| Time |
Nick |
Message |
| 01:43 |
|
VanessaE joined #minetest-dev |
| 02:31 |
|
VanessaE joined #minetest-dev |
| 02:36 |
|
VanessaE joined #minetest-dev |
| 02:39 |
|
VanessaE joined #minetest-dev |
| 02:46 |
hmmmm |
you know i was just thinking, it makes more sense to have the flag be BLOCK_EMERGE_FLAG_GENERATE instead of not-generate |
| 02:47 |
hmmmm |
you can update the flags by ORing them together.. if peer #1 OR peer #2 OR ... OR peer #n wants it as BLOCK_EMERGE_FLAG_GENERATE, it'll be set |
| 03:15 |
hmmmm |
I think I should try making conlock smaller... it nearly _always_ takes ~35ms to acquire which is ridiculous |
| 03:21 |
|
VanessaE joined #minetest-dev |
| 03:44 |
hmmmm |
phooooooooooooey https://github.com/kwolekr/minetest/commit/f7f3c5f4b9ef79b53f511b067d0b06bd003a0358 |
| 03:45 |
hmmmm |
at this point in time, it won't be usable for multiplayer, and it definitely seems slower in general (but it's not) since it doesn't drop any blocks it was asked to generate (at this point) |
| 03:51 |
VanessaE |
nice |
| 04:11 |
|
BackupCoder joined #minetest-dev |
| 04:20 |
|
ShadowNinja joined #minetest-dev |
| 04:47 |
|
DpEpsilon joined #minetest-dev |
| 05:10 |
hmmmm |
you know.. a lot of the data structures used in minetest aren't as appropriate as they could be |
| 05:11 |
VanessaE |
uh oh |
| 05:11 |
VanessaE |
<hmmmm> Screw it. I'm rewriting the whole engine in Ruby. |
| 05:11 |
VanessaE |
:-) |
| 05:11 |
hmmmm |
e.g. a "queue" is actually a list with work you duplicated to "push" and "pop" things, std::map is used absolutely everywhere when it's really not necessary at all |
| 05:11 |
hmmmm |
Ruby, great choice, good luck too |
| 05:11 |
VanessaE |
sorry, just trying to be funny |
| 05:12 |
hmmmm |
so I encourage all core developers to PLEASE USE UNORDERED_MAP WHEN YOU DON'T NEED THE MAP TO BE ORDERED, PLEASE USE SET AND UNORDERED_SET WHEN YOU DON'T ACTUALLY NEED TO ASSOCIATE ANOTHER VALUE, AND PLEASE, FFS, STOP USING IRRLICHT CONTAINERS |
| 05:12 |
hmmmm |
holy shit |
| 05:13 |
VanessaE |
heh |
| 05:35 |
|
ffoxin joined #minetest-dev |
| 07:45 |
|
ffoxin_work joined #minetest-dev |
| 09:07 |
|
ffoxin_work joined #minetest-dev |
| 09:16 |
|
ffoxin_work_ joined #minetest-dev |
| 09:25 |
|
SpeedProg joined #minetest-dev |
| 09:30 |
|
darkrose joined #minetest-dev |
| 10:16 |
|
proller joined #minetest-dev |
| 11:25 |
Taoki |
Looking into learning Inskape a bit. This made me curious: Is there any chance that support for svg textures could be added to MineTest? Considering the blocky and simple design, this sounds like something that would fit and be useful |
| 12:49 |
celeron55 |
it's not MineTest |
| 12:49 |
celeron55 |
it's Minetest |
| 12:49 |
Kray |
CamelCase |
| 12:49 |
celeron55 |
MineTest is just god-awful |
| 12:49 |
celeron55 |
GodAwful |
| 12:49 |
celeron55 |
8D |
| 13:10 |
dzho |
you can call it MineTest so that you can use CamelCase in a DesertBiome |
| 13:11 |
darkrose |
MINETEST, because uppercase is more awesome... and also what menulogo.png uses |
| 13:17 |
Kray |
iTest |
| 13:17 |
Kray |
iMine |
| 13:27 |
* celeron55 |
grins in amusement to hmmm's latest rant |
| 13:31 |
celeron55 |
how stable is the new movement code? |
| 13:33 |
celeron55 |
i promised to whack up a release this week and as upstream appears to have mostly followed the feature freeze (either accidentally or on purpose), i'll make that happen |
| 13:39 |
darkrose |
there was a feature freeze? |
| 13:39 |
celeron55 |
this could be a suitable time to roll out the 0.3 security fix too, assuming someone is interested in doing the little extra communication about it (which is, shout at the debian and ubuntu package maintainers) |
| 13:39 |
celeron55 |
darkrose: yes, mentioned only on this channel |
| 13:41 |
celeron55 |
about a week ago i guess |
| 13:42 |
darkrose |
movement code seems pretty good, works for those who've tried it anyway |
| 13:49 |
celeron55 |
if somebody wants to, they could try doing http://dev.minetest.net/Releasing_Minetest |
| 13:51 |
celeron55 |
or, well, not try, but do it |
| 14:06 |
darkrose |
I can bump the version and tag it, can't do a windows build, or your blog post though |
| 14:09 |
|
hmmmm joined #minetest-dev |
| 14:16 |
celeron55 |
i won't probably do a windows build either; too lazy to cobble up curl and possibly gettext for it |
| 14:20 |
darkrose |
hmmmm: able to do a windows build at all? |
| 14:20 |
hmmmm |
no |
| 14:20 |
hmmmm |
i stopped doing windows builds a while ago |
| 14:20 |
hmmmm |
but the good news is that i installed mac os x 10.8 the other day, i intend to get a build and testing environment on that perhaps |
| 14:21 |
celeron55 |
= too lazy, never |
| 14:21 |
celeron55 |
8) |
| 14:22 |
darkrose |
I don't have access to windows |
| 14:23 |
darkrose |
should I push the tags, and then we just have to volounteer someone for the windows build? |
| 14:24 |
hmmmm |
thexyz and plizadam are able to do windows builds |
| 14:24 |
hmmmm |
i'd rather wait for them to come on |
| 14:25 |
celeron55 |
maybe i'll just post that we're too lazy to do releases, download your stuff from these forum threads when you feel like it" |
| 14:25 |
celeron55 |
-" |
| 14:25 |
celeron55 |
8D |
| 14:26 |
celeron55 |
i guess linux distros like releases |
| 14:26 |
celeron55 |
and server operators |
| 14:30 |
darkrose |
it's not laziness, it's a "rolling release" |
| 14:32 |
celeron55 |
we'd still need to set a different version number when the protocol has an incompatible change, but it wouldn't really require anything else |
| 14:33 |
celeron55 |
still there should be a more official "builds" page with some kind of a way for multiple people to update it |
| 14:37 |
RealBadAngel |
hi |
| 14:37 |
RealBadAngel |
i made 6 directional + rotations facedir |
| 14:38 |
RealBadAngel |
now the question: enhance old one and keep the facedir name |
| 14:38 |
RealBadAngel |
or introduce new one |
| 14:39 |
RealBadAngel |
please note when im using old one and the name all old code and mods is working |
| 14:41 |
RealBadAngel |
just the definition changes: facedir means the direction top of the node is facing |
| 14:42 |
RealBadAngel |
old facedir is 1/6 cases of rotations over y+ |
| 14:42 |
RealBadAngel |
celeron55, any comments? |
| 15:30 |
|
SpeedProg joined #minetest-dev |
| 15:40 |
|
PilzAdam joined #minetest-dev |
| 15:45 |
celeron55 |
RealBadAngel: that didn't really explain what you have done |
| 15:46 |
RealBadAngel |
old facedir is only rotation over Y+ |
| 15:46 |
RealBadAngel |
top facing Y+ direction, and four rotations |
| 15:46 |
RealBadAngel |
got it? |
| 15:47 |
celeron55 |
what if you want an upside-down node pointing to X- |
| 15:47 |
celeron55 |
8) |
| 15:47 |
RealBadAngel |
ive added y-, x+, x-, z+, z- |
| 15:47 |
VanessaE |
in short, he's added the ability to rotate around all three axes. |
| 15:47 |
RealBadAngel |
facedir is now a value in range 0-23 |
| 15:47 |
RealBadAngel |
divide it by 4 to get direction |
| 15:48 |
celeron55 |
is the code fast? |
| 15:48 |
RealBadAngel |
module is the rotatation |
| 15:48 |
RealBadAngel |
same as yours |
| 15:48 |
RealBadAngel |
only not ifs just switch case |
| 15:48 |
RealBadAngel |
to define textures roatations |
| 15:49 |
RealBadAngel |
*modulo |
| 15:49 |
RealBadAngel |
lemme show you core code |
| 15:49 |
celeron55 |
by the way, are you not able to construct proper sentences on a single line? |
| 15:49 |
celeron55 |
because |
| 15:49 |
celeron55 |
reading this |
| 15:49 |
celeron55 |
is |
| 15:49 |
celeron55 |
fucking painful |
| 15:50 |
celeron55 |
in any case, there are, i think, exactly two things i am concerned about |
| 15:50 |
RealBadAngel |
http://pastebin.com/Kib3HkzT |
| 15:51 |
celeron55 |
1) the original facedir was intended to be used so that it can exist with other data in the parameter, consuming only few bits |
| 15:51 |
celeron55 |
s/the original/originally/ |
| 15:51 |
celeron55 |
2) if it is modified or a new thing is added, it should be made sure it really does do everything that is needed in the future |
| 15:51 |
RealBadAngel |
old 3 bits, new one 6 bits |
| 15:51 |
celeron55 |
because it is data that is saved on disk |
| 15:52 |
RealBadAngel |
code is propably a bit faster than original one |
| 15:53 |
celeron55 |
why aren't you using else ifs |
| 15:54 |
RealBadAngel |
it will speed up the code? |
| 15:54 |
RealBadAngel |
if so, no problemo can add "else" there |
| 15:55 |
RealBadAngel |
but im not quite sure |
| 15:55 |
celeron55 |
ehm... the only reason for it not speeding up the code is if the compiler is perfect at optimizing; altough the string concatenation is thousands of times slower than comparing an integer six times |
| 15:56 |
RealBadAngel |
so i will make switch inside then |
| 15:56 |
celeron55 |
so i don't really care; this is something hmmmm would point out 8) |
| 15:56 |
proller |
i want to use http://jsoncpp.sourceforge.net/ for anounces to new masterserver and building in game server list, is this good idea? |
| 15:57 |
RealBadAngel |
you know its not so important |
| 15:58 |
RealBadAngel |
moreover |
| 15:58 |
RealBadAngel |
if textures wasnt rotated on the very start |
| 15:58 |
RealBadAngel |
to fit just one case |
| 15:58 |
RealBadAngel |
the whole code would be much simpler |
| 15:59 |
celeron55 |
did you read what i said about your writing style? |
| 15:59 |
celeron55 |
i'm not even going to answer anything written like that anymore |
| 16:00 |
RealBadAngel |
hold on |
| 16:00 |
RealBadAngel |
what the heck youre refering to now? if-elseif? |
| 16:00 |
RealBadAngel |
i copied in this very routine original style |
| 16:01 |
celeron55 |
lol |
| 16:01 |
celeron55 |
i am talking about how you write on IRC |
| 16:01 |
RealBadAngel |
ah |
| 16:01 |
RealBadAngel |
sorry |
| 16:02 |
RealBadAngel |
please forgive me |
| 16:02 |
celeron55 |
proller: i hope somebody has some idea about this master server thing and answers you, because i don't have |
| 16:02 |
RealBadAngel |
this time i wanted to change it and started to ask for opinions |
| 16:02 |
RealBadAngel |
on what should i do |
| 16:03 |
proller |
celeron55, i want to port jsoncpp into src/json |
| 16:04 |
RealBadAngel |
celeron55, please take on account today is my birthday and give me some more credits :) |
| 16:04 |
celeron55 |
RealBadAngel: anyway, as long as it is perfectly backwards-compatible and performs equally well in the same things that it did before; i don't care |
| 16:05 |
RealBadAngel |
it is |
| 16:05 |
RealBadAngel |
have you checked pastebin? |
| 16:05 |
celeron55 |
but somebody else has to judge if it is suitable for upstream |
| 16:06 |
RealBadAngel |
i defined rotations such way, that old facedir is one of 6 cases of new one |
| 16:06 |
RealBadAngel |
only thing should change is the facedir definition |
| 16:07 |
RealBadAngel |
facedir = direction in which top face of the node is pointing |
| 16:07 |
celeron55 |
proller: i guess you can do that |
| 16:08 |
proller |
ok, i already do |
| 16:08 |
RealBadAngel |
facedir/4 = direction, modulo = rotation over this direction |
| 16:08 |
celeron55 |
proller: are you working with somebody on the thing, or is it completely your own? |
| 16:10 |
proller |
mine, idea here, http://forum.minetest.net/viewtopic.php?id=4780 , will pull when ready |
| 16:11 |
celeron55 |
okay; i guess that was planned somewhaat at some point 8) |
| 16:11 |
celeron55 |
-a |
| 16:13 |
RealBadAngel |
celeron55, i will publish the code tommorow so you will be able to check it out |
| 16:13 |
RealBadAngel |
now im goin off to get drunk. im turnin 40 today |
| 16:13 |
RealBadAngel |
c ya all |
| 16:14 |
darkrose |
using a simple csv format would reduce the data sent and processed, alternatively just use % encoded name/value pairs and the stored data would be exactly the same as the http post entity, either would result in less processing |
| 16:14 |
celeron55 |
lol, i won't check it; but somebody will |
| 16:15 |
RealBadAngel |
no problemo |
| 16:15 |
F00 |
Erk. |
| 16:15 |
F00 |
Mering the physics changes into master seems... premature. |
| 16:15 |
RealBadAngel |
i could make it way faster if i could drop manner of predefined texture rotations |
| 16:15 |
RealBadAngel |
but its not a problem |
| 16:16 |
celeron55 |
darkrose: i don't think using other than json for text formats makes much sense these days |
| 16:16 |
F00 |
At present if you fast move up a ladder, you shoot ten blocks off the top of it. Also, you seem to need dozens of ladders to stop a fall from being lethal from any significant height. |
| 16:16 |
celeron55 |
doing it that way makes it well interoperable with all modern web stuff |
| 16:16 |
F00 |
Deceleration curve is also massively different depending on whether fast move is enabled or not. |
| 16:17 |
celeron55 |
F00: lol, i'd call that "Taoki-grade" stuff |
| 16:18 |
PilzAdam |
F00, it works at least as good as the old system |
| 16:18 |
F00 |
I call shooting ten blocks off of a ladder 'bad'. |
| 16:18 |
RealBadAngel |
call it as you wish, but with her changes game plays amazing |
| 16:18 |
F00 |
The same with not being able to stop yourself with a block or two of ladder/water. |
| 16:19 |
F00 |
It may be unrealistic, but that change breaks pretty much every mine shaft/elevator going |
| 16:19 |
|
iqualfragile joined #minetest-dev |
| 16:19 |
darkrose |
they already broke doors, why not break mineshafts too? |
| 16:19 |
PilzAdam |
F00, ehm, thats not a valid point for me |
| 16:20 |
darkrose |
celeron55: I'm just not a fan of json for anything that isn't javascript |
| 16:20 |
F00 |
Inertia in water and fast move for climbing/swimming is great, shooting off of ladders and not being able to stop a lethal fall easily are not. |
| 16:20 |
F00 |
Also, try the fast move deceleration |
| 16:20 |
F00 |
It's like slipping around wearing teflon shoes on ice |
| 16:20 |
F00 |
Er |
| 16:20 |
F00 |
The non-fast-move decel, I mean. |
| 16:20 |
F00 |
(From fast move speed) |
| 16:20 |
|
Jordach joined #minetest-dev |
| 16:21 |
|
Jordach joined #minetest-dev |
| 16:21 |
celeron55 |
darkrose: the alternative really is XML, and... well, that one is bad |
| 16:21 |
F00 |
It also applies to vertical, it seems. |
| 16:21 |
RealBadAngel |
since when irl youre movin fast, or prevent fallin down to not splash yourelf down there easily? |
| 16:22 |
F00 |
Let's see. Ladder stops at 361 y - fast move shoots me to 372. |
| 16:22 |
VanessaE |
F00: these features are intentional |
| 16:22 |
F00 |
If I don't hold fast move, I slide down the ladder until 351 y |
| 16:22 |
F00 |
If I do hold fast move, I stop decelerating at 359 y. |
| 16:23 |
VanessaE |
except I doubt she meant for you to fly off the top of a ladder, but everything else is intentional |
| 16:23 |
F00 |
Having a massively different decel curve is... aggravating |
| 16:23 |
VanessaE |
trick: hold E after you let go of W. |
| 16:23 |
VanessaE |
you'll come to a halt immediatley. |
| 16:23 |
F00 |
Yeah, but that trick shouldn't be necessary |
| 16:23 |
VanessaE |
immediately* |
| 16:23 |
celeron55 |
lol |
| 16:24 |
celeron55 |
yeah i guess i'll just leave and let you completely screw up minetest |
| 16:24 |
celeron55 |
that's what keeps people happy and gets rid of the userbase so everything can be just shut down |
| 16:24 |
RealBadAngel |
i can see anythin what changes former behaviour is an enemy |
| 16:24 |
F00 |
Hitting a ladder after falling 10 blocks should not propel you down a ladder 10 blocks further. |
| 16:24 |
F00 |
No, I'm not saying that. |
| 16:24 |
F00 |
Fast move in water and for climbing is fantastic. That will please tons of people. |
| 16:25 |
PilzAdam |
the new movement system is realistic |
| 16:25 |
celeron55 |
"realistic" is *never* a thing i want from a game |
| 16:25 |
F00 |
celeron55: Agreed! |
| 16:25 |
celeron55 |
if a game says it's realistic, i'll go a different way |
| 16:25 |
celeron55 |
such is simply not worth playing |
| 16:25 |
F00 |
Personally I enjoy falling 25 meters and not dying |
| 16:25 |
RealBadAngel |
semi-realistic is the right description |
| 16:25 |
celeron55 |
of course, it could be right for minetest - but then minetest is not right for me |
| 16:25 |
F00 |
As well as being able to carry around 3200 m^3 of goods on my person |
| 16:26 |
VanessaE |
celeron55: how do you reconcile that with the fact that you don't play minetest anymore? |
| 16:26 |
PilzAdam |
but realistic means that players really get what they expect when moving |
| 16:26 |
VanessaE |
(the first part, re: "completely screwing up" the game) |
| 16:26 |
celeron55 |
VanessaE: in such a way that i usually try to keep from interfering with you or anyone else |
| 16:26 |
|
dimeshake joined #minetest-dev |
| 16:26 |
celeron55 |
or with minetest |
| 16:27 |
F00 |
I don't think there was anything wrong with the old decel curve. Back before anti-cheat I used to fly around at 160 nodes/s and it decel'd me adequately |
| 16:27 |
F00 |
Felt more like real air resistance than the current non-fast-move does, certainly |
| 16:28 |
|
Calinou joined #minetest-dev |
| 16:28 |
VanessaE |
celeron55: I could cite a few things where that wasn't the case, to be blunt. |
| 16:28 |
RealBadAngel |
F00, all settings are in config |
| 16:28 |
RealBadAngel |
feel free to define your own world |
| 16:28 |
celeron55 |
VanessaE: i am talking about this because there is one thing that still is dependent of me: releases |
| 16:29 |
celeron55 |
VanessaE: and this feature was added while a semi-official feature freeze was going on |
| 16:29 |
F00 |
My feeling is that certain mechanics (like stopping a fall with a ladder/bucket of water) along with sneak elevators are physics bugs that have been adopted by the majority of the player base. |
| 16:29 |
Calinou |
there are some physics bugs, yes |
| 16:29 |
celeron55 |
VanessaE: so i need to determine whether this needs fixing or not |
| 16:29 |
F00 |
And that it's bad to change that without asking the community what it wants. |
| 16:29 |
Calinou |
and the physics tweaks made them more obvious |
| 16:29 |
Calinou |
F00: mark shuttleworth says the same thing |
| 16:29 |
VanessaE |
celeron55: apparently hardly anyone was aware of the feature freeze |
| 16:30 |
celeron55 |
VanessaE: true, but quite irrelevant |
| 16:30 |
F00 |
Since there's no ceiling on falling speed, the current physics actually make it nigh impossible to definitively stop a fall from killing a player |
| 16:30 |
PilzAdam |
F00, actually there are some forum topics about the movement branch of Taoki and a win32 to get more people testing |
| 16:30 |
PilzAdam |
+build |
| 16:31 |
VanessaE |
F00: Taoki did some of her testing via my server in fact. |
| 16:31 |
celeron55 |
VanessaE: i can list people who were talking on the channel at the time: hmmmm, thexyz, PilzAdam, Jordach, Calinou; and hmmmm agreed to it |
| 16:31 |
celeron55 |
VanessaE: so it wasn't exactly anything that hidden |
| 16:32 |
VanessaE |
then it was evidently forgotten. |
| 16:32 |
celeron55 |
not really; darkrose merged the stuff and she wasn't here at that time |
| 16:32 |
PilzAdam |
what about creating a branch for a feature-freeze? |
| 16:33 |
PilzAdam |
so normal development can go on |
| 16:33 |
celeron55 |
no; too much work |
| 16:33 |
Jordach |
celeron55, agreed to what? |
| 16:33 |
celeron55 |
Jordach: nothing significant |
| 16:33 |
Jordach |
alrighty. |
| 16:34 |
celeron55 |
i suggest we just whip up a release at some time |
| 16:34 |
celeron55 |
but maybe not this time |
| 16:34 |
celeron55 |
sometime when thexyz is around, maybe |
| 16:35 |
PilzAdam |
I would wait until the ttf patch gets merged, its really necessary to get many languages working |
| 16:35 |
RealBadAngel |
i thought we are waitin for some more hmmm's work to be done |
| 16:35 |
RealBadAngel |
like emerge thread |
| 16:35 |
celeron55 |
RealBadAngel: that's not how you do releases |
| 16:35 |
celeron55 |
the best time to do a release is just before adding considerable things, not after it |
| 16:35 |
* F00 |
did another test. |
| 16:36 |
PilzAdam |
the only thing that is missing for ttf is a good font |
| 16:36 |
F00 |
Fell 39 blocks onto ladder. Slid 36 blocks down the ladder. |
| 16:36 |
RealBadAngel |
propably youre right, sorry |
| 16:36 |
VanessaE |
PilzAdam: What's wrong with Liberation Sans? |
| 16:36 |
PilzAdam |
its too big |
| 16:36 |
VanessaE |
too big? |
| 16:36 |
Calinou |
no need for a different font to make it smaller |
| 16:36 |
* Calinou |
suggests: font size configurability |
| 16:36 |
celeron55 |
RealBadAngel: it's boring for the developers, but it keeps everyone happy and non-stressed, and the releases are actually stable then |
| 16:36 |
PilzAdam |
its a monospaced font |
| 16:36 |
VanessaE |
um, no it isn't. |
| 16:37 |
thexyz |
I am. |
| 16:37 |
VanessaE |
that's Liberation Sans Mono |
| 16:37 |
VanessaE |
there are two. |
| 16:37 |
RealBadAngel |
celeron55, im not goin to hurry with facedirs neither, its way to big change for me alone to decide |
| 16:37 |
RealBadAngel |
*too |
| 16:38 |
celeron55 |
thexyz: 1) ttf?, 2) is there a problem in making a release exactly now?, 3) will you build an official windows release build if we release now? |
| 16:38 |
VanessaE |
er, 'scuse me, I guess they call them Liberation Sans and Liberation Mono. either way. |
| 16:38 |
thexyz |
VanessaE: my truetype branch uses Liberation Sans Mono by default |
| 16:38 |
thexyz |
or, yes, Liberation Mono |
| 16:39 |
RealBadAngel |
some of the developers even suggested drop the name "facedir". i defended it for sake of backwards compability |
| 16:39 |
thexyz |
celeron55: 1) why not; 2) ttf? 3) sure |
| 16:40 |
celeron55 |
thexyz: is ttf tested well on all OSes by now? |
| 16:41 |
F00 |
Oh god. |
| 16:41 |
F00 |
The deceleration isn't a curve |
| 16:41 |
F00 |
It's linear. |
| 16:41 |
F00 |
I just fell 81 blocks and took 78 blocks to slide to a stop. |
| 16:41 |
celeron55 |
it was linear before too |
| 16:41 |
F00 |
Really? |
| 16:41 |
VanessaE |
F00: so fix it? |
| 16:41 |
F00 |
It must have been very steep, then. |
| 16:41 |
thexyz |
celeron55: linux, windows, wine |
| 16:41 |
thexyz |
don't have mac to test on |
| 16:41 |
thexyz |
same for *bsd |
| 16:41 |
PilzAdam |
thexyz, https://github.com/PilzAdam/minetest/commit/ttf |
| 16:42 |
F00 |
I don't really have any opinions on fixing it, aside from 'sliding as far down a ladder as you've fallen is awful' |
| 16:42 |
F00 |
And since that seems to go against consensus, I'm just going to revert it locally |
| 16:42 |
celeron55 |
thexyz: i think it isn't really wise to put it now in upstream; wiser would be to release, and then put it, so it can mature for a good while |
| 16:42 |
F00 |
*as far |
| 16:43 |
thexyz |
celeron55: probably |
| 16:43 |
celeron55 |
thexyz: if you think otherwise, then you're free to do otherwise |
| 16:43 |
VanessaE |
it seems to me there are a few bugs that can at least be trivially fixed before 0.4.5 goes out the door (little stuff like disabling texture atlas by default) |
| 16:43 |
PilzAdam |
I actually would agree to put it upstream before realesing |
| 16:43 |
RealBadAngel |
me too |
| 16:44 |
PilzAdam |
since we have Chinese and other languages translated already, and they will simply not work |
| 16:44 |
celeron55 |
then we will move the feature freeze to a new point of time, and i request everyone to be damn sure to have everything in that they intend to |
| 16:44 |
celeron55 |
VanessaE's suggestions should NOT be done at the point of release, but many days before it |
| 16:45 |
celeron55 |
it's completely unproductive to note those at the time of release |
| 16:45 |
VanessaE |
celeron55: this was suggested at least a couple of weeks ago, just not officially. |
| 16:45 |
VanessaE |
and not by me initially, either |
| 16:45 |
RealBadAngel |
this is long known bug |
| 16:45 |
Calinou |
is there any reason not to merge this: https://github.com/celeron55/minetest/pull/409 (sorry for pointing at this 5 times or so but I'd like this to be merged in.) |
| 16:46 |
VanessaE |
I've lost count of the number of people I've told to disable their texture atlas which fixed their problem |
| 16:46 |
RealBadAngel |
nobody shall demand fixing it in any point of time, that just should be done asap |
| 16:46 |
VanessaE |
(and yes, I know my grammar sucked in that sentence) |
| 16:46 |
RealBadAngel |
no matter project timelines |
| 16:47 |
celeron55 |
so on what day are we going to have a feature freeze for real, and when after that should a release be made? |
| 16:47 |
RealBadAngel |
timelines are for adding new stuff, evolving. not the bugfixes. those shall be done and merged instantly |
| 16:47 |
celeron55 |
i don't care about when, as long as it is marked down and people actually act accordingly |
| 16:47 |
VanessaE |
ok how about this: |
| 16:48 |
Jordach |
go on. |
| 16:49 |
RealBadAngel |
celeron55, have you noticed logging to the disk slows down game by more than half? |
| 16:49 |
* darkrose |
sends VanessaE a SIGCONT |
| 16:49 |
VanessaE |
start the feature freeze on Monday. Bug freeze in two weeks - whatever is not actively being fixed and pushed by then gets left out. For the two weeks following that, fix bugs that were introduced in the previous two-three weeks during the freeze. |
| 16:49 |
celeron55 |
RealBadAngel: why are you asking me? |
| 16:49 |
VanessaE |
release 0.4.5 four or five weeks from today |
| 16:49 |
celeron55 |
RealBadAngel: it's not like i play it |
| 16:49 |
VanessaE |
then continue the usual development process after that |
| 16:49 |
celeron55 |
VanessaE: eh weeks? it should be 0.4.6 by then |
| 16:49 |
RealBadAngel |
because its so simple and stupid stuff not released yet that you wouldnt believe |
| 16:50 |
RealBadAngel |
hmms got it ready |
| 16:50 |
VanessaE |
celeron55: as a core dev, you of all people know how long it can take to hunt down and fix some random obscure bug. |
| 16:50 |
VanessaE |
so then call it 0.4.6, I don't really care *what* you call it. |
| 16:50 |
RealBadAngel |
folks have been reporting this issue |
| 16:50 |
VanessaE |
0.4.5 was just a suggestion, but releasing 0.4.5 NOW is not appropriate. |
| 16:51 |
celeron55 |
RealBadAngel: yes, and? don't waste your time telling it to me; tell it to somebody who actually can verify it, and merge it |
| 16:51 |
RealBadAngel |
hmmm is having this rdy |
| 16:52 |
celeron55 |
VanessaE: 0.4.5 is the next one; if it isn't now, then that doesn't change; but my opinion is that 0.4.5 should've gone out a month ago :P |
| 16:52 |
Jordach |
i actually agree with celeron55 |
| 16:52 |
VanessaE |
celeron55: well maybe if you put it *that* way sure |
| 16:52 |
RealBadAngel |
https://github.com/kwolekr/minetest/commit/3ff37437463cbfb6d5413d3a23df6f9b8532638c |
| 16:52 |
Jordach |
and we *dont* want to fragment the userbase from stable -> non-stable |
| 16:52 |
VanessaE |
I'm just saying a version bump now might be the wrong time - i.e. it's too late |
| 16:52 |
VanessaE |
(or too early depending on your point of view) |
| 16:53 |
RealBadAngel |
thx to it i do have 50fps instead of 25 |
| 16:53 |
RealBadAngel |
ofc i knew how to disable it before |
| 16:53 |
RealBadAngel |
point is that players didnt |
| 16:54 |
hmmmm |
is the feature freeze in effect _right now_? |
| 16:54 |
celeron55 |
but really; please somebody have enough of producer mentality to do this stuff without me shouting to everyone |
| 16:54 |
celeron55 |
hmmmm: it was removed because apparently it isn't appropriate... |
| 16:55 |
hmmmm |
well i didn't yet add in the debug log level setting |
| 16:55 |
RealBadAngel |
im using it already ;) |
| 16:55 |
celeron55 |
everyone gets only a fraction of the value out of their work if no releases are made |
| 16:55 |
celeron55 |
so work towards that happening |
| 16:55 |
hmmmm |
wouldn't you like to merge that in? |
| 16:55 |
RealBadAngel |
absolutely |
| 16:55 |
darkrose |
release now, if it's bad 0.4.6 can go out at the end of the month |
| 16:56 |
celeron55 |
0.4.x+1 releases can be even once per half an hour if needed |
| 16:56 |
hmmmm |
i'm sure it'd make a lot of people happy to not require a shortcut to run minetest without extra-verbose debugging logs |
| 16:56 |
RealBadAngel |
lets make then it in cycles |
| 16:56 |
* Taoki |
reads up part of the chat. Wonders why people are being upset about |
| 16:57 |
RealBadAngel |
weeks 1-3 code, 4th freeze, start of next release |
| 16:57 |
Taoki |
I'm very ha[ppy with the fact that new features and pull requests are being merged quickly now. MineTest advanced a lot recently thanks to that :) |
| 16:58 |
hmmmm |
so what's the final verdict? |
| 16:58 |
RealBadAngel |
and in week of freeze only bugfixes allowed |
| 16:59 |
RealBadAngel |
no new code |
| 16:59 |
celeron55 |
hmmmm: if you ask me, i still think a release should be made right now so people can start merging bigger things without fearing everything going bad |
| 16:59 |
darkrose |
better to release now and have a recent stable, then to delay a few weeks and have players wonder which version of 0.4.4 a server is using |
| 16:59 |
Taoki |
I agree with new releases being made frequently, so everyone can get the latest changes |
| 16:59 |
RealBadAngel |
celeron55, but the question will back in some time again |
| 16:59 |
celeron55 |
the changelog from 0.4.4 already is like multiple pages long; way longer than 0.4.3->0.4.4 |
| 16:59 |
hmmmm |
sure, that's fine by me. i really couldn't care either way, the people who wanted this feature the most already have access to it |
| 16:59 |
RealBadAngel |
lets make schedules |
| 17:00 |
hmmmm |
i sound bad in the changelogs because my changes aren't visible |
| 17:00 |
Taoki |
celeron55: Not sure if you're upset about my new physics. But I recomment trying them out, I'm sure anyone has to like them. It's a lot better to dive in water as you fall, rather than simply having your velocity disappear. Also, air acceleration wasn't removed... just slowed down so it doesn't feel as ugly and like you can walk on air. |
| 17:01 |
Taoki |
I think anyone can agree it feels a lot nicer now. IMO MineTest could never be complete without improvements like this |
| 17:01 |
RealBadAngel |
I agree, those changes have to be "felt" and played |
| 17:01 |
hmmmm |
are the physics like minecraft? |
| 17:01 |
RealBadAngel |
no |
| 17:01 |
Taoki |
hmmmm: Possibly. I didn't compare them up close |
| 17:01 |
hmmmm |
yeah, that would break some things... mob traps in particular |
| 17:01 |
Taoki |
Sinking in water is in MineCraft too. But I didn't put it in MineTest because of that, but because it's normal :P |
| 17:02 |
RealBadAngel |
hmmm, set the gravity to low |
| 17:02 |
Taoki |
normal != realistic necessarily |
| 17:02 |
hmmmm |
i really hope there's an option to disable this |
| 17:02 |
RealBadAngel |
and jump on the top of big sequoia, FUN! |
| 17:02 |
Taoki |
hmmmm: All new physics can be configured in minetest.conf. If anyone would want the old behavior they can do that with the current code |
| 17:02 |
RealBadAngel |
set the gravity just |
| 17:02 |
RealBadAngel |
default is 9,81 |
| 17:02 |
celeron55 |
ehm... so do the changes change the gravity to such as in MC? |
| 17:02 |
RealBadAngel |
is this familiar value? :) |
| 17:02 |
celeron55 |
well that is just horrible |
| 17:03 |
Taoki |
eg: You can set movement_water_fluidity_smooth to a high number, which should disable diving in water and work similar to before. But why would anyone want that :P |
| 17:03 |
hmmmm |
again, for resetting falls |
| 17:03 |
Taoki |
celeron55: Default gravity hasn't been changed. Only change to gravity is that it can be configured in minetest.conf |
| 17:04 |
Taoki |
And no, I did not open MineCraft in one window and MineTest in the other, and copy everything I saw in MC. The physics changes were done without any MineCraft reference. If MC didn't exist and only MT did, I would have done them the exact same way |
| 17:04 |
RealBadAngel |
but when it could be changed, imagine on far y+ outer space |
| 17:04 |
hmmmm |
nobody cares if they're the same as minecraft or not |
| 17:04 |
RealBadAngel |
or moon biomes |
| 17:04 |
hmmmm |
it's just be nice to know if it is or isn't |
| 17:04 |
Taoki |
Personally I don't. But it is a legit concern that MT might become too much like MC. Not my aim though |
| 17:05 |
RealBadAngel |
on the opposite heavy gravity close to the core |
| 17:05 |
hmmmm |
that's not a legitimate concern |
| 17:05 |
F00 |
I rather suspect that falling x blocks in Minecraft doesn't make you slide max(0, x-3) down a ladder. |
| 17:05 |
hmmmm |
i hate it when people complain about that |
| 17:05 |
Taoki |
HOWEVER: If a good feature exists in MineCraft (and other projects), I'm not going to say "never do this in MineTest cuz MineCraft has it too, so we'd be copying it" |
| 17:05 |
hmmmm |
minetest is an open source minecraft clone |
| 17:05 |
hmmmm |
people reasonably expect the same game mechanics |
| 17:05 |
RealBadAngel |
and yet mc copies mt ideas |
| 17:05 |
Taoki |
One example are mined items dropping to the ground instead of going to inventory: MC has that, but I'd like it in MT simply because it feels more right |
| 17:06 |
RealBadAngel |
they dropped texture atlas |
| 17:06 |
RealBadAngel |
splitted server and client |
| 17:06 |
Taoki |
If Mc didn't have mined items falling to ground, I'd still want that in MT never the less |
| 17:06 |
RealBadAngel |
me too |
| 17:07 |
RealBadAngel |
because of technic tools, explosives and mining machines |
| 17:07 |
hmmmm |
well now that torches fall when they're not attached to a block, it only feels right if you can pick them up by moving close to it |
| 17:07 |
hmmmm |
instead of having to point and click |
| 17:07 |
Taoki |
celeron55: But really, try out my physics changes. Nothing there copies MineCraft that I'm aware of, nor changes functionality in any bad way. Also, the new settings would allow functionality almost exactly like the physics before it, if someone wants the old ones on their server |
| 17:07 |
hmmmm |
i feel like at least that much should be added |
| 17:07 |
Taoki |
Except for Y smoothing which always happens now, and vertical acceleration works the same as horizontal one |
| 17:09 |
celeron55 |
Taoki: by the way, try going up the full length of ladders fast in an open space |
| 17:09 |
celeron55 |
upwards |
| 17:09 |
celeron55 |
8) |
| 17:09 |
celeron55 |
then try to not get hurt |
| 17:09 |
F00 |
I think the speed peaks in only a few blocks. |
| 17:09 |
Taoki |
Yes, fast climbing is possible now. But Fast itself is an exception (eg: A special permission for some players). So it's as realistic as running fast on the ground :P |
| 17:09 |
Taoki |
Yeah, falling damage is the same. If Fast isn't used properly you can get hurt |
| 17:10 |
F00 |
Many servers provide fast by default, if not fly (that may change with noclip) |
| 17:10 |
Taoki |
But it's just an optional new possibility. Previously you couldn't use fast on ladders or underwater |
| 17:10 |
F00 |
It's not really feasible to move around a 62000^3 world at 4 nodes/s |
| 17:10 |
Taoki |
One nice thing about Fast in water: You can now jump up like a dolphin, or ride on the water's surface like a jet boat. It's very fun ^_^ |
| 17:12 |
Taoki |
Which BTW, is not something you can do in MineCraft. |
| 17:14 |
Taoki |
It's not perfect of course, no code is. eg: Like you said, using Fast on ladders can get you hurt due to fall damage. Then again, using Fast can easily get you hurt as well, since you have less control and can fly off a cliff before you know it (happens often to me) |
| 17:14 |
RealBadAngel |
btw, what about fly and swim animations? |
| 17:14 |
Taoki |
RealBadAngel: Can be done and I should look into it |
| 17:15 |
Taoki |
BTW, something else that might be useful which my physics does: Try using Fast and Sneak at the same time. You will get the Fast velocity and acceleration, but can't fall of nodes. That's a neat little trick (I noticed it but considered it's great) |
| 17:15 |
VanessaE |
*scrolls back* |
| 17:15 |
Taoki |
If you need to go fast without the fear of falling |
| 17:15 |
RealBadAngel |
and one thing i noticed lately, step sounds when flyin or swimming shall not be played |
| 17:15 |
Taoki |
Yeah, can be fixed separately |
| 17:16 |
|
Jeija joined #minetest-dev |
| 17:16 |
RealBadAngel |
hi Jeija |
| 17:17 |
Jeija |
Hi RealBadAngel, hi everyone |
| 17:17 |
Taoki |
hi |
| 17:17 |
RealBadAngel |
Jeija, im ready with nodes, now defining nodeboxes roatations |
| 17:18 |
RealBadAngel |
so definitely will be ready very soon |
| 17:18 |
Jeija |
Sounds good, have you published it yet? |
| 17:18 |
Taoki |
RealBadAngel: Sounds awesome, can't wait |
| 17:19 |
RealBadAngel |
as i promised, i will publish it today with new branch |
| 17:19 |
RealBadAngel |
i got nodes well tested already |
| 17:19 |
PilzAdam |
thexyz, I have added a LiberationSans.ttf and added the ability to make the size configureable |
| 17:19 |
PilzAdam |
https://github.com/PilzAdam/minetest/commit/4a97c646fdff8a0eb4cff28c50a9164000959788 |
| 17:20 |
RealBadAngel |
but one thing shall be added to minetest game |
| 17:20 |
PilzAdam |
but it seems like the font breaks the "TAPE AND GLUE" and "SERVER" strings at the side of the main gui |
| 17:20 |
RealBadAngel |
screwdriver shall become default tool |
| 17:20 |
Jeija |
I just pushed my curl for lua branch: http://bit.ly/X7CU9W ; Allows you to download stuff from within lua |
| 17:21 |
RealBadAngel |
for those who dont know it yet, a tool to rotate nodes that use facedir |
| 17:22 |
Taoki |
BTW: What does everyone think about adding svg texture support (if that would be possible in Irrlicht)? Vector graphics could match the simplistic theme of MT very nicely, for custom games or mods |
| 17:22 |
Jordach |
Taoki, no, just no |
| 17:22 |
Taoki |
Why not though? |
| 17:22 |
Jordach |
because .svg can contain viruses |
| 17:23 |
Jordach |
(and has been known to) |
| 17:23 |
Taoki |
lol |
| 17:23 |
VanessaE |
if you have a problem with viruses in svg, I blame your SVG library |
| 17:23 |
Taoki |
That is kind of a silly reason. Any file can contain viruses |
| 17:23 |
RealBadAngel |
lua mods too ;) |
| 17:23 |
Jeija |
SVG support sounds good to me if it is not too much work and too much code. |
| 17:24 |
thexyz |
hm.. PilzAdam, any idea why? |
| 17:24 |
RealBadAngel |
even well written LOGO treegen axiom could hang your PC ;) |
| 17:24 |
PilzAdam |
thexyz, I tried different text aligns but it doesnt help, only the first char is printed |
| 17:29 |
VanessaE |
I'm sure Gambit would love SVG support. Seems his minetoon texture pack would be well-suited to that format |
| 17:31 |
|
Taoki_1 joined #minetest-dev |
| 17:34 |
PilzAdam |
thexyz, it works with wordWrap enabled |
| 17:35 |
PilzAdam |
this works now: https://github.com/PilzAdam/minetest/commit/926b3caa639fd331cd7c1717aec7fca21dfd8d42 |
| 17:38 |
|
pierreghz joined #minetest-dev |
| 17:40 |
thexyz |
PilzAdam: great |
| 17:40 |
thexyz |
can you merge it (with all your fixes)? |
| 17:40 |
PilzAdam |
sure |
| 17:47 |
PilzAdam |
done |
| 17:50 |
thexyz |
great |
| 17:52 |
|
Exio joined #minetest-dev |
| 17:52 |
|
proller joined #minetest-dev |
| 18:07 |
PilzAdam |
well, fuck |
| 18:07 |
PilzAdam |
win32 build is broken |
| 18:08 |
PilzAdam |
the unittest fail (they also did before, but now other unittest fail) and with --disable-unittests builtin throws some errors |
| 18:10 |
thexyz |
are you sure you're building "release" version? |
| 18:11 |
PilzAdam |
oh..... |
| 18:11 |
PilzAdam |
I had this setting set for some testing.... |
| 18:12 |
thexyz |
well, that should be obvious, "release" doesn't run any unit tests by default |
| 18:12 |
thexyz |
anyway, care to pastebin that error? |
| 18:13 |
PilzAdam |
https://gist.github.com/PilzAdam/4954879 |
| 18:14 |
hmmmm |
i should put some time into finding out why that unit test fails and probably fix it someday |
| 18:15 |
thexyz |
odd |
| 18:15 |
hmmmm |
so can anybody explain http://pastebin.com/Kib3HkzT to me? |
| 18:19 |
celeron55 |
it is an extension of this https://github.com/celeron55/minetest/blob/master/src/mapblock_mesh.cpp#L587 |
| 18:20 |
celeron55 |
it's a relatively horrible hack from the ground up |
| 18:20 |
hmmmm |
that code is quite a bit more reasonable though |
| 18:20 |
hmmmm |
i don't mean to be elitist, but minetest code should meet or exceed commercial quality |
| 18:20 |
celeron55 |
the original content of that function ends at "If not rotated or is side tile, we're done" |
| 18:21 |
celeron55 |
then i once hackily extended it with that texture modifier crap to rotate the top and bottom of a node when the node is rotated with facedir |
| 18:22 |
celeron55 |
it should really be done at the vertex level in... somewhere else |
| 18:22 |
celeron55 |
and RBA's stuff implements an extended facedir with the same thing (altough he replaces the few tricks in that with even more texture modifiers) |
| 18:24 |
|
ffoxin joined #minetest-dev |
| 18:24 |
celeron55 |
i am not sure if i need to say this, but the texture modifiers in that is... well, it's not really a thing to allow anyone to extend in any way, now that i think of it |
| 18:24 |
|
ShadowNinja joined #minetest-dev |
| 18:25 |
hmmmm |
so what do the (0,0,0) directions do exactly? |
| 18:25 |
hmmmm |
is that just nothing? |
| 18:26 |
celeron55 |
it's kahrl's code |
| 18:26 |
celeron55 |
i'm not... exactly sure |
| 18:26 |
hmmmm |
his code is kind of... interesting to say the least. |
| 18:27 |
celeron55 |
i think it like hashes the integer unit direction |
| 18:27 |
celeron55 |
for the table lookup, like it says |
| 18:27 |
celeron55 |
so it's the kind of performance stuff you should actually like |
| 18:27 |
hmmmm |
i just wonder if it could be done a better way |
| 18:28 |
hmmmm |
getTileInfo() sucks too |
| 18:28 |
hmmmm |
if you need over 5 arguments, that's a hint you should pass a structure instead |
| 18:30 |
celeron55 |
the whole cubic mesh generation code should probably be redone |
| 18:30 |
celeron55 |
it's really aged |
| 18:30 |
celeron55 |
again one of the few things that existed in the first minetest, and just extended from there |
| 18:32 |
hmmmm |
what's transformR90 and transformFXR90? |
| 18:33 |
celeron55 |
a texture modifier |
| 18:33 |
hmmmm |
what's the difference between* |
| 18:33 |
|
proller_ joined #minetest-dev |
| 18:33 |
hmmmm |
R obviously rotates, but not sure about FXR |
| 18:33 |
celeron55 |
well that is a good question |
| 18:33 |
PilzAdam |
FX = flip X |
| 18:33 |
hmmmm |
i see |
| 18:33 |
celeron55 |
https://github.com/celeron55/minetest/blob/master/src/tile.cpp#L1485 |
| 18:33 |
PilzAdam |
https://github.com/celeron55/minetest/blob/master/src/tile.cpp#L1495 |
| 18:34 |
hmmmm |
i figured someone knew off the top of their heads.. |
| 18:34 |
celeron55 |
it's all workarounds for inflexible vertex generation code |
| 18:34 |
celeron55 |
(in that usage) |
| 18:34 |
hmmmm |
i'm all for making minetest better, but this latest facedir hack is a bit too much to chew |
| 18:35 |
|
grondilu joined #minetest-dev |
| 18:35 |
PilzAdam |
RBA will rage if he hears that... |
| 18:35 |
celeron55 |
it's really a thing makeFastFace should do |
| 18:35 |
celeron55 |
but it doesn't have access to facedir as it is and so on |
| 18:35 |
celeron55 |
it wasn't designed for this kind of stuff 8) |
| 18:36 |
celeron55 |
(that is, rotatable nodes) |
| 18:37 |
grondilu |
Hello, after compiling the latest version on github, I crashed when trying to play on either single or multiplayer. I tried creating a brand new world, and now evrything works fine (including connecting to a server or to my local world that I couldn't play before). |
| 18:37 |
celeron55 |
(hmm... i guess maybe it could be made to work with just adding more parameters and moving stuff around...) |
| 18:37 |
hmmmm |
grondilu, that's a shame.. |
| 18:40 |
hmmmm |
so let me understand this... the direction input is how the node was placed, the facedir is up, down, left, right of the individual tile, and the tileindex is which of the 6 sides of the cube we're talking about |
| 18:41 |
celeron55 |
umm no |
| 18:41 |
celeron55 |
facedir is 0, 1, 2 or 3 |
| 18:41 |
celeron55 |
being the directions you can place eg. a furnace in |
| 18:41 |
hmmmm |
ah |
| 18:42 |
celeron55 |
tileindex is 0...5, being the faces of a cube |
| 18:42 |
hmmmm |
so the idea is to extend facedir to all 6 directions |
| 18:42 |
hmmmm |
this is nuts |
| 18:42 |
celeron55 |
no, the idea is to extend it to like million directions... being something |
| 18:42 |
celeron55 |
+like 20+ |
| 18:43 |
hmmmm |
erm.. w/e |
| 18:43 |
celeron55 |
apparently 24 |
| 18:43 |
hmmmm |
what is the purpose of this even |
| 18:43 |
hmmmm |
are they trying to rotate it in such a way that it's not in a grid or something..? |
| 18:43 |
celeron55 |
i'm not sure what they are, but RBA and VE said it gives full rotation freedom in all axis |
| 18:44 |
celeron55 |
everything that is in a grid i guess |
| 18:44 |
hmmmm |
full rotational freedom of individual NODES... |
| 18:44 |
celeron55 |
i mean, all 90 degree steps |
| 18:44 |
celeron55 |
in YZ, XZ and XY planes |
| 18:44 |
|
doserj joined #minetest-dev |
| 18:45 |
celeron55 |
at least that was the intention i believe |
| 18:45 |
hmmmm |
oh ah i see, 24 came from 6 faces and 4 directions for each tile |
| 18:45 |
celeron55 |
hmm, true |
| 18:45 |
hmmmm |
if it were up to me i would've done this in a completely different manner |
| 18:46 |
PilzAdam |
is this used for paramtype2 = "facedir" or "wallmounted"? |
| 18:46 |
celeron55 |
well first of all, i'm beginning to wonder how many textures get generated with that one when modders make all of their nodes use facedir |
| 18:46 |
|
ffoxin joined #minetest-dev |
| 18:46 |
hmmmm |
probably too many |
| 18:47 |
celeron55 |
the textures are gotten from a mapping from strings to cached textures |
| 18:47 |
celeron55 |
(obviously) |
| 18:47 |
celeron55 |
in a HD texture pack, that is going to kill just about everything |
| 18:48 |
ShadowNinja |
if you use the extended facedirs minetest makes 4 copies of each texture? |
| 18:49 |
ShadowNinja |
if so then that has to be possible another way |
| 18:49 |
celeron55 |
of course there is; just apply the facedir at the vertex generation level |
| 18:50 |
hmmmm |
i vote for: dump this code and apply it at the vertex level |
| 18:50 |
ShadowNinja |
aye |
| 18:50 |
celeron55 |
it requires modifying multiple functions but is the absolute right way |
| 18:51 |
PilzAdam |
doesnt it make even 8 copies? |
| 18:51 |
celeron55 |
PilzAdam: my internal sorting algorithm doesn't work for that long a list, but 8 sounds about right |
| 18:52 |
celeron55 |
of course it does them on-demand only when actually needed though |
| 18:52 |
celeron55 |
so probably closer to 0 than 8 |
| 18:52 |
celeron55 |
but anyway, worst case is pretty bad |
| 18:53 |
* celeron55 |
thinks it will be a good idea to not be around when RBA discovers this decision |
| 18:53 |
celeron55 |
8D |
| 18:53 |
PilzAdam |
yep |
| 18:53 |
hmmmm |
no doubt vanessa is going to complain about how it's so slow with the HD texture pack |
| 18:54 |
hmmmm |
i think this is a problem that'll fix itself |
| 18:55 |
celeron55 |
what does that even mean |
| 18:56 |
celeron55 |
i believe we haven't yet reached singularity, altough the complexity is clearly getting there |
| 18:56 |
celeron55 |
8D |
| 18:57 |
celeron55 |
bad joke, but worth it |
| 18:57 |
hmmmm |
when i say 'fix itself' i mean that part of the problem is the person who wrote it thinking that this was a good idea |
| 18:59 |
hmmmm |
aghk. looks like i was wrong, and QueuedBlockEmerge is going to have to return.. in a form however |
| 19:01 |
hmmmm |
instead of attaching it to the actual queue, i have it attached to the map that's used to check for existence, and instead of the peer ids being a map, i have it be a list, so it shouldn't be as horrible. |
| 19:02 |
doserj |
btw, are there plans to look at (and merge) https://github.com/celeron55/minetest/pull/439 and https://github.com/celeron55/minetest/pull/426? |
| 19:02 |
hmmmm |
shrug. |
| 19:04 |
PilzAdam |
the backface_culling serialization is useful |
| 19:04 |
celeron55 |
oh no no no no |
| 19:05 |
doserj |
? |
| 19:05 |
celeron55 |
don't in the name of god and other scary things change the version of TileDef |
| 19:05 |
celeron55 |
you'll be doomed |
| 19:06 |
|
celeron55 joined #minetest-dev |
| 19:06 |
celeron55 |
21:07:44 < celeron55> oh no no no no |
| 19:06 |
celeron55 |
21:08:20 < celeron55> don't in the name of god and other scary things change the version of TileDef |
| 19:06 |
celeron55 |
21:08:29 < celeron55> you'll be doomed |
| 19:06 |
doserj |
why? |
| 19:07 |
celeron55 |
or... wtf |
| 19:07 |
celeron55 |
if you connect an old client to a new server, what happens? |
| 19:08 |
doserj |
it throws an exception |
| 19:08 |
celeron55 |
it shouldn't |
| 19:08 |
celeron55 |
there is no real reason for it not to work |
| 19:08 |
PilzAdam |
it doesnt |
| 19:08 |
doserj |
how should old code be able to parse a new format? |
| 19:08 |
PilzAdam |
+throw an exception |
| 19:09 |
PilzAdam |
both versions are supported (0 and 1) and backface_culling is only read if version is >= 1 |
| 19:09 |
doserj |
PilzAdam: old clients only parse version 0 |
| 19:10 |
doserj |
and new servers only send version 1 |
| 19:10 |
PilzAdam |
oh, right |
| 19:10 |
|
grondilu left #minetest-dev |
| 19:11 |
celeron55 |
doserj: the versions in classes like that are silly because they are nearly useless in practice, because we want to maintain compatibility to both directions between versions (because that is... well, *very* useful and *very* doable if you know what you are doing) |
| 19:12 |
celeron55 |
the way it can be done is to just add it to the end of the parameters, and ignore the SerializationError (or whatever it is called) when deserializing |
| 19:12 |
celeron55 |
it's hacky but it keeps many people happy |
| 19:13 |
doserj |
that works as long as serialzation is not nested in something else's serialization |
| 19:13 |
celeron55 |
...true; is this such a case? |
| 19:14 |
doserj |
I don't know :) |
| 19:14 |
celeron55 |
aaaah, this one is, but! |
| 19:15 |
celeron55 |
(hell this serialization stuff is inconsistent) |
| 19:15 |
celeron55 |
ContentFeatures::serialize gets the protocol version the client is using as a parameter |
| 19:15 |
celeron55 |
(which should really happen everywhere, but it hasn't been applied to everywhere) |
| 19:16 |
celeron55 |
if you bump the protocol version and make it serialize/deserialize the version of TileDef with the value if protocol_version >= whatever_the_new_version_is, everything will magically work |
| 19:16 |
celeron55 |
because the server and client will negotiate a protocol version that both understand |
| 19:18 |
celeron55 |
just change TileDef::serialize(std::ostream &os) to be TileDef::serialize(std::ostream &os, u16 protocol_version) |
| 19:18 |
celeron55 |
and write the version 1 of TileDef if protocol_version is >= the bumped up one |
| 19:19 |
celeron55 |
i'd be happy if someone turned everything to work like that that doesn't yet do so |
| 19:19 |
celeron55 |
(beware of things that are written on disk though; this applies only to network-only data...) |
| 19:21 |
celeron55 |
(and to complicate things even more (i shouldn't do this), ContentFeatures will probably be changed to be writable on disk sometime in the future, because of things) |
| 19:21 |
doserj |
a change to the disk-format is in the other pull request :) |
| 19:21 |
doserj |
(for liquid updates) |
| 19:22 |
* celeron55 |
flees because of too much stuff -> |
| 19:26 |
celeron55 |
i apologize the bad management here; i know it sucks to have pull requests hanging like that |
| 19:26 |
celeron55 |
but i can't do anything to it without killing myself with too much thinking |
| 19:50 |
doserj |
like this? https://github.com/doserj/minetest/commit/75fe6d3e0c7d81405b833a476ca196e2c24eb8ab |
| 19:57 |
|
Jordach joined #minetest-dev |
| 20:13 |
|
doserj joined #minetest-dev |
| 20:19 |
doserj |
(pushed to https://github.com/celeron55/minetest/pull/439) |
| 20:31 |
|
ffoxin joined #minetest-dev |
| 20:32 |
|
SpeedProg1 joined #minetest-dev |
| 20:59 |
F00 |
celeron55: Any idea as to what you're doing with physics for the next release? |
| 21:00 |
F00 |
I feel like the greatly reduced air control and the inability to safely stop after a long fall are quite antagonist to people who build/use tall structures, particularly on servers without fast move |
| 21:05 |
F00 |
Also, I realized why the current deceleration feels so glacial. |
| 21:05 |
F00 |
The old walk_acceleration was 4.0 * BS, which was then multiplied by BS again. So accel/decel was 40/s (and I actually quite like it - feels responsive) |
| 21:07 |
F00 |
Looks like 52122c34 is responsible. |
| 21:07 |
F00 |
I think 40/s is good, anyhow. Otherwise it would take a full five seconds to get to 20/s fast move speed. |
| 21:24 |
proller_ |
again ready 8-) https://github.com/celeron55/minetest/pull/482 |
| 21:25 |
RealBadAngel |
wb, ive read all the discussion ;) |
| 21:26 |
RealBadAngel |
i agree it should be done better and i think i do know how to do that |
| 21:26 |
RealBadAngel |
same way as nodeboxes are rotated |
| 21:27 |
RealBadAngel |
to make my way (looks like it was kahrl's original idea) i had some hard time |
| 21:28 |
RealBadAngel |
to make this lookup table i was sitting there with wooden cube in hand, number of faces writen on it and rotated it |
| 21:30 |
RealBadAngel |
two pros of this way, its fully compatible with old code and adding all missing facedirs doesnt break any old code, core or mods |
| 21:30 |
RealBadAngel |
2nd is that i learned everything bout minetest facedirs, rotations |
| 21:31 |
RealBadAngel |
facedir applied to the nodeboxes is the way it should be done |
| 21:42 |
|
doserj joined #minetest-dev |
| 21:53 |
hmmmm |
i've been seeing much opposition to the new physics.. haven't tried it myself but if a noticable amount of people hate it, probably should get rid of it |
| 21:55 |
F00 |
I don't hate it as such, but I really dislike certain elements of it |
| 21:55 |
F00 |
I wish the fast move affecting climbing/water had been a separate commit |
| 21:55 |
F00 |
That's universally good in my eyes. |
| 21:58 |
F00 |
The lack of air control makes it quite hard to not fall to your death, sometimes. |
| 21:59 |
F00 |
Particularly if you're working on a sloping roof made out of whole blocks or some such |
| 22:03 |
RealBadAngel |
hmmmm, so try it first |
| 22:03 |
hmmmm |
yeah, that's what i'm doing. i just cloned the current upstream repository. |
| 22:04 |
RealBadAngel |
btw anythin new has always decent ammount of change haters |
| 22:04 |
hmmmm |
yeah but there are valid reasons to not like the new physics... |
| 22:04 |
hmmmm |
not really just a matter of taste |
| 22:04 |
RealBadAngel |
same was with new trees "omg i will have to work harder to chop a tree" |
| 22:04 |
hmmmm |
i don't know about you, but i'd rather err on the side of caution and be conservative |
| 22:05 |
ShadowNinja |
I like fast on water and ladders but shooting 10m in the air is a bit ridiculous |
| 22:05 |
RealBadAngel |
so maybe it shall be fixed |
| 22:07 |
RealBadAngel |
its a complex change to the players physics and can be tweaked |
| 22:07 |
hmmmm |
i feel like a motorboat |
| 22:07 |
F00 |
The decel rate varying basd on whether you have fast move pressed or not is... wonky |
| 22:07 |
ShadowNinja |
lol |
| 22:07 |
RealBadAngel |
its realistic |
| 22:07 |
F00 |
A lot of long mine shafts/towers just put a ladder at important stuff so people can AFK and arrive at the proper spot |
| 22:08 |
F00 |
Yes, but digging a one metre cube of stone out in half a second isn't, so I think realism isn't Minetest's goal |
| 22:08 |
ShadowNinja |
that is why there is a cushion in the jumping mod |
| 22:10 |
hmmmm |
i feel like i'm on the moon |
| 22:10 |
RealBadAngel |
set gravity to 0.2 |
| 22:13 |
RealBadAngel |
F00, so what is the goal? |
| 22:15 |
F00 |
Fun! |
| 22:16 |
hmmmm |
i'm gonna just say right now that i don't like the new physics |
| 22:18 |
F00 |
I'm attached to the older, accidental 40/s decel. |
| 22:18 |
RealBadAngel |
just because you used to it |
| 22:18 |
hmmmm |
OTOH i found a really nice seed |
| 22:19 |
F00 |
Eh, no. |
| 22:19 |
F00 |
With the new physics, if you fast move and hit the ground without holding it still, it's like you're on a slip-n-slide |
| 22:19 |
RealBadAngel |
i agree that ladders issue shall be tweaked, but sayin because of it whole idea is wrong is |
| 22:20 |
F00 |
I'm content with gravity-affected movement and free move having the only two different accel/decel sets |
| 22:20 |
RealBadAngel |
as you said ... wonky |
| 22:20 |
|
jin_xi joined #minetest-dev |
| 22:21 |
F00 |
I really like fast move in water and on ladders, I substantially dislike the much lower decel rate, and I loathe that a bit of water/ladder/rope/etc. is no longer good for stopping players that aren't holding down fast move. |
| 22:29 |
RealBadAngel |
imho just the ladder stuff needs to be tweaked |
| 23:31 |
VanessaE |
agreed. |
| 23:31 |
VanessaE |
after playing with it for a while, everything seems good except the ladders. |
| 23:34 |
hmmmm |
vanessa, you're just the guy i'm lookin' for |
| 23:34 |
hmmmm |
https://github.com/kwolekr/minetest/commits/new_emerge |
| 23:34 |
VanessaE |
guy? er |
| 23:34 |
hmmmm |
it's just a saying |
| 23:34 |
* VanessaE |
looks |
| 23:35 |
hmmmm |
at this point, the queue limits are readded, and it's usable in a multiplayer environment |
| 23:35 |
hmmmm |
you shouldn't experience any emerge freezing problems with this |
| 23:35 |
VanessaE |
nice |
| 23:36 |
hmmmm |
next up, i'll make the queue limits configurable, then i'm going to add in some debugging infrastructure so you can see how many entries the emergethread has, how much it's working, etc. |
| 23:38 |
VanessaE |
lets' see... /teleport 1000,1000,1000... //pos1 ... /teleport -1000,-1000,-1000.... //pos2 ... //set cobble |
| 23:38 |
VanessaE |
;-) |
| 23:38 |
hmmmm |
that has nothing to do with emerging honestly |
| 23:38 |
hmmmm |
aside from the teleportation |
| 23:38 |
VanessaE |
I was joking :-) |
| 23:39 |
|
proller joined #minetest-dev |
| 23:40 |
VanessaE |
large 5000+ node operations are a good way to choke the old code though - even long after worldedit is done, the engine is chewing on the results minutes later. |
| 23:40 |
VanessaE |
or was. |
| 23:42 |
VanessaE |
ok, your new_emerge branch is up and running. |
| 23:43 |
VanessaE |
yep, this is definitely faster as well |
| 23:43 |
hmmmm |
faster!? |
| 23:43 |
VanessaE |
well lemme rephrase that |
| 23:43 |
VanessaE |
more responsive |
| 23:43 |
hmmmm |
the code is certainly cleaner, in theory it shouldn't be faster |
| 23:43 |
hmmmm |
more responsive, i can see that |
| 23:43 |
VanessaE |
it seems to be a tad harder to "outrun" the map loader |
| 23:44 |
VanessaE |
anyhow we'll see how it performs in practice |
| 23:44 |
hmmmm |
tell me if anything comes up |
| 23:44 |
VanessaE |
willdo |
| 23:46 |
VanessaE |
yep by now what I'm doing should have already made it choke. ain't happening now. |
| 23:46 |
VanessaE |
this is a definite improvement |
| 23:48 |
VanessaE |
at worst I'm seeing a few seconds of latency every so often, followed by sudden bursts of map chunks loading up |
| 23:48 |
VanessaE |
(as seen from the client I mean) |
| 23:55 |
|
doserj joined #minetest-dev |
