| Time |
Nick |
Message |
| 00:07 |
Fixer |
RealBadAngel, just fps boost |
| 00:07 |
Fixer |
RealBadAngel, 3695 should be good |
| 00:07 |
RealBadAngel |
lookin at your comparisiong it shouldnt be even noticeable |
| 00:08 |
RealBadAngel |
the curves are exactly the same |
| 00:08 |
Fixer |
RealBadAngel, it is noticable, if it was 10ms per frame, now it is 5ms, big difference |
| 00:08 |
Fixer |
10 ms is 100 fps and 5 is 200 |
| 00:08 |
Fixer |
i mean that |
| 00:08 |
RealBadAngel |
so define small FPS boost |
| 00:10 |
Fixer |
RealBadAngel, +10% |
| 00:11 |
Fixer |
RealBadAngel, +7-10% |
| 00:11 |
Fixer |
RealBadAngel, roughly |
| 00:11 |
RealBadAngel |
that sounds like enough to polish that code and test more widely |
| 00:11 |
Fixer |
RealBadAngel, will be nice to have it |
| 00:11 |
Fixer |
RealBadAngel, i went near crazy with that stutter shit |
| 00:12 |
RealBadAngel |
7-10% doesnt walk along a street each day rdy to pick up |
| 00:13 |
RealBadAngel |
even if its not a solution to stutters, still nice find |
| 00:13 |
Fixer |
RealBadAngel, made automated test with mapgen, it was fine and good, and recording became corrupted after some time, shiiit, now back to just test copy, CRAZY stutters, closed everything, still stutters, test results thgrough the window |
| 00:15 |
Fixer |
RealBadAngel, this is funny, few days ago it was stuttering like hell on J-T copy with shaders, next day - same tests and build - no stutters, today... different build with your 3695, same test, shaders off/on - stuttering like shiiit |
| 00:15 |
Fixer |
i have no idea what is going on |
| 00:15 |
Fixer |
phase of the moon? |
| 00:15 |
RealBadAngel |
i already was suspecting sand storms on mars |
| 00:15 |
Fixer |
I think i need to remove Just-Test testing, it is suspicious world with strange bugs |
| 00:16 |
Fixer |
and make new ones, with mapgen and without it |
| 00:16 |
RealBadAngel |
what is different between yesterday and today? |
| 00:17 |
RealBadAngel |
its time |
| 00:17 |
Fixer |
RealBadAngel, yes |
| 00:17 |
Fixer |
RealBadAngel, hold on |
| 00:17 |
|
Warr1024 joined #minetest-dev |
| 00:17 |
RealBadAngel |
change date immediately on your system when you can see stuttering please |
| 00:18 |
Fixer |
RealBadAngel, 2 days ago and 1 day ago it was exactly the same time of test, same builds, same test, same programs in background, and yet it is stutters wildly in first day, and much less wildly next day %) |
| 00:18 |
Fixer |
o.O |
| 00:18 |
sofar |
ok so |
| 00:19 |
sofar |
someone needs to *fix* the b3d exporter |
| 00:19 |
Fixer |
RealBadAngel, good idea |
| 00:19 |
RealBadAngel |
maybe we do have some weird issues with some timers or what? |
| 00:19 |
sofar |
otherwise I'm going to convert the player model to .x format |
| 00:19 |
sofar |
GRRR |
| 00:19 |
sofar |
if we can't properly export B3D, then the format is fucking useless |
| 00:20 |
sofar |
.x is 4x as large :( |
| 00:20 |
RealBadAngel |
sofar, youre talking bout blender? |
| 00:20 |
sofar |
character.blend -> character.b3d export is completely broken |
| 00:21 |
sofar |
I can't make it work, it's either rotated, or the model is split up in two, or worse, left hand is digging instead of right hand |
| 00:21 |
|
Taoki joined #minetest-dev |
| 00:21 |
sofar |
relying on an unsupported export plugin that nobody can vouch for is broken |
| 00:21 |
RealBadAngel |
sofar, with exporting to .x i also had many issues before i learned about all the exporting options |
| 00:21 |
sofar |
.x works fine, first try |
| 00:22 |
sofar |
I just enabled all the check boxes :) |
| 00:22 |
RealBadAngel |
lol |
| 00:22 |
sofar |
b3d export is NOT working properly, no matter what |
| 00:29 |
sofar |
it's converting the model with left/right flipped, and 90 degrees rotated |
| 00:29 |
sofar |
b3dexport.py |
| 00:29 |
sofar |
if I rotate the entire model/pose in blender, it making the character sway sideways |
| 00:29 |
sofar |
plus digs with left hand |
| 00:29 |
sofar |
none of that happens with the .x export option |
| 00:30 |
|
est31 joined #minetest-dev |
| 00:32 |
kaadmy |
sofar: i use the one made for supertuxkart |
| 00:32 |
kaadmy |
works just fine for me |
| 00:32 |
kaadmy |
.x has big skeletal problems in mt for me |
| 00:32 |
sofar |
can you try re-exporting character.b3d and try it in-game? |
| 00:32 |
sofar |
from character.blend? |
| 00:33 |
kaadmy |
ok |
| 00:33 |
sofar |
also, what blender version? |
| 00:34 |
kaadmy |
2.69 |
| 00:34 |
sofar |
also, where exactly is the one made for "supertuxkart" ? |
| 00:34 |
kaadmy |
i can give you a file |
| 00:34 |
sofar |
hmm I'm on 2.76 |
| 00:34 |
kaadmy |
for some reason blender disabled the addon and complains when re-enabling it :(( |
| 00:35 |
kaadmy |
can't find script files even tho i know perfectly well they're there |
| 00:35 |
Fixer |
RealBadAngel, same performance with different dates/time |
| 00:36 |
est31 |
will soon push a fix for #3698 |
| 00:36 |
ShadowBot |
https://github.com/minetest/minetest/issues/3698 -- Private field 'm_selection_active' is not used |
| 00:36 |
kaadmy |
supertuxkart's blender .b3d exporter: http://supertuxkart.sourceforge.net/Installing_the_tools |
| 00:36 |
RealBadAngel |
Fixer, but can you observe stuttering now? |
| 00:36 |
kaadmy |
it has the d/l links |
| 00:36 |
est31 |
(10 mins) |
| 00:37 |
kaadmy |
wtf is supertuxkart still using SF for.. |
| 00:37 |
sofar |
if they're using SF that means the project is dead |
| 00:37 |
est31 |
well the project is not dead |
| 00:37 |
RealBadAngel |
est31, hold on a sec |
| 00:37 |
est31 |
and they use github afaik |
| 00:38 |
kaadmy |
yeah they do |
| 00:38 |
kaadmy |
still SF for the main page tho i think |
| 00:38 |
est31 |
RealBadAngel, do you want to keep the variable? |
| 00:39 |
RealBadAngel |
est31, theres 3678 waiting that edits the very same file |
| 00:40 |
est31 |
whats the pr doing RealBadAngel ? |
| 00:41 |
RealBadAngel |
finishing the simulating of item_image_buttons |
| 00:41 |
est31 |
so what bugs does it fix |
| 00:42 |
RealBadAngel |
before it item image button's image and text wasnt moving when button was pressed |
| 00:43 |
est31 |
ah that thing |
| 00:43 |
|
paramat joined #minetest-dev |
| 00:43 |
Fixer |
RealBadAngel, i see strange results with 3695 |
| 00:43 |
RealBadAngel |
Fixer, like? |
| 00:44 |
Fixer |
hold on |
| 00:44 |
est31 |
RealBadAngel, commented on the PR |
| 00:44 |
paramat |
RBA is game#834 correct? if so i'll merge it later |
| 00:44 |
ShadowBot |
https://github.com/minetest/minetest_game/issues/834 -- Remove minetest.inventorycube by Rui914 |
| 00:44 |
RealBadAngel |
paramat, yes |
| 00:45 |
paramat |
ok just water and glass |
| 00:45 |
Fixer |
RealBadAngel, so I compiled clean minetest-0.4.13-fefa148-win64 and minetest-0.4.13-fefa148-win64 + PR 3695, did 2 tests, one on just test copy, another mapgen v6 |
| 00:45 |
Fixer |
RealBadAngel, now strange parts |
| 00:47 |
est31 |
RealBadAngel, I've tested, your PR applies cleanly on top of master |
| 00:47 |
RealBadAngel |
paramat, and lava |
| 00:47 |
est31 |
its no big deal if two commits edit same file |
| 00:47 |
est31 |
as long as its not too close inside that file |
| 00:48 |
est31 |
git is smart tool :) |
| 00:48 |
paramat |
yes and lava |
| 00:48 |
Fixer |
RealBadAngel, with my default test config (which pretty much default + fixed 70 nodes range and no fps limit), mapgen test showed that PR3695 gave me +30% fps boost, BUT, here is interesting thing, on Just test testing it gave me -30% FPS boost o___0 |
| 00:49 |
paramat |
removing auto view range will be welcome. c55 was already going to remove it in farmap. many players disable it by setting min and max to same value. limiting it to 2000 will also help with #3659 |
| 00:49 |
ShadowBot |
https://github.com/minetest/minetest/issues/3659 -- Fix disappearing-world problem by OldCoder |
| 00:49 |
RealBadAngel |
Fixer, until you turn on bump mapping or relief mapping even shaders are workin with regular, non tangent vertices |
| 00:49 |
Fixer |
RealBadAngel, in mapgen test it was 132 to 175, in "just test" test - from 44 to 33 |
| 00:49 |
est31 |
#3659 |
| 00:49 |
ShadowBot |
https://github.com/minetest/minetest/issues/3659 -- Fix disappearing-world problem by OldCoder |
| 00:49 |
est31 |
that bug doesnt need view range limited |
| 00:50 |
est31 |
paramat just thinks again in his use case only |
| 00:50 |
RealBadAngel |
est31, theres another code |
| 00:50 |
RealBadAngel |
est31, http://irc.minetest.ru/minetest-dev/2016-02-12#i_4531932 |
| 00:50 |
RealBadAngel |
please do test that |
| 00:50 |
est31 |
he doesnt want to spend 0.0000006 s of his game loop with that |
| 00:50 |
est31 |
too wasteful for him |
| 00:51 |
Fixer |
RealBadAngel, settings were the same except that was saved map and that was map generated on the fly, same default settings, and yet in one big fps loss, in another - big boost |
| 00:51 |
RealBadAngel |
est31, according to fixer tests this change is worth 7-10% of FPS |
| 00:52 |
sofar |
kaadmy: so you can't even use that version of blender and b3d exporter? |
| 00:53 |
paramat |
it helps prevent the bug occurring in most of the world, and camera is already limited to 2000 nodes. and yes other code may be needed |
| 00:53 |
est31 |
RealBadAngel, I dont believe it |
| 00:53 |
est31 |
paramat, minetest is not supposed to be buggy at the world border |
| 00:53 |
est31 |
paramat, thats what we have a BORDER for |
| 00:53 |
Fixer |
sofar, will you ever finish your camera mod? :> |
| 00:53 |
RealBadAngel |
est31, so paste that thing and compile |
| 00:53 |
est31 |
its "work until the border", and after the border, it can be broken or something |
| 00:54 |
sofar |
Fixer: what's missing? what would you want next from it? |
| 00:54 |
est31 |
RealBadAngel, sorry my time is limited |
| 00:54 |
est31 |
you want to update #3678 ? |
| 00:54 |
ShadowBot |
https://github.com/minetest/minetest/issues/3678 -- Formspec: recreate item_image_button pressed state for its image and … by RealBadAngel |
| 00:54 |
est31 |
RealBadAngel, I can merge it once you address my comments |
| 00:54 |
RealBadAngel |
est31, yes, a sec |
| 00:54 |
Fixer |
sofar, well, that features you specified in Next Goals %) |
| 00:56 |
Fixer |
RealBadAngel, so what to do with those contradicting tests? |
| 00:58 |
|
rubenwardy joined #minetest-dev |
| 00:59 |
RealBadAngel |
est31, what do you mean exactly there? |
| 00:59 |
paramat |
:P |
| 00:59 |
RealBadAngel |
whats wrong with scale and parent_button initialization here? |
| 01:00 |
est31 |
RealBadAngel, initialize it like this https://github.com/minetest/minetest/pull/3678/files#diff-7dce35006f82e6eadf963514f5857a8bR255 |
| 01:00 |
est31 |
instead of parent_button = NULL; |
| 01:00 |
est31 |
do parent_button(NULL) |
| 01:00 |
RealBadAngel |
ive set them to NULL if not provided |
| 01:01 |
RealBadAngel |
but ok |
| 01:01 |
RealBadAngel |
editing it now |
| 01:01 |
est31 |
its just syntax |
| 01:03 |
Fixer |
sofar, it is pain to do tests with tinytask, camera mod will be helpful (loading recording via ingame command will be nice) |
| 01:08 |
RealBadAngel |
est31, need to just compile it to check, because its not syntax only. here order matters |
| 01:11 |
est31 |
yeah |
| 01:11 |
est31 |
order must be same as declared in class |
| 01:16 |
RealBadAngel |
est31, updated |
| 01:18 |
RealBadAngel |
Fixer, i dont know what to think about differences between servers, this change is a) nothing new b)not related in any fucking way to server or mods |
| 01:18 |
Fixer |
RealBadAngel, now, i connected to actual just test on the internet, and with your 3695 it has better fps |
| 01:19 |
RealBadAngel |
this is only client side change and both regular vertices and tangent vertices code is widely tested because it was used for so long |
| 01:20 |
RealBadAngel |
if you are getting mixed results that has to be something another |
| 01:20 |
Fixer |
could be |
| 01:21 |
Fixer |
hard to say, this "just test" world save is unpredictable, with stutters depending on phase of moon, and now this strange results |
| 01:21 |
Fixer |
maybe that crazy water bugs is a reason |
| 01:21 |
Fixer |
who knows |
| 01:21 |
RealBadAngel |
3695 is simple as that: use S3DVertex in almost all cases, enable usage of S3DVertexTangents only when bumpmapping or relief enabled |
| 01:22 |
Fixer |
yeah |
| 01:22 |
est31 |
RealBadAngel, you seem to have forgotten one place |
| 01:22 |
RealBadAngel |
est31, where? |
| 01:22 |
est31 |
and RealBadAngel I didnt say "remove the parent_button = NULL" |
| 01:23 |
est31 |
I said "use the initializer list" |
| 01:23 |
est31 |
thats different |
| 01:23 |
est31 |
you removed it |
| 01:23 |
est31 |
like here https://github.com/RealBadAngel/minetestHD/commit/000400686cc84219b7b028f1c5a54024cfff2fcf#diff-7dce35006f82e6eadf963514f5857a8bR142 |
| 01:23 |
RealBadAngel |
thats empty initialization |
| 01:24 |
RealBadAngel |
if nothing is set in constructor why button shall be different? |
| 01:24 |
est31 |
RealBadAngel, its a pointer |
| 01:24 |
est31 |
pointers arent initialized |
| 01:24 |
est31 |
RealBadAngel, in the empty constructor in C++ the constructor is run on every member variable |
| 01:24 |
est31 |
and the constructor for pointers does _nothing_ |
| 01:25 |
RealBadAngel |
good to know |
| 01:25 |
est31 |
it will get you read of uninitialized pointer warnings |
| 01:27 |
est31 |
RealBadAngel, alternatively, if the constructor is never used, you can just remove the whole constructor |
| 01:30 |
RealBadAngel |
est31, updated it |
| 01:31 |
est31 |
RealBadAngel, only remaining place is now https://github.com/minetest/minetest/pull/3678/files#diff-7dce35006f82e6eadf963514f5857a8bR258 |
| 01:35 |
RealBadAngel |
est31, aw, missed it because its another struct. done |
| 01:37 |
kaadmy |
sofar: no, but blender does this sometimss |
| 01:37 |
kaadmy |
disabled custom addons and has errors when re-enabling |
| 01:42 |
|
est31 joined #minetest-dev |
| 01:43 |
est31 |
okay merged |
| 01:43 |
est31 |
bye all |
| 01:45 |
|
Icedream joined #minetest-dev |
| 01:45 |
paramat |
i'll merge game#789 game#833 game#834 soon |
| 01:45 |
ShadowBot |
https://github.com/minetest/minetest_game/issues/789 -- New mesh door models, and extensive door API by sofar |
| 01:45 |
ShadowBot |
https://github.com/minetest/minetest_game/issues/833 -- Convert furnace ABM to node timer by sofar |
| 01:45 |
ShadowBot |
https://github.com/minetest/minetest_game/issues/834 -- Remove minetest.inventorycube by Rui914 |
| 01:48 |
|
ssieb joined #minetest-dev |
| 02:02 |
|
Warr1024 joined #minetest-dev |
| 02:04 |
|
Warr1024 joined #minetest-dev |
| 02:53 |
|
H-H-H joined #minetest-dev |
| 03:40 |
|
H-H-H joined #minetest-dev |
| 03:43 |
paramat |
now merging games 789 833 834 |
| 03:57 |
paramat |
merged |
| 04:05 |
|
linkedinyou joined #minetest-dev |
| 04:19 |
sofar |
neat, I'm throwing it on my personal server and updating the whole lot |
| 04:19 |
sofar |
if only jeija would pull my mesecons_doors patch :/ |
| 04:29 |
|
kaeza joined #minetest-dev |
| 05:34 |
|
leat joined #minetest-dev |
| 05:45 |
|
leat joined #minetest-dev |
| 05:53 |
paramat |
RBA https://github.com/minetest/minetest/blob/master/src/camera.cpp#L523 surely this should be 'if (viewing_range_max < 200)' because 'viewing range max' is in nodes? |
| 05:54 |
|
leat joined #minetest-dev |
| 05:57 |
paramat |
as you discovered, viewing range max is always < 2000 so cameranode farvalue is set to 20k |
| 06:07 |
paramat |
ok it makes sense, if view range < 2000 nodes fix camera farvalue to 2000 * BS, if above increase it as necessary |
| 06:08 |
paramat |
the comment mentions clouds and indeed clouds render out to 2000 nodes |
| 06:12 |
paramat |
perhaps i'll work on this function to also remove auto view range |
| 06:21 |
paramat |
'updateViewingRange' is called every camera 'update', if we remove auto view range this will remove a lot of processing |
| 06:28 |
paramat |
.. it was going to be done anyway in farmap |
| 06:47 |
|
paramat left #minetest-dev |
| 07:17 |
sofar |
hmm, I've been looking at blockmen's hunger and with a few tweaks I think it should be perfectly fine to put it in minetest_game, instead of merging something into core |
| 07:17 |
sofar |
it actually makes singleplayer/vanilla minetest_game an interesting game |
| 07:17 |
sofar |
I think it took me 3 game days to get my first bread |
| 07:18 |
sofar |
not enough to starve me, either, but I had to tend the crops well enough |
| 07:19 |
sofar |
I added a few tweaks to make it more interesting: hunger on crafting, and hunger healing reduces the hunger bar too (so if you took a lot of damage, expect to eat a lot of food) |
| 07:19 |
sofar |
I need to add hunger on fighting, but then I think it's pretty decent |
| 07:34 |
|
nolsen joined #minetest-dev |
| 07:50 |
|
Hunterz joined #minetest-dev |
| 07:56 |
|
paramat joined #minetest-dev |
| 08:04 |
paramat |
nore sfan5 any comments on my proposed improvements for game ABMs? game#827 |
| 08:04 |
ShadowBot |
https://github.com/minetest/minetest_game/issues/827 -- ABM improvements |
| 08:15 |
|
Krock joined #minetest-dev |
| 08:16 |
|
leat joined #minetest-dev |
| 08:56 |
paramat |
here's the PR game#839 |
| 08:56 |
ShadowBot |
https://github.com/minetest/minetest_game/issues/839 -- Consolidate ABMs by paramat |
| 09:01 |
|
paramat left #minetest-dev |
| 09:07 |
|
ud1_ joined #minetest-dev |
| 09:17 |
|
Obani joined #minetest-dev |
| 09:31 |
|
Calinou joined #minetest-dev |
| 09:59 |
|
jin_xi joined #minetest-dev |
| 10:00 |
|
VargaD joined #minetest-dev |
| 10:05 |
|
Lunatrius joined #minetest-dev |
| 10:08 |
|
everamzah joined #minetest-dev |
| 10:43 |
|
ud1_ joined #minetest-dev |
| 11:05 |
|
Amaz joined #minetest-dev |
| 12:17 |
|
ud1 joined #minetest-dev |
| 12:29 |
|
Fixer joined #minetest-dev |
| 12:45 |
|
ud1 joined #minetest-dev |
| 13:00 |
|
xeranas joined #minetest-dev |
| 13:00 |
|
Player_2 joined #minetest-dev |
| 13:07 |
|
Obani joined #minetest-dev |
| 14:05 |
|
dfelinto joined #minetest-dev |
| 14:09 |
|
ud1_ joined #minetest-dev |
| 14:30 |
|
EvergreenTree joined #minetest-dev |
| 14:45 |
|
STHGOM joined #minetest-dev |
| 14:48 |
|
H-H-H joined #minetest-dev |
| 14:49 |
|
kaadmy joined #minetest-dev |
| 15:00 |
|
jin_xi joined #minetest-dev |
| 15:23 |
|
kaeza joined #minetest-dev |
| 16:00 |
|
est31 joined #minetest-dev |
| 16:49 |
|
Alduin_ joined #minetest-dev |
| 17:15 |
|
Alduin_ joined #minetest-dev |
| 17:24 |
|
est31 joined #minetest-dev |
| 17:29 |
|
linkedinyou joined #minetest-dev |
| 17:44 |
|
hmmmm joined #minetest-dev |
| 17:58 |
sofar |
nore: ty for pulling 840 |
| 17:58 |
|
lubuntu_ joined #minetest-dev |
| 17:59 |
nore |
sofar: you're welcome, that was an important bugfix that had to be merged asap :) |
| 18:00 |
sofar |
eh, that's what you get when you merge API changes |
| 18:00 |
sofar |
happy to get them closed quickly, too |
| 18:05 |
|
kaeza joined #minetest-dev |
| 18:30 |
|
STHGOM joined #minetest-dev |
| 18:34 |
|
ud1 joined #minetest-dev |
| 19:05 |
|
est31 joined #minetest-dev |
| 19:11 |
|
est joined #minetest-dev |
| 19:25 |
|
est31 joined #minetest-dev |
| 19:42 |
|
ud1_ joined #minetest-dev |
| 19:43 |
Fixer |
RealBadAngel, have you seen smth like this in water: https://i.imgur.com/gl5W7MW.png config: http://pastebin.com/raw/5i52UCNW |
| 19:46 |
kaadmy |
i think it's caused when different liquid types are side-by-side |
| 19:46 |
kaadmy |
ie. river water beside normal water has z-depth fighting |
| 19:49 |
Fixer |
it is homedecor glowlight white top thing below the water |
| 19:49 |
kaadmy |
hmm |
| 19:55 |
Fixer |
https://i.imgur.com/X4rb74B.png :S |
| 19:55 |
ShadowNinja |
RealBadAngel, est31: You broke the build!!! |
| 19:55 |
kaadmy |
huh?? |
| 20:00 |
celeron55 |
let me guess: irrlicht 1.7 |
| 20:01 |
Fixer |
me? 1.8.1 |
| 20:01 |
est31 |
seems so |
| 20:01 |
est31 |
https://launchpad.net/~minetestdevs/+archive/ubuntu/daily-builds/+build/9006499 |
| 20:03 |
est31 |
/«PKGBUILDDIR»/src/guiFormSpecMenu.cpp:2438:19: error: 'EGDS_BUTTON_PRESSED_IMAGE_OFFSET_X' is not a member of 'irr::gui' |
| 20:08 |
Fixer |
RealBadAngel, looks like there is a problem with water when shaders enabled |
| 20:12 |
est31 |
#3701 |
| 20:12 |
ShadowBot |
https://github.com/minetest/minetest/issues/3701 -- Build with irrlicht 1.7 broken |
| 20:12 |
Fixer |
RealBadAngel, oh well, bug dissapeared again :/ |
| 20:12 |
est31 |
I do notice some problems with lava |
| 20:16 |
|
Darcidride joined #minetest-dev |
| 20:16 |
RealBadAngel |
hmm, looks i was right when wanted to use hardcoded offset |
| 20:18 |
Fixer |
est31, i got same problem you have https://github.com/minetest/minetest/issues/2780 |
| 20:18 |
RealBadAngel |
first the offset thing, gimme 5 minutes |
| 20:19 |
Fixer |
est31, i also feel like problem appears when shaders are enabled |
| 20:25 |
est31 |
#3703 |
| 20:25 |
ShadowBot |
est31: Error: Delimiter not found in "HTTP Error 404: Not Found" |
| 20:25 |
est31 |
err #3702 |
| 20:25 |
ShadowBot |
https://github.com/minetest/minetest/issues/3702 -- Liquids Rendering issue |
| 20:28 |
Fixer |
have not seen anything like thtat |
| 20:32 |
RealBadAngel |
est31, can you check the liquids with 3695 applied please? |
| 20:33 |
est31 |
RealBadAngel, moment pls |
| 20:34 |
|
Obani joined #minetest-dev |
| 20:37 |
est31 |
RealBadAngel, that PR fixes the bug |
| 20:38 |
RealBadAngel |
as i thought, that was that slow irrlicht method to convert mesh to tangent space |
| 20:40 |
RealBadAngel |
in 3695 ive switched to old way of making meshes but separate for S3DVertex and S3DVertexTangents |
| 20:40 |
RealBadAngel |
based on settings |
| 20:41 |
RealBadAngel |
so only bumpmapping and relief switches to tangent space ones |
| 20:43 |
sofar |
!seen jeija |
| 20:43 |
ShadowBot |
sofar: I haven't seen jeija in #minetest-dev. |
| 20:46 |
Fixer |
!seen Jeija |
| 20:46 |
ShadowBot |
Fixer: I haven't seen Jeija in #minetest-dev. |
| 21:10 |
RealBadAngel |
est31, #3703 |
| 21:10 |
ShadowBot |
https://github.com/minetest/minetest/issues/3703 -- Formspec: fix broken irrlicht 1.7 build by RealBadAngel |
| 21:11 |
RealBadAngel |
Fixer, ok, so what about those lamps? |
| 21:11 |
Fixer |
RealBadAngel, what lamps? %) |
| 21:12 |
RealBadAngel |
<Fixer> it is homedecor glowlight white top thing below the water |
| 21:12 |
Fixer |
RealBadAngel, yes, i stand inside in water, slightly move mouse and get those weird shapes |
| 21:13 |
est31 |
RealBadAngel, see comment |
| 21:13 |
est31 |
if you fix it I'll push the commit directly |
| 21:13 |
RealBadAngel |
a sec |
| 21:13 |
est31 |
it looks good, and buildbots are best way to find out whether build succeeds. |
| 21:14 |
Fixer |
RealBadAngel, you can test it on Liberty Land server and walk to coordinates on that screenshot |
| 21:14 |
RealBadAngel |
i dont trust them since i saw they triggered building for 3695 at least 4 or 5 times. sometimes failing sometimes not |
| 21:15 |
RealBadAngel |
it was happening yesterday |
| 21:15 |
RealBadAngel |
i had page open whole day and was watching the show lol |
| 21:17 |
RealBadAngel |
est31, done |
| 21:19 |
est31 |
RealBadAngel, thanks for the fix |
| 21:22 |
RealBadAngel |
Fixer, does 3695 fix that issue too or not? |
| 21:23 |
Fixer |
RealBadAngel, standing in water thing? nope |
| 21:25 |
RealBadAngel |
Fixer, on which conditions you can see the glitch exactly? |
| 21:29 |
Fixer |
RealBadAngel, hard to say, at least with shaders on/off, my config: http://pastebin.com/raw/5i52UCNW |
| 21:31 |
RealBadAngel |
ok, i can see the glitch |
| 21:31 |
RealBadAngel |
hold on a sec |
| 21:31 |
Fixer |
._. |
| 21:31 |
RealBadAngel |
its not about shaders or tangent space |
| 21:31 |
RealBadAngel |
lemme check something more |
| 21:33 |
RealBadAngel |
its not any recently born bug |
| 21:35 |
RealBadAngel |
it looks like a pretty ancient issue |
| 21:35 |
Fixer |
could be |
| 21:36 |
RealBadAngel |
at least it "works" the same in 1.5yrs old builds |
| 21:37 |
RealBadAngel |
so, please do open an issue on it but please dont put there sticker "rba's regression" :P |
| 21:37 |
RealBadAngel |
i will fix that later today |
| 21:37 |
Fixer |
one sec |
| 21:37 |
Fixer |
RealBadAngel, what about that square water thing? |
| 21:38 |
RealBadAngel |
well, if i wont get drunk today, in 1.5hours im turning 43 ;) |
| 21:38 |
Fixer |
nice |
| 21:38 |
RealBadAngel |
square water? |
| 21:39 |
Fixer |
RealBadAngel, this crap -> https://i.imgur.com/X4rb74B.png |
| 21:39 |
RealBadAngel |
also old like hell |
| 21:39 |
Fixer |
RealBadAngel, when shaders are enabled, old, yes |
| 21:39 |
Fixer |
RealBadAngel, for some reason my eye catched it again |
| 21:39 |
RealBadAngel |
if you will check tickets there are at least 3 issues open on that |
| 21:40 |
Fixer |
yes, looked at them |
| 21:40 |
RealBadAngel |
its general flaw of mapblockmesh animate method |
| 21:40 |
RealBadAngel |
bad by design |
| 21:40 |
RealBadAngel |
not my code |
| 21:42 |
RealBadAngel |
for such things to look consistent only solution is water surface shader that will do all visible surface at once |
| 21:43 |
RealBadAngel |
same effect you can see when changing rapidly daytime without shaders |
| 21:43 |
Fixer |
RealBadAngel, yes, but why water bug is off when shaders disabled? |
| 21:44 |
RealBadAngel |
shaders have nothing to do with texture animations |
| 21:44 |
Fixer |
ok |
| 21:46 |
VanessaE |
RealBadAngel: he's not talking about animations. |
| 21:46 |
RealBadAngel |
well, they can have something in common |
| 21:46 |
RealBadAngel |
transparency |
| 21:46 |
VanessaE |
he's talking about the opacity of the water, and it affects anything with alpha |
| 21:46 |
VanessaE |
including say stained glass or homedecor curtains |
| 21:46 |
Fixer |
yes |
| 21:47 |
RealBadAngel |
a sec |
| 21:47 |
VanessaE |
the effect is random, and affected by your view angle, distance, and pitch |
| 21:47 |
VanessaE |
s/angle/yaw/ |
| 21:48 |
RealBadAngel |
just checked, that lamps issue is not related to shaders |
| 21:48 |
RealBadAngel |
it "works" in both cases |
| 21:48 |
Fixer |
nope |
| 21:49 |
Fixer |
RealBadAngel, you mean this lamps https://cloud.githubusercontent.com/assets/8872119/8145588/dcc87e9e-120d-11e5-8499-1536abfafb58.gif ? |
| 21:49 |
RealBadAngel |
just checked 1.5yrs old build and current one |
| 21:49 |
RealBadAngel |
aw, we are switching topics |
| 21:49 |
VanessaE |
yeah it's an old-ass bug |
| 21:50 |
RealBadAngel |
i do have an idea |
| 21:50 |
Fixer |
now i'm confused |
| 21:50 |
|
Gael-de-Sailly joined #minetest-dev |
| 21:50 |
Fixer |
VanessaE, you mean that my screenshot with white homedecor windows? |
| 21:51 |
VanessaE |
Fixer: your screenshot shows an alpha glitch in water. that same glitch affects yes, white homedecor windows, curtains, fish tanks, stained glass (from the mod so named), ilights, and anything else with alpha. |
| 21:51 |
Fixer |
VanessaE, ok |
| 21:52 |
RealBadAngel |
that water thingy, mapblock sized |
| 21:52 |
RealBadAngel |
who is able to see it right now? |
| 21:52 |
VanessaE |
and depending on unknown variables, you can sit there unmoving and look at say, an ilight, and it'll sit there flickering between opaque and alpha for no reason at all, even though it is not animated. |
| 21:52 |
VanessaE |
RealBadAngel: last time I loaded minetest (been a couple weeks), the effect was still there. |
| 21:52 |
|
BrunoMine joined #minetest-dev |
| 21:52 |
RealBadAngel |
VanessaE, https://cloud.githubusercontent.com/assets/8872119/8145588/dcc87e9e-120d-11e5-8499-1536abfafb58.gif |
| 21:52 |
RealBadAngel |
you can see exactly that? |
| 21:53 |
VanessaE |
RealBadAngel: yes, I have seen precisely that effect, and again it affects EVERYTHING with alpha. if you had a big, wide field of ilights or stained glass, you'd see the same thing happen with it. |
| 21:53 |
VanessaE |
lemme check with my current build |
| 21:53 |
RealBadAngel |
prepare now working build with "working" glitch please |
| 21:54 |
RealBadAngel |
what version of irrlicht do you have? |
| 21:54 |
VanessaE |
just my luck, it doesn't want to glitch out for me now. |
| 21:54 |
RealBadAngel |
lol |
| 21:54 |
Fixer |
RealBadAngel, yes |
| 21:55 |
VanessaE |
Filename: pool/main/i/irrlicht/libirrlicht-dev_1.8.1+dfsg1-1+b2_amd64.deb |
| 21:55 |
VanessaE |
(my minetest build is commit 262dbe11 |
| 21:55 |
VanessaE |
) |
| 21:56 |
RealBadAngel |
one more question to be sure, does it happen with shaders enabled, right? |
| 21:56 |
|
BrunoMine joined #minetest-dev |
| 21:56 |
VanessaE |
I couldn't reproduce it just now either with or without shaders. I don't recall if shaders affected it for me when I last saw it. |
| 21:56 |
VanessaE |
lemme try loading up a heavier game |
| 21:56 |
RealBadAngel |
ok |
| 21:59 |
VanessaE |
can't reproduce it now. figures. I'm sure it'll show up later. |
| 21:59 |
RealBadAngel |
hehe propably when i go offline ;) |
| 21:59 |
VanessaE |
maybe my build already has some commit to fix it; I've been kinda not following minetest development for a while. |
| 21:59 |
Fixer |
hold on |
| 21:59 |
Fixer |
i will reproduce it for you |
| 21:59 |
VanessaE |
lemme try it with HDX |
| 22:00 |
RealBadAngel |
Fixer, so, does it pops up with shaders? |
| 22:00 |
Fixer |
RealBadAngel, with shaders |
| 22:00 |
RealBadAngel |
ok, and which ver of irr you do have? |
| 22:01 |
Fixer |
1.8.1 |
| 22:01 |
|
BrunoMine joined #minetest-dev |
| 22:02 |
RealBadAngel |
ok, open game.cpp then |
| 22:02 |
VanessaE |
can't reproduce it with commit 262dbe11 + HDX256 + dreambuilder, with or without shaders. |
| 22:03 |
Fixer |
RealBadAngel, wait, i need to find that house with curtains |
| 22:03 |
Fixer |
._. |
| 22:03 |
Fixer |
found it |
| 22:03 |
Fixer |
._. |
| 22:03 |
Fixer |
that was fast |
| 22:04 |
Fixer |
VanessaE, this one? https://i.imgur.com/vYsSnSr.png and https://i.imgur.com/RH4AiLP.png |
| 22:05 |
VanessaE |
RealBadAngel: come to think of it, my build is actually Dec. 7 commit a594963a plus an older version of your inventory meshes pull from Jan 24. |
| 22:05 |
VanessaE |
Fixer: that's the same glitch as the water, yes. |
| 22:05 |
RealBadAngel |
Fixer, https://github.com/minetest/minetest/blob/master/src/game.cpp#L909 |
| 22:05 |
Fixer |
ok, can reproduce, let me check with shaders disabled |
| 22:06 |
Fixer |
can't reproduce with shaders disabled |
| 22:06 |
VanessaE |
well Fixer's build is much newer than mine, so it's clearly still an issue up to that point anyway. |
| 22:06 |
Fixer |
RealBadAngel, ok |
| 22:06 |
Fixer |
RealBadAngel, what now? |
| 22:06 |
RealBadAngel |
find this and replace lines 909 - 922 with: http://pastebin.com/DmhAEH07 |
| 22:06 |
|
BrunoMine joined #minetest-dev |
| 22:07 |
Fixer |
RealBadAngel, ok |
| 22:07 |
Fixer |
compiling is +20 min as usual |
| 22:07 |
RealBadAngel |
sure |
| 22:07 |
VanessaE |
I guess I should update. |
| 22:08 |
RealBadAngel |
VanessaE, there are quite a few new nice things in |
| 22:09 |
VanessaE |
most of it I probably won't notice (like the halo stuff) |
| 22:09 |
RealBadAngel |
well, the halo works now way better than before |
| 22:10 |
RealBadAngel |
ive changed it to use bounding box instead of many overlapping boxes |
| 22:11 |
Fixer |
RealBadAngel, will compile on same commit |
| 22:11 |
Fixer |
fefa148 one |
| 22:14 |
VanessaE |
huh. no inventory meshes in current HEAD? |
| 22:14 |
VanessaE |
oh there it is. |
| 22:14 |
VanessaE |
animation disabled by default. that sucks :( |
| 22:15 |
VanessaE |
*turns that and halos on* |
| 22:15 |
VanessaE |
whose bright idea was it to turn those off by default? |
| 22:18 |
VanessaE |
I like the halo single bounding box idea, btw. |
| 22:18 |
VanessaE |
that looks 100000x better than a bunch of overlapping boxes. |
| 22:18 |
RealBadAngel |
told ya :) |
| 22:19 |
RealBadAngel |
btw its a tradition to turn off by default anythin i code ;) |
| 22:20 |
VanessaE |
speaking of alpha, RealBadAngel: take note of alpha in wielded objects like the aforementioned fishtank. |
| 22:21 |
VanessaE |
(there isn't any, it's completely transparent where it should have alpha) |
| 22:23 |
RealBadAngel |
wielded are setting other materials than used in world |
| 22:23 |
VanessaE |
http://daconcepts.com/vanessa/hobbies/minetest/screenshots/random/Screenshot%20-%2002132016%20-%2005%3A25%3A51%20PM.png |
| 22:24 |
|
DFeniks joined #minetest-dev |
| 22:25 |
VanessaE |
that aside, I find no obvious faults with current HEAD, but that was hardly a scientific test. |
| 22:26 |
VanessaE |
you still planning to do with cracks as you did with the halo? (a separate mesh or something wasn't it?) |
| 22:28 |
RealBadAngel |
yes |
| 22:28 |
VanessaE |
ok. |
| 22:29 |
RealBadAngel |
that has to be done no matter what |
| 22:29 |
RealBadAngel |
its a damn bottleneck |
| 22:29 |
VanessaE |
I know. |
| 22:29 |
VanessaE |
I just tested it with a(n animated) homedecor torch. |
| 22:29 |
VanessaE |
(just out of curiosity) |
| 22:33 |
Fixer |
compiled |
| 22:33 |
RealBadAngel |
well, ive fixed another old issue just meanwhile |
| 22:34 |
VanessaE |
cool |
| 22:34 |
RealBadAngel |
remember that agressive bloks culling in the screen corners? |
| 22:34 |
VanessaE |
yes |
| 22:34 |
RealBadAngel |
one liner |
| 22:34 |
Fixer |
yes |
| 22:38 |
Fixer |
RealBadAngel, did not help |
| 22:41 |
RealBadAngel |
Fixer, pity. but has to be merged anyway |
| 22:41 |
RealBadAngel |
meanwhile, #3705 |
| 22:41 |
ShadowBot |
https://github.com/minetest/minetest/issues/3705 -- Fix too agressive block culling by RealBadAngel |
| 22:49 |
|
kaeza joined #minetest-dev |
| 22:57 |
ShadowNinja |
RealBadAngel: Are you sure that doesn't just massively increase the cull angle? |
| 22:59 |
RealBadAngel |
ehm, angle was halved there |
| 23:00 |
RealBadAngel |
but i can dump number of blocks in range before and after |
| 23:00 |
|
est31 joined #minetest-dev |
| 23:03 |
ShadowNinja |
RealBadAngel: It seemed to be just about right the way it was though. I bit low perhaps, but not _that_ low. |
| 23:11 |
RealBadAngel |
ShadowNinja, just about right != right |
| 23:11 |
RealBadAngel |
i will check different fov now |
| 23:23 |
kahrl |
RealBadAngel: http://emoc.org/utile/camera_fov.png |
| 23:23 |
kahrl |
when you want to compute DIST given L or vice versa, you have to apply trig functions to FOV/2, not to FOV |
| 23:24 |
RealBadAngel |
ok, so whats wrong then with this? |
| 23:24 |
kahrl |
(and you get L/2 instead of L) |
| 23:24 |
kahrl |
RealBadAngel: I guess that it looks for the center of a mapblock |
| 23:25 |
kahrl |
but sometimes parts of a mapblock should be visible while the center is not |
| 23:25 |
kahrl |
s/guess/know, see line 202 |
| 23:31 |
RealBadAngel |
kahrl, looking at comments line 234 tries to take care of it |
| 23:34 |
kahrl |
right |
| 23:34 |
|
est31 joined #minetest-dev |
| 23:36 |
kahrl |
RealBadAngel: there is another problem with camera fov |
| 23:37 |
kahrl |
actually there should be two fovs, since the screen is rectangular: x_fov and y_fov |
| 23:37 |
kahrl |
in OpenGL fov usually means y_fov, while is typically the smaller of the two |
| 23:37 |
kahrl |
I think in Irrlicht it's the same |
| 23:38 |
kahrl |
maybe isBlockInSight should use max(x_fov, y_fov) |
| 23:38 |
kahrl |
s/while is/which is/ |
| 23:40 |
RealBadAngel |
changing screen proportions doesnt change that much |
| 23:41 |
RealBadAngel |
maybe slightly increase fov angle would do the trick? |
| 23:42 |
kahrl |
anything hacky like that will likely break in another situation again |
| 23:43 |
kahrl |
screens seem to get wider (but not higher) every year |
| 23:43 |
kahrl |
(and of course the user could change the window size themselves to be extremely wide) |
| 23:44 |
RealBadAngel |
5% is not enough, but increasing the angle by 10% works |
| 23:47 |
kahrl |
ah, the thing I said is already done |
| 23:47 |
RealBadAngel |
http://i.imgur.com/l5yr50s.png |
| 23:47 |
kahrl |
camera_fov is camera->getFovMax() which is the max of m_fov_x and m_fov_y |
| 23:48 |
RealBadAngel |
its with if(cosangle < cos((camera_fov*1.1) / 2)) |
| 23:51 |
RealBadAngel |
http://i.imgur.com/RSrFzEk.png |
| 23:51 |
RealBadAngel |
this is with 5% |
| 23:53 |
RealBadAngel |
note ive rotated screen by 90deg, but it shouldnt matter as you found out |
| 23:59 |
jomat |
win 32 |
