Time  Nick           Message
03:30 VanessaE       celeron55: is there any movement on that damn EmergeThread bug?  It's making it all but impossible to run a decent server now. :-/
06:12 hmmmm          eh, no movement, sorry
06:12 hmmmm          he made the fix for it, it's his problem now ;)
06:54 celeron55      VanessaE: the same reason for any complicated bug
06:55 celeron55      debugging it is almost like killing oneself
06:55 celeron55      nobody wants to do it as a hobby
06:55 celeron55      s/for/as for/
06:58 celeron55      VanessaE: if you want to help, you could try doing a slow and painful bisect
06:58 celeron55      dunno if there is anything to be found that way, but as the problem hasn't existed always, it's likely
07:13 hmmmm          i should probably add that it'll disappear when I add the multithread support
07:14 hmmmm          ...and it's likely that new bugs will appear in its place!
07:43 hmmmm          ugh, I just realized why mapgen_ aliases can't be used with the biome definitions in lua... nodedef aliases are updated after the lua scripts are initially run
07:44 hmmmm          maybe someday in the future, nodedef aliases will be updated as soon as they are registered
14:18 PilzAdam       thexyz, https://github.com/PilzAdam/minetest/commit/a27cdf16f40e3a3f4772b944ead8f1e4e5618359
14:19 sapier         @celeron55 button handling isn't working for lua formspecs. I've simplified them too much. I've created a fix to make them work again but it'll require a new protocol version and new message number. If not using a new number it'll break compatibility.
14:20 sapier         could anyone plz review this commit?: https://github.com/sapier/minetest/commit/e159443daa5fd328c9f92100b51d619dc5ed8d0a
14:22 PilzAdam       why do you put so many whitespaces in front of the "="s?
14:23 sapier         its to make assignments in one line to make them more readable
14:23 sapier         I'll change
14:25 sapier         oops .. no its a typo
14:26 PilzAdam       just one " = " is enough
14:26 sapier         are tabs allowed before =?
14:27 PilzAdam       never saw this before
14:27 sapier         e.g. scriptapi.cpp changes those whitespaces are tabs not blanks
14:28 sapier         in order to make = verticaly aligned for those 3 values
14:28 PilzAdam       http://www.minetest.net/wiki/doku.php?id=contrib see "Coding Style"
14:28 sapier         hmm seems to be not common coding style in minetest I'll change it too
14:30 VanessaE       PilzAdam: please stop being so anal about whitespace
14:30 VanessaE       For some of us, it greatly improves readability to add a few spaces here or there
14:31 PilzAdam       it has to fit Minetests conventions
14:32 sapier         @vanessae I'm with pilzadam whitespaces are a big problem if more than one person is working on code
14:33 darkrose       this isn't twitter, you don't need @ signs, and analness about whitespace is good
14:33 sapier         i think coding style example needs some additions, the only occurance of multiple assignments is somehow strange as it's accidenally end aligned
14:34 PilzAdam       "Refer to existing code written by celeron55 in case of stuff not shown here. If that does not help, refer to Irrlicht's code."
14:34 sapier         I don't even have twitter account ;-)
14:35 VanessaE       I can understand trying to stick to the conventions used in the main code, but there comes a point when you're being too picky.
14:35 VanessaE       that's all I'm saying.
14:39 sapier         is there something like githubs commit view for local usage?
14:39 PilzAdam       gitk
14:39 PilzAdam       what distro do you use?
14:39 sapier         debian
14:39 PilzAdam       apt-get install gitk should work
14:40 PilzAdam       gitk without arguments show the current branch history
14:41 darkrose       `git diff <commit>'
14:41 sapier         I know ;-) bit it's somehow ... inconveniant
14:42 PilzAdam       git diff highlights whitespaces at the end of lines wich is pretty useful
14:42 sapier         ok I think i got the whitespaces correct this time
14:43 sapier         yes but githubs browser based diff viewer is much more comfortable
14:44 PilzAdam       gitk is a nice GUI
14:45 sapier         tcl is just a little bit oldfashioned
14:45 darkrose       http://www.linuxjournal.com/content/convert-diff-output-colorized-html <- send output of git diff to that, then open in browser
14:46 sapier         comforable ;-)
15:02 thexyz         gitg
15:07 thexyz         PilzAdam: doesn't work at all
15:08 PilzAdam       what doesnt work?
15:09 thexyz         hm.. not at all, it works at random moments of time
15:09 thexyz         screen flashes instead of shaking
15:10 PilzAdam       the flashing doesnt shake
15:10 PilzAdam       shaking is just an addition
15:10 celeron55      sapier: hmm oh... yeah, indeed, there isn't a way to set callbacks to such temporary forms 8)
15:10 thexyz         oh, I see, but it's hard (for me) to notice it
15:11 PilzAdam       yea, maybe make the red less intense
15:11 thexyz         btw, shouldn't lava be non-renewable?
15:11 PilzAdam       I noticed that you see nothing when its dark and you get the maximum alpha
15:12 thexyz         yes
15:12 PilzAdam       thexyz, its hard to find consensus here
15:12 PilzAdam       (about the lava thing)
15:13 celeron55      why would lava be and water not be; it's inconsistent
15:13 PilzAdam       lava is rare and shouldnt be that easy to get
15:14 celeron55      hmm, well, maybe
15:15 celeron55      sapier: your patch is interesting
15:16 sapier         yes formname was already existing in code for future usage so why not use it
15:16 PilzAdam       thexyz, https://github.com/PilzAdam/minetest/commit/d2b1210376f6cc52fe031ac41d3bf54720ebd356
15:17 celeron55      sapier: i think there must be an enforced modname:formname naming convention because they share the namespace
15:18 sapier         I agree ... I'll have a look how to do this and push again
15:18 celeron55      also, don't change the protocol version, it doesn't matter for changes with only a day of time and no releases in between
15:18 sapier         ok don't skip number too? if reusing it clients may crash
15:19 celeron55      doesn't matter
15:19 sapier         ok
15:19 celeron55      if there was a release, or multiple days, before the last PROTOCOL_VERSION upping, then it would matter; not now
15:20 celeron55      s/before/after/
15:20 celeron55      sapier: fix also the "create_detached_formspec_raw" comment above l_show_formspec
15:21 thexyz         PilzAdam: seems ok
15:22 celeron55      hmm, somebody has removed the modname prefix checker from scriptapi.cpp
15:22 celeron55      it exists only in builtin/misc_register.lua now
15:22 PilzAdam       thexyz, both, the shaking and the red screen?
15:22 thexyz         yes
15:23 celeron55      sapier: well, rename minetest.show_formspec to minetest.show_formspec_raw and add a small wrapper that checks it in builtin/misc_register.lua
15:23 sapier         I was just about to ask :-)
15:23 celeron55      it's just like everything else that needs to check the prefix then
15:23 PilzAdam       celeron55, do you have any comments on this 2 commits: https://github.com/PilzAdam/minetest/commits/camera ?
15:24 celeron55      PilzAdam: i'm too lazy to check them
15:24 PilzAdam       okay, I just push it
15:25 sapier         is misc_register really correct place for show_formspec function?
15:25 celeron55      PilzAdam: you can do that if somebody else has agreed to it 8)
15:25 PilzAdam       thexyz did
15:25 VanessaE       damn it and I *just* pulled and built not 5 minutes ago. :)
15:26 celeron55      sapier: well, it's kind of odd but it doesn't make any sense to make a new file either or anything
15:26 sapier         misc.lua maybe?
15:26 celeron55      no
15:27 celeron55      it's worse :P
15:27 sapier         :-) ok I'm adding it to misc_register
15:30 celeron55      actually, it kind of could belong in builtin/detached_inventory.lua
15:31 thexyz         celeron55: anything about my STL branch?
15:31 celeron55      just add "minetest.check_modname_prefix = check_modname_prefix" after "local function check_modname_prefix" in builtin/misc_register.lua and put it there
15:31 sapier         ok
15:31 thexyz         https://github.com/minetest/minetest/commits/stl
15:32 celeron55      thexyz: umm... i'm not going to go through that :-D
15:32 thexyz         why?
15:33 celeron55      there's like 100000 lines of diff
15:33 celeron55      maybe we should make a competition of finding problems in it
15:33 PilzAdam       heh
15:34 thexyz         so, merge and wait for shitstorms?)
15:34 PilzAdam       should work
15:34 PilzAdam       have you tested it on windows?
15:35 celeron55      definitely needs a test on mingw and msvc first
15:35 celeron55      or, well, mingw is the same as gcc anyway
15:36 PilzAdam       it works on ubunut 12.10
15:37 PilzAdam       *ubuntu
15:37 sapier         wow :-) you're marvelous i didn't even get my mouse working in ubuntu 12.10 :-)
15:37 thexyz         huh
15:38 thexyz         seriously, PilzAdam, why shouldn't it?
15:38 PilzAdam       dunno
15:38 PilzAdam       just some random testing ;-)
15:47 sapier         is there any legit reason for get_current_modname not to return a value?
15:48 celeron55      it will return non-nil only at the time of the initial run
15:49 sapier         ok then prefix checking can't be done this way
15:49 celeron55      and in that run, it returns "__builtin" for builtin/*.lua sutff
15:50 celeron55      hmm... indeed
15:50 sapier         Im calling show_formspec from chatcommand resulting in an nil concatenation error as get_current_modname returns nil
15:51 celeron55      yeah, i didn't realize this 8)
15:51 celeron55      let's just mark it as a suggested convention in lua_api.txt
15:52 sapier         ok I'm reverting the lua changes too
15:53 celeron55      there is no way to know what mod is responsible for some running code in any later part of execution, because everything is callback-driven
15:53 celeron55      (except for some cases, but this is not one)
15:54 sapier         I've already missed this feature for profiling
15:55 celeron55      there are some things that could be done for profiling, but they are not useful for doing stuff like this
16:06 sapier         celeron there's already a notice that any names should be modname:name do we need an additional one?
16:07 sapier         wait its "registred names" only
16:20 sapier         https://github.com/sapier/minetest/commit/91ef4df8213e3c99d2d8353a70d7d285dc1fa9d7
16:45 celeron55      nah, don't add it at that part; just add "^ formname should follow "modname:<whatever>" naming convention"
16:46 sapier1        ok
16:47 celeron55      oh also, you failed the protocol message definition in clientserver.h
16:47 RealBadAngel   hi all
16:47 celeron55      there is supposed to be a "u16 len" before u8[len] formname too
16:48 RealBadAngel   celeron55, i dont want to argue again bout cave gen but i think you shall see this screenshot http://realbadangel.pl/clouds1.png
16:48 celeron55      RealBadAngel: i have already seen it many times
16:49 RealBadAngel   you still thinkin im "plain against" cave gen? :)
16:49 celeron55      yes?
16:49 celeron55      i would have noticed you if i thought otherwise
16:49 celeron55      noted*
16:50 celeron55      umm... whatever word that is supposed to be
16:50 RealBadAngel   it shall be fixed just imho, thats all
16:50 RealBadAngel   if you got an idea how, tell. i will code
16:50 celeron55      it could be fixed, but not so brutally as you are doing it
16:50 celeron55      but i am not sure how
16:51 RealBadAngel   lemme explain i thought it would be a quick fix before mapgen v7
16:51 RealBadAngel   but then hmmm told me he wants to leave cave gen intact
16:51 celeron55      that doesn't make any sense
16:51 celeron55      what is a good fix good for if it would be replaced anyway?
16:51 RealBadAngel   so i agree, for a long run, such fixes are not good
16:51 celeron55      eh
16:51 celeron55      bad fix*
16:52 celeron55      and then if it isn't replaced, then it is extra bad
16:52 celeron55      anyway
16:53 RealBadAngel   it just means cave gen shall be fixed with style, not hacks as i tried to do
16:54 celeron55      there is no good solution because of a variety of reasons, i just thought of one and just deleted it because it was actually bad
16:54 RealBadAngel   well, theres one good point to start with
16:55 celeron55      as far as i understand, the problem is that the cave generator can mess up trees when it edits ground slightly into nearby generated chunks just over the ground level
16:55 RealBadAngel   compute avg height in a block based on perlin and some points in cross
16:55 celeron55      because trees aren't generated to the main chunk before the cave generator is done
16:56 RealBadAngel   so allow cavegen replace only air/stone/dirt
16:56 RealBadAngel   not the added contens
16:56 celeron55      so the solution would be to restrict it to not touching nearby chunks trees when it is in the area a bit over ground height where it matters for trees
16:57 celeron55      nearby chunks or trees*
16:57 celeron55      either or, any of those would work for you
16:57 RealBadAngel   maybe just two groups of nodes
16:57 celeron55      no groups
16:58 celeron55      the mapgen already knows what is a tree node and what is not
16:58 celeron55      it knows because it generates trees
16:58 RealBadAngel   how?
16:58 celeron55      how could it generate a tree if it didn't know what is a tree and what is not 8)
16:59 RealBadAngel   we could have many kind of trees, plant life etc
16:59 RealBadAngel   even maybe later some dungeons, other structures
17:00 RealBadAngel   that if one piece of code uses some nodes doesnt mean other part is aware of it
17:00 celeron55      well, then make it such way that it will touch only things it knows is ground
17:00 celeron55      that it knows 100% because it is the one who makes all ground
17:00 sapier1        https://github.com/sapier/minetest/commit/40578c7fb1cf5480957f77bdc53e87ad6d5ca700
17:00 RealBadAngel   most simple? define group mapgen_basic or somethin like that
17:01 celeron55      never
17:01 hmmmm          do a check to see if it's mapgen_stone or mapgen_dirt before replacing the block 8)
17:01 celeron55      we don't need any groups
17:01 RealBadAngel   so we do need multiple checks
17:01 PilzAdam       hmmmm, also mapgen_desert_{sand, stone}
17:01 hmmmm          ya
17:01 celeron55      yes, just an std::set or something
17:01 RealBadAngel   for air/stone/dirt/sand/desert sand/desert stone
17:01 RealBadAngel   like this
17:02 celeron55      to which you put the stuff the mapgen knows at the start of the mapgen
17:02 RealBadAngel   i agree, but then imagine biomes introduced
17:02 celeron55      the mapgen knows of them
17:02 RealBadAngel   with own stuff
17:03 celeron55      mapgen always makes the ground
17:03 PilzAdam       so in mapgen v7 it needs to check every node that is passed to every biome=
17:03 PilzAdam       *?
17:03 hmmmm          bool MapgenV6::isGroundNode(content_t cid) { return m_groundnodes.find(cid) != NULL; }  would work for Mapgen V6 fine
17:03 PilzAdam       (as ground node)
17:03 hmmmm          MapgenV7 on the other hand will definitely do something smarter
17:04 hmmmm          MapgenV6 already does some relatively dumb stuff that's needlessly complex so I say let it burn
17:04 hmmmm          muwahah
17:04 RealBadAngel   :)
17:05 celeron55      as long as we don't have any viable alternative to it, maintaining it well is important
17:05 hmmmm          but mapgen v7 is my problem
17:06 RealBadAngel   and since all miners loves caves its our common problem ;)
17:06 RealBadAngel   just got an idea
17:06 hmmmm          I am going to figure out how to handle the boundary size better
17:07 RealBadAngel   cave gen already tries to mark air as cave air
17:07 hmmmm          really that's the core reason for the unlit tree problem, and the problem with the trees getting cut into
17:07 RealBadAngel   but its not workin right
17:07 celeron55      hmmmm: if you want to look into generating caves that are consistently continuable based on the seed with only noise lookup, see 0.3's cave generator
17:08 RealBadAngel   what if we could introduce to kinds of air simply?
17:08 RealBadAngel   *two
17:08 hmmmm          I probably will, but I don't remember 0.3's cave generator well, how nicely were the caves it generated?
17:08 celeron55      but there will always be such problems if the world is generated in pieces, UNLESS we do the exact same thing as minecraft does
17:09 hmmmm          s/nicely/nice/
17:09 RealBadAngel   when cavegen will reach regular air it will be forced to stop
17:09 celeron55      that is, use very solid seeding, generate everything near the generated area that could touch it and discard everything that does not go into the generated area
17:09 celeron55      it's well multithreadable and needs to touch in only the generated area
17:09 celeron55      there is a bit but to that though
17:09 celeron55      lighting.
17:09 celeron55      big*
17:10 celeron55      it's inherently non-linear
17:11 hmmmm          oh by the way
17:11 hmmmm          do you mind if i remove the now-redundant m_seed member of ServerMap?
17:12 hmmmm          i have a MapgenParams * in it
17:12 celeron55      why would i care
17:12 hmmmm          I don't know, maybe you liked it or something
17:13 RealBadAngel   celeron55, i will finish today update for l-systems trees and then will start to work on cave gen, im pretty sure now how to do it right
17:14 celeron55      RealBadAngel: just remember, the key is to never stop it, just discard results on a node-by-node basis
17:15 RealBadAngel   yup, cave can end in another cliff
17:15 RealBadAngel   just skip above ground contens
17:16 RealBadAngel   btw, completely offtopic
17:17 PilzAdam       -> #minetest
17:17 RealBadAngel   i watched recently videos on weekly snapshots of minecraft
17:18 RealBadAngel   this game is goin general direction of barbie and ken toys
17:19 RealBadAngel   those guys are tottaly burnt with ideas
17:20 iqualfragile   just one question: was minetest or minecraft first with give_initial_stuff?
17:20 celeron55      RealBadAngel: my intention is to steer clear of that as much as i reasonably can 8)
17:20 RealBadAngel   celeron55, i understand. and please dont go barbie way ;)
17:21 iqualfragile   celeron55: do you actualy recive donations?
17:21 RealBadAngel   yikes i can shoot a firework with a creeper smile on it
17:22 RealBadAngel   *milions of kids screams loudly*
17:24 celeron55      iqualfragile: very little, barely to pay the few tens of dollars of hosting per year
17:24 celeron55      i received somewhat more than that in 2011 though
17:25 hmmmm          minecraft is old and boring
17:25 Calinou        minecraft is boring and old, +1
17:25 hmmmm          i say we all quit minetest and start a call of doody clone
17:25 RealBadAngel   lol
17:25 RealBadAngel   no way
17:26 celeron55      if somebody feels like quitting minetest to do something else, i have a closed project that could interest such people 8) just ask if interested
17:26 RealBadAngel   im not askin :P
17:26 RealBadAngel   stayin here
17:27 iqualfragile   Calinou: actualy minetest is quite boring to play, but nobody just plays minetest, everybody writes mods/is a server-admin/does some other random stuff
17:27 sapier1        celeron is my latest commit fine?
17:27 celeron55      (i haven't intended to let any random people work on it though, but i'm open to opportunities :P)
17:27 Calinou        iqualfragile: redcrab's server has a lot of players
17:27 Calinou        iqualfragile: if we want a funnier minetest, we need to make it actually hard
17:28 RealBadAngel   iqualfragile, it needs more contens just
17:28 Calinou        almost everyone but tablet users like hard games
17:28 Calinou        we need hard to beat mobs, more dangers
17:28 RealBadAngel   hard but not to the square
17:28 Calinou        but of course there should be a mode where there is almost no danger
17:28 Calinou        (but not creative)
17:28 celeron55      one thing that would be interesting would be to make a minetest fork with a non-linear single-shot darf fortress like map generator
17:28 Calinou        keep 3 separate modes maybe
17:28 celeron55      it'd basically just need a good name
17:28 Calinou        2D minetest? heh :p
17:29 iqualfragile   thats gona be an interesting discussion (harder minetest, content, game-mechanics) but i dont think that it belongs in here
17:29 celeron55      (or, well, it could even just be a separate program to generate a minetest world)
17:29 darkrose       dwarftest
17:29 iqualfragile   maybe we should just join -delta and discuss about it for some minutes?
17:29 Calinou        people use -delta?
17:30 RealBadAngel   i thought once of makin a game mode, post apocalyptyc map, like in terminator movies, give players lasers, make two teams: humans vs machines, add some freakin music
17:30 RealBadAngel   and multiplayer team shooter ready
17:31 iqualfragile   Calinou: no, they dont, that why we could use it for this discussion/brainstorming
17:33 sapier1        https://github.com/sapier/minetest/commit/40578c7fb1cf5480957f77bdc53e87ad6d5ca700 sorry for spaming but I've got only about 30 minutes left and would like to have this commit in shape by time being
17:34 RealBadAngel   sure, lets stop spammin, some are codin here :)
17:34 * RealBadAngel signs off, need to rest a bit
17:42 hmmmm          celeron, i think i mentioned this sometime before, but I intend to have a mode of minetest that'll just generate the map
17:43 hmmmm          like minetest --mapcfg map_parameters.conf --outfile map.sqlite
17:43 hmmmm          it's actually quite trivial to do
17:44 Calinou        map generation settings in some .conf file would be cool
17:45 Calinou        sea level, frequency of lava and non-lava caves, terrain height...
17:45 hmmmm          it kind of already exists with my branch
17:45 Calinou        setting sea level as minetest.conf setting would be very easy
17:45 Calinou        i'd just like to see that in default game :p
17:45 hmmmm          dude calinou
17:45 hmmmm          i've got that holmes
17:46 hmmmm          i haven't pushed this to my fork yet but
17:46 hmmmm          http://ideone.com/CKZJDz all of this is configurable from the .conf
17:50 hmmmm          i originally had it configurable from lua but i decided to get rid of it completely because it was so complicated and messy
18:59 PilzAdam       thexyz, https://github.com/PilzAdam/minetest/commit/3acc3df46d6b394643724842985868d8e9193176
19:00 thexyz         aww, not this shit again
19:00 thexyz         PilzAdam: I mean, I dislike that idea in general
19:00 PilzAdam       its really handy
19:01 thexyz         at least, you should make it configurable
19:04 iqualfragile   i have got to admint: i dislike that idea, too
19:05 iqualfragile   its cloned from minecraft
19:05 PilzAdam       I actuall copied it from Terasology
19:05 PilzAdam       *actually
20:34 PilzAdam       thexyz, https://github.com/celeron55/minetest/pull/385 looks ok to me
20:35 celeron55      what is the point of such
20:36 PilzAdam       such little things make minetest look like an real game
20:36 celeron55      why does it make it look like a real game
20:36 thexyz         yes, same thoughts
20:37 celeron55      it's not like that would happen when you run, or anything
20:37 PilzAdam       a static camera makes the game look unfinished
20:37 thexyz         >Greater FOV when running or using free_move
20:37 celeron55      like, in real life
20:37 thexyz         oh
20:38 PilzAdam       game creators add such effects to make the animations/camera look more interesting
20:38 celeron55      but whatever, as long as you don't add cotton, i'm good with things you both agree to 8)
20:38 PilzAdam       such as blur when turning arround and such things
20:38 hmmmm          yeah, i think you're thinking of motion blur close to the corners of the screen
20:39 hmmmm          not changing the fov
20:39 hmmmm          i personally think it's ridiculous
20:40 PilzAdam       Terasology has it and it looks nice
20:40 PilzAdam       (only for free_move though)
20:42 hmmmm          terasology's mapgen looks sort of similar to that of minetest
20:42 * hmmmm        investigates
20:42 PilzAdam       they have 4 different types of mapgen IIRC
20:43 hmmmm          and it's java of course
20:43 celeron55      what's this trend for java and C# these days
20:43 hmmmm          man.... OOP already decreases the signal to noise ratio enough, but when you use java, it's way worse
20:44 celeron55      can't people code in native languages anymore
20:44 hmmmm          it's like 20% of the code in the codebase actually does stuff
20:44 hmmmm          i mean look at this https://github.com/MovingBlocks/Terasology/blob/develop/src/main/java/org/terasology/world/WorldProvider.java
20:44 hmmmm          they use an entire file and 120 LoC for basically nothing
20:44 celeron55      i started laughingg at "WorldProvider"
20:45 celeron55      -g
20:45 hmmmm          that's the same terminology that Minecraft uses
20:45 hmmmm          it's likely they just used the decompiled classes from Minecraft
20:46 celeron55      WorldProviderCore
20:46 thexyz         and AbstractWorldProviderDecorator.java
20:46 hmmmm          "I had a problem, so I thought to use Java..."
20:46 celeron55      WorldProviderFactoryAbstractorCore
20:47 celeron55      EnterpriseCodeFactoryFactoryFactoryFactory
20:47 hmmmm          we use factory methods in minetest
20:47 celeron55      i don't see any factories there though
20:47 celeron55      i know, but not excessively :P
20:48 hmmmm          the problem with Java is that it encourages the bullshitty parts of OOP; that is, overengineering your program that's actually about as complex as the little clicky thing on the end of ink pens
20:48 celeron55      but yeah, java is designed for making huuuuuuuge corporate programs using stupid code monkeys
20:48 celeron55      and you can see it in everything java does
20:48 hmmmm          it turns out a couple months later that those abstractions you made weren't too efficient and you end up having to recode it all
20:48 hmmmm          so how much work does that save in the long run.  hah.
20:49 hmmmm          you know why unsigned types were excluded from java?
20:49 celeron55      i think that is one of the few good things
20:49 hmmmm          gosling himself said "they would confuse programmers"
20:50 celeron55      they remove the stupid implicit unsigned-signed conversion bugs that C(++) programs are full of
20:50 hmmmm          for a language like C/++, though, a lack of unsigned types is simply unacceptable
20:50 celeron55      unsigned types are needed for bit operations
20:50 hmmmm          of course
20:51 celeron55      for anything else, not *that* much
20:51 celeron55      or, that one goes without the ** too
21:00 hmmmm          oh wow wait a minute, did you actually have a "generate the whole thing at once" generator in previous versions when ServerMap.m_chunksize == 0?
21:08 celeron55      nope