| Time |
Nick |
Message |
| 08:33 |
|
Calinou joined #minetest-dev |
| 09:02 |
|
PilzAdam joined #minetest-dev |
| 09:46 |
thexyz |
well, i've implemented those entities which don't unload https://github.com/minetest/minetest/commit/595511e1406147f01252e5f376e6b320ccabe51a but the problem now is that blocks near them don't seem to load (unlike blocks near players) |
| 09:57 |
|
OdnetninI joined #minetest-dev |
| 09:59 |
PilzAdam |
this looks interesting: https://github.com/celeron55/minetest/pull/243 |
| 10:00 |
thexyz |
i think i should reimplement this using groups |
| 10:01 |
PilzAdam |
but all the other liquid things are also done like this |
| 10:01 |
PilzAdam |
maybe move all the liquid params to groups |
| 10:02 |
thexyz |
yep, https://github.com/minetest/minetest/pull/1 |
| 10:10 |
|
OdnetninI left #minetest-dev |
| 10:33 |
PilzAdam |
this would be nice for people who want to start coding the engine: http://minetest.net/wiki/doku.php?id=dev:uml_diagrams |
| 11:30 |
|
celeron55 joined #minetest-dev |
| 11:32 |
PilzAdam |
thexyz, this fix works better: https://github.com/celeron55/minetest_game/pull/19 |
| 11:40 |
thexyz |
then merge it |
| 11:41 |
PilzAdam |
done |
| 11:43 |
thexyz |
btw, https://github.com/celeron55/minetest_game/pull/16 → http://irc.minetest.ru/minetest/2012-11-01#i_2601719 |
| 11:46 |
PilzAdam |
thexyz, the water improvements you mentioned above are very interesting |
| 11:46 |
PilzAdam |
are they stable? |
| 11:47 |
thexyz |
they work |
| 11:47 |
thexyz |
but they break API |
| 11:47 |
PilzAdam |
in what way? |
| 11:49 |
thexyz |
liquid_viscosity is removed |
| 11:49 |
thexyz |
> It is not backwards compatible with the old system, but I guess there are just few mods that registered new liquids |
| 11:50 |
PilzAdam |
can we seperate the other changes so we can keep the liquid_viscosity field and add the other things? |
| 11:51 |
thexyz |
better add workaround for liquid_viscosity in builtin/misc_register |
| 11:52 |
PilzAdam |
btw: it seems that you can swim up too fast in water |
| 11:52 |
thexyz |
oh, it also changes ContentFeatures (de)serialization |
| 11:53 |
thexyz |
feel free to tweak it & send patches |
| 12:03 |
PilzAdam |
thexyz, this is somehow broken: https://github.com/minetest/minetest/commit/9f02e3395e0df8896a65604fb1b8b79b2b7b9584 |
| 12:19 |
PilzAdam |
thexyz, https://github.com/minetest/minetest/pull/2 |
| 12:23 |
thexyz |
PilzAdam: why do you think it's broken? |
| 12:24 |
PilzAdam |
it looks weird |
| 12:24 |
thexyz |
? |
| 12:25 |
PilzAdam |
i have tested it by going into lava and i looks realy strange |
| 12:25 |
PilzAdam |
the head goes to the height of the knees |
| 12:27 |
thexyz |
oh! |
| 12:34 |
darkrose |
body melted because of the lava |
| 12:59 |
thexyz |
PilzAdam: ok, i reverted it |
| 13:00 |
PilzAdam |
we could bring this upstream: https://github.com/minetest/minetest/commit/f203b5efbfa88b82d3b5bac31cf718b189e0e6cb |
| 13:18 |
|
hmmmm joined #minetest-dev |
| 13:23 |
|
OdnetninI joined #minetest-dev |
| 13:30 |
|
rubenwardy joined #minetest-dev |
| 14:31 |
PilzAdam |
https://github.com/celeron55/minetest/pull/271 |
| 14:32 |
PilzAdam |
this can be merged, i guess |
| 15:08 |
PilzAdam |
^ thexyz what do you think? |
| 15:12 |
thexyz |
https://github.com/celeron55/minetest/pull/271/files#L0L2631 looks quite dirty (and in that case there shouldn't be that much speedup) |
| 15:12 |
PilzAdam |
https://github.com/minetest/minetest/commit/f203b5efbfa88b82d3b5bac31cf718b189e0e6cb#commitcomment-2046545 |
| 15:13 |
thexyz |
i mean, the first one, (statustext_time / (float)statustext_time_max)*(statustext_time / (float)statustext_time_max)); |
| 15:14 |
thexyz |
i think i'll revert it later |
| 15:19 |
PilzAdam |
https://github.com/celeron55/minetest/pull/271 |
| 15:21 |
darkrose |
rebase those into a single commit, for neatness |
| 15:25 |
PilzAdam |
https://github.com/celeron55/minetest/pull/272 |
| 15:26 |
PilzAdam |
merge it? |
| 15:26 |
darkrose |
if it works |
| 15:29 |
PilzAdam |
yep, it works |
| 15:37 |
celeron55 |
i wonder if the pow one performs badly because the variables are floats and pow takes doubles (powf would be the correct one) |
| 15:37 |
PilzAdam |
they are gone now, so who cares? |
| 15:37 |
celeron55 |
...actually, MSVC builds have always used powf, becaue MSVC overloads pow with a double and a float implementation |
| 15:38 |
celeron55 |
PilzAdam: anyone who likes readable code 8) |
| 15:39 |
PilzAdam |
https://github.com/PilzAdam/minetest_game/tree/animated_water |
| 15:40 |
PilzAdam |
ive implemented animated water but there is a bug |
| 15:40 |
PilzAdam |
one direction of the flowing water is wrong |
| 15:42 |
PilzAdam |
it is also wrong in lava |
| 15:42 |
PilzAdam |
the x- is the same like the z- direction |
| 15:45 |
celeron55 |
hmm... i wonder where the hell that is even calculated... |
| 15:46 |
* celeron55 |
continues search |
| 15:46 |
celeron55 |
ah, this block of code: https://github.com/celeron55/minetest/blob/master/src/content_mapblock.cpp#L464 |
| 15:51 |
celeron55 |
i don't see any terribly obvious errors |
| 15:53 |
celeron55 |
except that those two other ifs should be else ifs |
| 15:54 |
celeron55 |
(the conditions act as else ifs in practice though) |
| 15:57 |
celeron55 |
ehm... wtf |
| 15:58 |
celeron55 |
most likely the '<'s at the int dx, int dz part should be subtractions, in the reverse order |
| 16:00 |
celeron55 |
something like this might work https://gist.github.com/4002269 |
| 16:02 |
celeron55 |
apparently it does |
| 16:03 |
PilzAdam |
yep, it works |
| 16:04 |
celeron55 |
i'll commit it |
| 16:08 |
celeron55 |
oh and by the way, i would love if everyone would use the same grammatical tense in messages that begin with a verb as i have used (that is, the present tense) just for the sake of style |
| 16:08 |
celeron55 |
(so far that has happened) |
| 16:10 |
celeron55 |
done |
| 16:11 |
PilzAdam |
what do you think about the animated water by RealBadAngel? |
| 16:12 |
celeron55 |
nothing |
| 16:39 |
PilzAdam |
why doesnt minetest free the vide memory? http://minetest.net/forum/viewtopic.php?pid=50165#p50165 |
| 16:39 |
PilzAdam |
*video |
| 16:41 |
celeron55 |
there exists a patch called something like "don't leak textures all around the place" but not exactly, that i cannot currently find, that might address some of the issue |
| 16:43 |
celeron55 |
i thought i had already merged this, but i think i haven't: https://github.com/kahrl/minetest/commit/d85113bab5281eaa31ec4c662c265125f663df0e |
| 16:44 |
celeron55 |
(that it is) |
| 16:44 |
PilzAdam |
you cant merge this; it has no verb at the beginning... |
| 16:44 |
celeron55 |
that is not the rule |
| 16:45 |
celeron55 |
the rule applies IF there is a verb at the beginning |
| 16:45 |
PilzAdam |
okay |
| 16:45 |
celeron55 |
i'll test this |
| 16:46 |
celeron55 |
(after this, i'll go and work on other things; i just feel i have to get involved in bug fixing like these because they take much less time that way) |
| 16:49 |
celeron55 |
okay... |
| 16:50 |
celeron55 |
memory consumption, going repeatedly in and out of game: |
| 16:50 |
celeron55 |
107, 100, 155, 151, 202 <- before |
| 16:50 |
celeron55 |
85, 80, 101, 96, 116, 112 <- after |
| 16:51 |
celeron55 |
it does seem to accumulate some crap still, but it's way better |
| 16:54 |
celeron55 |
...i guess i'm pushing this now; i hope it's stable enough |
| 16:55 |
celeron55 |
and now off i go -> |
| 16:55 |
PilzAdam |
i think you will notice if its not stable |
| 16:55 |
* VanessaE |
peeks in |
| 16:56 |
celeron55 |
oh and also one thing: i need to whip up a new official build sometime, because there are already mods using stuff that is broken in 0.4.3 |
| 16:58 |
celeron55 |
...but that doesn't really affect anyone, except if somebody wants to do it instead of me |
| 16:59 |
VanessaE |
there are builds for ubuntu and win32 in the mipmap/aniso thread, current as of your sept. 9 commits anyway |
| 17:00 |
celeron55 |
i won't be touching those; somebody else feel free to |
| 17:09 |
PilzAdam |
thexyz, darkrose, merge this? https://github.com/celeron55/minetest_game/pull/20 |
| 17:11 |
darkrose |
it's messy, got commented out code in it |
| 17:12 |
PilzAdam |
https://github.com/celeron55/minetest_game/blob/master/mods/default/init.lua#L1101 |
| 17:12 |
PilzAdam |
torches have this too |
| 17:17 |
PilzAdam |
darkrose, so delete the old code? |
| 17:19 |
darkrose |
I think so, yeah... otherwise the code just gets messy |
| 17:19 |
PilzAdam |
do you want it all in one commit? |
| 17:19 |
PilzAdam |
for neatness? |
| 17:20 |
darkrose |
would be good |
| 17:20 |
PilzAdam |
grumble |
| 17:21 |
celeron55 |
VanessaE: actually, i somewhat misunderstood what you said; but whatever |
| 17:21 |
* VanessaE |
shrugs |
| 17:21 |
PilzAdam |
darkrose, anything else? I dont want to recommit 20 times |
| 17:23 |
darkrose |
nah, seems ok with the useless bits removed |
| 17:25 |
PilzAdam |
https://github.com/celeron55/minetest_game/pull/21 |
| 17:25 |
PilzAdam |
i guess i can merge it? |
| 17:25 |
VanessaE |
fwiw by the way, I left the commented out line in the torch code for reference |
| 17:25 |
darkrose |
PilzAdam: sure |
| 17:26 |
VanessaE |
in case someone wants to revert to non-animated torches or something. |
| 17:26 |
PilzAdam |
done |
| 17:29 |
darkrose |
VanessaE: for things like that, it'd be better to work a config setting into it somewhere, imo |
| 17:30 |
VanessaE |
true |
| 17:30 |
VanessaE |
no biggie, delete the offending line. :-) |
| 17:31 |
VanessaE |
bbl |
| 17:32 |
|
Calinou joined #minetest-dev |
| 17:40 |
|
Calinou joined #minetest-dev |
| 17:51 |
|
GTRsdk joined #minetest-dev |
| 18:21 |
|
OdnetninI joined #minetest-dev |
| 18:22 |
OdnetninI |
I've seen, than many mods, make the game lags. We need to fix them and improve mobs mod. |
| 18:46 |
Calinou |
no u |
| 18:47 |
|
OdnetninI left #minetest-dev |
| 19:00 |
|
nyuszika7h joined #minetest-dev |
| 19:01 |
hmmmm |
you guys are using the slower version? |
| 19:01 |
hmmmm |
you should make it: return t * t * t * (t * (6.0 * t - 15.0) + 10.0); |
| 19:02 |
hmmmm |
and to celeron: no, conversion from float to double or the other way around doesn't affect performance at all, since the floating point registers are internally 80-bit wide |
| 19:03 |
hmmmm |
the performance hit comes from the fact that pow() takes a floating point number as the power to raise it to which requires extra computation, that is, it doesn't just do repeated multiplication |
| 19:03 |
celeron55 |
so the matter of double vs. float is just how much from the register is copied back to RAM? |
| 19:03 |
hmmmm |
yes |
| 19:03 |
hmmmm |
there are way better reasons to use floats though |
| 19:04 |
hmmmm |
one, you really don't need the extra precision at all, and the fact that it's half as wide helps with cache locality |
| 19:04 |
hmmmm |
two, division is faster with floats than doubles |
| 19:05 |
hmmmm |
and of course if you're using SSE/AVX, you can pack twice as many floats and do twice as many operations. |
| 19:05 |
hmmmm |
the latter is the primary reason why i made the move, but the other two are added benefits |
| 19:06 |
hmmmm |
by the way, you're not offended if i removed all of the NoiseBuffer stuff, right? |
| 19:07 |
celeron55 |
i don't care; it's mostly up to other people than me to judge your work |
| 19:20 |
|
analphabet joined #minetest-dev |
| 19:21 |
|
analphabet left #minetest-dev |
| 22:34 |
|
SpeedProg joined #minetest-dev |
