots/Screenshot%20-%2003092013%20-%2001:46:02%20PM.png
01:11 VanessaE as it appears in the menu
01:11 VanessaE http://digitalaudioconcepts.com/vanessa/hobbies/minetest/screenshots/Screenshot%20-%2002262013%20-%2008:16:56%20PM.png
01:11 VanessaE in unified inventory
01:11 hmmmm is that the kerning issue you speak of
01:12 hmmmm how the r and the p is cut off for example in the tabs
01:12 Exio well, i'd personally change it, but i see pretty good :P
01:12 VanessaE hmmmm: a few chars lose one pixel off the right edge
01:12 hmmmm yeah i see it now
01:12 hmmmm hm
01:12 VanessaE loke the r in Multiplayer
01:12 VanessaE like*
01:12 VanessaE the A in Advanced, etc.
01:13 hmmmm i would add it if it were as big as the current one is
01:13 Exio but i see it *
01:13 hmmmm it does look way bigger and it has the kerning issues
01:14 VanessaE the current font is just too small imho
01:14 jojoa1997|Tablet joined #minetest-dev
01:14 VanessaE but my eyes suck and my new glasses still have not arrived yet
01:16 hmmmm what do you guys think about this https://github.com/minetest/common/pull/17
01:16 hmmmm i'd tell him to squash it
01:16 hmmmm but i mean the actual content
01:16 VanessaE bones?  ew.
01:16 VanessaE save that for survival only.
01:16 hmmmm what are bones?
01:17 VanessaE give_initial_stuff seems fine to me, though I don't us eit
01:17 jojoa1997|Tablet No
01:17 VanessaE bones takes away all your stuff on death
01:17 VanessaE and puts them in a specially-textures locked chest
01:17 hmmmm oh that's definitely survival
01:17 jojoa1997|Tablet In build you have safe bonea
01:17 hmmmm come on 0gb
01:17 jojoa1997|Tablet You could use bones for recipes
01:17 jojoa1997|Tablet Like black dye
01:17 Exio he, i was late to say "shouldn't it in survival?" :P
01:17 Exio like white dye you mean
01:18 VanessaE G*d no
01:18 VanessaE please no
01:18 VanessaE ANYTHING but that
01:18 RealBadAngel :)
01:18 VanessaE bonemeal to make dye?  that's just stupid
01:19 Exio i'd like to see flowers in common
01:19 jojoa1997|Tablet Same here
01:19 RealBadAngel why not? kill some players to get dye
01:19 VanessaE Exio: as would I, but plants_lib needs to be ported to C++ first.
01:19 sapier kill vombies they give bones ;-)
01:19 * RealBadAngel was joking :)
01:20 jojoa1997|Tablet What I would like to see is item_pickup and making cheats drop items when dug added
01:20 VanessaE I suppose I could take a crash course in C++ and do it myself, but I guarantee you it would not blend well with MT coding style guidelines.
01:20 VanessaE jojoa1997|Tablet: I thought someone just added that?
01:21 jojoa1997|Tablet No
01:21 Exio i think with a "generate-ores" function like in the engine but for that stuff would be pretty good
01:21 jojoa1997|Tablet Unless someone shows a commit
01:21 Exio then just generate_plant in default :)
01:22 Exio in that case
01:22 VanessaE hm.  it's only 429 lines of code.
01:22 Exio i really think before "we" should wait for a biome-based mapgen
01:22 * Exio looks around :P
01:23 Exio and then do that stuff with the landscaper
01:23 Exio and then do that stuff with the landscape *
01:24 VanessaE bbl, time to bowl
01:24 jojoa1997|Tablet Oh
01:24 hmmmm https://github.com/minetest/minetest/commit/c517215bcf3c92ccef1690ad871a234e547c8435
01:25 jojoa1997|Tablet An easy biome definer would be nice
01:26 sapier cam someone explain to me what exactly a biome is and what data is in there?
01:26 hmmmm a biome
01:26 Exio a biome is a biome
01:26 Exio like in real life
01:26 jojoa1997|Tablet atm there is desert biome and plains biome
01:26 VanessaE sapier: which definition of biome?  as in mapgen v7, or in plants_lib?
01:26 hmmmm it is an area of land that has a certain set of attributes
01:26 jojoa1997|Tablet a biome is a general area that is the same
01:26 hmmmm such as a snowland, a desert, a mountan range, plains, so on
01:27 hmmmm you know
01:27 VanessaE hmmmm: hm?
01:27 hmmmm it has a specific scientific meaning too
01:27 VanessaE oh
01:27 hmmmm Biomes are climatically and geographically defined as similar climatic conditions on the Earth, such as communities of plants, animals, and soil organisms,[1]
01:27 hmmmm http://en.wikipedia.org/wiki/Biome
01:28 VanessaE [citation needed]
01:28 VanessaE l(
01:28 VanessaE ;)
01:28 sapier I know I was aiming at the fact that mobf hase something like "environments" defining where a mob should be ... this requires e.g. air to be in there too ... as far as I understood this information isn't within current minetest biome definition ... is it?
01:28 hmmmm mm yes
01:28 hmmmm for mapgen v7, the intent is to tell lua what area is what biome
01:29 hmmmm so you'd be able to spawn mobs in specific biomes
01:29 sapier it's not only for spawning i need this information for movement too
01:29 Exio with the mapgen v7 there will be a "whatbiome" function for nodes/blocks, etc? :P
01:29 sapier e.g. a fish should never swim to air ;-)
01:30 VanessaE ok, now I'm off.
01:30 sapier great I'm looking forward to it :-)
01:30 Exio good night VanessaE
01:31 sapier hmm ... wait ... today night clock is skrewd up .... oops I think I should be in bed for an hour now ... gtg :-)
01:32 sapier left #minetest-dev
01:32 Exio hahaha
01:37 hmmmm what do you guys think about https://github.com/minetest/minetest/pull/579
01:38 hmmmm it looks good enough
01:38 Exio if it works, why not?
01:38 hmmmm i don't know if it works, i didn't test it.... but it doesn't look from the code like there's any reason for it to not work
01:39 Exio it seems to work, for me too
01:39 jojoa1997|Tablet hmmmm i would love that
01:39 hmmmm most patches, especially small ones like that, it's much less effort to just read the changes rather than test it
01:39 jojoa1997|Tablet also
01:39 jojoa1997|Tablet how do you give singleplayer all privs
01:39 Exio ./grant singleplayer all
01:41 hmmmm there's a minetest.conf setting so you don't have to keep doing that on new worlds
01:52 Exio what is the next thing in your todo list hmmmm?
01:53 hmmmm flags field for oregen, base proller's latest modification on that
01:54 VanessaE ok, back now.  bowled 215 twice in a row.  sucky.
01:55 Exio hmmmm: some info for you in #mt
01:56 hmmmm hm? you mean #minetest?
01:56 hmmmm where
01:56 Exio he, yep
01:56 VanessaE hmmmm: 579, +10000
01:57 Exio i got used to say "#mt" "#mt-dev" "##se" "#fn" and so on :P
01:57 VanessaE the code looks benign enough, too
01:57 hmmmm i am doing like 8 pull requests at once.
01:57 hmmmm get on my level
01:58 VanessaE ok fine. 30/56.  99 are correct.  Limited 4 and 8 are missing.
02:08 hmmmm first things i want to do after 0.4.6 is released is finally add the LuaJIT, then add leveldb, then add sapier's huge change
02:09 hmmmm what are your opinions on this https://github.com/ShadowNinja/minetest/commit/e7d9527e5185f647ee376e65190dac3c22b96009 ?
02:10 VanessaE hmmmm: seems fair to me.
02:11 hmmmm oh woa woa
02:12 hmmmm https://github.com/PilzAdam/minetest/commit/38a6a7f5af3572d3100701f1d342e3a9238cce7b  pilzadam, do you really mean to do this?  you're copying an entire std::list for the sake of copying it again to newlist which probably copies it a third time
02:13 VanessaE hmmmm: #minetest, your input is needed
02:29 Exio just a random thing, what about this? https://github.com/minetest/minetest/pull/539
02:29 Exio a cosmetic and "just for completeness" change
02:29 Exio stacks are already infinite, why not tools too?
02:30 hmmmm no idea
02:30 hmmmm need more of a consensus on that
02:30 Exio yep, i just throw it
02:30 VanessaE I would say yes.
02:30 VanessaE even if the hand breaks everything in half a second, some tools are still faster
02:31 VanessaE when you get used to it, suddenly half a second seems like forever.
02:31 Exio i have changed the half second in my "local" repo to 0.1
02:34 VanessaE [03-30 22:30] <sdzen> does anyone think that a server should announce the *** Slowpoke joined the game, after it leaves the loading screen?
02:34 VanessaE [03-30 22:33] <VanessaE> sdzen: I agree
02:34 VanessaE [03-30 22:33] <VanessaE> it should not announce you have joined or show you in the world to other players until you yourself are able to play
02:34 VanessaE (for after 0.4.6 I think)
02:34 Exio just looking in the pull requests, https://github.com/minetest/minetest/pull/453 wasn't this already merged?
02:35 VanessaE yes
02:35 hmmmm plizadam forgot to close that
02:44 Exio // WTF is this? It can't be right
02:44 Exio lovely comments
02:56 Exio lol, started geany for checking something in the config and segfaulted
03:08 Exio what the hell?
03:08 Exio changed one line in player.h because it didn't compile
03:08 Exio now i need to recompile 80% of minetest, other time
03:10 VanessaE hmmmm: got a second?
03:15 jojoa1997|Tablet joined #minetest-dev
03:30 Exio does anyone like this? https://github.com/EXio4/minetest/commit/b448b431a0e47bc2f5d95bf99435509431eb07b1
03:31 Exio just tested it a bit, and does 95% of what the "optifine zoom" thing does
03:31 VanessaE Exio: can you bind that to the middle mouse button?
03:32 VanessaE (as in Quake, OpenArena, et.al)
03:32 Exio hm... no for what i can see, (i just use the "same way" as other keybinds are working)
03:32 ShadowNinja joined #minetest-dev
03:32 VanessaE ok, well someone can do that later
03:33 jojoa1997|Tablet Allowing you to have to be zoomwd in for certain areas would be nicw
03:33 jojoa1997|Tablet Like walking in radius of a special node
03:33 VanessaE it would be quite handy in fact.
03:33 Exio do you mean something like a "set_fov" from lua?
03:35 Exio i use it with left control, but i don't see if that key "should be default" for that
03:35 Exio i will take two screenshots for jojoa1997|Tablet now :)
03:41 Exio jojoa1997|Tablet: http://exio4.com.ar/_/mt/zoom_out.png http://exio4.com.ar/_/mt/zoom_in.png
03:44 jojoa1997|Tablet Wow that is a nice zoom distance
03:44 Exio now i need to recall how you squash commits
03:48 khonkhortisan something like git rebase -i HEAD~5
03:51 Exio k, done
03:51 Exio https://github.com/EXio4/minetest/commit/3390906f3aad519c9ac5dbb19f3b407946f52b06
03:56 Exio i https://github.com/minetest/minetest/pull/583
03:56 Exio my work is done *
03:57 Exio i'll remote the (like optifine) in the commit message, that is already in the pull request
03:57 VanessaE why>
03:57 VanessaE ?
03:57 VanessaE just leave it
03:58 Exio for making the commit message "shorter"?
03:58 Exio but ok
03:58 VanessaE naw
03:58 Exio i still didn't do anything :P
03:59 VanessaE because your commit will be what gets added, so that part won't be redundant
03:59 Exio hehe
04:03 Exio the only things what i have coded in C++ for minetest are pure "cosmetic" things
04:03 Exio the crosshair/selectionbox color and this
04:03 VanessaE that's all I've done too
04:03 VanessaE for C++ I mean
04:04 Exio kaeza is going to like this change, a bit
04:05 Exio s/change/option/
04:18 jojoa1997|Tablet It would be notice for zooming in with his gun mod
04:20 Exio a proper way to do that would be with the lua_api, but i will not touch it until sapier's pull request gets implemented
04:26 hmmmm yeah?
04:26 hmmmm exio, could've sworn you've done more than that
04:27 Exio hehe
04:28 hmmmm i like that zoom thing
04:28 hmmmm someone else had a patch that never got merged where they changed the FOV while running, but that was just weird
04:28 hmmmm this is actually useful though
04:29 Exio jeija
04:29 Exio https://github.com/minetest/minetest/pull/385
04:31 hmmmm yeah... there was never a consensus on people liking it, so..
04:31 Exio well, i like higher FOVs when playing
04:31 hmmmm i'll add your zoom though
04:31 Exio if it is an option disabled by default (like finite_liquid is actually)
04:32 Exio thanks? :P
04:32 hmmmm not for 0.4.6, we're in a feature freeze right now
04:32 Exio i got that
04:32 Exio when 0.4.6 gets released there will be like 10 big changes in MT
04:33 Exio the leveldb, luajit, sapier's api change
04:33 hmmmm minetest is really developing fast, contrary to what people think
04:34 hmmmm it's like we have more than enough material for a version change every month
04:34 kaeza hm?
04:34 Exio kaeza: https://github.com/minetest/minetest/pull/583
04:35 Exio hmmmm: minetest: http://en.wikipedia.org/wiki/Exponential_function
04:35 hmmmm lol
04:36 kaeza Exio, that is great
04:58 Exio hmmmm: just a random question what i have (and i wonder), what is your age?
04:58 hmmmm 21
04:58 Exio that makes me feel like a child,
04:58 Exio meh :P
05:03 kaeza 21? /me feels so old now :(
05:23 jojoa1997|Tablet Off to bed cya
05:33 hmmmm wait a minute
05:33 VanessaE ?
05:33 hmmmm proller, why don't you just register two more ores for your height ranges?  why do you need an absolute value anyway
05:35 khonkhortisan rainbow skycaves
05:38 RealBadAngel eh, cmon
05:39 hmmmm well i am still adding it
05:39 RealBadAngel diggin shall be fun
05:39 hmmmm under the premise of saving another ore registration
05:39 RealBadAngel even with 10+ ores caves are not rainbow ones
05:40 RealBadAngel ofc with wise conditions
05:40 RealBadAngel different heights, rarity etc
05:42 Exio well, playing with the new minetest.register_ore
05:42 Exio i made pseudo-floating-islands with dirt and so on
05:45 RealBadAngel hmmmm, what are the plans for v7?
05:46 hmmmm http://dev.minetest.net/Mapgen_V7
05:48 RealBadAngel so we are facin the problens hihglighted now?
05:48 hmmmm well my first approach, yeah..
05:48 hmmmm i'm planning on scrapping it
05:48 hmmmm the new approach can work out better
05:48 RealBadAngel all?
05:48 hmmmm yeah
05:48 hmmmm i don't think it'll be too hard to come up with something better
05:48 RealBadAngel you surely have too much spare time
05:48 RealBadAngel lol
05:49 hmmmm mapgen v7 is being made to be good
05:49 hmmmm i'm going to put the node coloration thing off to the side for now, along with the block sending code
05:49 RealBadAngel ok, so now make it PERFECT
05:49 RealBadAngel not just good
05:49 hmmmm if you're wondering about the node coloration, see the TODO
05:50 hmmmm i want to add a new type of node, where you can define a table of colors and change node colors that way
05:51 RealBadAngel can you say how would u like to achieve that?
05:52 hmmmm i did
05:52 RealBadAngel in TODO?
05:52 hmmmm yea
05:52 RealBadAngel link?
05:52 hmmmm lol
05:52 hmmmm http://dev.minetest.net/TODO#Add_colorlike_to_node_param_types
05:53 RealBadAngel (sorry, bit tired)
05:53 hmmmm calinou messed up that page by changing all the occurances of "chunk" to "block" when i really mean "chunk"
05:53 hmmmm ..grr
05:54 Exio chunks -> mapblocks -> nodes
05:54 Exio no? i mean, a chunk has mapblocks, and mapblocks "has" nodes?
05:55 hmmmm a chunk is an NxNxN cube of mapblocks
05:55 RealBadAngel hmmmm, i can see one problem already
05:55 hmmmm a sector is a 1xNx1 selection of mapblocks
05:55 hmmmm RBA, ?
05:55 RealBadAngel new facedir is using more bits now
05:56 hmmmm i thought it's only using 3?
05:56 RealBadAngel was
05:56 hmmmm er 5
05:56 hmmmm how many is it using now
05:56 RealBadAngel mask 1F
05:56 RealBadAngel 0-23 dec
05:56 hmmmm that's 5 bits
05:56 hmmmm 1 | 2 | 4 | 8 | 16
05:56 RealBadAngel yup
05:57 hmmmm so yeah, you can still do 8 colors with facedir
05:57 hmmmm what are you talking about then
05:57 RealBadAngel really?
05:57 hmmmm 8 bits in a byte
05:57 hmmmm you use 5
05:57 hmmmm i use 3
05:57 RealBadAngel its a wrong approach imho
05:57 hmmmm what should i do then?
05:57 VanessaE 8 colors?  ew.
05:57 Exio how does MT handle the endian-ness in maps / so?
05:57 khonkhortisan 1F -> 00011111
05:57 RealBadAngel no no no
05:57 hmmmm exio, it's big endian
05:58 Exio always?
05:58 hmmmm yeah
05:58 RealBadAngel hold on all of you
05:58 Exio ok
05:58 hmmmm see serialization.cpp
05:58 khonkhortisan holding on...
05:58 Exio k, thanks
05:58 RealBadAngel store modified data in nodes meta
05:58 hmmmm can't do that at all
05:58 hmmmm do you realize how heavy node metadata is?
05:59 RealBadAngel c55 did that
05:59 hmmmm stored node colors in node metadata?
05:59 RealBadAngel more
05:59 hmmmm node metadata is for things like storing items in chests
05:59 hmmmm definitely not this
05:59 hmmmm node metadata wouldn't be fast enough for what i want to do either
06:00 RealBadAngel seen this code workin?
06:00 hmmmm i haven't even started it yet
06:00 hmmmm i just wrote on the todo what needs to be done, roughly
06:00 RealBadAngel it allows to change textures, nodeboxes on the fly
06:00 hmmmm i'll code it up and show you, and prove you wrong
06:00 hmmmm it'll work great
06:01 hmmmm but not now
06:01 hmmmm mapgen v7 needs to get done
06:01 hmmmm it's so anticipated
06:02 RealBadAngel if you do that ill owe you big johny waker bottle, black labeled.
06:02 hmmmm ok :)
06:02 RealBadAngel *walker
06:03 RealBadAngel deal?
06:03 hmmmm deal
06:03 Exio haha
06:03 RealBadAngel cool :)
06:06 RealBadAngel so, by now v7 will be next step yes?
06:06 hmmmm yes
06:07 RealBadAngel after 046 i mean
06:07 hmmmm of course.. i don't know, mapgen v7 is a big thing
06:07 hmmmm it'll stay experimental for some time too
06:07 hmmmm but if i were to actually sit down and work on it, i'd probably get it finished in less than a week
06:08 RealBadAngel so move your lazy ass and get it done ;)
06:08 hmmmm i'm going to bed soon
06:09 RealBadAngel sleeping time (1 hour) granted
06:09 RealBadAngel bonus! youre the winner ;)
06:10 hmmmm and i thought i was getting relaxation on my break
06:10 RealBadAngel me too, and im rewritting whole technic
06:12 RealBadAngel and whats most funny im gettin annoyed seein some parts of it
06:12 Exio that happens when you code stuff, always
06:12 Exio i've coded stuff a year ago, or just some months ago, and when i see it now
06:12 Exio i just say "WTF, HOW CAN THIS EVEN WORK"
06:12 RealBadAngel "fuck! who did that???"
08:24 darkrose joined #minetest-dev
09:51 PilzAdam joined #minetest-dev
11:04 darkrose joined #minetest-dev
11:33 troller joined #minetest-dev
12:53 PilzAdam joined #minetest-dev
14:20 serengeor joined #minetest-dev
14:36 Zeg9 joined #minetest-dev
14:41 sapier joined #minetest-dev
14:45 PilzAdam joined #minetest-dev
14:57 celeron55_ hmmmm: I'll just note this: your approach to management will work, as long as you are around doing stuff - but if you leave (well, probably you will, given infinite timespan), then much more will be dead than with a more liberal approach
15:01 celeron55_ i respect you as long as development overally works though
15:03 celeron55_ i have seen many people like prestidigitator when i have been in the middle of coding some of my own bigger minetest things
15:03 celeron55_ they never understand that communication is key to everything
15:04 celeron55_ it's a shame
15:05 jojoa1997|Tablet joined #minetest-dev
15:27 Rick_ joined #minetest-dev
15:30 Rick_ Hey guys, could I get a bit of help here? I'm trying to compile my own version of the source and am getting some linker errors.
15:31 Rick_ I have the required libraries in a directory called DIR (currently building with sound disabled) but it seems to be having trouble with irrlicht
15:31 Rick_ Here is the output I'm getting:
15:31 Rick_ Linking CXX executable ..\bin\minetest.exe Creating library file: ..\bin\libminetest.dll.a CMakeFiles\minetest.dir\sky.cpp.obj:sky.cpp:(.text$_ZN3irr5scene10ISceneNode11ge tMaterialEj[__ZN3irr5scene10ISceneNode11getMaterialEj]+0x1): undefined reference  to `_imp___ZN3irr5video16IdentityMaterialE' CMakeFiles\minetest.dir\content_cao.cpp.obj:content_cao.cpp:(.text$_ZN10GenericC AO16updateTexturePosEv[__ZN10GenericCAO16updateTextur
15:31 jojoa1997|Tablet What are your os
15:31 Rick_ I'm running on a windows laptop
15:31 Rick_ widowns 7
15:33 Rick_ hmm it seems to have cut off the end of my paste, anyways there are 3 similar messages to that one output after followed by these error messages:
15:34 Rick_ collect2.exe: error: ld returned 1 exit status make[2]: *** [bin/minetest.exe] Error 1 make[1]: *** [src/CMakeFiles/minetest.dir/all] Error 2 make: *** [all] Error 2
15:36 sapier the functions written in your error are irrlicht functions maybe library is not correct
15:36 Rick_ Thats what I was thinking, so I've been messing with it a little bit
15:37 Rick_ I have Irrlicht 1.7.3 in the parent directory of my mintest folder
15:37 Rick_ and I've been referring cmake to the Irrlicht.lib in irrlicht's win32-visualstudio folder
15:38 sapier :-) if you've been already messing around most likely starting from scratch will be better than trying to find what went wrong ;-)
15:38 sapier do you compile with visual studio?
15:38 Rick_ Well, all I've changed is a bit in the cmake findIrrlicth file
15:38 Rick_ no I am compiling with MinGW
15:38 sapier mingw and visual studio libs don't matc
15:38 sapier h
15:39 Rick_ oh really?
15:39 sapier at least they didn't when I was using windows some years ago
15:39 Rick_ so I have to compile my own version of the Irrlicht library then?
15:39 celeron55_ Rick_: are you using the mingw version of Irrlicht.lib?
15:39 sapier or find a precomiled one for mingw
15:40 sapier (16:39:15) Rick_: and I've been referring cmake to the Irrlicht.lib in irrlicht's win32-visualstudio folder
15:40 Rick_ I dont think that I am. I had wanted to use the precompiled version that came with the irrlicht library to save some time
15:40 celeron55_ irrlicht stopped distributing prebuilt mingw versions after 1.7.3
15:40 celeron55_ i mean, after 1.7.2
15:40 celeron55_ just get irrlicht 1.7.2 if you're in a hurry
15:40 Rick_ Okay, I'll do that then
15:41 Rick_ thanks a lot guys!
15:41 sapier good to know too :-) I assume I would have discovered this once updating my win32 build env
15:41 celeron55_ http://downloads.sourceforge.net/irrlicht/irrlicht-1.7.2.zip
15:43 celeron55_ (all this assuming my memory actually works; i might be a version off or something)
15:44 sapier https://github.com/minetest/minetest/pull/585 delete texture once it's rebuilt
15:44 Rick_ I just looked in the directory in the zip, the win32-gcc contains a .dll and a .dll.a
15:44 Rick_ I'm assuming those are the ones I want
15:46 Zeg9 Yeah, that's the ones mingw32 uses (at least under linux...)
15:46 Rick_ Actually, theres a Irrlicht.dll in the bin/Win32-gcc, and then which of the lib files do I use?
15:46 VanessaE sapier: how much of a fix does that result in?
15:46 Rick_ in lib/win32-gcc theres libIrrlicht.dll.a and libIrrlicth.dll
15:47 sapier don't know but overwriting a pointer containing an allocated texture is always bad :-)
15:47 celeron55_ Rick_: .dll.a = lib
15:48 Rick_ ah cool, thanks a lot! Sorry, I'm a bit inexperienced with writing code outside of school, we tend to avoid these kinds of things.
15:48 celeron55_ that's why people who come from school are so useless 8)
15:49 celeron55_ half of actual coding is not writing code but fighting with all kinds of... things!
15:49 Rick_ haha you can look at it that way I guess.
15:50 Rick_ I am also much more focused on the embedded space though, debugging microcontrollers and hardware interactions is easy
15:50 Rick_ software meanwhile seems to be a hodge podge of random tools and names
15:51 celeron55_ well, software is extremely prone to this: http://xkcd.com/927/
15:51 celeron55_ in hardware making something differently usually at least costs something
15:54 sapier "debugging microcontrollers and hardware interactions is easy" unless you've got a 20 year old asic with no/wrong documentation :-)
15:59 hmmmm joined #minetest-dev
16:00 rubenwardy joined #minetest-dev
16:00 Rick_ Haha, thats why you shoot for new versions of certain products. TI microcontrollers are nice, even new SI labs controllers will make you pull your hair out.
16:07 Rick_ This makes me so happy =): Linking CXX executable ..\bin\minetest.exe Creating library file: ..\bin\libminetest.dll.a [100%] Built target minetest
16:08 sapier you can't use "new" products if you're developing industrial grade embedded devices ;-)
16:15 PilzAdam hmmmm, the ore gen of the mese blocks with scarcity=38^3 and num_ores=3 results in: "register_ore: clust_scarcity and clust_num_ores must be greater than 0"
16:16 hmmmm ah yes
16:16 hmmmm it must be clust_scarcity and clust_num_ores
16:17 hmmmm not scarcity and num_ores
16:17 PilzAdam ehm, it has it, I just shortened it
16:17 PilzAdam it appeard today the first time
16:17 PilzAdam (the message)
16:17 hmmmm erm, can i see your code?
16:17 PilzAdam its the upstream code
16:18 hmmmm common?
16:18 PilzAdam yes
16:18 PilzAdam common/mods/default/mapgen.lua:122
16:20 hmmmm uh oh
16:20 hmmmm clust_scarcity is an s16 isn't it
16:21 PilzAdam make it an u16
16:22 hmmmm i think it's better to have that as a u32
16:22 hmmmm now why the hell did i do that...
16:26 hmmmm https://github.com/minetest/minetest/commit/35ed5198fe45eceb62b33dfb1b85f019ee4016e1
16:33 iqualfragile joined #minetest-dev
16:33 Rick_ So after finally linking successfully, I tried to run the produced executable. I put Irrlicht.dll and zlibwapi.dll in the same folder as the executable for it to actually run (compiled without sound).
16:33 Rick_ The problem is it only launches a blank window and then crashes
16:34 Rick_ crashes right after it prints GLSL  version: 0.0 on the cmd it opens when executing
16:34 Rick_ Anyone have any helpful hints for me =(?
16:34 sapier disable shaders?
16:35 VanessaE why are shaders even enabled by default anyway?
16:35 VanessaE too many people have trouble with those
16:35 VanessaE (and they cost fps for some people, too)
16:36 Rick_ umm how do I go about that
16:36 Rick_ I'll google it
16:36 VanessaE Rick_: in the "Settings" menu of the game
16:37 PilzAdam VanessaE, because its a benefit for a lot more people
16:37 VanessaE disable it by default anyway
16:37 VanessaE for the same reason the filtering stuff I added is disabled by default
16:37 celeron55_ enable_shaders = 0 to minetest.conf (how are you going to go to a Settings tab if the window just crashes? 8))
16:37 VanessaE always assume such features will break
16:38 VanessaE oh hell, I didn't see that
16:40 celeron55_ there should be some kind of autodetection for shaders (http://irrlicht.sourceforge.net/docu/classirr_1_1video_1_1_i_video_driver.html#adde468368b77441ada246e1603da4f47)
16:49 Rick_ So I went to the minetest.conf file in the parent directory to the executable, uncommented and changed to enable_shaders = 0, still getting the same crash though.
16:49 sapier did you build RUN_IN_PLACE?
16:49 VanessaE Rick_: that wasn't the .example file was it?
16:50 Rick_ no, I made sure it was just the minetest.conf
16:50 Rick_ yes RUN_IN_PLACE is set true
16:51 hmmmm it might be helpful if you compile a debug version and try to get a backtrace
16:53 Rick_ ok, just changed build type to Debug, building now.
17:11 Rick_ so after the rebuild in debug and running it I get this:
17:11 Rick_ 11:10:16: ERROR[main]: Test (hand_server.count == 1) failed: UASSERT 11:10:16: ERROR[main]: Test (hand_server.last_id == 2) failed: UASSERT 11:10:17: ERROR[main]: run_tests(): 1 / 12 tests failed. 11:10:17: ERROR[main]: run_tests() aborting.
17:11 PilzAdam run it with --disable-unittests
17:12 Rick_ when I do that it fails in exactly the same way as it did before
17:12 Rick_ this is the full output after running:
17:12 Rick_ Irrlicht Engine version 1.7.2 Microsoft Windows 7 Home Premium Edition Service Pack 1 (Build 7601) Using renderer: OpenGL 3.0.0 Intel(R) HD Graphics Family: Intel OpenGL driver version is 1.2 or better. Warning: OpenGL device only has one texture unit. Disabling multitexturing. GLSL version: 0.0
17:14 celeron55_ run it under gdb
17:15 celeron55_ (i'd have initially suggested just setting full logging and looking at the log first though; dunno why people went directly for a backtrace)
17:18 Rick_ seems irrlicht hates me
17:18 Rick_ Program received signal SIGSEGV, Segmentation fault. 0x634c9fee in irr::CIrrDeviceWin32::setResizable(bool) ()    from C:\Users\Ricky\My Documents\Academic\ECEN\Code\DIR\minetest\bin\Irrlicht .dll
17:18 jojoa1997|Tablet joined #minetest-dev
17:19 jojoa1997|Tablet Pilzadam good news
17:19 jojoa1997|Tablet Oh wrong channel
17:24 Rick_ Seems like I'm going to need to compile a 64 bit version myself. Doesn't like the 32 bit .dll I guess?
17:30 hmmmm perhaps
17:31 hmmmm are you mixing 32 bit and 64 bit binaries?
17:31 hmmmm how does that even work
17:31 Exio well, https://github.com/minetest/minetest/pull/583
17:31 hmmmm exio, not until 0.4.6 is released
17:31 Exio should i add the mouse_sens option with that, too?
17:31 Exio hmmmm: yes, not talking about that
17:31 hmmmm no idea what that is
17:31 Exio mouse sensitivity, when you are zooming it is "slower" now
17:32 hmmmm oh yes
17:32 hmmmm definitely
17:33 Exio actually the value is "real_value / 4" when zooming. i think a "proper fix" should be adding a config and, "by the way" fix this issue and some older one what i don't recall https://github.com/minetest/minetest/issues/517
17:33 Exio but, i should start other pull request or making two commits in that pull request?
17:33 PilzAdam Exio, I dont want to handle the zoom in Lua, I just want to tell Lua about it
17:33 Exio ah
17:34 hmmmm exio, i think there's a point where there's too much configurability
17:34 hmmmm the zoom mouse slowdown factor, i think, is the point where it gets crazy
17:34 hmmmm just leave it at 4
17:34 Exio i didn't mean that, just the sensibility "raw value", the /4 seems perfect (imo) for this
17:35 PilzAdam still,  Im kinda against the whole idea
17:36 Exio hm
17:37 PilzAdam I could agree with it, if the setting was server side
17:38 Exio that is what i wanted to avoid by making it fully client-side, i don't like messing with the protocol
17:38 hmmmm wait a minute
17:38 hmmmm the address of rick's segfault
17:38 hmmmm it looks familiar
17:39 Rick_ ?
17:40 hmmmm hrmm, nevermind, unless irrlicht's binary really spans that far
17:40 hmmmm https://github.com/minetest/minetest/issues/534
17:40 hmmmm just kinda looked like that, but that's 0x63 and this is 0x69
17:41 hmmmm see, sapier and i figured the address in that issue was openal
17:41 hmmmm we don't really know unless we take a look at irrlicht's build script and see if it sets 0x63 or something as its preferred base address
17:47 Rick_ I thought the issue was that the dll is 32 bit and I was compiling for 64 bit which doesn't really fly. I dont think that was it though, I think it was using my 32 bit compiler so I'm not sure...
17:48 sapier hmm valgrind did show me lots of openal warnings about memcpy with overlapping areas
17:48 sapier on linux
17:49 Rick_ I actually am not using openal
17:49 Rick_ I decided to just do it without sound first
17:49 sapier maybe there is something wrong in general with openal
17:50 celeron55 joined #minetest-dev
18:16 hmmmm it might be worth investigating the crash deeper and fixing it for irrlicht
18:17 hmmmm but first it'd be better to check if 1.8 has the same problem
18:17 hmmmm it's rare that we have somebody (like Rick) who can reproduce that problem consistently and has a decent debugging environment
18:21 jojoa1997_Tablet joined #minetest-dev
18:33 jojoa1997|Tablet joined #minetest-dev
18:35 Rick_ I'm going to try my own build of Irrlicht when I get a little less lazy and see if for some reason that fixes the problem. I'll get it working eventually =)
18:38 Calinou joined #minetest-dev
18:40 hmmmm sapier, are you still around?
18:40 hmmmm you attempted to fix the texture leak i see
18:41 sapier yes I am
18:41 hmmmm you reference the atlas there, but does that work if the texture atlas is disabled? (it usually is)
18:41 sapier it's not fixing the full problem just a small step toward it
18:42 hmmmm karhl tried to fix the problem a while back but there was something wrong with it, i forget
18:42 sapier not quite I discovered texture attlas structures are used for caching textures too. Don't mix this with texture atlas itself
18:42 sapier I only fixed textures being recreated on client connect to server overwriting old ones
18:43 sapier obviously the old pointer is lost but not removed
18:43 sapier but there hast to be at least another leak
18:44 hmmmm okay
18:45 hmmmm anyway the check if it already existed was being done in LoadTextureFromFile() or something like that (forget the actual name)
18:45 hmmmm er, it was put there, anyway
18:46 sapier did I add a duplicate check?
18:49 hmmmm hmmm no
18:50 hmmmm nevermind, i can't find what i was talking about
18:50 hmmmm you should check that t_old is valid before passing it along to removeTexture() like that
18:50 sapier deleteTexture with 0 is working fine ... not sure if this is intentionally
18:50 hmmmm or are we actually sure it's valid because it's in the cache
18:51 hmmmm hrm
18:51 sapier and I can't distinguish between valid and invalid memory addresses
18:51 sapier so we need to assume it's either 0 or valid
18:53 hmmmm passing NULL to removeTexture doesn't seem documented
18:53 sapier ok so I better add a check
18:53 hmmmm but then again, if it's in the atlas cache, it'd never be NULL to begin with (just need to be sure that it's not getting deleted by something else)
18:54 hmmmm i haven't really worked with tile.cpp before so i'm not sure, just being cautious
18:55 sapier I've seen 0 when debugging so this seams to be a valid case
18:55 hmmmm indeed, it may be NULL, see how t is assigned only if img?
18:55 hmmmm so in the case of a texture that couldn't be created, why is it still added to the texture cache?
18:56 hmmmm wth
18:56 hmmmm could this be the reason why people see screwed up inventory textures if too many nodes are loaded?
18:57 sapier maybe
18:58 sapier I was about to remove atlas when I discovered this leak
18:58 hmmmm evidently, adding NULL textures is intended
18:58 kaeza joined #minetest-dev
18:58 hmmmm where are the textures actually fetched from the atlas cache?
18:59 hmmmm we need to make sure everything which calls that actually checks for a NULL texture
18:59 hmmmm such as Formspec
18:59 sapier yes but that's an addition to fixing the leak
19:00 hmmmm true
19:00 sapier if pushed the 0 checked version
19:00 hmmmm oh yeah.. another thing that probably isn't the problem, but worth looking at:
19:01 hmmmm addTexture is never checked for success
19:01 hmmmm if addTexture failed, then img should be freed as well, should it not?
19:01 hmmmm or is img deleted along with the NULL texture
19:01 hmmmm bah nevermind, i'm just spouting nonsense at this point
19:01 hmmmm don't listen to me
19:02 sapier :-) no its good to have even wrong thoughts so I don't have to sort them out myself ;-)
19:02 hmmmm i'm just looking at one line of the source and not the rest
19:02 sapier are there any objections to removing texture atlas remainings?
19:03 hmmmm eh
19:03 hmmmm i'd rather not
19:03 hmmmm celeron doesn't care for it though
19:03 Exio the only time i used it i had glitches everywhere
19:04 sapier you know removing code always results in fixing bugs :-)  .... pure statistics :-)
19:04 Exio so, unless someone(?) fixes those glitches, i'd like to see it removed :P
19:04 Exio just for that, exactly
19:05 troller Exio, and not forgot to disable zoom in deathmatch mode (in future)
19:05 Exio hehe
19:06 Exio my main idea for the zoom-stuff was not "fully that", but i will wait for the sapier's patch with the lua<->core api change for playing with that stuff
19:06 sapier I'll create a atlas removal branch maybe I detect the texture leaks while trying to understand what to remove
19:07 troller but maybe allow controlled zoom for sniper riffle
19:07 Exio yep, set_fov?
19:23 kaeza troller, I have the sniper rifle ready, awaiting a set_fov() function :P
19:24 ecube joined #minetest-dev
19:26 troller cool!
19:26 hmmmm there already is a lua set fov function if i recall
19:26 troller but for dm its must be only one way to change fov & sens - snipe riffle
19:27 hmmmm i really don't want people to get in the habit of changing too many client-side things with lua
19:30 * VanessaE peeks in
19:31 troller and we need in auto-respawn option
19:31 * troller trying to buld tower from self bones 8)
19:32 kaeza hah also trying to turn MT into an FPS I see
19:32 * kaeza approves
19:36 hmmmm for some reason i am having a real mental block on the absolute height ore ranges
19:36 hmmmm i need to invert height_min and height_max
19:36 hmmmm but i can't come up with a clean way to do this
19:36 hmmmm without like 100 if statements
19:37 hmmmm proller does the same sort of thing too with his code
19:37 VanessaE gevalt!
19:37 VanessaE my eyes bleed!
19:39 VanessaE hmmmm: well I presume you want some kind of linear falloff in ore density, right?
19:39 hmmmm i think proller's intention by making height_abs a signed 8 was for a tri-state, so he can specify 0 for off, -1 for completely inverted coordinates, and 1 for both absolute
19:39 hmmmm vanessae, not at all, we dumped that idea, remember?
19:39 VanessaE I thought you had, I just wasn't sure.
19:39 VanessaE so what then are you trying to do?
19:40 hmmmm if (nmin.Y > height_max || nmax.Y < height_min), we don't want to generate the ore
19:40 hmmmm but then if we're doing this in terms of absolute height, we want to make sure it's in neither range
19:40 hmmmm you can't simply do if (nmin.Y > abs(height_max) || nmax.Y < abs(height_min))
19:40 hmmmm since abs(height_max) < abs(height_min)
19:41 hmmmm so you need to flip the two arguments around
19:41 hmmmm the obvious solution is to add an && (inverted conditions here)
19:41 hmmmm that's ugly though
19:41 hmmmm i am looking for a clean, one-liner solution that won't force me to add like 2 more conditions
19:42 VanessaE hm, good question.
19:43 hmmmm obviously in order to get the abs() to work, i need to flip the inequality signs
19:43 VanessaE and all you really want to do is check if min/max are flipped in the calling code?
19:43 hmmmm i am looking at taking the difference between the two values and testing if that result is >= 0
19:43 VanessaE s/flipped/out of order/
19:44 hmmmm so if i am comparing signs instead of using different operators, i can negate it
19:44 hmmmm look
19:44 sapier is there any reason to create a media fetch thread if curl is disabled?
19:44 hmmmm if (nmin.Y - height_max > 0 || nmax.Y - height_min < 0) will do the same thing
19:45 hmmmm sapier, not really
19:45 sapier client does this nevertheless curl is enabled or not :-)
19:46 sapier without curl enabled thread count setting most likely is 0 but if not we'll have multiple nonsense threads :-)
19:47 hmmmm yeah, i really don't like having nonsense threads.. i felt a lot better after i stopped having emergethreads uselessly being created every time something's queued
19:48 sapier I wonder if the USE_CURL define isn't used for to small parts of code ... unless anyone wants to implement another mechanism to fetch data instead of curl maybe whole media fetch thread declaration could be compilerswitched with USE_CURL
19:53 hmmmm ugh no good
19:53 hmmmm the only solution i can think of requires using a goto and it's more complicated than just adding an extra condition
19:54 sapier sometimes I feel like goto is chacing me :-)
19:56 sapier whoever added the curl media fetch thread added a lot of additional code even in non curl case
20:02 jojoa1997|Tablet joined #minetest-dev
21:00 sapier1 joined #minetest-dev
21:04 jojoa1997|Tablet joined #minetest-dev
21:07 sapier1 argh there are so many missing delete/free/img->drop() in destructors you don't even have a chance to see where memory really is lost :-/ I'll fix them first
21:08 serengeor joined #minetest-dev
21:09 hmmmm i took a 20 minute nap and i came up with a better idea of how to do what i wanted to do
21:09 hmmmm you'll see it when i commit
21:09 hmmmm i think it's more clever than having 30 if statement conditions
21:09 Exio write it before you forget that! :P
21:10 sapier1 and don't forget the destructor ;-)
21:52 rsiska joined #minetest-dev
21:55 jojoa1997|Tablet joined #minetest-dev
22:18 sapier can someone plz add a rule to our coding guidelines that allocated memory has to be freed?
22:24 khonkhortisan that should be a given, but it's not being followed
22:25 sapier it should, but there are so many memory leaks in minetest it's almost impossible to finde the very big one with valgrind
22:25 khonkhortisan they just all clump together?
22:27 sapier no it's just such a big report you don't see the relevant one ... and you can't run minetest very long within valgrind so it isn't the "big one" in small probes
22:28 sapier I'll clean up as much as possible but for some things there's required additional investigation
22:28 VanessaE which means they have to be fixed the hard way - one at a time.
22:29 sapier ore and biome features are new am I right?
22:30 VanessaE ore gen is in the engine, yes, but biome stuff is for mapgen v7
22:30 sapier not even those new features are done in a clean way
22:36 sapier at least I'll see parts of minetest I never had a look at
22:50 Exio blame hmmmm for those
22:51 Exio (yes, i want to blame hmmmm for everything :D)
22:55 jojoa1997|Tablet joined #minetest-dev
22:55 VanessaE well considering how much work he's put into the engine, any blame at him is a good compliment anyway :)
22:56 jojoa1997|Tablet what is the blame
22:57 Exio culpar in spanish
23:11 jojoa1997|Tablet joined #minetest-dev
23:44 jojoa1997|Tablet joined #minetest-dev

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