Minetest logo

IRC log for #minetest-dev, 2014-05-18

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

All times shown according to UTC.

Time Nick Message
00:01 VanessaE I run it as:  (cd to the mapper dir, then...)  ./minetestmapper --drawscale --geometry -5000:-7500+12000+15000 -i $worldpath"/"$worldname"/" -o $imagepath"/"$worldname".png"
00:35 VanessaE ShadowNinja: comments on the above?
00:42 ShadowNinja VanessaE: Last I used it it shouldn't have been much slower, faster even.
00:42 VanessaE by "FOR BLOODY EVER" I mean roughly 5x slower than your branch
00:42 VanessaE take a look at my server's stats page, you can see the results.
00:43 VanessaE that large spike happening now is me re-generating the overview maps with your fork (and I'm nearly done), the big long run starting at 0500 is the official branch
00:43 ShadowNinja VanessaE: git bisect maybe.
00:44 VanessaE that will be difficult to do.  production server and very large map files.  But I can try.
00:45 ShadowNinja Try with something smaller.  I think that all of my commits have increased or not affected performance though.
00:56 VanessaE bisecting now.
00:57 VanessaE taking a baseline time measurement with your branch first though, using a copy of the nostalgia map and the default colors.txt.
00:58 VanessaE ok, about 56 seconds for that.
01:01 OldCoder joined #minetest-dev
01:06 VanessaE *sigh*
01:06 VanessaE nostalgia map either is too small or does not contain whatever data is needed to trigger the problem
01:07 VanessaE best time, 56s with yours.  worst, ~66s at commit 925de1bb (and that's probably just statistical noise because the server isn't idle)
01:07 VanessaE I'll have to do this on my home PC
01:08 VanessaE this might invalidate the result, which is why I didn't want to do it
01:11 Taoki joined #minetest-dev
01:37 stormchaser3000 joined #minetest-dev
01:46 stormchaser3000 joined #minetest-dev
02:02 us`0gb joined #minetest-dev
02:20 VanessaE ok. bisect failed miserably.  doing this the hard way.  3m27s for the baseline time for yours, with Creative and my colors.txt.
02:32 Megaf_ joined #minetest-dev
02:32 Piggybear87 joined #minetest-dev
02:48 VanessaE ShadowNinja: the "good" news is I can reproduce the super-long times with minetestmapper HEAD.
02:48 ImQ009 joined #minetest-dev
02:49 VanessaE (as of right now I'm still waiting for the first run of "the hard way", going on ~21 minutes now)
02:49 * Megaf is glad his not the only one
02:49 VanessaE pardon?
02:50 * Megaf is glad his not the only one who has the long wait for mapgen to gen a map
02:50 Megaf I mean
02:50 Megaf mapper
02:50 Megaf my brain isnt working properly lately
02:50 OldCoder joined #minetest-dev
02:50 Megaf VanessaE; minetest mapper takes very long here too
02:51 Megaf and I'm not talking about an ARM board
02:53 VanessaE Megaf: this is an image that should take 3m37s to generate (I timed it with ShadowNinja's fork about half an hour ago)
02:53 Megaf VanessaE; want me to try it?
02:53 Megaf how big is the world you are trying to map?
02:53 VanessaE compare it against this one, https://github.com/ShadowNinja/minetest-mapper-cpp
02:54 VanessaE Megaf: about 5 GB.  it's my Creative map.
02:54 Megaf oh
02:54 VanessaE no, strike that.  my Survival map.
02:54 khonkhortisan As big as this! *runs off in one direction, and runs back from the opposite direction*
02:54 Megaf ok, so I will try to map a slightly old version of my world
03:01 Megaf VanessaE; what are the options you are using for minetest-mapper?
03:02 VanessaE time ./minetestmapper --drawscale --geometry -5000:-7500+12000+15000 -i /path/to/world_folder o imagename.png
03:02 VanessaE oops, make that -o
03:03 Megaf ok, running the Shadow's mapper
03:04 Megaf minetestmapper using 18% of my CPU
03:04 Megaf ah, there you go, 100% now
03:04 VanessaE It'll use all of one core.
03:06 Megaf ok, Im runnign two mappers at the same time, each using 99.9% of CPU
03:06 Megaf VanessaE; Shadow's mapper real1m56.731s
03:06 Megaf mireq's mapper real1m7.536s
03:06 Megaf is there other mappers?
03:07 Megaf VanessaE; what mapper are you using?
03:07 VanessaE don't run them at the same time, tun them separately.   the one I'm having trouble with is the official one, https://github.com/minetest/minetestmapper
03:10 Megaf VanessaE; minetest's mapper real1m5.384s
03:11 Megaf running the other two again now
03:11 Megaf one by one
03:11 Megaf mireq's mapper real1m7.319s
03:12 Megaf Shadow's mapper real1m5.887s
03:13 Megaf the official one is the fastest here
03:13 VanessaE you have to run it on a big map with my colors.txt to trigger the bug apparently.
03:14 Megaf lol, minetest forked from shadow, shadow forked from sfan, sfan forken from mireq
03:14 VanessaE yup
03:14 VanessaE that's how forks work
03:14 VanessaE that's notmal.
03:14 VanessaE normal.
03:14 Megaf VanessaE; pastebin your colors.txt
03:15 Megaf or send me in another way
03:15 Megaf maybe DCC
03:15 VanessaE http://digitalaudioconcepts.com/vanessa/hobbies/minetest/colors.txt
03:20 Megaf VanessaE; well, I'm testing first the official mapper with that colors.txt and its taking quite long already
03:20 VanessaE define "quite long"
03:20 VanessaE so far I'm working on 33 minutes here.  O_O
03:21 VanessaE I'm letting it finish just so I can see the output :P
03:21 Megaf Im going to run other mappers side by side and will still have a core to spare
03:21 VanessaE don't.
03:21 Megaf why not?
03:21 VanessaE don't make them compete for memory and disk.
03:22 Megaf ok..
03:23 Megaf just killed minetest mapper
03:23 Megaf gonna try with shadows and mireqs mappers first
03:23 VanessaE when you're taking timing measurements, you should try not to do *anything* else that could potentially cost speed, even if you have cores to spare.
03:24 Megaf VanessaE; Shadow's fork with your colors.txt 1m6.436s
03:33 VanessaE ok that's just too long to wait.  35+ minutes now.
03:34 * VanessaE cancels the run.
03:34 VanessaE real73m37.504s
03:34 VanessaE shit, over an hour?
03:34 VanessaE I wasn't even paying attention
03:35 VanessaE ok well obviously that's wrong :P
03:35 * VanessaE backs it off by 2 commits
03:35 VanessaE (I know one commit behind HEAD is also broken)
03:35 * VanessaE adds her colors.txt, recompiles, and tries again
03:39 Megaf VanessaE; well, Im not going to wait that long
03:40 Megaf we both confirmed the bug
03:40 VanessaE yes, now I'm tracking down the commit that broke it
03:47 VanessaE ok 12 minutes in already.  *cancels and backs down another commit*
03:50 Megaf VanessaE; I found what triggers the bug
03:50 Megaf VanessaE; if you remove those lines "...# CONTENT_..." it takes the normal amount of time to map
03:52 VanessaE I was afraid of that.
03:52 VanessaE I'll still track down the bad commit anyway
04:21 Anchakor joined #minetest-dev
04:22 RealBadAngel hi again
04:23 RealBadAngel was just reading logs, celurean256's idea of drop consolidation is just wrong
04:23 RealBadAngel when one is mining, there are no drops like stacks
04:23 VanessaE wb RBA
04:24 RealBadAngel there are bazilions of single items that player dug one by one
04:24 RealBadAngel idk from where and how he got that idea, weird
04:56 VanessaE RealBadAngel: at any rate it's a nice feature.  Have you given any thought on how to implement that feedback feature we talked about?
05:08 VanessaE sfan5: you bwoke it! :'(
05:08 VanessaE https://github.com/minetest/minetestmapper/commit/84d46ab8eb3205d47605817e390b529398a72ff5
05:08 VanessaE It looks like that ^^^ commit is what broke the mapper per Megaf and I's conversation.  See above where he mentions about the legacy "...# CONTENT_..." entries at the start of the colors.txt file.  Not that I'm against removing them if they aren't used anymore.
05:09 nore joined #minetest-dev
05:14 RealBadAngel VanessaE, feedback feature?
05:15 VanessaE RealBadAngel: the thing with making the newly formed stack grow and then shrink, but stopping at a new, larger size than the default stack, when it collects a new item, plus the brief flash
05:15 VanessaE the thing you proposed?
05:15 RealBadAngel ah that
05:15 RealBadAngel i figured out that animation shall be made in core
05:16 RealBadAngel so atm i made another trick
05:16 VanessaE *nod* it would have to be to have any chance at being decently-performing
05:16 VanessaE what trick?
05:16 RealBadAngel item stack which got increased... jump ;)
05:16 VanessaE oh
05:16 VanessaE I saw that in the video
05:16 VanessaE honestly, I don't like it
05:17 VanessaE it looks strange
05:17 VanessaE feels...glitchy
05:17 VanessaE I think the item should smoothly inflate/deflate like a balloon, but without changing its position at all
05:17 RealBadAngel but that means core
05:18 VanessaE yeah I know
05:18 VanessaE simply directly setting a new size should be doable in Lua though
05:18 RealBadAngel lua is good at cleaning carpets etc, but not at things like that
05:18 VanessaE (without an animation)
05:19 RealBadAngel propably yes, but that code is somehow weird
05:19 RealBadAngel i mean built in item code
05:21 VanessaE right
05:21 RealBadAngel i will try to do that in lua, but chances are next to none
05:22 VanessaE ~tell ShadowNinja Found the bad commit and megaf found a workaround for the mapper problem.  check the #minetest-dev log for 23:35 through 01:08 EDT
05:22 ShadowBot VanessaE: O.K.
05:24 VanessaE RealBadAngel: well wait, I don't mean trying to make it change size over time.  I mean just setting a new size once, when the item is added.
05:25 VanessaE don't try to fake the animation, just change the size to reflect the size of the stack, forget about trying to do any animation until something in-core can be coded.
05:26 RealBadAngel im not quite sure about different sizes
05:26 RealBadAngel have to think about it
07:03 CraigyDavi_ joined #minetest-dev
07:04 Calinou joined #minetest-dev
07:22 darkrose joined #minetest-dev
08:01 cerulean256 joined #minetest-dev
08:21 PenguinDad joined #minetest-dev
08:23 OldCoder joined #minetest-dev
08:24 jin_xi joined #minetest-dev
08:47 alexxss joined #minetest-dev
09:09 arjuna joined #minetest-dev
09:43 sapier joined #minetest-dev
09:49 ImQ009 joined #minetest-dev
09:56 restcoser joined #minetest-dev
10:11 [Pavel_S] joined #minetest-dev
10:31 Jordach joined #minetest-dev
11:11 kahrl joined #minetest-dev
11:12 tomreyn joined #minetest-dev
11:29 RealBadAngel joined #minetest-dev
12:16 jin_xi joined #minetest-dev
12:40 sapier joined #minetest-dev
12:43 PilzAdam joined #minetest-dev
12:43 iqualfragile joined #minetest-dev
12:44 kaeza joined #minetest-dev
12:53 pitriss joined #minetest-dev
12:54 iqualfragile joined #minetest-dev
13:05 kaeza sapier, >+#     "topbottom"  = split screen top boton,
13:05 kaeza (#1269 @ L133)
13:06 kaeza (+minetest.conf.example)
13:06 sapier oh :-) two typos in one word
13:06 kaeza heh :)
13:07 RealBadAngel hi sapier
13:07 RealBadAngel http://i.imgur.com/wSlpLo2.png
13:09 sapier different sizes based uppon amount?
13:10 RealBadAngel yeah
13:10 sapier very interesting :-)
13:10 RealBadAngel code is damn simple, just a few lines
13:11 sapier btw I just finished checking worldcraft fixes ... sadly they didn't fix any performance related things. there are almost only gui things in there. especially the variable font size idea would be worth adding too ... but they did it in a way we can't use much of it
13:12 iqualfragile joined #minetest-dev
13:12 sapier most of the differences are because their code is about end of november 2013
13:13 RealBadAngel huh, now the merging works quite nicely
13:13 sapier I didn't find a exact match, they seem to have merged a few later fixes
13:13 RealBadAngel entities do grow when merged
13:14 Zeitgeist_ joined #minetest-dev
13:25 kaeza comments? https://gist.github.com/kaeza/73c680e11cd8b312280c
13:26 kaeza RealBadAngel, may want to add this to yours ^
13:26 RealBadAngel i was just reading it, i will try how it works
13:26 RealBadAngel may be a good addition to item tweaks
13:27 RealBadAngel i think i will push to git repo initial code in half an hour or so
13:27 RealBadAngel polishing it now
13:34 ^EOF joined #minetest-dev
13:34 ^EOF joined #minetest-dev
13:59 RealBadAngel http://youtu.be/OivpaSPY5v8
14:02 kaeza joined #minetest-dev
14:05 Megaf joined #minetest-dev
14:23 BlockMen joined #minetest-dev
14:31 Bhima_ joined #minetest-dev
14:33 Krock joined #minetest-dev
14:35 Krock sapier! I sent you a forum message and haven't had the idea of joining this channel. (with this stupid message I try to reach an instant acceleration of answer time)
14:36 Krock left #minetest-dev
14:37 sapier lol :-) hmm maybe I should read forum messages more regular
14:43 RealBadAngel BlockMen, http://youtu.be/OivpaSPY5v8
14:45 BlockMen RealBadAngel, nice feature for a mod ;)
14:46 RealBadAngel yup, thats gonna be a mod, polishin it now
14:46 Bhima joined #minetest-dev
14:46 RealBadAngel and core part of technic game too, ofc
14:49 sapier #1313 ... my first pull request not adding a single line ;-)
14:49 ShadowBot https://github.com/minetest/minetest/issues/1313 -- Remove settings not used any longer by sapier
14:51 PenguinDad a pull request for proller ;)
15:00 Megaf Morning all
15:02 Megaf sapier; Hi, what happens if one change those lines?
15:02 sapier nothing
15:09 hmmmm joined #minetest-dev
15:31 rubenwardy joined #minetest-dev
15:45 Calinou joined #minetest-dev
15:47 ShadowNinja joined #minetest-dev
15:53 ShadowNinja Megaf: Actually, I forked from mireq, sfan forked from me, amd minetest forked from sfan.
16:26 Warr1024 joined #minetest-dev
16:32 zat joined #minetest-dev
17:44 diemartin joined #minetest-dev
18:03 Piggybear87 joined #minetest-dev
18:10 RealBadAngel btw, i just got hit by "core" namespace change
18:10 RealBadAngel item tweaks mod doesnt work with that
18:11 RealBadAngel that was definitely "great" idea
18:11 RealBadAngel now one have to write mods for minetest. and for core.
18:17 cheapie joined #minetest-dev
18:21 EvergreenTree joined #minetest-dev
19:01 EvergreenTree A segmentation fault I got while playing minetest + (the game, not the mod):
19:01 EvergreenTree http://paste.ubuntu.com/7485069/
19:02 EvergreenTree It might be something with minetest + but it shouldn't crash the engine
19:03 stormchaser3000 joined #minetest-dev
19:03 PilzAdam EvergreenTree, can you reproduce that with a build with debug symbols compiled in?
19:04 EvergreenTree Alright
19:11 RealBadAngel joined #minetest-dev
19:19 EvergreenTree http://pastebin.ubuntu.com/7485175/
19:19 EvergreenTree with debugging symbols
19:22 EvergreenTree any idea what the problem is?
19:24 sapier "i = 1404952" seems to be a little bit big
19:25 PilzAdam maybe the game is calling calc_lighting with wrong positions
19:25 EvergreenTree also, here is the issue on github: https://github.com/minetest/minetest/issues/1314
19:27 BlockMen PilzAdam, but that should caught somehow to not crash the engine
19:27 BlockMen EvergreenTree, https://github.com/BlockMen/minetest_plus/tree/fix <- happens with this aswell?
19:28 EvergreenTree I'll try it
19:34 EvergreenTree BlockMen: that appears to have fixed it
19:36 BlockMen EvergreenTree, ok
19:37 BlockMen But it should be caught somehow to prevent crashes, so #1314 can stay open IMO
19:37 ShadowBot https://github.com/minetest/minetest/issues/1314 -- Crash caused by calc_lighting()
19:51 smoke_fumus joined #minetest-dev
20:05 iqualfragile joined #minetest-dev
20:21 crazyR Ìû
20:38 werwerwer_ joined #minetest-dev
20:43 diemartin joined #minetest-dev
20:57 Bhima joined #minetest-dev
21:29 ImQ009 joined #minetest-dev
22:20 Bhima joined #minetest-dev
22:20 iqualfragile joined #minetest-dev
22:20 smoke_fumus joined #minetest-dev
22:20 RealBadAngel joined #minetest-dev
22:20 EvergreenTree joined #minetest-dev
22:20 cheapie joined #minetest-dev
22:20 Piggybear87 joined #minetest-dev
22:20 zat joined #minetest-dev
22:20 Warr1024 joined #minetest-dev
22:20 hmmmm joined #minetest-dev
22:20 Zeitgeist_ joined #minetest-dev
22:20 pitriss joined #minetest-dev
22:20 sapier joined #minetest-dev
22:20 jin_xi joined #minetest-dev
22:20 tomreyn joined #minetest-dev
22:20 alexxss joined #minetest-dev
22:20 OldCoder joined #minetest-dev
22:20 cerulean256 joined #minetest-dev
22:20 darkrose joined #minetest-dev
22:20 Anchakor joined #minetest-dev
22:20 Taoki joined #minetest-dev
22:20 blaise joined #minetest-dev
22:20 troller joined #minetest-dev
22:20 grrk-bzzt joined #minetest-dev
22:20 deltib joined #minetest-dev
22:20 khonkhortisan joined #minetest-dev
22:20 Exio4 joined #minetest-dev
22:20 lanxu joined #minetest-dev
22:20 catninja joined #minetest-dev
22:20 ffoxin_work joined #minetest-dev
22:20 VargaD joined #minetest-dev
22:20 SoniEx2 joined #minetest-dev
22:20 crazyR joined #minetest-dev
22:20 Ritchie joined #minetest-dev
22:20 harrison joined #minetest-dev
22:20 puhfa joined #minetest-dev
22:20 Amaz joined #minetest-dev
22:20 dsimon joined #minetest-dev
22:20 Fresh_me_ joined #minetest-dev
22:20 nyuszika7h joined #minetest-dev
22:20 whiskers75 joined #minetest-dev
22:20 Robby joined #minetest-dev
22:20 dzho joined #minetest-dev
22:20 Anchakor_ joined #minetest-dev
22:20 Guest42825 joined #minetest-dev
22:20 mrtux joined #minetest-dev
22:20 Sokomine joined #minetest-dev
22:20 ShadowBot joined #minetest-dev
22:20 daswort joined #minetest-dev
22:20 psedlak joined #minetest-dev
22:20 CiaranG joined #minetest-dev
22:20 eeew joined #minetest-dev
22:20 sfan5 joined #minetest-dev
22:20 Kray joined #minetest-dev
22:25 BlockMen left #minetest-dev
22:26 sapier left #minetest-dev
22:30 Taoki joined #minetest-dev
22:47 kaeza joined #minetest-dev
22:58 cerulean256 left #minetest-dev

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