Minetest logo

IRC log for #minetest-dev, 2013-12-12

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

All times shown according to UTC.

Time Nick Message
00:22 Megaf joined #minetest-dev
00:23 Megaf Can I speak here?
00:23 Megaf Yay!
00:23 Megaf celeron55, ping
00:29 Megaf celeron55, well, just wanted to pass by and say thanks =)
00:37 VanessaE I have to side with pitriss, I don't want the grey chat shadowing feature
00:37 VanessaE I'd prefer to be able to turn it off, as I use a bitmapped font that already has built-in shadowing
00:37 VanessaE so maybe his idea of mapping to a keypress is best.  just put it on F2.  Chat off -> Chat on -> Chat on, shaded -> Chat off -> ...
00:38 VanessaE simple, elegant, one key.
00:41 VanessaE also, the screwdriver mode RBA mentioned was at my insistence by way of what it does, not by way of the code that makes it work (he probably means either mode 1 or 3).
01:05 us}0gb joined #minetest-dev
01:11 Megaf us}0gb, are you member of Free Software Foundation?
01:11 Megaf I'm sorry, let's take this to other channel
01:12 us}0gb Megaf: Yes, and which channel?
01:26 kaeza Program received signal SIGSEGV, Segmentation fault.
01:26 kaeza 0x0833337a in ServerMap::~ServerMap (this=0x90603b0, __in_chrg=<optimized out>) at /home/diego/src/minetest/src/map.cpp:2569
01:26 kaeza 2569delete m_mgparams;
01:27 kaeza anybody else getting this?
01:28 VanessaE kaeza: you bwoke it!
01:28 us}0gb kaeza: Client or server?
01:28 kaeza singleplayer, so both
01:28 us}0gb I don't know then, I only have the server.
01:31 kaeza lel now it crashes at startup
01:32 kaeza http://pastebin.com/mR6LjL9a
01:35 kaeza it also seems it's not generating new terrain; already generated terrain is loaded OK
01:36 kaeza may be DB corruption
01:37 EvergreenTree joined #minetest-dev
02:05 e1z0 joined #minetest-dev
02:12 ShadowNinja That's silly. If you want to see the chat enable it, if you don't disable it. The chat is invisible when empty, so if you want activity notifications you get them with the text. And adding that bloats a three-line patch to a 50 liner or so.
02:15 VanessaE ShadowNinja: I think the problem is that the area will remain shaded long after you've looked at it, and it obscures the width of the screen even if you have five lines of two words each.
02:16 VanessaE hence my suggestion to rotate it in/out via F2.
02:16 VanessaE if it takes 50 lines of code to do that, someone needs to re-learn how to code :)
02:17 ShadowNinja VanessaE: It's mostly transparent. (80/255) And you look mostly at the middle and lower sections if you aren't checking chat.
02:18 VanessaE still doesn't address the other issue":
02:19 VanessaE some of us have already covered the readability problem by switching to a better font.
02:19 VanessaE please, don't put the shaded chat in.  it's not wanted if it can't be disabled.
02:19 us}0gb I'd have to agree. Any feature becomes a problem if it cannot be disabled.
02:24 ShadowNinja VanessaE: It won't conflict, and freetype is the better way...
02:26 ShadowNinja us}0gb: You can always modify the code, but adding options for many things, especially little things like this will only give you a 1GB example config, and a really slow step due to config checks. This would have to be checked on every frame, unless you do clever things that make this much longer and confusing.
02:35 VanessaE then make it a compile-time flag?
02:38 ShadowNinja ...
02:38 VanessaE freetype is NOT a better way
02:38 VanessaE freetype sucks for some of us
02:38 VanessaE if you need i18n, fine
02:39 VanessaE it's better then and only then
02:41 ShadowNinja Then tweak your texture if it looks ugly.
02:42 ShadowNinja If something's fixed you remove the workarounds.
02:52 VanessaE eh?
02:53 VanessaE I'm complaining here because the proposal is to add something that *has already been dealt with in another, better way* with *no option to decline it*
02:55 ShadowNinja VanessaE: No it hasn't. Your texture isn't merged and doesn't work with freetype.
03:12 ShadowNinja john_minetest: Hmmm, I could set up something to get the URL for #1337.  But it would need to guess which repo it was from. perhaps >500 = minetest <500 = minetest_game.
03:19 Weedy_lappy joined #minetest-dev
04:12 ShadowBot joined #minetest-dev
04:12 ShadowNinja #123
04:12 ShadowBot https://github.com/minetest/minetest_game/pull/123
04:12 ShadowNinja !part
04:12 ShadowBot left #minetest-dev
04:13 ShadowNinja I'll leave it out untill someone agrees.
04:19 ShadowNinja That works mid-sentence and has the above minetest/minetest_game selector.  Unfourtunately due to what appears to be a python bug it doesn't require a preceding word boundary, so test#1234 works, but #1234test doesnt.
04:20 ShadowNinja (The second is intentional)
04:20 VanessaE what's the cutoff for autodetecting which repo is being referenced?
04:21 ShadowNinja VanessaE: 500.
04:21 VanessaE (it's a good idea, btw, as long as it doesn't spam the channel..  maybe one de-reference no more often than every 10 mins or something)
04:21 VanessaE 500 sounds good, yeah
05:06 VanessaE sigh
05:06 VanessaE more unlogged crashes
05:07 hmmmm weeeoow
05:07 hmmmm my electricity bill is twice as much this month as it was last month
05:09 hmmmm what's the bare minimum code necessary to add items to a chest?
05:11 VanessaE no clue
05:11 VanessaE check the glooptest mod "treasure chest" module, there's some code there that may be of use
05:12 hmmmm well I already have code that works
05:12 hmmmm it's just a matter of finding the bare minimum that I have to do in order to make it work
05:12 hmmmm http://pastebin.com/1zQ8uzNc
05:12 VanessaE https://github.com/GloopMaster/glooptest/blob/​master/glooptest/othergen_module/init.lua#L95
05:12 hmmmm ahh that's better
05:12 VanessaE I guess that ^^^ is what gloo dpes
05:13 VanessaE goop does*
05:13 VanessaE ...
05:13 hmmmm so
05:13 VanessaE his code is known to be slow though
05:13 hmmmm I think I'm going to fill some chests with treasure stuff
05:13 hmmmm what are some good things aside from gear from minetest_game that I could use
05:13 VanessaE probably because he iterates through the whole chest whether it needs to be totally full or not
05:14 VanessaE diamonds, mese?  maybe rare stuff like a nc rainbow?
05:14 VanessaE oh wait
05:14 VanessaE aside from stuff from*
05:14 VanessaE hm...
05:15 hmmmm I suppose I could put in gold ingots and diamonds
05:15 hmmmm and mese chips or whatever
05:15 hmmmm anyway I want to do something for christmas
05:15 VanessaE well yeah the basics of course I guess.  but if you wanna venture outside of mt_game a bit, maybe stuff from moreores, technic, pipeworks, homedecor, that sorta thing
05:15 us_0gb joined #minetest-dev
05:15 us_0gb joined #minetest-dev
05:15 hmmmm I haven't decided whether it should be a hard coded santa hat on players for christmas eve and christmas day only
05:16 hmmmm or like a snowman block that's found in treasure chests
05:16 hmmmm or something like that
05:16 VanessaE heh
05:16 VanessaE cute
05:17 us_0gb The day should be configurable. Some of use don't celebrate Christmas.
05:17 hmmmm then that's too bad
05:17 us_0gb Just a suggestion. I don't mean to get in your way.
05:17 hmmmm configurable easter eggs
05:17 hmmmm so then they can become kuanza eggs
05:18 VanessaE haha
05:18 hmmmm they are a specific kind of the superclass holiday eggs, so really we should do class EasterEgg : public HolidayEggs
05:19 hmmmm and of course you can configure the date from minetest.conf or from lua
06:01 celeron55 it's somewhat hard to do any easter eggs in the minetest main project :P
06:01 celeron55 in independent mods and games on the other hand, nobody might notice until they pop up
06:01 VanessaE I dunno, some engineers managed to hide a few in the ROM of the old C128 back in the old days :)
06:02 VanessaE (well, one)
06:05 celeron55 the thing is that from C128 probably nothing else than marketing materials were given to customers
06:06 celeron55 but in minetest, basically normal users read the commit logs
06:07 us_0gb That and people talk a lot here. Developers and users alike discuss things, so it may slip.
06:07 VanessaE nope. it was a credits screen, plus a pleasant wish, something like "link arms, don't make them".  you had to SYS some obscure address to make it show up (equivalent to a randomly-named function call)
06:09 VanessaE but my point there was they managed to hide it in what was (as I recall) 16k of ROM intended for the OS and debugger/monitor
06:09 VanessaE but yeah, I see your point
06:09 VanessaE you just have to be more creative (read:  sneaky ;) )
06:10 VanessaE (and C128 users were routinely given programming manuals and such with their systems)
06:15 VanessaE so....the unlogged error issue...
06:16 VanessaE um...
06:17 VanessaE too bad hmmmm already signed off, I think his theory that it's EmergeThread triggering them may be correct - they always seem to happen when someone signs on.  I've got worlds bugging out that haven't had mods added or changed in months, and which are not logging their crashes anymore.
06:37 darkrose joined #minetest-dev
07:00 NakedFury joined #minetest-dev
07:59 kaeza request for comments: https://github.com/kaeza/minetest/commit/​b9985f152f3a72c76da00500a51f9725baaf35e0
08:03 kaeza basically, a set of callbacks (in Lua) are called before the player is actually initialized. possible uses include banning by name, wordlist, IP range, etc
08:03 kaeza simple example usage: http://pastebin.com/xQwgnt1y
08:04 VanessaE ...or perhaps displaying a message to the user such as a notice that this server rejects "Guest" names.
08:04 kaeza that too
08:05 VanessaE (I can predict that nearly every server operator in the community will install such a mod in a heartbeat :D )
08:05 kaeza hah
08:05 kaeza I know for sure the inchra network will be very please
08:05 kaeza +d
08:05 VanessaE G*d yes
08:37 djdduty joined #minetest-dev
08:48 darkrose joined #minetest-dev
08:59 ecube joined #minetest-dev
10:08 john_minetest joined #minetest-dev
10:10 zweipktfkt joined #minetest-dev
10:19 proller joined #minetest-dev
10:28 troller joined #minetest-dev
10:48 Megaf joined #minetest-dev
10:51 Megaf joined #minetest-dev
10:56 zweipktfkt joined #minetest-dev
10:59 Gethiox3 joined #minetest-dev
11:12 proller__ joined #minetest-dev
11:34 iqualfragile joined #minetest-dev
11:36 EdB joined #minetest-dev
11:42 specing joined #minetest-dev
11:51 smoke_fumus joined #minetest-dev
11:53 EdB_ joined #minetest-dev
11:58 NakedFury joined #minetest-dev
12:36 ImQ009 joined #minetest-dev
13:00 zat joined #minetest-dev
13:08 hmmmm joined #minetest-dev
13:10 jojoa1997 joined #minetest-dev
13:19 proller__ https://play.google.com/store/ap​ps/details?id=com.vicono.zombie
13:30 emptty joined #minetest-dev
13:31 emptty Hello there.
13:32 emptty I just stumbled over https://github.com/fogleman/Craft and it seems to me that the rendering engine is faster than in minetest
13:32 emptty you may want to get inspiration or something
13:43 kaeza I've tested it, and seems quite unresponsive on my hardware
13:43 kaeza didn't have time to fiddle much with it though
13:45 emptty strange. It really stresses le my machine than minetest
13:45 emptty it may be because it leverages my GPU? I've no idea at all
13:45 emptty *stress less than minetest
13:46 kaeza could be
13:46 kaeza my CPU is mid-to-low-end (Intel E5700 @ 3GHz x2), but the gfxx card is crap (Intel integrated)
13:46 kaeza -x
13:48 emptty I don't even know what GPU I have. nvidia or something. The machine is 3+ years old so that's all rather well supported now
13:49 kaeza "Intel Corporation 4 Series Chipset Integrated Graphics Controller (rev 03)", says lspci
13:50 emptty Chipset: "NVIDIA NVa8" says the nouveau driver in X.log
13:51 emptty so, it may well be the case that the GPU code of Craft is better than the one of minetest, but that the comparison is otherwise when the CPU must be used because the GPU is not enough
13:56 iqualfragile proller__: https://play.google.com/store/ap​ps/details?id=com.vicono.hungry https://play.google.com/store/a​pps/details?id=com.vicono.build
13:57 proller__ we neeed sooources!
13:58 emptty that's a clear GPL violation, yep
14:00 emptty or not, maybe that's just a game on top of minetest without any change to the engine
14:01 proller__ its have lot of changes about porting to android. disabling settings and optimize
14:03 emptty did you guys ask for the sources already?
14:04 kaeza one thing I just noticed is I get consistent low FPS (I'd say about 20FPS) in Craft no matter where I look (even at the sky), while in minetest it varies
14:04 kaeza and still it isn't using even 10% of ONE CPU
14:05 kaeza err core or whatever
14:10 john_minetest press and hold "F"
14:10 john_minetest awesome
14:14 PilzAdam joined #minetest-dev
14:15 emptty that too uses only 5% of my CPU. They are leveraging the GPU in a very impressive way
14:17 john_minetest yeah
14:18 john_minetest Framerate drops extremely when mining
14:19 emptty iqualfragile: proller__: did you ask for the sources of hungrygame and everything already?
14:21 proller__ im not, but somebod y ask
14:21 john_minetest There are a few pull requests open that should be merged. https://github.com/minetest/minetest/pull/817 for example. It got +1s by many people and looks cool.
14:25 iqualfragile emptty: no
14:25 thexyz emptty: I did ask
14:26 thexyz celeron55: you got any reply from google?
14:26 emptty thexyz: what was the answer?
14:26 thexyz no answer yet
14:27 emptty when did you ask?
14:27 john_minetest Who did that server list icons thing again?
14:27 emptty (Im wondering of wether I should ask myself too, trying to be candid)
14:27 thexyz emptty: yesterday; I've also asked celeron55 to submit a copyright violation to Google
14:32 celeron55 i have thrown a DMCA notice to google about com.vicono.build and com.vicono.hungry
14:32 celeron55 others shouldn't do it for now
14:32 celeron55 we'll leave that for later if google doesn't care about my notice
14:33 emptty but google will remove it. We would love to have it instead, wouldn't we?
14:33 celeron55 also, the devs of those know for sure that they are breaking the law (they previously replied to an e-mail where i said it to them)
14:34 celeron55 so don't expect anything on that side
14:35 emptty I fail to see whether they sell it or give it for free (because I have no compatible device myself)
14:35 proller__ just save apk + cache
14:35 thexyz emptty: yes, we've already tried to contact them in the past
14:35 celeron55 emptty: the price is a couple of USD
14:35 thexyz they give it for free but that doesn't matter
14:35 celeron55 oh, it's free now? well whatever; it was paid previously though
14:35 emptty so, that's a clear GPL violation
14:35 proller__ celeron55, now its free but with lot of adv
14:35 thexyz or no?
14:35 emptty I feel like I should contact debian's lawyers
14:36 thexyz yes, it looks like it's free, and yes, it's LGPL violation
14:36 celeron55 emptty: you can do that, but make sure to not spread any rumors; tell only facts
14:36 celeron55 and make sure you know what LGPL means
14:37 emptty celeron55: :) what you mean?
14:37 celeron55 most people are utterly clueless how these licenses work and what is copyright means
14:37 emptty for the record, I write LGPL code since 10+ years
14:37 celeron55 -is
14:37 john_minetest emptty: Write a letter to Stallman.
14:37 celeron55 (and i don't know your background)
14:38 emptty celeron55: if you can write down a statement, I could get it discussed with lawyers
14:38 emptty we have a debian-private mailing list to get advices on such issues without spreading rumors, if you want to go under radar for now
14:39 celeron55 emptty: the main issue here is that the minetest engine is distributed under LGPL by it's copyright holders and those things on play.google.com use a modified version of it without distributing modified source code
14:40 thexyz I think for now we should wait for Google reply
14:40 celeron55 or, that's not 100% sure; but it's very probable
14:41 thexyz why's that not 100%? even the description page says it is
14:42 john_minetest They could just be trying to use our good reputation.
14:42 emptty you guys downloaded the software?
14:42 emptty very simple foresenic on the binary could prove it, to
14:44 emptty thexyz: I'm not sure I agree
14:44 emptty google will kill that project, and I'd prefer to free that project, and merge it into the main project
14:46 thexyz that'd be good; but it's not that important
14:46 thexyz the android port already is being developed
14:46 thexyz and yes, we've already asked for the source code
14:47 thexyz if you missed that my line
14:47 thexyz like, a month ago, I guess?
14:56 john_minetest A question: http://i.imgur.com/EMGMDW2.jpg
14:57 ShadowNinja kaeza: Add "lua_remove(L, -2);" After the second getfield/getglobal.
14:58 emptty john_minetest: haha ! :)
14:58 ShadowNinja kaeza: Also log the reason.
14:58 emptty john_minetest: do you know http://dev.minetest.net/Git_Guidelines ?
15:03 john_minetest emptty: there are pull requests out there that are really useful, tested and add a small feature.
15:04 proller__ i know one place. where they already merged ;)
15:05 emptty john_minetest: I don't have the commit right myself
15:05 john_minetest proller__: That one yet? -> https://github.com/RealBadAngel/minetest/com​mit/d56214cb097ff0297e479501dd21d6d86c573b6d
15:11 proller__ john_minetest, we trying to wait 1-3 days before merge, to not fight with rebases and sync once with minetest/master
15:12 Taoki Ohh, RBA's shaders are finally in Minetest master!
15:12 Taoki I wonder if there's a list on which settings to use to enable the all
15:15 ShadowNinja kaeza: Otherwise it looks good.
15:15 john_minetest RBA requested that one being pulled yesterday. Seems it isn't ingame yet. It only adds 12 lines of code and the only thing it does is allowing one single normal map for all block.
15:16 ShadowNinja Taoki: They're in the settings menu.
15:16 Taoki ShadowNinja: Oh, good then
15:18 kaeza ShadowNinja, alright
15:18 kaeza what about the function name as sapier pointed?
15:19 us|0gb joined #minetest-dev
15:19 us|0gb joined #minetest-dev
15:33 Megaf joined #minetest-dev
15:35 ShadowNinja kaeza: prejoin seems right. But you definitely shouldn't be using RUN_CALLBACKS_MODE_FIRST as it only lets the first registered callback disconnect players.  Use MODE_OR.
15:36 kaeza ShadowNinja, that is intended
15:36 kaeza why would you want to call another callback when one errors?
15:36 kaeza s/errors/returns a meaningful value/
15:36 ShadowNinja kaeza: Um, so only one mod can disconnect players? (Or builtin)
15:37 kaeza oh wait, misread that
15:37 ShadowNinja kaeza: You cal log the IP and the like. But the right mode for that is MODE_OR_SC.
15:40 ShadowNinja IMO it should be MODE_OR.  The first reason will be used but other mods and log and announce things.
15:40 ShadowNinja can*
15:43 ShadowNinja And this makes a whitelist childishly easy to implement.
15:44 ecube joined #minetest-dev
15:50 salamanderrake joined #minetest-dev
15:57 Calinou joined #minetest-dev
16:04 specing Why is minetest constantly stat()ing /etc/localtime?
16:04 specing stat("/etc/localtime", {st_mode=S_IFREG|0644, st_size=1931, ...}) = 0
16:04 specing stat("/etc/localtime", {st_mode=S_IFREG|0644, st_size=1931, ...}) = 0
16:04 specing stat("/etc/localtime", {st_mode=S_IFREG|0644, st_size=1931, ...}) = 0
16:05 specing grep found 34! stat() calls on that file before it crashes
16:05 specing write(2, "17:05:29: ERROR[main]: ERROR: An"..., 8817:05:29: ERROR[main]: ERROR: An unhandled exception occurred: Cannot open database file) = 88
16:05 specing Minetest has some weird directory management
16:05 specing strace -f ./bin/minetestserver --config minetest.conf --world home --logfile ''
16:06 specing [pid   434] open("/data/specing/programs/minetest/minetes​t/bin/../games/minetest_game/minetest.conf", O_RDONLY) = -1 ENOENT (No such file or directory)
16:06 specing Wait what, I give it a config and it can't look into the CWD?
16:06 specing open("minetest.conf", O_RDONLY)         = 3
16:06 specing open("/data/specing/programs/minetest/minet​est/bin/../games/minetest_game/game.conf", O_RDONLY) = 3
16:06 specing [pid   463] open("/data/specing/programs/minetest/minetes​t/bin/../games/minetest_game/minetest.conf", O_RDONLY) = -1 ENOENT (No such file or directory)
16:07 ShadowNinja specing: Pastebin please.
16:07 specing I don't understand this game.
16:09 specing http://sprunge.us/CiiC
16:10 thexyz you are thinking too much
16:11 thexyz and in the wrong direction too
16:12 thexyz for instance, nothing wrong with it looking for minetest.conf in game's path
16:12 thexyz because that's how things work; the game can have minetest.conf too
16:12 thexyz so can you please clarify your problem a bit?
16:16 specing There is none atm, I was just wondering about those needless stat()s
16:20 john_minetest joined #minetest-dev
16:25 specing lol @ mod workers
16:27 EvergreenTree joined #minetest-dev
16:27 kaeza ShadowNinja, are the lua_remove/pop strictly required? I have checked SCRIPTAPI_PRECHECKHEADER, and it seems to handle that stuff conveniently
16:28 ShadowNinja celeron55: So what do you think about the #num -> url bot?
16:30 ShadowNinja kaeza: You need to remove the minetest table after retrieving it.  Unless that resets the stack when the function returns.
16:31 kaeza ShadowNinja, SCRIPTAPI_PRECHECKHEADER seems to do that itself, other wise all the other functions would be in error, as none of them bothers to clean up anything
16:32 Jordach joined #minetest-dev
16:32 celeron55 ShadowNinja: enable it and we'll see what people think
16:32 celeron55 i'm sure we know after a few weeks if it's good or not
16:37 VanessaE joined #minetest-dev
16:38 ShadowBot joined #minetest-dev
16:40 Warr1024 joined #minetest-dev
16:41 Warr1024 ShadowNinja: looks like that shaded chat commit didn't make it in?
16:43 ShadowNinja Warr1024: I haven't pushed it yet. But it's ready as soon as someone else agrees.
16:46 Warr1024 oh, ok, I thought you had had the necessary agreements.
16:47 ShadowNinja Well, I could push it I guess. But apparently it's a little controversial...
16:47 VanessaE "a little" ...
16:48 Warr1024 really?  I thought there was only 1 against.
16:48 Warr1024 we could try to do the text outlining thing.
16:48 emptty you guys develop a game using methodologies that would be needed for nuclear power plants :)
16:49 Warr1024 I'd hope that making a change to a nuclear power plant requires the agreement from at least 3 core engineers...
16:49 ShadowNinja Warr1024: Do you have any ideas on how to do that well that will work with both freetype and bitmaps?
16:51 NakedFury joined #minetest-dev
16:51 us`0gb joined #minetest-dev
16:52 specing Hmm they dont handle lava
16:57 kaeza https://dl.dropboxusercontent.com/u/100008207/Sc​reenshot%20from%202013-12-12%2013%3A53%3A51.png <-- 13:13:45: ERROR[main]: Access denied. Reason: Guest accounts are disallowed. Please choose a proper username and try again.
16:57 kaeza ShadowNinja, ^
16:58 kaeza related: http://pastebin.com/akMiaR6e
16:58 VanessaE someone needs to fix that ^^^ stupid error dialog to be capable of showing more than ~3 lines of text.  even with the default font you can see it gets chopped off, despite being hugely oversized.
16:58 VanessaE ~2*
16:59 us^0gb joined #minetest-dev
16:59 Warr1024 If you render text into the bounding box (0,0,100,100), it should produce the exact same pixels as if rendered into (1, 1, 101, 101) using any font engine.
17:00 Warr1024 so we can outline the text by just rendering it a handful of times with shifted bounding boxes, in black, before rendering the final one in white over top.
17:00 kaeza ShadowNinja, also changed to MODE_OR
17:00 VanessaE Warr1024: but do that only if freetype is being used.
17:00 Warr1024 VanessaE: why?
17:01 VanessaE Warr1024: because a bitmap font doesn't need it.  https://forum.minetest.net/viewtopic.php?id=4436
17:01 kaeza anything else?
17:01 Zeitgeist_ joined #minetest-dev
17:01 Warr1024 VanessaE: ah, that's pretty neat, actually.
17:02 Warr1024 in that case it sounds like this is only a problem for freetype.
17:02 VanessaE nevermind whether that font is actually included with Minetest (it probably never will be because no one can be bothered)...the fact that it can be done without changes to the engine is all that shoudl matter.
17:02 VanessaE that font has some kerning issues, but I can't be sure whether the font is the problem or the engine is
17:03 VanessaE and no one seems to be bothered enough to look into it
17:03 Warr1024 I'd love to have a nice narrow sans monospace font as the standard...
17:03 Warr1024 unfortunately it looks like a lot of UI elements aren't sized right for that font.
17:04 VanessaE yeah I know
17:05 VanessaE between the kerning glitches (which I think are somewhere in the engine) and the fact that its i18n support is limited, it will probably never be merged
17:05 VanessaE but that font is, imho, the best look minetest's fonts will ever get short of your idea of multiple drawing passes.
17:06 thexyz https://github.com/minetest/minetest/issues/1048
17:06 Warr1024 hm, yeah, I forgot about i18n.
17:06 thexyz i18n matters.
17:06 Warr1024 yeah, maybe freetype font outlining is the best long-term approach.
17:07 Warr1024 projects that undervalue i18n are f4d.
17:07 ShadowNinja Warr1024: Yes, but that's a extremely hacky and slow way to do it. It isn't a good way.
17:07 Warr1024 yes.  There are no good ways.
17:07 VanessaE thexyz: I'm sure it does, to a large part of the minetest userbase for sure, but a much larger part of the userbase doesn't, honestly, give two shits about it :)
17:08 thexyz I personally think that bitmap fonts should be dropped
17:09 Warr1024 ShadowNinja: so what would be the non-hacky way?
17:09 VanessaE freetype has outlining effects avaailable.
17:09 VanessaE -a
17:09 VanessaE you could draw in two passes
17:09 Warr1024 oh?  that'd be pretty neat
17:09 VanessaE one in black, outlined mode, one in white, normal mode on top of that
17:09 thexyz VanessaE: where?
17:09 ShadowNinja ^ Or the background.
17:09 VanessaE http://www.wildfiregames.com/f​orum/index.php?showtopic=17365
17:09 VanessaE saw it on this page
17:09 VanessaE "Glyph outline" mode
17:09 Warr1024 I don't know if irrlicht exposes anything like that in its interfaces, and I definitely don't know the engine well enough to end-run around irrlicht to access freetype directly...
17:10 kaeza https://github.com/kaeza/minetest/commit/​e7a9468cd907a3b44eb34c0ddbec5499290f94dd
17:11 thexyz VanessaE: I think font shadow will look better, tbh
17:11 VanessaE indeed so.
17:12 Warr1024 stroke effect would be the most readable.
17:12 Warr1024 unless the shadow is tunable
17:12 Warr1024 i.e. so we can have some shadow in the upper-left, even if it's heavier in the lower-right.
17:12 OldCoder joined #minetest-dev
17:16 Warr1024 also, VanessaE, it looks like your example is using shaders.
17:17 VanessaE what example?
17:17 Warr1024 your wildfiregames link.
17:17 VanessaE wildfiregames?
17:17 VanessaE o.O
17:17 Warr1024 (12:09:35) VanessaE: http://www.wildfiregames.com/f​orum/index.php?showtopic=17365
17:17 VanessaE oh that
17:18 VanessaE oh right, well pay that no mind anyway
17:18 VanessaE there are freetype calls to do the job too
17:18 Warr1024 ah, ok
17:18 VanessaE http://www.freetype.org/freetype2/do​cs/reference/ft2-glyph_stroker.html
17:18 VanessaE http://www.freetype.org/freetype2/docs​/reference/ft2-outline_processing.html
17:18 VanessaE probably more
17:18 Warr1024 I guess that would mean using freetype to do text instead of irrlicht, though, which could be a bit bigger of a task...
17:19 VanessaE bbl
17:19 thexyz we already use CGUITTFont
17:19 thexyz I feel like modifying it won't be a hard task
17:19 Warr1024 feel like taking a shot at it?
17:20 thexyz playing with it right now
17:20 Warr1024 sweet
17:20 Warr1024 maybe we can have readable chat AND keep spacebar heating :-)
17:20 Warr1024 bbl
17:20 VanessaE http://xkcd.com/1172/
17:20 VanessaE oblig.
17:25 thexyz after hacking into cguittfont sources http://i.imgur.com/swRRhjy.png
17:25 NakedFury so much better
17:25 NakedFury how is it ingame?
17:26 thexyz http://i.imgur.com/9PTeUZb.png
17:26 NakedFury yeah thats much better
17:27 thexyz sun http://i.imgur.com/VDd6iTn.png
17:27 VanessaE indeed that is better
17:27 VanessaE what method did you use?
17:28 thexyz ok, then I'll add a setting and submit a pull
17:28 thexyz just draw it twice
17:28 thexyz https://gist.github.com/xyzz/e465463​d2834a390bca0/raw/e77b56cd2dbf673863​b762a12011042b546acdcc/gistfile1.txt
17:28 pitriss thexyz: please can you try screenshot with bigger font? :)
17:29 thexyz shadow offest's gonna be configurable; anyway, here's a patch, you can test it yourself
17:31 pitriss thexyz: okay.. if it will be configurable, then ok.. and if it will be offset setting then offset = 0,0 is "no effect" in fact.. so great for me:)
17:31 ShadowNinja Irrlicht and CGUITTFont don't seem to provide any methods to draw a outline.
17:32 jin_xi joined #minetest-dev
17:32 Warr1024 argh with the configuration again
17:33 VanessaE ok, bbl for real now.
17:34 Warr1024 if I had my way, I'd probably draw a copy to the upper-left as well, semi-transparent
17:34 Warr1024 that'd make the text pop out completely, and mean a bit less brain-work decoding certain characters from their shadows
17:34 Warr1024 but the way it is right now is already a big improvement
17:35 Warr1024 though if I'm feeling particularly picky, I'd probably just maintain a patchset.
17:36 Warr1024 damn thexyz, that was some nice work.
17:37 thexyz 5-lines patch isn't anything remotely nice work
17:38 Warr1024 I'd be impressed if you got it to work in 3, but given the time it took you, it's not bad.
17:40 thexyz okay
17:41 Warr1024 maybe I'll mess around with this thing some time, see if I can come up with some different effects, and maybe a reasonable way to configure the thing without becoming a maint nightmare...
17:43 Warr1024 does calling that draw 2d image batch method actually cause freetype to render the text again, or was it already pre-rendered to some off-screen surfaces and it just gets blitted?
17:50 thexyz https://github.com/minetest/minetest/pull/1049
17:50 thexyz no, it doesn't call freetype
17:50 Warr1024 ah, cool
17:51 Warr1024 so you can add more copies of the text at relatively low cost?
17:51 thexyz looks like it
17:53 Warr1024 It'd be pretty neat if we could configure it by specifying an array of (dx, dy, r, g, b, a) tuples for each additional copy to draw, with some common example(s) in minetest.conf.example.
17:59 emptty joined #minetest-dev
18:08 ShadowNinja https://github.com/minetest/minetest/pull/1047 looks good.  Anyone to agree?
18:17 ShadowNinja thexyz: ^
18:21 specing pastebin.com
18:21 specing Come on
18:21 specing Out of all the pastebins
18:23 BlockMen joined #minetest-dev
18:25 BlockMen thexyz, the shadow is enabled by default in #1049?
18:25 ShadowBot https://github.com/minetest/minetest/issues/1049
18:26 Warr1024 oh, mt issue # -> link converter, nice
18:27 thexyz BlockMen: yes
18:28 Warr1024 thexyz: wow, you even managed to make it pitriss-compatible.
18:29 BlockMen thexyz, i would prefer to leave it disabled by default.
18:29 BlockMen and how about an option for the shadow color like for selection box https://github.com/minetest/minetest/b​lob/master/minetest.conf.example#L150
18:29 thexyz I don't think it should be turned off by default
18:30 Warr1024 it's a double-edged sword, without reliable data on what people actually want overall...
18:30 thexyz not sure if there should be an option either; having options for every single things isn't the best idea
18:30 thexyz people want readable fonts
18:30 thexyz this pull gives them readable fonts
18:30 Warr1024 disabling it by default could cause people not to know it even exists.
18:31 ShadowNinja Whatever is added has to be enabled by default.
18:32 ShadowNinja BlockMen: Does #1047 look good?
18:32 ShadowBot https://github.com/minetest/minetest/issues/1047
18:33 ShadowNinja Oh, kaeza: The semicolon should be on the last code line, not it's own line.
18:33 celeron55 thexyz: throw that in
18:34 kaeza ShadowNinja, :I
18:35 ShadowNinja kaeza: Or I can do it...
18:35 kaeza too late
18:36 thexyz celeron55: done
18:38 ShadowNinja thexyz, celeron55: Can you check that?
18:38 BlockMen ShadowNinja, seem good for me
18:42 thexyz uh, I guess it's ok? not sure why are you asking me because I've never worked with lua api
18:45 BlockMen i still see not the point of not adding a setting for the shadow color
18:46 thexyz add it then
18:47 * BlockMen will do
18:47 BlockMen *but not now :P
18:57 OldCoder joined #minetest-dev
19:01 troller joined #minetest-dev
19:01 proller joined #minetest-dev
19:03 EvergreenTree joined #minetest-dev
19:11 djdduty joined #minetest-dev
19:28 * BlockMen wonders if color or alpha makes more sense
19:44 pitriss BlockMen: and is there any big difference to use RRGGBB or RRGGBBTT ?:)
19:44 troller TT
19:45 BlockMen he means AA i guess
19:45 pitriss TT as transparency leve
19:45 pitriss l
19:45 BlockMen and yes, but since its a shadow alpha makes more sense
19:45 pitriss yep
19:45 BlockMen so i will use alpha now
19:50 Calinou joined #minetest-dev
20:06 BlockMen is that ok? https://github.com/BlockMen/minetest/commi​t/58ff3aa6468c0b83ebb44e7d12c304e38d1c8924
20:07 BlockMen but i would prefer to make default alpha ~120 instead of 255
20:08 celeron55 why not read the whole color from the file?
20:08 celeron55 wasn't a function for that added anyway for some formspec stuff
20:09 BlockMen i considered that but then i thought its a shadow so its ment to be dark
20:09 BlockMen maybe we can add a function for fontcolor, that would make more sense IMO
20:09 BlockMen s/function/option
20:14 celeron55 eh
20:15 celeron55 "dark" does not mean "colorless"; shadows are simply areas that are lit by some other light source than the major light source
20:15 celeron55 not that this really matters but you started
20:21 BlockMen ok, lets say "natural shadows" are ment to be colorless itself but seem/look like to have different colors depending on their intensity
20:22 BlockMen but since there are graphical shadows they can have every color ofc
20:29 BlockMen but beside the shadow-definitions...can someone agree?
20:32 ShadowNinja Does https://github.com/minetest/minetest/pull/1046 look good?
20:38 BlockMen ShadowNinja, seems fine
20:38 EvergreenTree joined #minetest-dev
20:39 Miner_48er joined #minetest-dev
20:41 celeron55 ShadowNinja: i've basically agreed to it already
20:49 BlockMen since it seem to need the longer way: https://github.com/minetest/minetest/pull/1050
20:51 BlockMen left #minetest-dev
21:07 Miner_48er joined #minetest-dev
21:10 proller joined #minetest-dev
21:19 us{0gb joined #minetest-dev
21:20 Akien joined #minetest-dev
21:53 bas080 joined #minetest-dev
22:04 Weedy_lappy joined #minetest-dev
23:10 book` joined #minetest-dev
23:11 werwerwer joined #minetest-dev
23:24 iqualfragile feature request function to remove crafting recepies
23:28 sapier joined #minetest-dev
23:29 iqualfragile sapier, core devs: wouldnt it be a good idea to put most of mobfs code into core?
23:34 VanessaE iqualfragile: that general idea has been discussed to death
23:34 iqualfragile to which outcome?
23:34 sapier iqualfragile not by now
23:34 VanessaE to THAT outcome.
23:35 sapier I don't want to start that discussion again there's no chance to merge any of the two competing mob implementations for now as both aren't ready for this ... mobf is to complex while sm is to simple
23:35 sapier oops forgot , again sorry
23:36 iqualfragile ah, ok, just wondering
23:38 sapier I try to get mobf's features to core where usefull and possible but it'll take some time ... in best case this process will render mobf useles
23:38 sapier +s
23:39 sapier but don't expect that to happen within next year ;-)
23:41 djdduty_ joined #minetest-dev
23:42 iqualfragile but a remove crafting recepie function would be nice
23:42 sapier great minetest now is capable of precognition
23:42 sapier oh no it isn't
23:43 sapier it needs to be cleant up prior merge
23:44 iqualfragile sapier: precognition?
23:45 sapier right now it's just a proof of concept. Any final version souldn't just duplicate 90% code for remove and add but use same code to do it for both
23:45 sapier forget about it :-)
23:45 VanessaE cleant? :)
23:45 sapier hmmm guess clean isn't a irregular verb?
23:46 sapier I don't know how often I did this wrong ... guess I'll never learn it
23:47 VanessaE no grammar lessons today :)
23:48 bas080 joined #minetest-dev
23:50 john_minetest There should be a mob framework that is partly hardcoded.
23:50 john_minetest So it is easy to make own logics and mobs without having to lose so much performance.
23:51 sapier you don't loose "so much" performance, mobf is quite performant in latest version. at least if you don't use those features that can't be done in a performant way
23:52 john_minetest The last time I tried mobf it was fat and laggy.  Simplemobs are much lighter. But this was quite a long time ago and I didn't check out the new mobf yet.
23:53 john_minetest Btw, are there ridable bears yet?
23:53 PandemoniuM joined #minetest-dev
23:53 john_minetest Since we got the right to own and arm bears.
23:53 sapier e.g. if you want to have 200 mobs active, each one attacking next available enemy, each mob has to check all other 200 mobs around
23:54 john_minetest I don't think that mobs should really fight each other.
23:54 PandemoniuM mobs vs mobs = immersion
23:54 john_minetest I think that a TP can change immersion MUCH MORE than mobs
23:55 sapier if you want to have a guard, that's the only way to do
23:55 john_minetest I tried misa's TP and it changed the gameplay so much. It was amazing.
23:55 proller 20 years ago doom mobs was able to attack other mobs
23:55 PandemoniuM TPs are game ruiners imo
23:55 PandemoniuM like i said, my opinion lol. not a popular view, i know
23:55 john_minetest Misa's TP is amazing. imo it is the best looking TP I have ever tried in minetest
23:56 john_minetest Heh, I am looking forward for the texture pack store being made.
23:56 sapier of course 200 mobs is a quite extreme usecase, another slow option is lot's of mobs trying to do pathfinding same time
23:57 proller sapier, no need to make 200^2 lookups, you can find lot of optimized algorithms
23:57 PandemoniuM i really like this project because its so much more developed and well written than Manic Digger, but i threw MT on my 4gb rackspace server and im encountering a lot of lag issues..
23:57 john_minetest Is it A*?
23:57 proller some doom2 maps have 1500+ mobs
23:57 sapier A* is only good in average case
23:57 sapier doom2 is scripted
23:58 sapier I've never seen more than a couple of mobs same time there ... and they do have precompiled worlds
23:58 sapier btw doom2 is 2d map only
23:59 proller 2.1d and it can run on 486
23:59 PandemoniuM is there a lifetime on the blocks that havent been picked up?
23:59 proller now you have 1000++x more cpu power
23:59 proller and lua, it eat all ;)
23:59 sapier and a real 3d dynamic map
23:59 sapier that's most important thing
23:59 sapier doom2 can use precompiled movemaps

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