Minetest logo

IRC log for #minetest-dev, 2014-04-07

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

All times shown according to UTC.

Time Nick Message
00:53 salamanderrake joined #minetest-dev
01:29 Taoki joined #minetest-dev
02:21 ShadowNinja celeron55: Typo in minetestmapper description: Generates aN overview...  (I can't edit it)
04:46 ImQ009 joined #minetest-dev
04:55 darkrose joined #minetest-dev
05:22 kaeza joined #minetest-dev
05:31 cj joined #minetest-dev
06:23 Miner_48er joined #minetest-dev
08:19 Jordach joined #minetest-dev
08:25 kaeza joined #minetest-dev
08:56 RealBadAngel joined #minetest-dev
08:56 ImQ009 joined #minetest-dev
08:57 RealBadAngel hi
08:58 RealBadAngel celeron55, i am aware of the water lighting problem, i will pull fix for it in a day or two
08:59 RealBadAngel strange thing it was (and is) workin correctly here
09:00 RealBadAngel and i wasnt noticed on such problems before, while many were testing that pull
09:01 RealBadAngel kahrl, online?
09:49 Calinou joined #minetest-dev
10:03 Calinou https://github.com/minetest/minetest/issues/1044 why closed?
10:07 celeron55 hard to know because it doesn't show who closed it
10:10 proller PilzAdam closed this 4 months ago
10:11 proller гззук
10:11 proller upper
10:11 celeron55 oh it's in the middle of it
10:12 celeron55 i think that could be reopened
10:55 proller joined #minetest-dev
10:59 smoke_fumus guys
10:59 smoke_fumus question
11:00 smoke_fumus what's most efficient way to push changes which are shared between multiple chunks?
11:27 werwerwer joined #minetest-dev
11:33 Shardvex joined #minetest-dev
11:34 e1z0 joined #minetest-dev
11:44 Shardvex joined #minetest-dev
11:58 ImQ009 joined #minetest-dev
12:06 celeron55 smoke_fumus: from where?
12:30 PilzAdam joined #minetest-dev
12:38 smoke_fumus celeron55: generic execution in code, not related to minetest. i am pushing change in 2d chunk divided by marching cubes algorythm. on wipe it should set x,y; x+1,y; x,y+1; x+1,y+1 to  zero. How to determine which chunk was target x,y in?
12:38 smoke_fumus i just need a formula x.x
12:39 smoke_fumus *marching squares
12:40 celeron55 uh, you basically divide the coordinates by the chunk size, and make sure that it behaves sanely when it crosses zero
12:41 celeron55 like this https://github.com/minetest/minetest​/blob/master/src/util/numeric.h#L48
12:42 celeron55 then you have the chunk's position and probably have some way to get the chunk based on that
12:42 smoke_fumus hmhmhmhmhm
12:42 celeron55 if not, then you're doing something else than i think you are
12:46 smoke_fumus yes you are right. i found that it sorta works - just not pushing textre on the result
12:46 smoke_fumus celeron55: http://www.youtube.com/watch?v=b​fxsLD4Xzc0&feature=youtu.be
12:46 smoke_fumus here's where i am at currently
12:46 smoke_fumus or rather was 2 hours ago
12:53 smoke_fumus marching squares is rather simple algorythm which in layments terms - takes "let's split a round-edged square on 9 slices" idea and makes it into "4 rather than 9")
12:54 smoke_fumus although, i think after all this idiocity i might just return to 9 splices
13:03 hmmmm joined #minetest-dev
13:34 troller joined #minetest-dev
13:42 proller joined #minetest-dev
14:08 Garmine joined #minetest-dev
14:19 proller joined #minetest-dev
14:33 Exio4 joined #minetest-dev
14:41 zat joined #minetest-dev
14:44 werwerwer joined #minetest-dev
14:45 rsiska joined #minetest-dev
14:50 Jordach joined #minetest-dev
14:53 rubenwardy joined #minetest-dev
14:53 rubenwardy joined #minetest-dev
15:23 Calinou joined #minetest-dev
15:49 sapier joined #minetest-dev
15:54 sapier ~ iqualfragile what reason do you have to believe this might be less performant then the original one? shift operations are way more efficient on any architecture I know then multiplications are. I'm not even sure compiler can really optimize this as bool true castet to (int) may not even be 1 but could be 0xFFFFFFFF on some compilers too. And yes I know "standard" defines true to be one ... but especially for this one some (smart) comp
15:57 tomreyn joined #minetest-dev
15:58 sapier hmm shadowbot is offline?
15:58 sapier no
15:58 VanessaE you forgot the ~tell part.
15:58 sapier argh
15:58 sapier ~ tell iqualfragile please read log ;-)
15:58 ShadowBot sapier: O.K.
15:59 VanessaE and your message got cut off.
15:59 sapier really?
15:59 VanessaE "[...] but especially for this one some (smart) compi--"
15:59 sapier I didn't even know about a size limit on irc
15:59 sapier ... compilers ignore it for optimization reasons.
16:00 sapier actually those compilers aren't really smart but a pain in the ass ... but that doesn't change anything
16:09 luizrpgluiz joined #minetest-dev
16:09 luizrpgluiz hi
16:09 luizrpgluiz hi
16:12 luizrpgluiz left #minetest-dev
16:16 proller sapier, why you make lot of changes about nothing and 0 changes about game? you newer start game?
16:16 PenguinDad joined #minetest-dev
16:17 sapier because someone has to do the infrastructure work not giving any credits but providing the base for all of your nice and shiny "I'm so cool I added feature X"
16:17 sapier "your" not beeing proller but generally spoken
16:19 proller now game have lot of problems , more important than main menu and slow * operayion
16:19 sapier what problems are you talking about?
16:21 proller try to join in game, you can find 10+ in 5 min
16:21 proller but you newer join..
16:21 sapier write an issue and I may look at it but I'm not the only developer here so why do you expect ME to fix anything?
16:22 proller unstable YOUR connection with lot of added bugs ?
16:22 sapier that's your claim without any proof
16:23 sapier almost all of those (not obvious) bugs I've been accused to be responsible for haven't been mine but just got more obvious due to other things not beeing broken any longer
16:25 sapier Don't get me wrong proller if you know about bugs that need to be fixed tell me.
16:25 * VanessaE looks at the master server list and notes that her up-to-date-and-patched-with-sapier's-code "almost vanilla" server is inexplicably at the top again
16:26 proller lot of "unknown" players in first report after join
16:26 VanessaE whatever the problems are, proller, they aren't keeping people from *playing*
16:26 proller lot of draw hud errors
16:26 scientistnik joined #minetest-dev
16:27 proller sapier, you can find most of bugs if you try join to any server
16:27 sapier well proller none of my hud changes are merged by now so this is not my fault for sure ;-)
16:27 proller its init sequence problem
16:27 proller appear after your changes
16:28 sapier of course, joining with outdated client and broken media on server is a problem
16:28 sapier fix your broken media and the problem will occur only once
16:28 VanessaE sapier: 1143 needs to go in by now I think.
16:28 VanessaE it's been long enough hasn't it?
16:28 proller its happen on singleplayer
16:29 sapier as I said broken media
16:30 proller broken media in singleplayer ?
16:30 sapier you're right VanessaE that would fix the issue ... but for what I remember there have been objections
16:30 VanessaE sapier: I thought the delay was only because you were worried about the whole "do we just break it all for 0.4.10" issue?
16:31 sapier nope problem is 1143 would fix it but is opposit of poll result
16:31 VanessaE hm
16:31 sapier according to pull result we should break compatibility while 1143 has compatibility support
16:31 VanessaE maybe just ignore the poll then.
16:32 sapier proller could you try 1143?
16:33 proller nobody need in _jitter and _rtt info in lua, except you
16:34 sapier Proller jitter and rtt are preparations for benchmarking later protocol replacements
16:34 proller in lua ???
16:34 sapier of course or do you wanna write a c++ plugin interface to create a benchmark?
16:34 sfan5 https://github.com/minetest/minetest_game/pull/250
16:35 sfan5 ^ comments please
16:35 sapier I want to do real benchmarks not some theoretical maximum values
16:35 VanessaE sfan5: +1
16:36 sapier I like the idea ... but not the texture ... but I'm not sure if my taste is suitable for deciding things like that ... I guess not :-)
16:36 proller but minetest_game is frozen and PilzAdam regect any change to it ?
16:37 sfan5 PilzAdam: https://github.com/minetest/minetest_game/pull/250 opinion?
16:37 PilzAdam sfan5, basically what proller said
16:39 PilzAdam someone could release a desert mod that adds all kind of cool stuff to deserts
16:39 proller good progress >:)
16:39 VanessaE such a mod exists.
16:40 VanessaE https://forum.minetest.net/viewtopic.php?id=3761
16:40 sapier well I'm not sure if this really is something for a mod, it's basicaly a symmetry thing, :-)
16:41 sfan5 PilzAdam: I suggest you to read https://forum.minetest.net/vie​wtopic.php?pid=136403#p136403
16:41 PilzAdam sfan5, I already read that
16:41 sfan5 the last sentence too?
16:42 proller good pull for freeminer default game ;_
16:44 PilzAdam sfan5, the goal is to get alternative games
16:44 VanessaE bullshit
16:45 PilzAdam merging the desert cobble pull request would conflict / break the mod VanessaE linked
16:45 VanessaE those alternative games have had many months now to be proposes and added.
16:45 VanessaE where are tey?
16:45 sfan5 We'll get nowhere with Minetest if you keep hindering innovation like this
16:45 VanessaE they*
16:45 sapier ~ tell iqualfragile hmmm well looking at assembler code at least gcc is capable of optimizing so new and old version look quite similar ... but ... if I didn't remove the temporary variable the new version would become faster then the old one ... guess I need to do another fix.
16:45 ShadowBot sapier: O.K.
16:48 PilzAdam sfan5, what exactly is wrong with installing mods?
16:48 sapier is mmdb still broken?
16:48 sapier if yes ... that's wrong ;-)
16:49 sapier back to #1143 is making versions available to debug builds only enough to avoid modders from using those information in regular mods or not?
16:49 ShadowBot https://github.com/minetest/minetest/issues/1143 -- Protocol version <= 22 compatibility mode + clean new client init by sapier
16:51 sapier can everyone please check 1143? Proller is right, the network fixes need to be completed and that one IS the missing part. It's crazy to let it wait for some minor issues that could be easyly fixed.
16:53 sapier especially as we've got a working android port now we can point tablet users to
16:53 sfan5 PilzAdam: how long to you want to keep minetest_game the same?
16:54 sfan5 people like it better when they have new content by default than having to install mods
16:54 sfan5 also desert cobble is more for symmetry with cobble
16:54 sapier what about merging minetest_game to minimal?
16:54 sfan5 not an entirely new block
16:54 sfan5 sapier: no, why?
16:54 OldCoder joined #minetest-dev
16:55 sapier this way minimal could be base (which is already defacto standard) and game could switch to a more dynamic development state
16:55 sapier it's just an idea
16:55 VanessaE well I need not check 1143, I know it works :P
16:56 VanessaE sapier: no, keep minimal the way it is
16:56 VanessaE that really *should* remain frozen
16:56 VanessaE but minetest_game should be allowed to grow
16:56 sapier ok another idea creating minetest_evolution ? :-)
16:57 VanessaE if it breaks a seldom-used mod here or there, oh well -- just so long as doing so is in the spirit of incorporating THAT mod's functionality.
16:57 PilzAdam sfan5, we had this discussion about minetest_game several times now, please read the logs
16:58 sfan5 link?
16:58 sapier well PilzAdam I understand you but imho it can't stay frozen forever
16:58 sapier at some point in time it has to be updated
16:58 rubenwardy !title https://www.youtube.com/watch?v=fSp68ySdy1E
16:58 rubenwardy oh, old
16:59 sfan5 lol
17:00 rubenwardy Googling "minetest worst mod" comes up with technic for me.
17:00 rubenwardy and I agree
17:00 rubenwardy :P
17:00 rubenwardy Nah
17:00 VanessaE rubenwardy: um, offtopic.....
17:00 rubenwardy !title https://www.youtube.com/watch?v=5P-o5wnwBgU
17:01 sfan5 rubenwardy: did you notice MinetestBot is not here?
17:01 rubenwardy no
17:03 sfan5 rubenwardy: videos that begin with a person critcizing how the clouds reset back one in a while like that is the most bad bug ever are not worth to be watched
17:03 sfan5 </offtopic>
17:03 sfan5 PilzAdam: link?
17:03 VanessaE PilzAdam: minetest_game needs to grow and develop.  Enough of this freeze bullshit.  people aren't gonna balk at new stuff being added, they're only gonna complain if existing stuff is changed too much.
17:09 rubenwardy +1
17:09 sfan5 +1
17:10 rubenwardy But only if it is logical
17:10 rubenwardy oh damn
17:10 rubenwardy I meant to post those videos in #minetest
17:11 rubenwardy Logical not meaning realistic, but meaning not adding stuff like pokemon.
17:11 VanessaE right
17:12 proller and remove nyancat to mod
17:12 sfan5 no not the nyan cat!
17:12 VanessaE fuck no!
17:12 VanessaE blasphemer!
17:12 VanessaE heretic!
17:12 rubenwardy Don't remove from game
17:12 VanessaE :)
17:12 proller why nyan better than pokemon >
17:12 VanessaE why mese, then?
17:12 rubenwardy But moving to another mod is fine, in the game
17:13 VanessaE why wooden picks?
17:13 VanessaE because we already have them
17:13 proller wooden pick - okay
17:13 rubenwardy proller, heritage
17:13 rubenwardy keep our cats!
17:13 VanessaE rule #1 of a game:  don't remove content.  ever.  unless there is a really good G*d damn reason for doing so.
17:14 VanessaE the existence of an illogical item in a game does not give license to add more illogical elements, but it also doesn't give license to remove that element either.
17:15 sfan5 PilzAdam does not seem to be there
17:15 VanessaE (on the other hand, I've advocated for removal of wooden picks because they're just stupid :) )
17:16 VanessaE he's ignoring us now, sfan5
17:16 proller how to dig first stone ?
17:16 sfan5 ^
17:16 VanessaE proller: I had proposed a while back to have a replacement tool that can be crafted from sticks and loose stones lying on the ground, similar to what's done in Realtest.
17:17 VanessaE dirt could very rarely drop a loose stone to be used to make said tool, as well
17:18 VanessaE that way it works on old maps
17:18 rubenwardy How about dirt pick axes?
17:18 proller shit pick
17:18 VanessaE a dirt pick? er.. no
17:18 rubenwardy I troll
17:18 proller eat apple, produce shit, make pick
17:19 VanessaE (apropos:  https://forum.minetest.net/viewtopic.php?id=4129 )
17:19 VanessaE rubenwardy:  "There's a mod for that"
17:19 rubenwardy heh
17:19 rubenwardy Since everyone is using Windows 8 now, shall we rename mods to 'Apps'?
17:20 VanessaE no.
17:20 VanessaE but getting back to the point...
17:20 VanessaE if you want to add something to the damn game, just add it.
17:20 VanessaE I see no legit reason to block.
17:21 rubenwardy How about adding a nodebox for bones?
17:21 VanessaE purpose of same?
17:21 rubenwardy The bone texture is currently ugly :P
17:21 rubenwardy (IMO)
17:22 proller +1
17:22 VanessaE agreed, but how would a nodebox change this?  you're thinking of doing this from the standpoint of a modeled pile of bones then?
17:22 rubenwardy yeah
17:22 VanessaE mmm
17:22 VanessaE well that would certainly look nicer
17:23 VanessaE wouldn't add a great deal of load to the rendering engine either, since bones are so infrequently found in the world anyway
17:23 VanessaE but I can't see that going over too well because it would need a metric assload of very small nodeboxes to look good.
17:23 VanessaE such a thing really cries out for a proper 3d model
17:29 proller joined #minetest-dev
17:38 Zeitgeist_ joined #minetest-dev
17:38 Zeitgeist_ joined #minetest-dev
17:50 sapier and some crazy one would create a bone pile from hundreds of bone nodes and then loudly complain about lag :-)
17:51 VanessaE of course.
18:02 sfan5 why don't we have smooth lightning for nodeboxes yet?
18:03 sapier because smooth lightning is already a performance hog ... ok ok guess most people aren't aware of this
18:03 sapier more correct reason is most likely because of manual light calculation for nodeboxes is a quite complex task
18:04 sapier I'd not be surprised if this would make minetest unplayable on quite a lot of machines
18:04 VanessaE sapier: actually smooth lighting doesn't seem to affect my machine versus "simple" lighting
18:05 sapier did you try on arm cpu too? ;-)
18:05 VanessaE who cares about arm :P
18:05 sapier any android user
18:05 VanessaE (really though, for such machines, there's the 'simple' lighting mode)
18:06 VanessaE that does not, however, explain why things like *signs* also have bad lighting.....
18:06 sapier well if someone wants to implement the quite complex task of manually calculating model lighting we'll see how much of performance this will cost
18:07 sapier signs are a bad example as they're not much more then a dissorted node  ... others like fences are way more complex ... and even fences are simple nodeboxes
18:08 VanessaE well I guess RBA plans on doing hardware-assisted smooth lighting some time soon
18:08 sapier I'm looking forward to it
18:09 sapier as I said if someone implements smooth lighting for nodeboxes I'd be quite positive about merging (as long as it's configurable) ... but for what I know light calculation this is not a easy task
18:10 sapier but true I don't know light calculation very well
18:10 Jordach anyone noticed an inability to generate stone in MGV7 using custom nodes (it appears to generate unknown nodes)
18:11 Jordach wait
18:12 sapier guess that's something to fix prior v7 can be made default
18:12 Jordach sapier, apparently i forgot to enclose texture.png within curly brackets
18:13 Jordach which is interesting as Minetest only takes a table format for images, yet "texture.png", doesn't raise an error
18:14 Jordach http://paste.ubuntu.com/7218156/
18:14 sapier I remember looking for this error for hours too :-)
18:16 VanessaE seems to me that the engine should be throwing in a proper "dummy" texture in that case.
18:20 sapier isn't unknown our dummy texture?
18:21 VanessaE no, a dummy texture is usually a solid color
18:21 VanessaE some random, often randomly-flashing color
18:22 sapier ok
18:35 rsiska joined #minetest-dev
18:55 Shardvex joined #minetest-dev
19:33 sapier as noone wants to comment about #1143 I'm gonna delay merge till tomorrow, if still noone has valid reasons not to merge it I'm gonna merge it
19:33 ShadowBot https://github.com/minetest/minetest/issues/1143 -- Protocol version <= 22 compatibility mode + clean new client init by sapier
19:34 sapier ok noone except vanessae
19:34 zat joined #minetest-dev
19:41 * VanessaE hides.
20:08 scientistnik joined #minetest-dev
20:09 exoplanet joined #minetest-dev
20:11 Jordach anyone noticed a bug where fill_ratio values less than 0.0155 will not display the decoration
20:16 khonkhortisan joined #minetest-dev
20:16 sapier 0.01?
20:16 Jordach sapier, same there too
20:16 sapier that's most likely a pure rounding effect
20:17 Jordach it looks *like* sidelen has something to do with it, but not really sure
20:17 sapier how many pixels are left for 0.015? ;-)
20:17 Jordach 0.016 still is too many :PO
20:18 sapier is this a scaling value ? maybe for a 16 pixel texture? ;-)
20:18 Jordach sapier, decorations using mgv7
20:19 sapier ok then maybe nodes per block?
20:19 Jordach sapier, https://cdn.mediacru.sh/pMEe6zgoW1DJ.png
20:19 Jordach 1.0 when sidelen is 4
20:20 sapier ok
20:21 sapier well doesn't change the fact small floats may be rounded in this case ... or a classical float int calculation missmatch bug
20:23 Jordach up or down, that is the question
20:24 sapier open up debugger and check the calculated values ;-)
20:34 nyuszika7h joined #minetest-dev
21:22 iqualfragile joined #minetest-dev
21:25 _BrandonReese joined #minetest-dev
21:27 iqualfragile !tell sapier I do not think the previous version would be faster, i just think it might be wasted effort and could lead to less readability
21:27 ShadowBot iqualfragile: O.K.
21:28 iqualfragile sapier: oh, you are still here
21:29 sapier usually abusing multiplications for shift operations is not considered "more readable" ;-)
21:29 sapier most important issue is a silly compiler could really do a multiplication there which is quite slow compared to a shift
21:40 Exio4 joined #minetest-dev
21:42 e1z0 joined #minetest-dev
21:42 proller joined #minetest-dev
21:51 hmmmm people seem to assume that something that's inferior and slower is more readable
21:51 hmmmm how on earth are math tricks more readable than bit operations, which is what's ACTUALLY intended
21:52 hmmmm no.  you're wrong.  stop it.  all the people who think that "optimize" means "less readable" are plain wrong, there's no other way to put it.  you all need to reexamine your prejudices.
21:54 sapier hmmmm whom are you talking to right now? :-)
21:55 VanessaE hmmmm: maybe leftovers from the old OLD days when you really DID have to hand-optimize to the point that the code was horribly unreadable?
21:56 sapier well on performance critical things looking at assembler code is still necessary ;-)
21:57 iqualfragile hmmmm: you are right, for this example it does not really matter
21:58 iqualfragile or, well, its actually easier understandable this way, but it should not be done on any occasion
21:58 sapier it does ... in this location there ain't a number calculated but a bitmap created where each bit is independent and not a number ;-)
22:00 e1z0 left #minetest-dev
22:00 sapier left #minetest-dev
22:02 iqualfragile_ joined #minetest-dev
22:03 rsiska joined #minetest-dev
22:05 AndChat-93396 joined #minetest-dev
22:11 Exio4 joined #minetest-dev
22:14 paramat joined #minetest-dev
22:17 paramat hi, i know minetest game is frozen, but this is an exception, the wieldhand should match the character skin =) https://github.com/minetes​t/minetest_game/issues/249
22:28 Taoki joined #minetest-dev
22:30 eeew joined #minetest-dev

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