Time  Nick         Message
03:46 hmmmm        lotsa devving going on
03:46 hmmmm        ahhhh i want to jump in :(
03:47 VanessaE     so jump in, blindly merge a few requests, and jump out before the shit hits the fan ;)
03:47 hmmmm        merging pull requests isn't fun and games...
03:47 hmmmm        it's probably one of the least desirable things to do
03:48 VanessaE     well you didn't say it had to be fun :)
03:48 hmmmm        erm
03:48 hmmmm        looking at the very first pull request.... a hud bugfix...?
03:48 hmmmm        i was not aware of any bugs....
03:49 hmmmm        oh i see, he wants to use the scale specified, before it was ignored
03:49 hmmmm        this is an enhancement and not a bugfix, it is completely mislabeled.
04:08 ShadowNinja  hmmmm: Pull requests may not be fun, but we have 60+. And LevelDB support has to be merged eventualy.
04:08 hmmmm        i guess celeron was right, 0.4.7 needs to be pretty soon
04:09 VanessaE     yeah.
04:09 VanessaE     oh,
04:09 hmmmm        it's really not necessary for there to be some big new feature added
04:09 VanessaE     put sapier's scriptapi split thing through already
04:09 VanessaE     he's been ....griping again
04:09 hmmmm        that's basically all up to him
04:09 VanessaE     :)
04:09 hmmmm        i don't think anybody's been changing scriptapi
04:09 VanessaE     he said he's waiting on you guys actually.
04:09 hmmmm        so here's a good chance
04:09 hmmmm        erm
04:09 hmmmm        who put me in charge of this
04:10 hmmmm        i thought it was mostly sapier and celeron
04:10 VanessaE     [05-08 21:48] <sapier> rebasing is finished for about 3 weeks now
04:10 VanessaE     just saying
04:11 VanessaE     you guys have been kinda hemming and hawing over it.
04:13 ShadowNinja  Quite a few pull requests that are ready and waiting...
04:14 hmmmm        again; i'm not doing anything until all my things for school are done
04:14 VanessaE     understandable
04:37 VanessaE     can someone please fix the rate at which the view_range autotuner makes its adjustments?  it's painfully slow
08:52 sfan5        <messa4> i know python and asm etc. but i dislike c++ :(
08:52 sfan5        wat?
13:42 kahrl        hmm if you run minetest --speedtests it doesn't drop the IrrlichtDevice
13:42 kahrl        which means if you have fullscreen=true and some screenW,screenH set in minetest.conf it doesn't reset the video mode to its original value
13:54 kahrl        bugfix: https://github.com/minetest/minetest/pull/708 enhancement: https://github.com/minetest/minetest/pull/709
13:58 kahrl        I just keep putting work on you guys who accept pull requests, sorry haha
13:59 PilzAdam     well, I would give you push access to the repo so you can push little things like this https://github.com/minetest/minetest/pull/708 directly
14:00 kahrl        hmm, I guess
14:00 kahrl        I don't trust myself to figure out what's small enough to push without discussion
14:04 kahrl        would it be okay if I post such things as pull requests, ask for one dev's approval and then push it myself?
14:05 celeron55    that's how people often do it (even I)
14:06 celeron55    the rule for pushing little things without asking is for times when there simply is nobody around to comment so that fixes don't get bogged down because of that
14:06 PilzAdam     you dont even have to open a pull request, just point to the commit in your repo
14:06 kahrl        celeron55: well, nothing I do is time critical, I'm okay if it takes days or weeks to get merged
14:07 kahrl        PilzAdam: ah, even simpler
14:07 celeron55    whatever, i'll put kahrl in the core team; kahrl: the single most important and immediate rule is: don't push merge commits upstream
14:07 kahrl        yup, I know
14:07 PilzAdam     http://dev.minetest.net/Git_Guidelines
14:08 PilzAdam     I also have tested this now in valgrind: https://github.com/minetest/minetest/pull/635
14:08 kahrl        mmh I usually update my local master by git checkout master; git fetch upstream; git merge upstream/master
14:08 kahrl        does that create a merge commit if it's fast forward?
14:09 celeron55    i don't know
14:09 PilzAdam     fast forward does not create merge commits
14:09 PilzAdam     use git merge --ff-only to be sure to not have merge commits
14:10 PilzAdam     kahrl, do you want your e-mail adress be listed in the website?
14:10 kahrl        I don't really care
14:10 kahrl        though I don't check it often so maybe not
14:11 PilzAdam     you also can get a @minetest.net adress if you want
14:11 PilzAdam     website updated
14:12 kahrl        as a forward to kahrl@gmx.net?
14:12 kahrl        or a separate mailbox?
14:12 PilzAdam     ehm, ask thexyz
14:13 PilzAdam     (for me its a seperate mailbox, but I use KMail anyways)
14:13 kahrl        ah
14:14 kahrl        well I guess I'll get one later if I want
14:14 PilzAdam     you can push this device drop thing now
14:15 kahrl        okay, let's try this :)
14:15 kahrl        nothing can go wrong <-- last famous words
14:15 PilzAdam     :D
14:17 thexyz       of course you can use it as a forward as well
14:18 kahrl        thexyz: ah I think I'll use a forward one then
14:19 PilzAdam     why isnt c55 voiced?
14:19 kahrl        did it work?
14:20 PilzAdam     seems so
14:20 thexyz       PilzAdam: i don't have enough permissions to set +V flag on him
14:21 celeron55    i'm so high up that it has overflown to zero
14:22 celeron55    i guess i can add that flag
14:23 celeron55    hmm, actually no, i'm usually not even identified on freenode so that's useless
14:27 sfan5        I guess https://github.com/minetest/minetest/pull/709 can be merged
14:28 kahrl        should I do it?
14:28 PilzAdam     if it works
14:28 sfan5        kahrl: i'm already ready to push it
14:28 sfan5        also, does it work?
14:29 kahrl        I have only tested it on my linux machine so far, and it works here
14:29 sfan5        i won't test it on non-linux either
14:29 sapier1      https://github.com/minetest/minetest/pull/677 is still ready for review
14:29 sfan5        so, time to merge it
14:31 kahrl        btw, am I able to close pull requests or do I need more permissions?
14:31 kahrl        I don't see the Close Pull Request button
14:31 sfan5        kahrl: you should be able to close any pull
14:32 kahrl        I don't think I can
14:32 celeron55    kahrl: does your CSS hack break it?
14:34 kahrl        nope
14:35 kahrl        ah! the help at https://help.github.com/articles/closing-a-pull-request is just outdated
14:36 kahrl        I closed https://github.com/minetest/minetest/pull/709
14:39 PilzAdam     ummm
14:40 PilzAdam     wine segfaults when running bin/minetest --videomodes
14:40 PilzAdam     after printing the availible modes
14:40 kahrl        ugh
14:41 thexyz       but do we support wine?
14:41 PilzAdam     yes
14:41 kahrl        does it segfault on windows too, or just wine?
14:41 sfan5        in case someone wants to test on win: http://sfan.sf.funpic.de/minetest-builds/c55/minetest-0.4.6-3b1c3ac-win32.7z
14:42 thexyz       PilzAdam: i mean, do we really need to have everything work correctly in wine?
14:42 PilzAdam     it doesnt really hurt since you get the output you want
14:43 kahrl        thexyz: it could be a coding error that is hidden in linux but only shows in wine
14:43 kahrl        so it should ideally get fixed in that case
14:43 kahrl        on the other hand if it's an irrlicht bug under wine... can't really do much
14:46 kahrl        PilzAdam: do you have debug info from the segfault?
14:48 thexyz       PilzAdam: it works fine for me under W7
14:49 celeron55    we want most things to work correctly in wine because that way windows cross-builds are way easier to test on linux
14:50 celeron55    and yes, that is very important
14:50 iqualfragile is there even anyone able to compile on windows?
14:51 thexyz       same for winxp
14:53 PilzAdam     kahrl, wine outputs this: https://gist.github.com/PilzAdam/5547957
14:53 PilzAdam     and it only appears randomly
14:55 kahrl        huh. what is __clone and why does it call itself recursively so often?
14:55 thexyz       oh, well
14:55 thexyz       it also works for me in wine
14:55 thexyz       (sfan5's build)
14:56 kahrl        thexyz: how often have you tried? PilzAdam said it appears randomly
14:56 thexyz       ~10 times
14:56 Zeg9         PilzAdam, and other that asked me to do that: having clouds on every loading screen is quite a pain to make. And results aren't really good since some like creating server just freeze the clouds.
14:57 celeron55    sounds like uninitialized memory
14:57 PilzAdam     Zeg9, freezing is still better than blackscreen
14:58 sapier1      clouds on loading screen? plz be carefull the clouds in menu have been one of the more complicated to fix memory leaks
14:58 PilzAdam     kahrl, hm.. seems like this happens _after_ minetest has shut down correctly
14:58 PilzAdam     its a bug in wine, then
14:58 Zeg9         Well...
14:58 celeron55    PilzAdam: how can you tell that?
14:58 Zeg9         sapier1: does clouds->drop() deletes the clouds (free the memory) ?
14:59 PilzAdam     celeron55, look at the modules and the threads
14:59 PilzAdam     there is no mention of minetest
14:59 PilzAdam     (in a segfault that happens in minetest, there is)
14:59 sapier1      for some reason sky/cloud deletion order is quite relevant if you drop too soon you'll get a double free crash
15:00 Zeg9         Guess I'll need to make another scene manager for those damn clouds
15:00 PilzAdam     maybe I shouldnt use the unstable ppa of wine...
15:01 Zeg9         Else I see map blocks inside the shutdown screen...
15:01 sapier1      no just be carefull maybe run valgrind after adding your changes
15:02 sapier1      oh something completely different
15:02 PilzAdam     sapier1, I have tested your itemdef leak fix in valgrind
15:02 PilzAdam     it seems to work
15:03 sapier1      I'm glad it's same for you
15:03 sapier1      still I'm a little bit unhappy about that error messages in current core version due to trying to fetch a texture for deletion that already is deleted
15:04 PilzAdam     any objections? https://github.com/minetest/minetest/pull/635
15:05 PilzAdam     (actually only the second commit is relevant
15:05 PilzAdam     +)
15:06 sapier1      first one is already included
15:06 kahrl        looks good to me
15:06 sapier1      but obviously buggy
15:06 PilzAdam     Ill set up a proper commit message
15:06 Zeg9         What is the type minetest uses for dtime variables?
15:06 Zeg9         Ok never mind found it was just a float
15:07 kahrl        Zeg9: it's almost always in seconds
15:07 PilzAdam     sapier1, another thin I noticed in valgrind is that mapgen and shaders seem to leak
15:07 kahrl        unless the variable is called msec or something
15:07 PilzAdam     *thing
15:07 sapier1      shaders can't be fixed
15:07 sapier1      irrlicht doesn't support unloading shaders
15:08 sapier1      mapgen is work in progress and has been changed quite a lot while I was fixing the memory leaks
15:08 Zeg9         sapier, is that why I get strange lava lakes since a recent update?
15:09 sapier1      don't know I'm not involved in mapgen rewrite
15:09 PilzAdam     sapier1, and CGUITTFont seems to leak too
15:09 Zeg9         Ok, never mind then
15:10 PilzAdam     Zeg9, I guess the lava lakes are intentional
15:10 sapier1      I assume this means I'm now officially appointet minetest memory leak plumber? :-)
15:10 Zeg9         Yeah but, well, look at some screenshots in zmobs
15:10 PilzAdam     sapier1, you started to fix them :-)
15:10 Zeg9         Some are generated in mid-air of caverns
15:11 sapier1      ok but I want authority to clap those ones fingers adding new memory leaks ;-)
15:12 PilzAdam     merged the itemdef leak fix
15:14 PilzAdam     can someone else updated the changelog at the dev wiki?
15:15 Zeg9         Q: does it matter if the shutting down stuff... screen is still black, or you really need clouds there?
15:15 PilzAdam     I guess I dont care about that
15:16 PilzAdam     (would be cool, though)
15:16 * Zeg9       thinks it would introduce some leaks having clouds there.
15:16 * Zeg9       needs to learn how to use valgrind to check for these by the way
15:17 PilzAdam     valgrind --leak-check=full --track-origins=yes bin/minetest
15:18 Zeg9         Does it require a debug build?
15:18 PilzAdam     then search for "leak" in the output
15:18 PilzAdam     yes, a debug build is needed to show the origin AFAIK
15:27 Zeg9         PilzAdam, should I show the progress bar background on Connecting to server ?
15:27 PilzAdam     yea
15:28 Zeg9         Okay, what about other ones, creating server etc
15:28 PilzAdam     just the string in the clouds doesnt look good (I think)
15:28 Zeg9         No, it doesn't
15:29 Zeg9         But with the progress bar I have to remove the rotating / when connecting... or put it inline
15:31 PilzAdam     maybe use the progress bar for connecting to server too
15:31 PilzAdam     (like 25% resolving adress, 50% connecting etc.)
15:32 Zeg9         waht about: 5% Creating server, 5% Creating client, 5% Resolving address, 5% connecting to server, and others for media?
15:32 PilzAdam     also remove the "(*%)..." from media
15:33 Zeg9         I will
15:33 PilzAdam     Zeg9, no, better reset the bar when loading media
15:33 Zeg9         Ok
15:40 PilzAdam     Zeg9, well, use "Media..." (same as nodedef and itemdef)
15:40 Zeg9         I changed them to Loading nodes, Loading items and Downloading media, but I can change it back if you wish
15:41 Zeg9         These are better for translation (should I add wgettext btw?)
15:41 PilzAdam     loading nodes is wrong
15:41 PilzAdam     yes, translation is good
15:41 PilzAdam     (dont run updatepo.sh yet)
15:41 Zeg9         Well, Node definitions and Item definitions... what about media?
15:42 PilzAdam     also show a 50% progress bar after "Item definitions"
15:42 PilzAdam     hm?
15:43 Zeg9         Let's keep "Media..."
15:43 PilzAdam     yea
15:43 PilzAdam     (with wgettext())
15:44 Zeg9         Ok
15:45 Zeg9         Is it wgettext(L"...") or wgettext("...")
15:45 Zeg9         nvm
15:46 PilzAdam     wgettext(std::string)
15:50 * Zeg9       added wgettext to all loading screens
15:51 Zeg9         The only left problem is that clouds only move for "connecting to server" and node,item,media step
15:52 PilzAdam     that not that bad
15:53 Zeg9         Ok, I'll check my code and commit
16:16 * Zeg9       is back
16:17 celeron55    http://forum.minetest.net/viewtopic.php?id=5713
16:17 celeron55    has anyone considered this for minetest_game?
16:18 celeron55    or, well, PilzAdam
16:18 PilzAdam     we wait for DecorationDef
16:18 PilzAdam     but these trees are nice
16:18 celeron55    but i think what that mod does means we don't need moretrees-sized trees - it makes nice trees that are reasonably sized
16:19 celeron55    so no rendering performance problems
16:19 celeron55    and gameplay doesn't change in unwanted ways
16:20 celeron55    anyways, indeed that is worth waiting (assuming it's coming this month or so)
16:23 celeron55    PilzAdam: do you have time for some continued design issues from the build/survival attempt?
16:23 celeron55    i'm concerned about the name of minetest_game
16:24 PilzAdam     yes, I have lots of time
16:24 celeron55    i just realized a particularly confusing thing about it; it is the mods/minetest/ directory
16:24 PilzAdam     well, https://github.com/minetest/minetest/pull/690
16:25 celeron55    that implies the mod namespace is the same between games
16:26 Zeg9         PilzAdam, https://github.com/minetest/minetest/pull/710
16:26 Zeg9         Maybe not the good moment...
16:27 PilzAdam     Zeg9, s/Connected !/Connected!
16:28 Zeg9         k
16:28 celeron55    i'm trying to understand why on earth we named the main game minetest_game and it's main mod default
16:28 celeron55    it's nonsense
16:28 celeron55    :P
16:29 Zeg9         PilzAdam, fixed
16:29 PilzAdam     first of all I would not remove the "_game"
16:29 PilzAdam     so people actually see what minetest_game is
16:30 celeron55    where would you remove it?
16:30 celeron55    everywhere that isn't related to persistent names?
16:31 celeron55    ...i mean, not remove it
16:31 celeron55    remove removal
16:31 PilzAdam     well, its kinda annoying that its already written to the world files
16:31 PilzAdam     so people's maps will be broken
16:32 celeron55    well that compatibility must be kept
16:32 PilzAdam     maybe provide a fallback
16:32 celeron55    either as a special case or by keeping the _game removal always there
16:33 PilzAdam     removing _game just doesnt make sense for minetest_game, since Minetest is the name of the engine
16:33 celeron55    is there any other place that has a persistently stored "minetest" instead of "minetest_game" than world.mt?
16:34 PilzAdam     mods/minetest
16:36 celeron55    adding special handling for those isn't hard
16:37 celeron55    can anyone point anything that will make this change explode?
16:45 PilzAdam     Zeg9, why do the clouds change when switching from main menu to loading screen?
16:45 PilzAdam     (same as when deleting a world or connecting to server fails)
16:45 sfan5        PilzAdam: a new "Cloud" object is created
16:45 Zeg9         PilzAdam, not in the pull request version
16:46 Zeg9         or there is some problem...
16:46 Zeg9         Wait, there IS some problem -_-
16:48 Zeg9         Trying to fix....
16:51 Zeg9         Thanks to my brother, I can't work on it anymore, sorry... I'll be back in like 1 hour -_-
17:24 Zeg9         How can I declare a class from a namespace (irr::scene::ISceneManager) from within a .h ? (without including irrlicht.h)
17:24 Zeg9         is namespace irr{namespace scene{class ISceneManager;}} correct ?
17:32 celeron55    Zeg9: yes, C++ doesn't have other syntax for that
17:32 celeron55    (which sucks though)
17:32 celeron55    it's one of the devil's plan to increase compile times by the order of 100
17:41 celeron55    i can imagine how painful the clouds are to add to all the screens
17:42 Zeg9         ...
17:42 Zeg9         Just read what I just wrote in #minetest -.-
17:42 celeron55    i did 8)
17:43 celeron55    if it's too hard or hacky, i think simply the background color with the progress bar would be enough, but in the end i'm leaving this to PilzAdam
17:43 Zeg9         Just have to fix them different on the loading screen than on the main menu
17:45 PilzAdam     sapier1, valgrind says that your sky fix creates 2 new (minor) memory leaks
17:46 * Zeg9       begins to hate clouds :-)
17:46 Zeg9         I think when I'll be done with this, I'll disable both menu_clouds and enable_clouds lol
18:18 hmmmm        now how will the cloud in the loading screen affect render-to-texture
18:18 hmmmm        that's what i originally feared of doing that
18:19 PilzAdam     the screen just freezes
18:19 sapier1      pilzadam which one? sky fix has been createt multiple times don't know if last version is really correct
18:19 PilzAdam     sapier1, https://github.com/minetest/minetest/pull/627
18:20 sapier1      I'll check again
20:23 sapier       any chance we could add a "pull requests have to be warning free" requirement somewhere?
20:24 PilzAdam     sapier, I guess thats part of "code style"
20:24 sapier       is it? then enforcement isn't working ;-)
20:27 celeron55    different compilers give different warnings so that's not easily enforceable
20:30 sapier       didn't think about that problem true, but I assume that annyoing "non virtual destructor" warning is shown by all compilers
20:32 celeron55    yes, everyone is waiting for hmmmm to fix them as the made them
20:32 celeron55    he*
20:32 hmmmm        somebody has a pull request to fix that, and gcc 4.2.2 doesn't have that warning fyi
20:34 sapier       hmmmm do you use 4.2.2?
20:34 hmmmm        yes
20:34 sapier       ok :-) what os do you use?
20:34 hmmmm        freebsd 9
20:36 sapier       an i thought debian is using old compilers :-)
20:36 hmmmm        versions after that conflict have GPLv3 which conflicts with the BSD licesning
20:37 sapier       hmm is llvm bsd compatible?
20:37 hmmmm        as for the hud not showing up on player join; i have no idea there.  there's no reason why that should happen.. again, can't do anything related to this right now
20:37 hmmmm        yes, llvm is
20:37 hmmmm        clang+llvm is the base compiler for freebsd 10 in fact
20:38 sapier       i wonder if llvm will be next standard compiler
20:38 hmmmm        kinda wish i could use gcc47 but it'll break other ports when i go to install those
20:46 sapier       pilzadam I that mapgen memory leak you mentioned I think I already fixed it in all in one patch but didn't create a separate pull request because hmmmm told he'd fix it himself
20:48 hmmmm        yes, i haven't forgotten about that
20:51 sapier       who did mess around in gui in area of favourites?
20:53 PilzAdam     hm?
20:53 sapier       doesn't matter whoever did this should sit on her/his desk and read some chapters about memory management
20:54 sapier       wgettext("some text") allocates memory this memory is required to be free if you pass it to gui directly the memory is lost
21:05 sapier       hmmmm there are at least two different memory leaks in mapgen one related to noise params and one in emerge manager
21:14 sapier       plz declare scriptapi freeze till saturday 11pm gmt to rebase scriptapi folder separation change request on order to merge it
22:05 hmmmm        a mapgen related leak in emergemanager.... i suppose you mean i mishandle a MapgenParams somewhere?
22:05 hmmmm        FYI, MapgenParams is deleted by Map
22:06 hmmmm        the only exception to this is if a MapgenParams never gets associated with a Map in the first place (i.e. invalid parameters or whatever when creating it)
22:07 hmmmm        the Map has these parameters associated with it that the Mapgen uses; it's not really like they're attached to the Mapgen, although there is a specific MapgenParams for each Mapgen
22:07 hmmmm        a specific MapgenParams class, i should clarify, not object
22:45 sapier       https://github.com/minetest/minetest/pull/677 rebase complete plz have a look if I missed something due to files beeing moved I had to do some manual merges
22:46 sapier       scriptapi folder separation