| Time |
Nick |
Message |
| 00:56 |
|
jap__ joined #minetest-dev |
| 01:35 |
ShadowNinja |
A lot of things that should be in verbosestream are in infostream, and we should probably have another level like trace. Aditionally there should be a warningstream. |
| 01:35 |
|
psedlak joined #minetest-dev |
| 02:06 |
|
CraigyDavi joined #minetest-dev |
| 02:41 |
|
Eater4 joined #minetest-dev |
| 02:54 |
|
werwerwer joined #minetest-dev |
| 04:34 |
|
kaeza joined #minetest-dev |
| 04:35 |
|
diemartin joined #minetest-dev |
| 04:38 |
|
Exio4 joined #minetest-dev |
| 05:18 |
|
BrandonReese joined #minetest-dev |
| 05:25 |
|
nore joined #minetest-dev |
| 05:30 |
|
nore joined #minetest-dev |
| 05:37 |
|
Kray joined #minetest-dev |
| 05:43 |
|
ShadowNinja joined #minetest-dev |
| 07:06 |
|
restcoser joined #minetest-dev |
| 07:18 |
|
darkrose joined #minetest-dev |
| 08:10 |
|
rsiska joined #minetest-dev |
| 08:15 |
|
ImQ009 joined #minetest-dev |
| 09:17 |
|
darkrose joined #minetest-dev |
| 09:38 |
|
jin_xi joined #minetest-dev |
| 09:46 |
|
proller joined #minetest-dev |
| 09:48 |
|
blaise joined #minetest-dev |
| 11:24 |
|
smoke_fumus joined #minetest-dev |
| 11:40 |
|
grrk-bzzt joined #minetest-dev |
| 11:48 |
|
PenguinDad joined #minetest-dev |
| 11:54 |
|
nore joined #minetest-dev |
| 12:56 |
|
restcoser joined #minetest-dev |
| 13:03 |
|
hmmmm joined #minetest-dev |
| 14:25 |
|
Hiradur joined #minetest-dev |
| 14:38 |
Taoki |
hmmmm: I was wondering. Is it possible for mapgen v7 to generate lakes which are composed of water, but the top layer is ice? Would be useful for generating snow biomes directly with ice on them rather than water. |
| 14:40 |
|
grrk-bzzt joined #minetest-dev |
| 14:45 |
|
sapier joined #minetest-dev |
| 15:00 |
sapier |
ShadowNinja: we already have a trace level log |
| 15:00 |
ShadowNinja |
sapier: Oh. |
| 15:01 |
ShadowNinja |
sapier: I only see verbosestream in log.h. |
| 15:01 |
sapier |
It's easy to miss because it seems to be added in a quite ugly way |
| 15:01 |
sapier |
log.h L74 |
| 15:01 |
ShadowNinja |
Oh, yes, I see. |
| 15:02 |
ShadowNinja |
The log streams are kind of ugly to begin with though. |
| 15:03 |
sapier |
but I don't understand why it's been done that way, trace level is directly below error so it'd have been most obvious to be handled same way ... no idea why it wasn't done that way |
| 15:09 |
|
BlockMen joined #minetest-dev |
| 15:10 |
ShadowNinja |
sapier: 1070 Seemed good. It isn't done as nicely as it would be nice to do, but it's better than crashing. |
| 15:11 |
sapier |
what's #1070 |
| 15:11 |
ShadowBot |
https://github.com/minetest/minetest/issues/1070 -- Fix support for "null" videodriver e.g. for bots by sapier |
| 15:12 |
sapier |
oh, celeron is right it's really ugly |
| 15:12 |
sapier |
ShadowNinja: does script_error print backtrace? |
| 15:14 |
|
Anchakor_ joined #minetest-dev |
| 15:21 |
|
Jordach joined #minetest-dev |
| 15:26 |
ShadowNinja |
sapier: It prints the string value on the top if the stack. If you used lua_pcall with script_error_handler that string will contain a traceback. |
| 15:26 |
sapier |
I don't know what's used above I wanna call it from various locations |
| 15:27 |
Taoki |
Hi. Does anyone know if 6D facedir is possible for nodeboxes at this day? So a node with a single nodebox definition can be rotated in any way based on placement angle... rather than having to define 6 items for each rotation of a nodebox thing |
| 15:27 |
ShadowNinja |
Without a error handler it just prints the error message. (attempt to index a nil value at /foo/bar/mod/init.lua:1337) |
| 15:27 |
sapier |
are all our lua calls surrounded by error handlers yet? |
| 15:28 |
ShadowNinja |
sapier: Well, it's only supposed to be called after a failed lua_pcall. |
| 15:28 |
ShadowNinja |
sapier: Yes, all use a error handler, unless I messed one (unlikely). |
| 15:28 |
ShadowNinja |
(Or someone added a new one without adding a error handler. |
| 15:28 |
ShadowNinja |
) |
| 15:31 |
BlockMen |
Taoki, look at the screwdriver (_game) |
| 15:32 |
Taoki |
Right... I remember that now. Thanks |
| 15:32 |
Taoki |
Does this means the wires in technic_game are finally defined as one entity each, and there are no longer 6 entities for every corner and the like? |
| 15:33 |
ShadowNinja |
Taoki: s/entity/node/ and no. |
| 15:33 |
BlockMen |
idk anything of technic_game, sry |
| 15:41 |
Jordach |
interesting bug: create a glasslike node and then give it different textures for each face instead of the expected result, you'll get the same texture all over |
| 15:43 |
|
proller joined #minetest-dev |
| 15:43 |
Taoki |
Ugh... I wish someone documented register_biome. I know some still consider it experimental, but that would help |
| 15:43 |
Jordach |
do it with a nodebox and the effect works AS IS. |
| 15:44 |
Taoki |
Anyone know where I can find the latest biome script for MGV7 |
| 15:44 |
Jordach |
Taoki, i did consider making MVG7 to work with MTG |
| 15:44 |
Taoki |
And no, do NOT tell me again to not use v7 because it's not finished! :P |
| 15:44 |
Taoki |
MTG? |
| 15:44 |
Jordach |
Minetest_Game |
| 15:44 |
Taoki |
Oh |
| 15:45 |
Taoki |
It does work with minetest_game very nicely :P You just need the biome script |
| 15:45 |
Taoki |
Which gets downdated often. And the latest version I have doesn't seem to define biomes properly any more |
| 15:45 |
Jordach |
Taoki, which is why i was going to do myself |
| 15:45 |
Taoki |
Oh. hmmmm maintains a biome script for minetest_game IIRC |
| 15:45 |
Taoki |
Just not in the same place |
| 15:49 |
|
Calinou joined #minetest-dev |
| 15:52 |
sapier |
https://github.com/minetest/minetest/pull/1261 |
| 15:52 |
sapier |
#1261 |
| 15:52 |
ShadowBot |
https://github.com/minetest/minetest/issues/1261 -- Add proper lua api deprecated handling by sapier |
| 15:52 |
sapier |
ShadowNinja can you make ShadowBot to lookup full pull links too? |
| 15:53 |
ShadowNinja |
sapier: Yes. |
| 15:53 |
sapier |
thanks |
| 15:54 |
sfan5 |
sapier: mimicry (that is not a word I know) -> mimic, defailt -> default |
| 15:55 |
Jordach |
sfan5, that is |
| 15:56 |
sfan5 |
oh well, ignore the first thing I said then |
| 15:57 |
sapier |
ok the question is does mimicry realy mean same thing in english I wanted to say |
| 15:57 |
sapier |
no it's anoun |
| 15:58 |
sfan5 |
mimicry is the act of mimicing something |
| 15:58 |
sfan5 |
you used it like a verb |
| 15:58 |
sapier |
yes so it's a word but still wrong ;-) |
| 15:59 |
sapier |
fixed it |
| 16:10 |
|
PilzAdam joined #minetest-dev |
| 16:14 |
sapier |
any comments to #1241? I'd like to merge it today if noone finds bugs in it |
| 16:14 |
ShadowBot |
https://github.com/minetest/minetest/issues/1241 -- Add formspec toolkit and refactor mainmenu to use it by sapier |
| 16:16 |
sapier |
Just to make it clear, the "game" buttons may not be exactly identical to previous ones, in exchange the game bar is now scrollable |
| 16:23 |
ShadowNinja |
sapier: I've made some. |
| 16:24 |
|
rsiska joined #minetest-dev |
| 16:25 |
|
[PavelS] joined #minetest-dev |
| 16:26 |
sapier |
well imho it's wrong lua files not having license headers, it's been reasonable while lua code have been minor parts but it's grown to a degree where license does make sense |
| 16:28 |
ShadowNinja |
Perhaps. But it so, later and to all files. |
| 16:28 |
sapier |
You've got a pull out there touching all the other files you can fix it there too |
| 16:29 |
ShadowNinja |
I haven't touched the mainmenu files, other than init.lua, yet. |
| 16:29 |
sapier |
and the this/self issue ... I don't see any benefit in doing that change, most of us are c++ coders and will understand this even better |
| 16:31 |
ShadowNinja |
sapier: But this is Lua. And you're doing classes wrong, every object will have a copy or every member function, rather than Lua looking up the main one. |
| 16:31 |
sapier |
I'm gonne check the files I touched for missing headers and add them but I wont check other files too |
| 16:31 |
ShadowNinja |
of* |
| 16:31 |
ShadowNinja |
FWIW Python uses self too. I don't think anyone will have trouble understanding what it is. |
| 16:31 |
sapier |
if it was wrong it didn't work. And I wont discuss another 10h about this topic like the default size, if you believe it to be wrong you can change it later |
| 16:33 |
ShadowNinja |
sapier: I won't agree to this way of doing classes. Even functions that all take a table parameter would be better. |
| 16:34 |
sapier |
ok then I'm gonna drop that pull and whole android port with it |
| 16:34 |
ShadowNinja |
I even gave you an example. |
| 16:34 |
sapier |
I'm done with silly style guide discussions |
| 16:35 |
ShadowNinja |
It's not style, it's design. |
| 16:35 |
|
diemartin joined #minetest-dev |
| 16:36 |
sapier |
it's not design it's your personal cursade for your codestyle dogma |
| 16:36 |
sapier |
no need to continue this discussion |
| 16:36 |
sapier |
what about #1254? |
| 16:36 |
ShadowBot |
https://github.com/minetest/minetest/issues/1254 -- Add download rate to media progress bar (non-HTTP mode only!) by sapier |
| 16:38 |
ShadowNinja |
sapier: That's like saying using (std::map) obj["getX"]() instead of obj.getX() is a style decision. It affects what the code does. |
| 16:38 |
ShadowNinja |
You're passing around copies of every menber function. |
| 16:39 |
ShadowNinja |
sapier: s/disable_remote_media_server/enable_remote_media/ and it seems fine. |
| 16:40 |
sapier |
if you can proove self is anything different then this except of name I agree. |
| 16:42 |
ShadowNinja |
sapier: self is a special value in Lua, just like this is a special value in C++. It's passed implictly when you use member function notation. |
| 16:43 |
sapier |
that'S a claim proove it |
| 16:43 |
ShadowNinja |
But the main point is that you should use metatables. |
| 16:43 |
ShadowNinja |
sapier: Um, this runs in a Lua interpreter: |
| 16:43 |
sapier |
if you believe this to be the only way to do it you're gonna have to do it yourself anyway because I won't spend the time to do it |
| 16:44 |
ShadowNinja |
local t = {} local function t:f() print(self.x) end t.x = "Hello world" t:f() --> "Hello World" |
| 16:44 |
sapier |
I've once said I'm gonna support minetest as long as it's fun and it's started to not beeing very funny recently |
| 16:46 |
ShadowNinja |
sapier: How about #1258? |
| 16:46 |
ShadowBot |
https://github.com/minetest/minetest/issues/1258 -- Organize builtin into subdirectories and use "core" namespace by ShadowNinja |
| 16:47 |
sapier |
you're gonna merge it anyway so I don't need to comment it |
| 16:48 |
BlockMen |
ShadowNinja, why renaming the namespace? |
| 16:49 |
ShadowNinja |
BlockMen: It's awkward for forks. core is also shorter. :-) |
| 16:50 |
BlockMen |
ShadowNinja, then use engine, not core (since its minetest engine) ;) |
| 16:51 |
ShadowNinja |
BlockMen: No, a mod could use that. |
| 16:51 |
ShadowNinja |
It's known as both the core and the engine. |
| 16:51 |
ShadowNinja |
That's why we're known as *Core* developers. |
| 16:52 |
BlockMen |
ShadowNinja, wut? core could also used by a mod |
| 16:52 |
BlockMen |
and for me core developer was like the core of a large thing |
| 16:52 |
sapier |
Shadowninja according to docs there's not a single difference except of name of parameter between your and my code |
| 16:52 |
ShadowNinja |
BlockMen: Yes, any namespace could. But core is less lakely to be used. |
| 16:52 |
BlockMen |
show me one mod that uses engine |
| 16:53 |
ShadowNinja |
sapier: Huh, what docs? For what code? |
| 16:53 |
sapier |
lua doc |
| 16:53 |
sapier |
the only difference is that notation table:function(a,b) implicitly calls first paramerter "self" |
| 16:54 |
BlockMen |
i think renaming the namespace it not necessary now and should be done within a larger switch |
| 16:54 |
BlockMen |
like for minetest 0.5.x |
| 16:54 |
sapier |
so table:function(a,b) is exactly identical to table.function(self,a,b) |
| 16:54 |
ShadowNinja |
sapier: That's the only difference to the code inside the function. But the object and the way that it is called is very different. |
| 16:54 |
sapier |
no it isn't |
| 16:54 |
ShadowNinja |
sapier: Nope, table:function(table, a, b) |
| 16:54 |
sapier |
because I DO call it by using object:(a,b) |
| 16:55 |
ShadowNinja |
sapier: The same function gets called, but THE WAY THAT IT IS CALLED is different. |
| 16:55 |
sapier |
I don't call it that way |
| 16:55 |
sapier |
I call it table:function(a,b) |
| 16:56 |
ShadowNinja |
One calls a (copied) value, annother looks up the function in the class table and then calls it. |
| 16:56 |
sapier |
READ WHAT I'M TELLING |
| 16:56 |
ShadowNinja |
Er, table:function(a, b) == table.function(table, a, b) rather. |
| 16:56 |
ShadowNinja |
I am. |
| 16:56 |
sapier |
no it isn't |
| 16:56 |
|
proller joined #minetest-dev |
| 16:57 |
ShadowNinja |
I don't think you understand what I'm saying. |
| 16:57 |
sapier |
table:function(a,b) doesn't lookup table again while table.function(table,a,b) does |
| 16:57 |
ShadowNinja |
sapier: See how this does classes: https://github.com/ShadowNinja/LuaIRC/blob/master/init.lua |
| 16:57 |
ShadowNinja |
I know, but I'm not talking about that. |
| 16:57 |
sapier |
do it yourself I don't believe you |
| 16:58 |
ShadowNinja |
Do what? |
| 16:58 |
sapier |
your're so smart if you really believe your way is the only way to do it you're gonna have to do it yourself or proove you're right |
| 16:58 |
ShadowNinja |
BlockMen: It's fully compatible. |
| 16:59 |
ShadowNinja |
There's no reason to wait for this. |
| 16:59 |
BlockMen |
ShadowNinja, i know. but tell me a good reason to do that now |
| 16:59 |
ShadowNinja |
BlockMen: Because it clears up the minetest/engine confusion. |
| 16:59 |
sapier |
becaus it's ShadowNinjas will and codingsytle interpretation isn't that enough BlockMen? |
| 17:00 |
sapier |
Hmm I start to understand proller |
| 17:00 |
ShadowNinja |
sapier: Calm down. |
| 17:00 |
BlockMen |
ShadowNinja, and start new confiusion why it is now core, wheather there is any difference for no plus on the other side |
| 17:01 |
BlockMen |
i still think this should be done within 0.5.x |
| 17:02 |
sapier |
No I wont calm down ShadowNinja because yesterday discussion was about fixing mod bugs in core which is absolutely insane and you kept silent all the time |
| 17:02 |
ShadowNinja |
BlockMen: Just say that core is a new name for minetest, and that they are both the same and supported. |
| 17:02 |
proller |
... in this time in [censored]miner speed of liquid flow ~300000 nodes per second ;) |
| 17:03 |
sapier |
no need to be proud of avoiding all the hard and time consuming tasks by doing all your own proller ;-P |
| 17:03 |
ShadowNinja |
sapier: I don't know what you mean. |
| 17:03 |
BlockMen |
ShadowNinja, just say that minetest != minetest_game ? |
| 17:04 |
ShadowNinja |
BlockMen: Huh? minetest_game isn't involved. Although it should be changed to use the new namespace eventualy. |
| 17:04 |
proller |
sapier, i prefer code instead of shittalk |
| 17:05 |
sapier |
well as I said I'm starting to understand you at least partial |
| 17:05 |
BlockMen |
ShadowNinja, i thought you ment that with minetest/engine confusion. what confusion are you talking about then= |
| 17:05 |
BlockMen |
*? |
| 17:05 |
sapier |
10h of discussion to 2h of coding is absolutely inacceptable |
| 17:06 |
ShadowNinja |
BlockMen: Builtin used both names, and a lot of sections had things like `local tbl = engine or minetest` in them. |
| 17:06 |
proller |
sapier, and solve real problem instead of imagined |
| 17:07 |
BlockMen |
engine is only used by clientside lua IIRC |
| 17:07 |
sapier |
proller once you start to write clean code I'm gonna explain to you what's different between imagined and hidden problems |
| 17:07 |
ShadowNinja |
BlockMen: It's only used by the main menu. But some files are used by both the server and main menu. |
| 17:08 |
BlockMen |
so there is no problem for 99% of modders |
| 17:09 |
BlockMen |
IMO it will cause more confusion than there is now, so a change with a huge version change makes more sense |
| 17:09 |
ShadowNinja |
BlockMen: Yes, this should only affect main menu script writers (if which there are about two, including sapier). |
| 17:10 |
ShadowNinja |
BlockMen: Just make an announcement saying "We now use the core table, you can continue using the minetest table without issues, but mods should switch to the new table. Both tables are identical, so you can mix them.". |
| 17:11 |
BlockMen |
there is no need to switch, ShadowNinja. Im against it. |
| 17:11 |
sapier |
well you're gonna have a hell of a work to cleanup current mainmenu because it's become really messy |
| 17:12 |
|
proller joined #minetest-dev |
| 17:12 |
ShadowNinja |
BlockMen: There is. The two tables have made the code messy. We could even leave `core` as an implementation detail and not tell modders about it until 0.5. |
| 17:12 |
|
[PavelS] joined #minetest-dev |
| 17:13 |
BlockMen |
ShadowNinja, you make confusion for many many moders just because in a few files there is "minetest or engine" |
| 17:14 |
sapier |
I'm positive about the name switch but I wont rebase my cleanup to it it's already way too much code moving |
| 17:15 |
sapier |
and as usual it's gonna be me to be acused for any bug no matter if it's caused by my changes or ShadowNinjas |
| 17:15 |
ShadowNinja |
BlockMen: Don't tell moders about 'core' then. And using one table alows me to do other things with the core to make it cleaner, such as moving table creation to initialization. |
| 17:16 |
ShadowNinja |
sapier: I intentionally didn't touch anything more than the header of the main mainmenu file so that it would rebase with minimal conflicts. |
| 17:17 |
BlockMen |
ShadowNinja, fine. do it "in the background", leave documentation as is and im fine (in general) |
| 17:17 |
BlockMen |
although im still for engine instead of core |
| 17:17 |
sapier |
if you do a rename you have to rename it everywhere otherwise I don't agree to it. That's what's been done fpr luahud add hud functionality and keep same thing in c++ code resulting in two things to maintain |
| 17:18 |
BlockMen |
sapier, as long we have 0.4.x it should stay minetest. |
| 17:18 |
BlockMen |
* in documentation |
| 17:19 |
sapier |
thought he did rename engine to minetest? am I wrong about it? |
| 17:19 |
BlockMen |
he changed everything to "core" |
| 17:20 |
sapier |
good god ... ShadowNinja ... wtf are you doing? |
| 17:20 |
ShadowNinja |
lol |
| 17:20 |
ShadowNinja |
I already explaned this. |
| 17:21 |
sapier |
isn't it enough to have one compatibility name you need to add two compatibility modes? for what changing a generic name as engine to another generic name like core? why don't you just call it hansottofriedrichwillhelm |
| 17:22 |
Megaf |
Hi all |
| 17:22 |
Megaf |
I go no bugs to report |
| 17:22 |
Megaf |
got* |
| 17:22 |
sapier |
celeron55 did express his dislike of generic names because of namespace collisions. I admit I didn't think about that issue when adding engine, and I agree to him. But your're changing things for what? sake if changeing? |
| 17:23 |
sapier |
I could understand if you didn't agree to this issue but then the only sane rename would be minetest->engine |
| 17:24 |
sapier |
not because engine is superior to core, but because it's equal good/bad and engine is already there |
| 17:25 |
ShadowNinja |
<celeron55> ShadowNinja: also the issue i mentioned is not about confusing a minetest file with a generic lua file; the issue is not being able to use a generic term locally in each mod's code <celeron55> i'd like to hear everyone's opinion on this though <celeron55> i'm in favor of "core" because it's shorter <celeron55> (compared to "engine") |
| 17:26 |
* BlockMen |
still prefers engine |
| 17:26 |
sapier |
me too because it's one less compatibility layer |
| 17:53 |
sapier |
ShadowNinja: did I understand you correct that you believe variant b is superior to variant a? https://gist.github.com/sapier/11407336 |
| 17:55 |
ShadowNinja |
sapier: Yes, but that's not what I was talking about before. I'll show you how I want it... |
| 17:56 |
sapier |
yo don't need to show me what you want if you don't manage to explain to me why this is really better I don't care what you want or not if there's no real reason for it |
| 17:59 |
ShadowNinja |
sapier: https://gist.github.com/sapier/11407336#comment-1219777 |
| 18:01 |
ShadowNinja |
sapier: Also: assert(someobject.func == new_someobject().func) |
| 18:02 |
sapier |
ok and now what't the benefit of adding a fallback table that may or may not be overwritten? |
| 18:03 |
sapier |
for what I know functions aren't copied by assignment but referenced only so what's the difference? exept of direct referencing may be even faster |
| 18:03 |
ShadowNinja |
One copy of the function obviously. And it can be overwritten, yes, but you can also overwrite new_someobject. There's no point in supporting that. |
| 18:04 |
sapier |
you expect me to believe that table.someting = somefunc does create a copy of that function? |
| 18:04 |
ShadowNinja |
Yes. Lua only passes tables by reference. |
| 18:05 |
Taoki |
Ugh... did biome temperatures also get removed with the removal of finite liquid? I hope I'm not understanding right... cuz that will break a lot of things! |
| 18:05 |
ShadowNinja |
Also, pairs() on the table will iterate over functions too, whether or not they're copied. |
| 18:05 |
Taoki |
Can anyone explain what's happening? Why would anyone remove such an important feature for no reason? |
| 18:06 |
ShadowNinja |
The only mention I found of copying Lua functions in a web search was copying functions between lua_States. |
| 18:06 |
Taoki |
There are many mods that likely depended on temperature and humidity and more |
| 18:06 |
Taoki |
Finite liquid is fine, don't think anyone cared much... but why remove support for temperature system and the like? |
| 18:06 |
ShadowNinja |
Taoki: Weather was removed. It was buggy and nobody supported it. |
| 18:06 |
Taoki |
Mods could have used temperature for other things |
| 18:06 |
Taoki |
That's the worst decision I've ever seen taken in this project so far |
| 18:07 |
BlockMen |
Taoki, you know even one mod that uses get_heat()? |
| 18:08 |
Taoki |
BlockMen: If none were mentioned publicly that doesn't mean it's ok to remove such features after they were added |
| 18:09 |
Taoki |
Anyway, I'm seeing you can still get the heatmap of biomes... |
| 18:09 |
Taoki |
Maybe that still works. At least with v7 mapgen |
| 18:09 |
BlockMen |
Taoki, just because something was introduced it does not mean it has to stay in forever |
| 18:10 |
Taoki |
If you have a reliable project, features that can be used like that aren't typically removed |
| 18:11 |
BlockMen |
when things are buggy and get not fixed/improved then there is no reason to keep |
| 18:11 |
sapier |
according to all doc I find ShadowNinja you're wrong tables strings AND functions are allways passed by reference |
| 18:11 |
Taoki |
What was buggy about area temperatures though? |
| 18:11 |
sapier |
only plain datatypes are passed by value |
| 18:12 |
sapier |
I've got doubts about this doc beeing true for strings too but that'd be easy to chck |
| 18:12 |
ShadowNinja |
sapier: Lua strings are interned, so they are internally a reference. It fact `a="test"` and `b="test"` both point to the same memory location. They act like seperate values though. |
| 18:12 |
ShadowNinja |
I assume functions are interned too. |
| 18:13 |
Taoki |
sapier: Any opinion on this too? IMO it's not normal to remove Lua functions which themselves worked well and mods might be depending on |
| 18:13 |
Taoki |
I'm thinking of asking enough developers if parts of the weather system can be reverted... |
| 18:13 |
Taoki |
Without finite liquid itself, but the parts that were good and shouldn't have been removed |
| 18:13 |
sapier |
this results for a = func() end b = a, b beeing same as a |
| 18:14 |
sapier |
so table.myfunc = somefunc doesn't create a copy of that func but a first level reference compared to your second level reference by using metatables |
| 18:14 |
ShadowNinja |
sapier: Yes, but if you change one of them the other won't be updated. |
| 18:14 |
sapier |
one of what? |
| 18:14 |
ShadowNinja |
sapier: One of the functions. |
| 18:15 |
sapier |
if those functions are changed runtime something is terribly broken |
| 18:15 |
VanessaE |
Taoki: if you need temperature/humidity, use plants_lib. |
| 18:15 |
Taoki |
VanessaE: Looking or what lua functions I can use to still get them in latest master |
| 18:15 |
Taoki |
I plan to use them myself later on |
| 18:15 |
Taoki |
Well, PLANNED to use features that got blindly removed for no reason. Now I need to find an alternative |
| 18:16 |
sapier |
I dont want my library code to be rewritten on the fly that horrific style |
| 18:16 |
Taoki |
eg: I wanted players not wearing armor (good clothing) to take damage where it's too cold |
| 18:16 |
ShadowNinja |
sapier: No, functions are often overridien by mods (I beleive this is called decorating) to catch values passed to them. |
| 18:17 |
BlockMen |
Taoki, all i can say that i got no sane values for get_heat() when testing it |
| 18:17 |
ShadowNinja |
sapier: That's also harder to maintain, since you have to add a copy for every new member function, rather than just adding it to the class table. |
| 18:17 |
sapier |
you don't get it if those functions are overwritten it's gonna be broken |
| 18:17 |
VanessaE |
Taoki: plants_lib can do that. |
| 18:17 |
Taoki |
BlockMen: Heat and humidity would oscilate per area |
| 18:17 |
VanessaE |
(it sounds unintuitive, but it does have that ability oddly enough) |
| 18:17 |
Taoki |
VanessaE: Is plants_lib a md or code for the engine? I need mine to work with the original engine in master |
| 18:18 |
VanessaE |
it's a mod. |
| 18:18 |
BlockMen |
Taoki: https://github.com/minetest/minetest/commit/0279f32db742f06d24a9d01c46f7d70c5e570eb2#diff-1305560bd8befb32862f0feeefabd02eL1446 |
| 18:18 |
sapier |
that's not some custom code it's basic like 1+1 = 2 ... if you change this behaviour don't expect any calculation to be correct |
| 18:18 |
Taoki |
ok |
| 18:19 |
Taoki |
BlockMen: Yeah, a very bad choice IMO. I'm hoping other developers will agree to revert some of it |
| 18:19 |
ShadowNinja |
sapier: Eg, local x = core.register_node function core.register_node(name, ...) print("Node", name, "registered") return x(name, ...) end |
| 18:20 |
ShadowNinja |
Prints all node names as they are registered. |
| 18:20 |
sapier |
we're not talking about some code to be replaced by mods |
| 18:20 |
Taoki |
Ok... minetest.get_mapgen_object should be able to retreive humidity and temperature of a biome. Can anyone confirm? |
| 18:21 |
ShadowNinja |
sapier: Regardless, metatables are cleaner. |
| 18:21 |
sapier |
no they aren't their just your style no more no less |
| 18:22 |
sapier |
they add another language feature causing additional side effect where that feature isn't required and even potentially harmfull |
| 18:23 |
sapier |
if someone messes up a single object it's gonna be broken for all objects of same type |
| 18:23 |
Taoki |
great... now everyone's ignoring me |
| 18:25 |
sapier |
to much bad blood in here right now taoki it's not your fault |
| 18:25 |
Taoki |
makes sense |
| 18:25 |
Taoki |
Where can I find a list with all core developers? I want to petition re-adding weather without finite liquid |
| 18:26 |
sapier |
I guess shadow is writing at a lua styleguide right now to force all others to follow his will |
| 18:27 |
BlockMen |
Taoki, https://github.com/orgs/minetest/members |
| 18:27 |
BlockMen |
all core devs |
| 18:27 |
Taoki |
Thanks |
| 18:32 |
|
Anchakor_ joined #minetest-dev |
| 18:34 |
|
Miner_48er joined #minetest-dev |
| 18:43 |
|
Megaf_ joined #minetest-dev |
| 18:43 |
sapier |
I'm gonna push #1254 once travis build of fixed version is completed |
| 18:43 |
ShadowBot |
https://github.com/minetest/minetest/issues/1254 -- Add download rate to media progress bar (non-HTTP mode only!) by sapier |
| 18:43 |
|
Megaf_ joined #minetest-dev |
| 18:48 |
sapier |
ShadowNinja: what to use instead of script_error, errorhandling is your baby |
| 18:50 |
|
Amaz_ joined #minetest-dev |
| 18:52 |
ShadowNinja |
sapier: errorstream or an exception. |
| 18:52 |
ShadowNinja |
(Errorstream would be more appropriate in that case) |
| 18:53 |
sapier |
errorstream will flood in game log that's not intended |
| 18:54 |
sapier |
what to use instead of script_error, that's way more important then what stream to use for fatals |
| 18:57 |
ShadowNinja |
sapier: It isn't? Don't you want to log depreciation errors? |
| 18:58 |
sapier |
I want to log them not to flood game with it |
| 18:58 |
sapier |
verbosestream is usually written to log imho that's enough |
| 18:59 |
sapier |
and mod developers should use error anyway where they're gonna see the reason too |
| 18:59 |
ShadowNinja |
errorstream -> log and console. |
| 18:59 |
sapier |
isn't script error supposed to do that too? |
| 18:59 |
ShadowNinja |
It might also go to the chat in singleplayer. |
| 18:59 |
ShadowNinja |
sapier: script_error is for an entirely different purpose. |
| 19:00 |
sapier |
that's exactly what I don't want it to behave |
| 19:00 |
sapier |
those messages (at log level) shall be there but not stop minetest from working at all |
| 19:00 |
ShadowNinja |
sapier: script_error dumps the string on the top of the stack into a exception. |
| 19:00 |
ShadowNinja |
Which is caught and sent to errorstream. |
| 19:00 |
sapier |
ok what do i use to get the current backtrace? get backtrace and throw luaerror? |
| 19:01 |
sapier |
pushing 1254 now |
| 19:01 |
ShadowNinja |
Getting the backtrace is only successfull if the stack hasn't unwound. And you could use a LuaError. |
| 19:02 |
sapier |
ok lets precisely specify what I want to have, I want to see the backtrace of current lua call where the deprecated function is called. how do i do this with current code |
| 19:03 |
sapier |
don't tell me it's not possible that'd be strange |
| 19:04 |
Taoki |
celeron55, PilzAdam, hmmmm, BlockMen, sapier, and anyone else in the dev team: https://forum.minetest.net/viewtopic.php?f=7&t=9215 Please vote and post your opinion here if you wish to have a say on the matter. |
| 19:04 |
Taoki |
Sorry for highlighting many devs also... I really wish for someone to take a look at this however |
| 19:05 |
sapier |
taoki do you have anyone to do the work? |
| 19:05 |
Taoki |
sapier: Reverting a GIT commit is the only work that would be needed :P Apart from separating finite fluid from weather perhaps, if only one is to be re-added |
| 19:06 |
ShadowNinja |
sapier: script_get_backtrace should work there. |
| 19:06 |
ShadowNinja |
Taoki: Git isn't an acronym. |
| 19:06 |
Taoki |
sapier: If people want, I can try to pick the GIT diff and extract weather from finite liquid |
| 19:06 |
sapier |
but that one doesn't throw the error |
| 19:07 |
sapier |
taoki that code was reverted because it's been broken and quite ugly we wont merge it again that way. in order to get it merged it needs some rework |
| 19:07 |
sapier |
ok actually it was reverted because of proller didn't want it to be there but noone did reject his request because of upper reason |
| 19:08 |
Taoki |
sapier: What was broken? I didn't see any problem with it |
| 19:08 |
sapier |
ShadowNinja so throw LuaError(script_get_backtrace(L)) wont cause double stacktrace? |
| 19:09 |
ShadowNinja |
sapier: Yes. LuaError doesn't collect one anymore. |
| 19:09 |
|
tomreyn joined #minetest-dev |
| 19:10 |
sapier |
ok ... btw can you stop quoting same code shifted by 2 lines over and over again and use a single quote and line numbers? |
| 19:11 |
ShadowNinja |
sapier: Do you mean make one comment for multiple lines? |
| 19:11 |
sapier |
if those lines are L10 L11 and L12 yes ;-P |
| 19:16 |
sapier |
ok guess using errorstream in error mode wont cause any harm so I'm gonna change it to what you requested here too |
| 19:27 |
|
harrison joined #minetest-dev |
| 19:28 |
sapier |
ok ok verbosestream is wrong for everything, actually I'd like it to be at actionstream ... but that's pretty wrong because it ain't a action, why don't we have any real loglevel written to log but not shown in char |
| 19:32 |
sapier |
ShadowNinja: throw LuaError is causing a double stacktrace |
| 19:36 |
|
Hiradur joined #minetest-dev |
| 19:37 |
|
proller joined #minetest-dev |
| 19:39 |
sapier |
any other comments to #1261? if not I'm gonna merge it in about 2h |
| 19:39 |
ShadowBot |
https://github.com/minetest/minetest/issues/1261 -- Add proper Lua API deprecated handling by sapier |
| 19:50 |
Taoki |
Oh... one more thing. Don't wanna sound ironic, but seeing this it's actually a real concern; I'm going to use mapgen v7 as part of a game I will be making. Is there a certainty that at least this mapgen will never be removed? At least as long as there isn't another one suporting all of its features... particularily Lua biomes |
| 19:50 |
Taoki |
Like I dunno... if hmmmm doesn't fix the remaining bugs in the next years or so (just theoretically, am sure he will) |
| 19:51 |
sapier |
usually we keep compatibility, I'll try to make sure n-1 compatibility is kept but I'd not bet on it |
| 19:54 |
Megaf_ |
sapier: did you get my message? |
| 19:55 |
sapier |
no megaf |
| 19:55 |
PilzAdam |
Taoki, if you want heat and humidity in mods just use the noise functions of mapgen |
| 19:55 |
sapier |
was lost withing heavy fire |
| 19:56 |
Taoki |
PilzAdam: How do I get the humidity and heat at a certain position at any moment however? I'm seeing it's only possible to read in on_generate, using the mapgen object |
| 19:56 |
PilzAdam |
Taoki, thats basically what the API function did when you had weather=false in minetest.conf |
| 19:56 |
Taoki |
I know all mapgens have heat / humidity map, but no idea if there's still a way to read them from Lua |
| 19:56 |
Taoki |
Is there? |
| 19:56 |
PilzAdam |
I guess a callback for that can be added |
| 19:57 |
sapier |
wasn't that one removed with weather? |
| 19:57 |
PilzAdam |
we don't need to re-add the whole dynamic weather stuff for that |
| 19:57 |
Taoki |
That would be very appreciated yes |
| 19:57 |
Taoki |
If a Lua function is added to get mapgen heat and humidity at any posision (x, y, z), I'm happy enough |
| 19:57 |
sapier |
of course if that data is still there it's not a big deal to provide access to it |
| 19:57 |
Taoki |
Though I miss humidity dependentfog density :/ |
| 19:58 |
PilzAdam |
was that even in Minetest? |
| 19:58 |
PilzAdam |
I thought it was only in freeminer |
| 19:59 |
Taoki |
Or well, any Lua function that can provide the complete heatmap and humiditymap. But possibly with a builtin function to get value per position |
| 19:59 |
Taoki |
PilzAdam: Yes, for quite some time |
| 20:00 |
PilzAdam |
Taoki, have you set weather=true in minetest.conf? |
| 20:00 |
Taoki |
PilzAdam: Yes. Though it no longer matters now since the system was removed. |
| 20:01 |
PilzAdam |
the fog is only changed if you have that to true |
| 20:01 |
PilzAdam |
*was |
| 20:01 |
Taoki |
proller decided to remove the whole weather system with finite liquid. Which I disagree with. |
| 20:01 |
Taoki |
I understand finite liquid was buggy, but I haven't heard the same about weather myself |
| 20:03 |
PilzAdam |
proller implemented weather in a way that it depends on finite liquids |
| 20:03 |
PilzAdam |
also all these hard-coded ABMs aren't good for Minetest as a game engine |
| 20:04 |
Taoki |
PilzAdam: My personal suggestion is to add back only part of the weather code. Particularily per-position temperature / humidity, as well as fog density depending on humidity |
| 20:04 |
Taoki |
If that's done, any game can do its own weather in Lua |
| 20:05 |
Taoki |
And freezing / unfreezing can be done the right way, without any hard-coded ABM (I didn't even know about that). Just have an ABM executing every X seconds and checking temperature at its location |
| 20:05 |
PilzAdam |
so simply add functions to read the heat and humidity that is used in mapgen |
| 20:06 |
sapier |
Taoki everyone agrees to having weather in there but you need to find someone to write/fix the code |
| 20:06 |
Taoki |
Sure. Though I also enjoyed seeing them in the debug text (F5) and fog depending on humidity. Perhaps that can be kept? |
| 20:06 |
PilzAdam |
the dynamic change of that can be done in Lua too |
| 20:06 |
Taoki |
sapier: What's so badly broken though :P |
| 20:07 |
PilzAdam |
bringing back the fog thing is a problem though; since I don't want to make the heat/humidity functions dynamic |
| 20:07 |
PilzAdam |
that part should be done in Lua |
| 20:07 |
Taoki |
PilzAdam: IIRC you can't set a player's fog density in Lua (or draw distance). That would also solve it, since I mostly care to use this for non-minetest_game stuff |
| 20:07 |
Taoki |
ok |
| 20:07 |
Taoki |
Later we can add a function to force draw distance, which will also include fog |
| 20:07 |
PilzAdam |
btw, was that weather stuff removed from minetest_game? |
| 20:08 |
Taoki |
If I consider anything urgent though, it's the ability to get heat / humidity from mapgen per position. |
| 20:08 |
Taoki |
PilzAdam: I think so. Freezing is obviously gone now |
| 20:08 |
PilzAdam |
the freezemelt field is still there; also in lua_api.txt |
| 20:09 |
PilzAdam |
that needs to be cleaned up |
| 20:09 |
proller |
but minetest_game now read only! |
| 20:11 |
Taoki |
proller: I understand the weather code at least is yours, so you know best what's up with it. Any way in which you consider it could be added back without finite liquid? Or more specifically, and parts of it? |
| 20:11 |
PilzAdam |
#1263 |
| 20:11 |
ShadowBot |
https://github.com/minetest/minetest/issues/1263 -- Remains of weather |
| 20:12 |
proller |
Taoki, do what you want |
| 20:12 |
Taoki |
I don't have GIT access :P If it's just selecting parts of the code someone who can also merge them would be best at that |
| 20:12 |
Taoki |
Asking you since you did the reverting |
| 20:13 |
Taoki |
Also I'm bad with mapgen stuff... never looked at that part of the code |
| 20:20 |
|
Exio4 joined #minetest-dev |
| 20:21 |
Taoki |
Either way, I usually never ask other developers to code something for me. But since this is something I made plans based on, and the revert removed this feature, I'd really appreciate a re-implementation of heat / humidity @ position reading in Lua. Past that I don't care... humidity based fog can be done in Lua later perhaps. |
| 20:22 |
Taoki |
But someone please re-add this ASAP. I mean, if anyone has the time and wishes to. |
| 20:23 |
proller |
it implemented in freeminer, faster and less buggier |
| 20:24 |
sfan5 |
that is not a solution |
| 20:24 |
sfan5 |
Taoki: simply use your own perlin noise 3d to get such things |
| 20:24 |
sfan5 |
s/simply// |
| 20:24 |
Taoki |
proller: Can it be pulled in Minetest too? |
| 20:24 |
Taoki |
sfan5: I need to know the mapgen's temp / heat map. Can perlin noise be used for that? |
| 20:25 |
sfan5 |
hm |
| 20:25 |
sfan5 |
yes |
| 20:25 |
sfan5 |
but you may need to specilize yourself to mgv6 |
| 20:25 |
proller |
Taoki, its not small changes, everybody will against this |
| 20:25 |
sfan5 |
mgv7 has different noise params and biomes |
| 20:26 |
Taoki |
ok. Kinda needing such a thing for v7 |
| 20:26 |
Taoki |
But if it's possible, someone who understands perlin noise might add the function in builtin |
| 20:26 |
sfan5 |
I think biomes affect this thing too |
| 20:26 |
sfan5 |
but I'm not sure |
| 20:26 |
sfan5 |
you don't need a function |
| 20:26 |
Taoki |
What's the current way to do it in Lua in that case? |
| 20:26 |
sfan5 |
you just need to have the data mgv6 uses for the perlin noise 2D/3D that is used to compute heat & humidity |
| 20:27 |
Taoki |
Sadly I don't understand anything this fancy, when it comes to perlin maps and all that |
| 20:28 |
sfan5 |
you just need to understand the bit of code in mgv6 that calculates heat / humidity and port that to lua |
| 20:28 |
sfan5 |
then you have a (mgv6 specific) solution |
| 20:28 |
Taoki |
I think the old code (back when weather was in) worked for all mapgens |
| 20:29 |
Taoki |
That's why I suggested porting back only that piece of code |
| 20:29 |
sfan5 |
why? |
| 20:29 |
sfan5 |
heat / humidity has no use for minetest or it's core features |
| 20:29 |
Taoki |
sfan5: It can be used for many things. And I plan to use it myself a lot |
| 20:29 |
sfan5 |
if a mod needs it you may have to port the code to lua and then use |
| 20:29 |
Taoki |
When I get started on my game... hopefully soon |
| 20:30 |
sfan5 |
may I know for what? |
| 20:30 |
sapier |
ShadowNinja: I admit string.format is really faster, it doesn't make a huge difference but as it's supposed to be a base lib I'm gonna fix it |
| 20:31 |
Taoki |
For one thing, I want to add freezing / unfreezing of water sources in my game, even without finite liquid. There's other ideas too... such as players taking damage when it's really cold / hot and they aren't dressed properly. Or tinging the screen blue / red based on how hot or cold it is |
| 20:31 |
sfan5 |
are we using things like http://msgpack.org/ yet? |
| 20:31 |
Taoki |
**tinting |
| 20:31 |
sfan5 |
Taoki: you can use a perlin 3D map with parameters you like for that |
| 20:31 |
Taoki |
But temperature must represent the biome :P |
| 20:32 |
Taoki |
I have no idea how to do that, and apparently it's no longer possible since the old weather system is gone |
| 20:32 |
Taoki |
That's the problem |
| 20:32 |
sfan5 |
then somehow get the biome by looking which nodes are there |
| 20:33 |
Taoki |
That's an ugly hack, and wouldn't always work... especially not properly |
| 20:33 |
Taoki |
The right way is a basic Lua function that can return heat / humidity at position |
| 20:34 |
ShadowNinja |
sfan5: We're using JSON. |
| 20:34 |
sfan5 |
isn't msgpack faster? |
| 20:34 |
ShadowNinja |
freeminer uses msgpack for a lot of things. |
| 20:35 |
ShadowNinja |
sfan5: Maybe. I haven't done benchmarks. |
| 20:35 |
Taoki |
proller: What are the changes needed for that, and why would anyone be against it? |
| 20:35 |
sapier |
I am against it |
| 20:35 |
ShadowNinja |
Deserialization should definitely be faster. Serialization might be comparable. |
| 20:35 |
Taoki |
:/ |
| 20:35 |
proller |
;) |
| 20:36 |
sapier |
we just added json because someone wanted this super cool new format ... that's been half a year ago |
| 20:36 |
Taoki |
Ah, thought aspier meant me lol |
| 20:36 |
Taoki |
**sapier |
| 20:36 |
proller |
remove json now! |
| 20:36 |
sapier |
next one in another half a year? |
| 20:36 |
sapier |
we're gonna switch serialization format more often then we do a release this way |
| 20:37 |
proller |
ShadowNinja will implement masterserver with marchal serialize |
| 20:37 |
proller |
on php |
| 20:37 |
Taoki |
proller: Anyway, can you please try to find a solution for this, since you know that code best? Simplest way works if it returns the proper value, so any implementation the devs find best will do |
| 20:37 |
sfan5 |
proller: php is bad |
| 20:37 |
sapier |
no if I know ShadowNinja rigth he's gonna do base lua dump serialization ;- |
| 20:37 |
ShadowNinja |
sapier: JSON was added because a standard format was needed. Lua format isn't easy to read. |
| 20:38 |
sapier |
You really believe this true? |
| 20:38 |
proller |
Taoki, fm. |
| 20:38 |
Taoki |
fm? |
| 20:39 |
proller |
sapier, xml much better! |
| 20:39 |
ShadowNinja |
sapier: Yes. Do you think that we should have implemented minetest.deserialize in perl? |
| 20:39 |
proller |
Taoki, freeminer |
| 20:39 |
Taoki |
I don't plan to develop something which isn't compatible with core Minetest |
| 20:39 |
Taoki |
Need it to work here too |
| 20:40 |
sapier |
no but I know that json wasn't added because it's better but because it's been prefered language of a single person ... the one promising to rewrite the json parsing code ... which isn't even done yet |
| 20:40 |
proller |
json is not language |
| 20:41 |
sapier |
call it whatever noun you want |
| 20:41 |
ShadowNinja |
sapier: JSON is standardized and has a easy to use library for just about every language in existence. It's better. |
| 20:42 |
sapier |
as well as how many other languages/formats out there? ... I don't say it may not have been chosen if having done a proper evaluation too but that didn't happen |
| 20:43 |
sapier |
but if was done we could now answer to requests for another serialization format "no we don't need it because of x y z ..." |
| 20:44 |
ShadowNinja |
sapier: Tell me of another standardized format that has so many implementations. |
| 20:44 |
sapier |
if you only want "so many implementations" xml is the obvious answer ... not telling xml is better for what we need |
| 20:46 |
sapier |
as I said json may have been choosen to but don't tell me it's been chosen based upon solid evaluation that's simply not true |
| 20:47 |
sapier |
and if there is good reason to use msgpack now that'd be a proove the way json was chosen was wrong |
| 20:50 |
sapier |
I'm pushing #1261 in a few minutes anyone wanna tell about a bug beeing in there? |
| 20:50 |
ShadowBot |
sapier: Error: Delemiter not found. |
| 20:51 |
|
grrk-bzzt joined #minetest-dev |
| 20:55 |
sapier |
I wonder why github decided to print 20 lines for each line comment is a good idea, if someone like shadowninja abuses this feature to mark every second line the comments get quite useless. |
| 21:16 |
|
BlockMen left #minetest-dev |
| 21:18 |
ShadowNinja |
sapier: XML does have a fair amount of implementations, but not as many as JSON. And it's very difficult (and slow) to use. |
| 21:19 |
ShadowNinja |
msgpack is in some ways better that JSON. |
| 21:20 |
sapier |
I already regret telling about the insanenness of adding a new serialization twice per year .. |
| 21:26 |
sapier |
no open pull requests by sapier ... I wish I could keep it that way ... but I wanna get android port done so back to fighting |
| 21:29 |
sapier |
ShadowNinja: may I delete the comments I for #1241 I fixed in order to see whats left? |
| 21:29 |
ShadowBot |
sapier: Error: Delemiter not found. |
| 21:56 |
|
EvergreenTree joined #minetest-dev |
| 21:59 |
sapier |
any concerns about #1264 |
| 21:59 |
ShadowBot |
sapier: Error: Delemiter not found. |
| 22:00 |
sapier |
Fix usage of deprecated functions in builtin #1264 |
| 22:00 |
ShadowBot |
sapier: Error: Delemiter not found. |
| 22:00 |
* VanessaE |
kicks ShadowBot in the bolts. |
| 22:02 |
CiaranG |
The world needs more delemiters, whatever the hell they are |
| 22:02 |
CiaranG |
ShadowBot: try looking for delimeters, that could work ;) |
| 22:05 |
sapier |
guess ShadowBot doesn't like me any longer ;-) |
| 22:05 |
diemartin |
delimiters work too :P |
| 22:06 |
sapier |
I'm still gonna push 1264 it's a syntactical fix only |
| 22:12 |
sapier |
#1265 please check it does only change heart and bubblebars to scale correctly |
| 22:12 |
ShadowBot |
sapier: Error: Delemiter not found. |
| 22:14 |
|
RealBadAngel joined #minetest-dev |
| 22:15 |
RealBadAngel |
hi folks, whats up? |
| 22:15 |
Taoki |
hiii |
| 22:15 |
Taoki |
Not much |
| 22:16 |
VanessaE |
hey RBA |
| 22:16 |
Taoki |
Pissed about a feature removal related to something I was going to use, trying to get the attention of more devs. Also I wanted to ask you something and toptally forgot what :P |
| 22:16 |
RealBadAngel |
lemme guess |
| 22:16 |
sapier |
I'm not involved ;-P |
| 22:16 |
RealBadAngel |
some fluid thingy? |
| 22:17 |
Taoki |
RealBadAngel: Close. The removal of the weather system, which also removed the ability to read mapgen humidity and temperatures |
| 22:17 |
Taoki |
Ah, and I think I remembered at least one thing I wanted to ask |
| 22:17 |
VanessaE |
maybe I should rename plants_lib to biome_lib and add a few extra functions to make it possible to replace the now-defunct weather code. anyone wanna help? ;) |
| 22:18 |
RealBadAngel |
sapier, i will push the changes tommorow i think, i took day off so will have lotsa time to finish that. and i want it to be complete, so no1 will ever touch it again |
| 22:18 |
Taoki |
RealBadAngel: 6D facedir for nodeboxes is possible now. Is Technic finally using it for wires? So only one form of each type of wire can be defined as a node, to not bloat and slow things |
| 22:18 |
VanessaE |
RealBadAngel: do not. |
| 22:19 |
sapier |
RealBadAngel: I just added proper lua api deprecation handling, maybe having this mechanism in place is a way to settle the dispute ... doesn't seem so if you're already at confrontation ... |
| 22:19 |
sapier |
+again |
| 22:19 |
RealBadAngel |
oh, what have you changed? |
| 22:20 |
sapier |
added a way to notify mod developers about deprecated api in a more sane way then flooding screen with error messages for everyone |
| 22:20 |
RealBadAngel |
Taoki, only reasonable thing for wires is wirelike |
| 22:21 |
Taoki |
RealBadAngel: Wouldn't 6D facedirs for nodeboxes do till then? Wirelike would be nice, but doesn't seem to be in yet |
| 22:21 |
RealBadAngel |
or that branch that uses meta to store node defs |
| 22:21 |
sapier |
basically it's how minetest will react to deprecated api, in release mode it's set to legacy --> mimic old behaviour |
| 22:21 |
sapier |
debug build defaults to logging usage of deprecated api |
| 22:22 |
sapier |
and for mod developers there's an additional configuration level "error" causing minetest to print a backtrace and stop execution on usage of deprecated api |
| 22:23 |
RealBadAngel |
sapier, what i started to code is fixing statbar to behave same way as it before, but also allowing the way you have used to (if needed) |
| 22:23 |
sapier |
why do you start coding things already done ... nih syndrom? |
| 22:23 |
RealBadAngel |
started 2 days ago, but had no time to finish |
| 22:24 |
Taoki |
RealBadAngel: Ahhh, now I remembered what I was going to ask you! :) |
| 22:24 |
sapier |
as far as I remember you didn't ever want to keep the statbar behaviour |
| 22:24 |
RealBadAngel |
sapier, you have forced me to want it |
| 22:25 |
Taoki |
RealBadAngel: About the texturable sky. Is it possible to have the 6 sky textures contain transparent areas, but have the normal sky be visible behind it (colors, sun / moon, stars and clouds)? |
| 22:25 |
RealBadAngel |
so you will have ur baloons ;) |
| 22:25 |
sapier |
but that's pointless please check 1265 |
| 22:25 |
Taoki |
The Lua API seems to suggest you can only use bgcolor behind transparent areas |
| 22:25 |
RealBadAngel |
sapier, ok |
| 22:25 |
sapier |
no it's just what taoki suffers from right now |
| 22:25 |
sapier |
some behaviour beeing changed |
| 22:25 |
RealBadAngel |
Taoki, atm no |
| 22:25 |
Taoki |
Ahh, ok. I was hoping to use that |
| 22:26 |
RealBadAngel |
using skybox disables other elements |
| 22:26 |
Taoki |
sapier: It's just a feature I was going to use being removed, but yeah |
| 22:26 |
sapier |
yes I admit I was somehow sleeping when proller did remove it, didn't realize about the consequences |
| 22:26 |
RealBadAngel |
this has to be changed |
| 22:26 |
Taoki |
RealBadAngel: Any chance you might be able to correct that sometime? Would be very nice and helpful... might make good stuff with it |
| 22:26 |
RealBadAngel |
we do need at least two layers of skyboxes |
| 22:26 |
RealBadAngel |
and regular sun/moon |
| 22:27 |
RealBadAngel |
with that we can emulate weather |
| 22:27 |
RealBadAngel |
and day/night cycle |
| 22:27 |
RealBadAngel |
fire up terasology and see how folks have done it |
| 22:27 |
Taoki |
sapier: Just worried some of my ideas might no longer be possible, which is why I'm poking devs to find a solution. To me though, any of the 3 things is OK: Re-add everything back, re-add only part of proller's code (the weather), or add a completely new Lua function to retreive heat / humidity |
| 22:28 |
Taoki |
But I hope one will happen |
| 22:28 |
Taoki |
RealBadAngel: The current sky and sun / moon is good. But yeah, it would prolly require two layers to allow both sky textures and regular sky |
| 22:28 |
sapier |
if heat/humidity is stored in map, it's not gonna be a big deal to readd it |
| 22:29 |
RealBadAngel |
shouldnt that be a part of mapgen? |
| 22:29 |
Taoki |
RealBadAngel: If you can though, put it on your TODO list please. Thinking of how this could be used to put things like giant spaceships on the sky or much more... but without removing the current normal sky |
| 22:29 |
VanessaE |
RealBadAngel: and the G*d damned skybox needs to be available BY DEFAULT without a separate mod. |
| 22:29 |
VanessaE |
that means adding the necessary code to minetest_game, even if the skybox images themselves aren't actually being supplied. |
| 22:29 |
Taoki |
sapier: IIRC all mapgens use heat and humidity. At least v7... not sure if v6. So it should be possible to add a Lua function to get thgem again |
| 22:30 |
RealBadAngel |
VanessaE, propably c55 will scream out loud again that im replacing so damn good code with my revelations |
| 22:30 |
Taoki |
problem is if and when anyone will do that. Since most devs seem to be against re-adding proller's code |
| 22:30 |
RealBadAngel |
as with sun gradients ;) |
| 22:30 |
Taoki |
Sun gradients were a great idea |
| 22:30 |
VanessaE |
RealBadAngel: yeah but you got the sun gradient to a point where he was at least conducive to the idea. |
| 22:30 |
Taoki |
RealBadAngel: BTW... do sun gradients color the directional fog too now? |
| 22:36 |
Taoki |
No worries if not. It should be easy to add that also, since the sun and moon colors both exist at once in the sky code, and are applied based on camera direction and blending value. So just replace the colors themselves and everything else will work by itself |
| 22:37 |
RealBadAngel |
well, i was watching its behaviour closely |
| 22:37 |
RealBadAngel |
it jumps |
| 22:37 |
RealBadAngel |
tonemaps makes the transitions smooth |
| 22:37 |
RealBadAngel |
propably c55 just put too much time into that code so he just love it ;) |
| 22:37 |
RealBadAngel |
one thing i wish some1 could start to code |
| 22:37 |
RealBadAngel |
using meshes as nodes |
| 22:37 |
RealBadAngel |
Taoki, you could use tonemaps |
| 22:37 |
sapier |
isn't that the one taboo minetest is never supposed to do? |
| 22:38 |
RealBadAngel |
simple if (m_sun_tonemap) |
| 22:38 |
* VanessaE |
lags |
| 22:38 |
Taoki |
Not familiar with them myself sadly |
| 22:38 |
sapier |
considering we already have mesh/polygon handling issues I'd ask if it'd be wise to use compilcated meshes for nodes too ;-) |
| 22:38 |
RealBadAngel |
Taoki, just read my latest changes to sky cpp |
| 22:39 |
RealBadAngel |
sapier, i have already saw (made) irrlicht dwarf being displayed in mt |
| 22:39 |
RealBadAngel |
thats not a problem |
| 22:40 |
sapier |
I don't say it doesn't work for a single mesh but thinking about vanessae's spawn issue I'm almost sure we've got a serious issue |
| 22:40 |
sapier |
what's gonna happen if you've got 200 meshed nodes each beeing 500 polygons? |
| 22:40 |
sapier |
and that'd not even be a complicated mesh |
| 22:40 |
RealBadAngel |
that is also not a problem |
| 22:41 |
RealBadAngel |
you know why? |
| 22:41 |
RealBadAngel |
its gonna be only client side |
| 22:41 |
sapier |
ohhhh I seee 1fps on client side is not an issue |
| 22:41 |
RealBadAngel |
the problem, the real one is the number of nodes |
| 22:42 |
sapier |
If you know the problem why didn't you fix the one happening for vanessae's server yet? |
| 22:42 |
VanessaE |
well no one sane would use meshes for something that there will be hundreds of displayed |
| 22:42 |
sapier |
I don't know what exactly happens there, I just know basics |
| 22:42 |
RealBadAngel |
we do have thousands of multiplied node definitions |
| 22:43 |
RealBadAngel |
for wires/tubes and the like |
| 22:43 |
sapier |
I don't believe in features being used in a sane way, especially if it's that easy to abuse it |
| 22:43 |
VanessaE |
but something like pipes where 99% of them are never actually *visible* (e.g. buried underground) it shouldn't matter if they have 50 or 50,000 polys each. If you can't see it, it shouldn't contribute to the rendering pipeline and shouldn't slow your fps down. |
| 22:43 |
VanessaE |
on Creative I mean |
| 22:43 |
VanessaE |
that specific instance is where they're all hidden |
| 22:43 |
VanessaE |
and yet FPS is poor there |
| 22:44 |
VanessaE |
everyone blames the underground pipes |
| 22:44 |
Taoki |
VanessaE: Since you run a popular server, I thought to suggest you too: https://forum.minetest.net/viewtopic.php?f=3&t=156&p=139975#p140000 See the mod in this screenshots post I made. It adds a beautiful environment Minetest doesn't have by default, and which would add an amazing feel to it |
| 22:44 |
sapier |
I agree to VanessaE and until we know why this happens we should be very very carefull to not increase polygon count even more then already possible |
| 22:44 |
Taoki |
(and sorry, I forgot this isn't for the dev channel) |
| 22:44 |
RealBadAngel |
sapier, polygon count is and will be high no matter what |
| 22:45 |
RealBadAngel |
even dirt is made out of it |
| 22:45 |
Megaf |
dev people, please do remove the "Preload Item Visuals" from settings and set it to off, since it doest actually do anything at all on the gameplayer and it only makes connection time a really painful and slow thing |
| 22:45 |
sapier |
maybe, but by now it seems to be best explanation, you can provide a better one. I assume you've got the knowledge to find that issue, I don't have that much knowledge about ittlicht |
| 22:45 |
Megaf |
^ sapier |
| 22:46 |
VanessaE |
Taoki: see /msg |
| 22:46 |
RealBadAngel |
Megaf, that should be replaced with some well thought rendering code |
| 22:47 |
VanessaE |
Megaf: when turned on, preload item visuals pre-renders all of the nodes/items ahead of time so that the game doesn't have to hang when new items show up. this isn't too noticable with default textures, but if you use a high-rez texture pack it is VERY noticable. |
| 22:47 |
hmmmm |
ern |
| 22:47 |
VanessaE |
it's a trade-off, what's needed is a way to turn off item texture extrusion entirely, like with this patch I use that kahrl gave me |
| 22:47 |
Taoki |
hi hmmmm |
| 22:47 |
hmmmm |
are you suggesting that objects aren't occlusion culled already? |
| 22:47 |
sapier |
I agree to megaf about switching the default, maybe vanessae but with high res textures you usually give up connection kill minetest and disable preloading in order to be able to enter |
| 22:48 |
VanessaE |
it makes leaving preload item visuals turned off almost entirely lag-free during normal gameplay |
| 22:48 |
VanessaE |
(at the expense of no 3d wielded items) |
| 22:48 |
VanessaE |
hmmmm: others are alluding to this, yeah. |
| 22:48 |
hmmmm |
err rather |
| 22:49 |
hmmmm |
it seems that occlusion culling in irrlicht doesn't cull things outside of the view frustrum |
| 22:49 |
hmmmm |
err does* |
| 22:49 |
hmmmm |
just doesn't cull nodes that are inside but occluded |
| 22:49 |
VanessaE |
hmmmm: correct |
| 22:49 |
hmmmm |
that's the purpose of occlusion queries which we probably don't make use of |
| 22:49 |
VanessaE |
there's an issue report from AGES ago about this |
| 22:49 |
hmmmm |
oh wait a minute |
| 22:49 |
hmmmm |
occlusion happens on a per-node basis |
| 22:50 |
hmmmm |
isn't the entire scene in minetest a single node? |
| 22:50 |
VanessaE |
https://github.com/minetest/minetest/issues/184 |
| 22:50 |
VanessaE |
yes, it is. |
| 22:50 |
VanessaE |
a single irrlicht node that is |
| 22:50 |
hmmmm |
well gosh i think that's the problem |
| 22:50 |
hmmmm |
why is the entire map a node |
| 22:50 |
hmmmm |
what problems are there with making a mapblock an entire node? |
| 22:51 |
hmmmm |
(question directed to people who know more about irrlicht and 3d graphics than I do *cough taoki celeron55 kahrl*) |
| 22:51 |
hmmmm |
RBGA too |
| 22:51 |
hmmmm |
RBA rather |
| 22:51 |
VanessaE |
hmmmm: beats the hell outta me |
| 22:51 |
VanessaE |
I'd have figured one mapblock per irrlicht node would make more sense |
| 22:51 |
Taoki |
hmmmm: Not sure if you saw the long debate above. But in essence, weather support was removed with finite liquid, and there's no way to read humidity / temperature per position any more. So I was wondering: Would it make sense to add a function for v7, which can return temperature / humidity at specific position? |
| 22:51 |
Taoki |
Since mgv7 is what I'm planning to use either way |
| 22:51 |
VanessaE |
or one node per material per material per mapblock maybe |
| 22:52 |
hmmmm |
erm |
| 22:52 |
hmmmm |
taoki that exists already to a point |
| 22:52 |
Taoki |
oh? |
| 22:52 |
hmmmm |
they're called "humiditymap" and "heatmap" mapgen objects |
| 22:52 |
Taoki |
I looked in the Lua API but didn't find much |
| 22:52 |
Taoki |
Ah, |
| 22:52 |
Taoki |
Yes, saw those |
| 22:52 |
hmmmm |
the problem is |
| 22:52 |
Taoki |
You can only get them in on_generated however |
| 22:52 |
hmmmm |
yes |
| 22:52 |
Taoki |
I need to always be able to get temp / humidity at any position |
| 22:52 |
hmmmm |
I'm working on the biome stuff |
| 22:52 |
Taoki |
Any way to add that? |
| 22:53 |
Taoki |
Awesome :) |
| 22:53 |
hmmmm |
right NOW, no |
| 22:53 |
hmmmm |
i do plan on making a query biome at point sort of thing for lua though |
| 22:53 |
hmmmm |
and the other associated functions |
| 22:53 |
Taoki |
ok |
| 22:53 |
VanessaE |
from what I've read about how people prefer irrlicht to be used, that's how it "should" be done anyway |
| 22:53 |
VanessaE |
oops, too many "per material" there. |
| 22:53 |
hmmmm |
i've been blowing a lot of time polishing up my perlin noise generator for quick map previewing |
| 22:53 |
Taoki |
hmmmm: Also, I suggest seeing this: https://forum.minetest.net/viewtopic.php?f=3&t=156&p=139975#p140000 Congrats on what v7 is able to do so far, that looks amazing! |
| 22:54 |
VanessaE |
one node per material, per mapblock maybe |
| 22:54 |
hmmmm |
this is how i can figure out optimal values for perlin noise without starting up minetest 500 times with a new map |
| 22:54 |
Taoki |
It's another mod BTW, not mine |
| 22:54 |
hmmmm |
mapgen v7 as-is sucks |
| 22:55 |
Taoki |
Paradoxally, given you're developing it and care most, I disagree! |
| 22:55 |
Taoki |
v7 is awesome as-is |
| 22:56 |
Taoki |
I know it's incomplete. But still awesome :) |
| 22:56 |
RealBadAngel |
gosh i was almost called a color system ;) |
| 22:56 |
RealBadAngel |
and yes, whole mt world is a single node |
| 22:56 |
RealBadAngel |
sometimes two |
| 22:56 |
hmmmm |
and the first iteration was great too |
| 22:56 |
VanessaE |
hmmmm: I mean, in a given mapblock, stone might be assigned to one irrlicht node, dirt with grass to another, etc etc. then do the same thing again for the next mapnode |
| 22:56 |
VanessaE |
fuck me, LAG CITY. |
| 22:56 |
RealBadAngel |
because meshes are restricted in irrlicht for 64k count for a scene node |
| 22:56 |
hmmmm |
in certain parts |
| 22:56 |
Taoki |
Hahaa! RGBA does sound with RBA :D |
| 22:56 |
Megaf |
Hey, Im getting a lot of out of memory errors on my client right now |
| 22:56 |
Taoki |
Ironic since you work in graphics so much too |
| 22:56 |
Megaf |
they are scrolling on its "console" |
| 22:56 |
hmmmm |
i should save these as like, mapgen v7.1, mapgen v7.2, so on |
| 22:56 |
hmmmm |
RealBadAngel, and that's a problem |
| 22:56 |
Megaf |
ERROR[main] WARNING .... out of memory |
| 22:57 |
Megaf |
first time I seethat |
| 22:57 |
|
VanessaE joined #minetest-dev |
| 22:57 |
hmmmm |
alright |
| 22:57 |
RealBadAngel |
jeez, irc lags, i just wrote 2 lines and got like 40 back |
| 22:57 |
sapier |
what did you do megaf? |
| 22:57 |
VanessaE |
ditto |
| 22:57 |
hmmmm |
so while we're playing with scene nodes it'd probably be a good idea to split off translucent things into a separate one |
| 22:57 |
Megaf |
http://paste.debian.net/96484/ |
| 22:57 |
VanessaE |
hmmmm: agreed. |
| 22:58 |
Megaf |
sapier: nothing, a friend of mine was having very same issue yesterday |
| 22:58 |
RealBadAngel |
we could start with making groups of nodes |
| 22:58 |
hmmmm |
so, is it plausible or not to have two scene nodes per mapblock (one opaque and one translucent materials)? |
| 22:58 |
sapier |
maybe we shold check this with valgrind again |
| 22:58 |
VanessaE |
hmmmm: plausible, sure. absolutely. |
| 22:58 |
Megaf |
and dont say that Im out of memory, 64 bit and 8 GB of ram here |
| 22:58 |
hmmmm |
why don't we do this... :/ |
| 22:58 |
hmmmm |
christ |
| 22:58 |
hmmmm |
i already have so many other things to work on |
| 22:58 |
Megaf |
sfan5: out of memory on client too, |
| 22:58 |
hmmmm |
i need to just start hurrying up |
| 22:59 |
VanessaE |
hmmmm: because lazy? random? |
| 22:59 |
RealBadAngel |
like water scene node for example |
| 22:59 |
VanessaE |
honestly because no one wants to attack the client |
| 22:59 |
RealBadAngel |
we can use then what irrlicht can offer by default |
| 22:59 |
VanessaE |
I expect this sort of stuff is not trivial |
| 22:59 |
RealBadAngel |
like reflections, refractions and so on |
| 23:00 |
VanessaE |
hmmmm: and like you just said, maybe simply a lack of time |
| 23:00 |
RealBadAngel |
also if you would like to know all the stuff ive coded (i mean parallax, bumpmapping is aviable in irrlicht by default) |
| 23:01 |
RealBadAngel |
just as material type |
| 23:03 |
|
caneves joined #minetest-dev |
| 23:04 |
RealBadAngel |
hmmmm, i will try to make just two groups of nodes for a start |
| 23:04 |
Megaf |
RealBadAngel: Im compiling minetest from you shaders branch |
| 23:04 |
Megaf |
Im not sure what to expect from it |
| 23:05 |
RealBadAngel |
Megaf, definitely you shall expect mesh updates to be WAY faster |
| 23:05 |
Megaf |
RealBadAngel: and what about those new shaders? |
| 23:06 |
RealBadAngel |
they do not add anything new in sense of features |
| 23:06 |
RealBadAngel |
all what was there before is put into ONE universal nodes shader |
| 23:07 |
RealBadAngel |
also the shaders system is now targeted |
| 23:07 |
RealBadAngel |
so nodes have own shader, water another one |
| 23:07 |
RealBadAngel |
each element can have in fact own shader |
| 23:07 |
|
sapier left #minetest-dev |
| 23:07 |
RealBadAngel |
even the air |
| 23:08 |
RealBadAngel |
whatever evelement have tile defined, can have set shader to show that tile |
| 23:09 |
Taoki |
Sounds awesome! |
| 23:10 |
Taoki |
Your works on graphics are quite lifesaving RBA :) Especially since MT sorta suffers from low FPS since it existed... faster mesh updates or shaders should mean a lot |
| 23:10 |
Taoki |
If only... if only... VBO code... grumble |
| 23:10 |
Megaf |
well, I just enabled a lot of waving and shader and smoothing thing on minetest.conf |
| 23:10 |
Megaf |
and bumpmaping and stuff |
| 23:10 |
Megaf |
Im expacting to get 4 FPS :P |
| 23:11 |
RealBadAngel |
megaf, please do download sphax's tp |
| 23:11 |
RealBadAngel |
and then turn generatin maps/bumpmapping |
| 23:11 |
Megaf |
RealBadAngel: FPS is pretty good, but I see all reddish |
| 23:11 |
RealBadAngel |
ah, then youre failed to compile shaders |
| 23:12 |
RealBadAngel |
can i have logs? |
| 23:12 |
|
rsiska joined #minetest-dev |
| 23:12 |
celeron55 |
01:53:48 <+hmmmm> what problems are there with making a mapblock an entire node? |
| 23:12 |
celeron55 |
none, assuming you have infinite time for implementing it |
| 23:12 |
Taoki |
Sounds like a weirder video driver. That's usually what causes shaders to fail for a minority of users |
| 23:12 |
celeron55 |
or, well |
| 23:13 |
VanessaE |
Megaf: what video hardware? |
| 23:13 |
celeron55 |
you have to somehow build into it the occlusion culling and so |
| 23:13 |
RealBadAngel |
celeron55, main issue? makin use of different materials |
| 23:13 |
celeron55 |
but it shouldn't be a big deal |
| 23:13 |
Megaf |
GLSL shader failed to compile |
| 23:13 |
Megaf |
Vertex shader failed to compile with the following errors: |
| 23:13 |
RealBadAngel |
for example water should be separate scene node |
| 23:13 |
Megaf |
VanessaE: Radeon HD 4200 using AMDs driver |
| 23:14 |
Megaf |
I saw this happens on windows machines with nvidia cards too |
| 23:14 |
Taoki |
Megaf: More completely, what video hardware, video driver and version, platform (windows and linux) and version of irrlicht |
| 23:14 |
Megaf |
using sfan5's build |
| 23:14 |
Taoki |
Ah |
| 23:14 |
Taoki |
(also irrlicht versino might matter less actually) |
| 23:14 |
Megaf |
so, it doesnt really matter my conf |
| 23:14 |
Megaf |
latest stable |
| 23:14 |
Megaf |
from irrlight website |
| 23:15 |
Megaf |
Taoki: it happens only with RealBadAngel branch |
| 23:15 |
VanessaE |
Megaf: integrated graphics? |
| 23:15 |
VanessaE |
(and those have shaders?) |
| 23:15 |
Megaf |
VanessaE: yep |
| 23:15 |
|
EvergreenTree joined #minetest-dev |
| 23:15 |
Megaf |
ERROR: 0:75: error(#71) Syntax error incorrect preprocessor directive |
| 23:15 |
Megaf |
ERROR: error(#273) 1 compilation errors. No code generated |
| 23:15 |
Megaf |
those kind of errors |
| 23:15 |
RealBadAngel |
integrated gfx is good to fire up office |
| 23:15 |
RealBadAngel |
not the shaders ;) |
| 23:16 |
Megaf |
RealBadAngel: well, this GPU had been a pretty damn good GPU for me, till I try your build |
| 23:16 |
Megaf |
Im not changing GPU because of you |
| 23:16 |
Taoki |
Megaf: I know. Asking because I've had compilation errors too, on MESA in my case. Sometimes a video driver doesn't like some little things being written one way rather than another. Like one might prefer x = x + 1 and not agree with x += 1 |
| 23:16 |
RealBadAngel |
please consider buying something that is capable of displaying graphics not just cells and rows ;) |
| 23:17 |
Megaf |
RealBadAngel: you should better know that card before talking shit |
| 23:17 |
RealBadAngel |
which is? |
| 23:17 |
Megaf |
RealBadAngel: would you pay me a new GPU? |
| 23:18 |
RealBadAngel |
i can send you my old hd 4660 if you want to |
| 23:18 |
Megaf |
RealBadAngel: it can run normal minetest shaders without any problem |
| 23:18 |
RealBadAngel |
it does 40fps on 1,8ghz dual core system |
| 23:18 |
VanessaE |
Megaf: HD4000-series cards are very old and very slow compared to even my HD6800-series card (and it's not exactly top-of-the-line these days either). Integrated graphics devices are NOT meant for playing 3d games! |
| 23:19 |
Taoki |
RealBadAngel: That does sound like an old card. But like Megat said, not everyone has money to get a new GPU |
| 23:19 |
Taoki |
As much as it sucks... since I like big and fast video cards |
| 23:19 |
Megaf |
RealBadAngel: my does 60 FPS at 1080p |
| 23:19 |
Megaf |
on minetest |
| 23:19 |
VanessaE |
maybe RBA's code has a bug, but your code is ass-old. |
| 23:19 |
RealBadAngel |
with shaders? |
| 23:19 |
Taoki |
but yeah, 4xxx sounds very old indeed >.< |
| 23:19 |
Taoki |
Especially on the official drivers! Since the latest for it are prolly very old too |
| 23:19 |
VanessaE |
60 fps at 1080p...with what view range? |
| 23:19 |
Megaf |
yep, but withou mipmapping and smooth lighting |
| 23:19 |
VanessaE |
and what how much stuff displayed? |
| 23:19 |
Megaf |
VanessaE: 40 |
| 23:19 |
* Taoki |
wonders if it can even support OpenGL 3.0 or 3.1 or 3.2 |
| 23:20 |
VanessaE |
yeah. thought so. try again with a view range of say, 120 |
| 23:20 |
VanessaE |
and complex terrain with lots of structures. |
| 23:20 |
RealBadAngel |
btw, i just realized one thing |
| 23:21 |
VanessaE |
s/but your code/but your card/ |
| 23:21 |
RealBadAngel |
how sfan5 is supposed to have build of my experimental branch? |
| 23:21 |
RealBadAngel |
even i dont have win build for that |
| 23:21 |
Megaf |
what do you think about the AMD R7 240? |
| 23:21 |
Megaf |
RealBadAngel: he didnt] |
| 23:22 |
RealBadAngel |
so what have you tried exactly? |
| 23:22 |
Megaf |
I just said I had the same issue on totally different configuration |
| 23:22 |
RealBadAngel |
can you link me that? |
| 23:22 |
Megaf |
and build |
| 23:23 |
RealBadAngel |
i just want to see client/shaders folder |
| 23:23 |
RealBadAngel |
when theres more than one shader, its out of date |
| 23:26 |
Megaf |
VanessaE: I think I will buy an AMD R7 240 |
| 23:27 |
VanessaE |
Megaf: nice |
| 23:28 |
Megaf |
its new, at least |
| 23:28 |
VanessaE |
no wait |
| 23:28 |
VanessaE |
misread that |
| 23:28 |
VanessaE |
240? forget it |
| 23:28 |
Megaf |
why? |
| 23:28 |
VanessaE |
that's way slow |
| 23:28 |
Megaf |
much better than my hd 4200 |
| 23:29 |
VanessaE |
that card rates a 944 on passmark. my HD 6870 rates 2580. |
| 23:29 |
VanessaE |
get something modern. |
| 23:29 |
VanessaE |
something like an R9 270 would be good |
| 23:29 |
VanessaE |
those are reasonably-priced, around $180 |
| 23:30 |
Taoki |
HD 6870 here. Had it for over 3 years I think, but still very good |
| 23:30 |
Taoki |
Though 1024 Vram can sometimes get too little at this day |
| 23:30 |
VanessaE |
yep, HD6870 is good if used properly. |
| 23:30 |
Megaf |
the R7 240 costs here 143.08 US Dollar |
| 23:30 |
RealBadAngel |
anything integrated is shit, no matter what |
| 23:30 |
Megaf |
equivalent to 143.08 US Dollar |
| 23:30 |
VanessaE |
the R9 270 rates 4238 on passmark |
| 23:30 |
Megaf |
not cheap already |
| 23:30 |
VanessaE |
R7 240 is dog slow |
| 23:30 |
VanessaE |
do NOT buy it |
| 23:31 |
VanessaE |
if you spend money on a video card, buy a good one |
| 23:31 |
* Taoki |
still wonders if I'm the only one running Minetest and other games on the free video drivers tho (Gallium / MESA) |
| 23:31 |
VanessaE |
you will kick yourself later for wasting your money if you do not |
| 23:31 |
RealBadAngel |
first of all, most of integrated ones share memory with cpu |
| 23:31 |
Megaf |
what about a HD5450 ? |
| 23:31 |
RealBadAngel |
which is the main problem |
| 23:31 |
RealBadAngel |
and that can kill even the fastest gpu |
| 23:32 |
Megaf |
I dont really have a lot of options in Brazil... |
| 23:32 |
VanessaE |
HD5450 rates 233. |
| 23:32 |
VanessaE |
screw that. |
| 23:32 |
RealBadAngel |
can you check 4660? |
| 23:32 |
RealBadAngel |
i just wonder ;) |
| 23:32 |
VanessaE |
passmark has no result for that one. |
| 23:32 |
VanessaE |
why are we discussing this in -dev? |
| 23:33 |
RealBadAngel |
out of boredom? ;) |
| 23:33 |
Taoki |
Right. Discussion here tends to deviate to non-dev stuff easily sometimes :P |
| 23:33 |
RealBadAngel |
oh, cmon, we are socializing ourselves :) |
| 23:35 |
RealBadAngel |
thats what differs us from bots ;) |
| 23:36 |
RealBadAngel |
megaf, i will make a proper win build for you tommorow |
| 23:36 |
RealBadAngel |
then you will check how it does |
| 23:36 |
Megaf |
RealBadAngel: Its linux here... |
| 23:36 |
Megaf |
in looking for a new GPU |
| 23:37 |
RealBadAngel |
huh? then why you have mentioned sfan builds? |
| 23:37 |
Megaf |
because I had the same issue on a totally different setup |
| 23:37 |
Megaf |
VanessaE: RealBadAngel: is this any good http://www.newegg.com/Product/Product.aspx?Item=N82E16814202094 ? |
| 23:37 |
RealBadAngel |
ah i see |
| 23:38 |
VanessaE |
--> #minetest |
| 23:38 |
RealBadAngel |
idk if its any good but it has cool red fans :) |
| 23:41 |
Taoki |
RealBadAngel: While the chat here is idle anyway. There's... actually another idea I had in mind. Which could be done with shaders, and I'm thinking it might be really cool |
| 23:42 |
Taoki |
What if some solid blocks, such as dirt and sand, could be subdivided and noise added to them? So they won't look like full blocks any more, but noisy and crushed... to look more random like real dirt is. |
| 23:43 |
RealBadAngel |
hmmm |
| 23:43 |
VanessaE |
Taoki: maybe a texture? |
| 23:43 |
Taoki |
Something like this http://russersminecraft.files.wordpress.com/2011/08/minecraft-brownies.jpg lol |
| 23:43 |
VanessaE |
say you supply a noise texture with several "frames" |
| 23:43 |
RealBadAngel |
but with much higher dpi |
| 23:43 |
VanessaE |
and random frames from it are added randomly to the world |
| 23:43 |
Taoki |
VanessaE: Nah. Parallax wouldn't be enough to get a good enough effect. Such would likely need geometry deformation |
| 23:43 |
Taoki |
Similar to the wavery water, but of course static... and noisier |
| 23:44 |
VanessaE |
nono not for parallax. I mean literally added to the nodes' surfaces. |
| 23:44 |
RealBadAngel |
Taoki, have you tried Tiled texture packs? |
| 23:44 |
Taoki |
Wait, found a better example |
| 23:44 |
Taoki |
http://gamesloveres.com/wp-content/uploads/2014/04/minecraft-grass-block-texture-topminecraft--claying-grass-blocks-----how-to-by-regi-on-cut-out---keep-hkx2tzer.jpg |
| 23:44 |
VanessaE |
ohhhh |
| 23:44 |
Taoki |
Photoshopped actually, but well |
| 23:44 |
VanessaE |
interesting idea but that would be GPU-intensive, no? |
| 23:45 |
RealBadAngel |
ouch, thats ugly |
| 23:45 |
VanessaE |
what about my idea though? |
| 23:46 |
Taoki |
Don't think it would. Depends how many times it would subdivide block meshes |
| 23:46 |
RealBadAngel |
http://i.imgur.com/Da5e49B.png |
| 23:46 |
Taoki |
But I think it could give really awesome results to use this idea in some way |
| 23:46 |
VanessaE |
you provide an image, say default_dirt_with_grass_noise.png and this file has, let's say, 10 vertical frames. It gets added to all dirt with grass topsides randomly, so that no two neighboring nodes look exactly the same |
| 23:46 |
Taoki |
Subdivide the mesh of solid blocks, and add some noise. To make the ground feel truly grass-like and noisy, not so much lego with perfectly flat surfaces |
| 23:46 |
VanessaE |
er, one randomly-chosen frame from it gets added, that is |
| 23:46 |
Taoki |
RealBadAngel: Nice! |
| 23:47 |
RealBadAngel |
this feature is already in |
| 23:48 |
RealBadAngel |
https://forum.minetest.net/download/file.php?id=264 |
| 23:48 |
RealBadAngel |
just try those texture packs with shaders enabled |
| 23:49 |
RealBadAngel |
and the bumpmapping ofc |
