Time  Nick        Message
05:46 Zeno`       some of this bugs are like trying to find the end of a rainbow :(
05:53 Zeno`       hmmmm, you said the other day smoothlighting might be broken again. Is this confirmed?
05:54 hmmmm       no, smooth day/night transitions
05:55 hmmmm       we're not going to be able to release tomorrow at this rate
05:56 Zeno`       ah ok
05:56 * Zeno`     looks
05:57 Zeno`       what is "broken" though?
05:57 hmmmm       it doesn't gradually make the transition
05:57 hmmmm       instead each block is updated to the new day/night color at some cutoff time
05:58 hmmmm       this seems to keep happening for some ungodly reason
05:58 hmmmm       I mean the same bug
05:58 hmmmm       then it gets fixed, time passes, it breaks again
05:58 Zeno`       I'm not seeing it. This is another one of those "chase the rainbow bugs" :/
05:59 hmmmm       i'll have to do it
05:59 Zeno`       darn it
05:59 hmmmm       i can replicate it just fine
05:59 Zeno`       well, don't break mine when you fix it :)
05:59 Zeno`       because my game is fine :P
06:00 hmmmm       btw is wayward_one able to compile his own minetest?
06:00 Zeno`       yeah
06:00 hmmmm       oh good
06:01 hmmmm       that narrows things down, it might suffice to have him add a print statement in the one exception handler for Server::Receive()
06:02 Zeno`       yeah he'd be able to do that
06:03 hmmmm       yeah, Zeno, are you checking for the day/night transition to begin with?
06:03 hmmmm       let me try changing it back to Debug mode
06:03 Zeno`       I've even had him apply gist patches for me
06:03 hmmmm       maybe it's release that has the problem
06:04 Zeno`       I'm wonder if maybe (just maybe) stuff like "Optimize bumpmapping mathematics" is the issue that cause intermittent regressions; since, as you know, floating point statements/operands are not communicative (unless fast math is enabled as a compiler switch, but this wouldn't apply to shaders)
06:05 Zeno`       not "necessarily" communicative*
06:05 Zeno`       I was looking at finalColorBlend() for another reason and remembered your statement from the other day
06:06 hmmmm       yup it's broken for me still
06:06 hmmmm       let's see here
06:08 Zeno`       actually fastmath doesn't make floating point ops commicative; it allows the compiler to refactor statements that it normally wouldn't because of the non-commutativity
06:10 Zeno`       I made a change to finalcolorblend not that long ago, btw
06:11 Zeno`       https://github.com/minetest/minetest/commit/5e58a954910d9a9a3368bee71ab3aa0b3baf7c3a
06:11 Zeno`       Can't see how that would break anything, but *shrug*
06:11 hmmmm       that's pretty innocent looking
06:12 hmmmm       nah it's older than that
06:12 hmmmm       I just tried a commit from 15 days ago
06:12 hmmmm       hrmmm
06:16 hmmmm       i think it's been happening since before 0.4.11 and nobody noticed
06:16 VanessaE    it has been.
06:17 hmmmm       you knew about this??
06:17 VanessaE    yeah, figured everyone else did too:
06:17 VanessaE    RBA broke and fixed it once:  https://github.com/minetest/minetest/issues/1705
06:17 Zeno`       I thought that was due to RBA's removal of stuff from shaders, but that that got added back just before 0.4.11 was released
06:17 hmmmm       oh thefucidfa
06:17 hmmmm       no
06:17 Zeno`       yeah that one
06:17 hmmmm       Zeno`, that was my own hack to disable shading of the wieldhand with shaders
06:18 hmmmm       God dammit RBA
06:18 hmmmm       alrgiht
06:18 VanessaE    but I've noticed it since then as well so I guess his fix got undone somehow.
06:18 Zeno`       how would it get undone?
06:18 hmmmm       he removed it from the shaders again
06:18 Zeno`       oh dear
06:18 VanessaE    beats me
06:18 VanessaE    ohhhhh
06:19 hmmmm       i'll probably have to get that one
06:19 hmmmm       so before 0.4.12, we need to fix:
06:19 VanessaE    everything.
06:19 VanessaE    ;)
06:19 hmmmm       day/night transitions with shaders, finish the minidump thing, the single player serverthread hangups
06:21 hmmmm       the get_craft_recipie thing i think, but i won't be the one to do that
06:21 VanessaE    there are a few bugs marked as "high priority" that might be worth considering before 0.4.12 also
06:22 VanessaE    (but two of those are "unconfirmed" also)
06:22 hmmmm       schematic crash?
06:23 hmmmm       why don't people ever respond to these
06:23 VanessaE    that's one of them but idk what that's even about
06:23 hmmmm       that one's old... I remember Jordach also having a schematic crash that we solved in the channel and he didn't even need to make a problem report for
06:23 hmmmm       could be the same
06:24 VanessaE    another that should be sorted before 0.4.12 goes out is when sapier broke the size of the HUD
06:24 VanessaE    (default window size, it's too big now, causes the health bar to overlap the hotbar in e.g. dreambuilder, probably carbone too)
06:24 hmmmm       https://github.com/minetest/minetest/issues/2024 ouch
06:24 hmmmm       yeah
06:24 hmmmm       I don't really know how to fix that one
06:24 hmmmm       I'll try though
06:50 Zeno`       hmmmm, is it still an issue?
06:51 Zeno`       I know it was reported on 30th dec... but on 24th Nov there was https://github.com/minetest/minetest/commit/26cf98ccfc70909be916d29f0cc480f6929cbb08
07:28 * Zeno`     tears his hair out
07:37 Zeno`       the size of the hotbar looks fine (to me) at 0.75% (I got that number from 800/600) but... it's all messy
07:38 Zeno`       I mean the relationships between the sizes of font, display density, gui element size, layout... it's all very odd
10:57 Zeno`       Does anyone understand #2024?
10:57 ShadowBot   https://github.com/minetest/minetest/issues/2024 -- Local map saving runs on locally started server, overwriting another world.
11:02 Zeno`       I mean the description... not where the bug is ;)
11:15 Zeno`       sfan5, https://gist.github.com/Zeno-/b2e9c9f8b183ab38f74a    <--comments?
11:27 Zeno`       I doubt it fixes 2024, but *shrug*
11:28 Zeno`       I can't really understand the description of 2024 so I don't really know
12:06 Zeno`       Others please look at https://gist.github.com/Zeno-/b2e9c9f8b183ab38f74a as well, of course
12:07 Zeno`       I still cannot reproduce 2024 though ;/
13:29 kilbith     sfan5: a bit of media optimization before the release please : https://github.com/minetest/minetest_game/pull/415
13:29 twoelk      only slightly related problem; when saving to local mt seems to not be able to save to local if the game stated in the world.mt is not installed
13:33 * twoelk    notes he might have had to inspect some of the recent links before stating the obvious :(
13:35 * T4im      already amended a more detailed description of reproduction for Zeno
13:35 T4im        bug reproduction*
13:39 T4im        only missed him by an hour or two :(
13:42 twoelk      he might be server testing on the inchra network
13:46 T4im        on a different bug, might I raise awareness of #2222 in hope a fix would still make it into the release? ;) lua unit test was added yesterday too
13:46 ShadowBot   https://github.com/minetest/minetest/issues/2222 -- minetest.get_craft_recipe occasionally returns wrong recipes
13:57 shadowzone  Wasn't there suppose to be a release today?
13:58 sfan5       the day hasn't ended yet
13:58 sfan5       releases mostly happen in the evening
13:59 shadowzone  Oh.
13:59 shadowzone  What time is it then? I just woke up.
14:00 sfan5       15:00 here (CET)
14:03 shadowzone  Okay.
14:12 twoelk      sfan5: have been playing withthe copy-to-local feature for some time now. Could there be an info.txt file added into the world folder with stuff from the server list related to the downloaded world, such as a modlist and other stuff?
14:40 est31       I'm currently trying to bisect an issue, and I'm getting "Assertion 'hand_server.count == 1' failed." for 0.4.0. What could I have done wrong?
14:41 T4im        where is the assertion failing?
14:44 casimir     I did not notice that there is a release planed for today. What will be the version number?
14:44 est31       ~topic
14:44 ShadowBot   est31: **FEATURE FREEZE FOR 0.4.12** **TENATIVE RELEASE DATE: 2015-01-29** Minetest core development and maintenance. Chit-chat goes to #minetest. Consider this instead of /msg celeron55. http://irc.minetest.ru/minetest-dev/ http://dev.minetest.net/
14:45 est31       T4im: http://pastebin.com/BkCE98t4
14:46 est31       0.4.4 works, 0.4.3 is the latest release the assertion fails
14:46 est31       and unfortunately my problem already exists with 0.4.4
14:47 T4im        hmm try running with --disable-unittests
14:47 T4im        provided you don't need em ;)
14:48 est31       works thanks :D
14:50 T4im        great :)
14:51 est31       ok, I'm sure now this is not caused by minetest. Some driver or irrlicht misbehaves or something like that,  but not MT. Never used 0.4.0, and 0.4.0 already fails.
14:52 est31       weird error though. Happens when you enable sound, and stacktrace points to some irrlicht vector code
14:52 T4im        sound can be located and directed
14:56 est31       bt for master: http://pastebin.com/nDFipXRc
14:57 est31       and the funny thing is, the bt changes from crash to crash
14:58 T4im        maybe find the common tracestep then?
14:59 T4im        if there is one…
14:59 est31       ??
14:59 T4im        but might have to run that through valgrind
14:59 est31       oh gonna disable clouds then
14:59 * est31     knows valgrind is slooooooow
15:00 * T4im      is uncomfortable with c++ memory issues though as java dev
15:00 T4im        very much so
15:01 est31       oh man, valgrind already crashes at startup
15:01 est31       src/script/cpp_api/s_base.cpp:75: ScriptApiBase::ScriptApiBase(): Assertion 'm_luastack' failed.
15:02 T4im        that would have been the next question… what mods are you running?
15:02 T4im        especially one that plays sound
15:02 est31       minimal
15:02 T4im        O_o
15:02 T4im        oh
15:03 est31       I mean, valgrind crashes when I start the mainmenu
15:03 est31       without valgrind, the crash only occurs when I load a world, and MT would play a sound (at least I guess so)
15:06 T4im        so it doesn't load lua when running with valgrind? O_o
15:08 est31       seems so
15:08 est31       when I disable luajit, it works
15:08 est31       at least mainmenu
15:09 est31       although valgrind sais "invalid read of size 8"
15:11 est31       oh FPS of 0
15:11 est31       drawtime 3649
15:14 est31       yeah, its complaining about libopenal...
15:14 est31       gonna install libopenal debug symbols....
15:15 T4im        well that sounds fitting
15:17 T4im        and valgrind slowness is normal :P it instruments so much code… and from what I've seen it really has problems when including the graphicsdriver as they seem to rewrite their code in memory
15:19 est31       and now... lets gonna make a minimal test example for upstream...
15:20 est31       debug symbols dont contain line numbers :(
15:20 est31       just func names
15:36 est31       also funny the func name isnt even /found/ in src dir
15:39 est31       http://pastebin.com/QPuKV1Ph
15:39 est31       ^ Thats the log
15:44 est31       seem to be not the only one: http://www.soldak.com/forums/showthread.php?p=72247
15:45 hmmmm       did you compile a debug build of openal?
15:46 est31       not yet will try
15:46 hmmmm       oh how'd you get symbols
15:47 hmmmm       they obviously weren't full symbols though
15:47 est31       btw found the func in 1.15.1, the one I checked last was 1.14.1
15:47 hmmmm       so basically it seems their MixDirect_SSE is a little messed up
15:47 est31       yeah
15:48 hmmmm       i can see how that error might occur..
15:48 hmmmm       i doubt you'll be able to make more sense until you compile a debug version
15:49 * est31     searches for that cool dpkg-buildpackage stuff
15:55 Wayward_One hmmmm, still need me to add that print statement?
15:56 hmmmm       oh yes
15:56 hmmmm       i didn't realize you were in this channel
15:56 Wayward_One yep, i'm always here :P
15:58 hmmmm       so on src/server.cpp, line 1175, add a print statement
15:58 hmmmm       if my theory is correct, you should see a LOT of that being printed out constantly in singleplayer
16:02 Wayward_One ok, and what exactly should i put? i'm not too familiar with print statements
16:03 hmmmm       okay nevermind
16:03 hmmmm       I realized why I haven't been seeing any
16:03 Wayward_One oh
16:03 hmmmm       i'm getting a flood of "No incoming data"
16:04 hmmmm       you should be too... except in your case it's eating a metric shitton of cpu time
16:04 hmmmm       okay this to the end:
16:04 hmmmm       catch(std::exception &e)
16:04 hmmmm       {
16:04 hmmmm       // Do nothing
16:04 hmmmm       printf(">>>>>>>>> %s\n", e.what());
16:04 hmmmm       }
16:05 hmmmm       in the console you'll see a huuuuuuuuge amount of output
16:06 hmmmm       thanks sapier'
16:08 hmmmm       in fact I'm not even sure how this doesn't cause a warning
16:08 hmmmm       Connection::Receive doesn't return any value..
16:42 Zeno`       wtf is this: https://forum.minetest.net/viewtopic.php?f=7&t=5432&p=169046
16:43 sfan5       Zeno` possibly wanted to link to https://forum.minetest.net/viewtopic.php?p=169046#p169046
16:43 Zeno`       quoting gregorycu: "The new standards may be outside the understanding of some developers"
16:43 sfan5       also this is users
16:43 Zeno`       this is not about users (completely)
16:43 Zeno`       see my response
16:44 Zeno`       just about every server operator compiles their own minetest
16:44 Zeno`       I don't know of any that don't
16:44 Zeno`       and none of us understand c++11?
16:44 Zeno`       s/us/coredevs
16:47 Zeno`       dammit I am sick of him. I gave him nothing but support and encouragement (including rebasing his PRs, fixing his whitespace errors and fixing his repo when he broke it)
16:48 hmmmm       well I certainly am not aware of all the new features of C++11
16:48 hmmmm       I support its adoption more than i used to
16:48 Zeno`       I do as well. But apparently it's beyond our comprehension, hmmmm
16:49 Zeno`       Too hard for us to understand
16:49 hmmmm       the specific wording used indicates arrogance
16:49 hmmmm       frighteningly similar to prestidigitator for us failing to adopt his super-convoluted OOP patterns
16:50 T4im        …or frustration venting
16:50 hmmmm       this seems to be a common theme among language fanboys
16:50 sfan5       is the what-not-to-do example still in the style guidelines?
16:51 hmmmm       yeah i added it back
16:51 hmmmm       it's very telling
16:51 sfan5       Shadowninja also removed the second "Do not uses spaces to indent."
16:51 hmmmm       we don't want code where 85% of the code is framework
16:52 hmmmm       well, at least *I* don't
16:52 hmmmm       and I would imagine that I speak for the majority of others
16:52 Zeno`       Apparently, though, this is why he put me on ignore and when the personal attacks started
16:52 hmmmm       read this http://sebastiansylvan.com/2013/08/16/the-perils-of-future-coding/
16:53 hmmmm       I should probably add a link to this article to the coding style guidelines
16:53 hmmmm       I can't beat it into peoples' heads good enough
16:53 hmmmm       there is a happy medium between proller and prestidigitator and it's quite easy to achieve
16:53 hmmmm       the majority of us already do
16:55 Zeno`       i.e. right here: http://irc.minetest.ru/minetest-dev/2015-01-24#i_4129295
16:56 Zeno`       I can't see how that conversation warrants the treatment directed towards me by him
16:56 Zeno`       He told me today that's why he put me on ignore
16:56 hmmmm       that's ridiculous
16:57 hmmmm       reentrant locks are poor design anyway.  i can't understand what's so bad about passing the new value ..
16:57 sfan5       hmmmm: that's an interesting article
16:57 hmmmm       we explicitly avoid reentrant locks already in Settings and Map
16:58 hmmmm       sfan5 yeah, it does a good job at articulating the issues i have with certain people
16:58 sfan5       http://irc.minetest.ru/minetest-dev/2015-01-24#i_4129385 wat
16:59 hmmmm       the most infuirating thing about future coders is their superiority complex and how they'll often presume we're not using their methods because we don't understand them
16:59 Zeno`       sfan5, apparently that's not a personal attack
17:00 sfan5       Zeno`: that _is_ a personal attack
17:00 Zeno`       anyway, this is where gregorycu's issue with me apparently started
17:01 Zeno`       And I still don't know what I said or did wrong
17:01 hmmmm       regardless, this doesn't mean we should stop accepting his contributions
17:02 hmmmm       for the most part they're good
17:02 sfan5       oh also
17:02 sfan5       Zeno`: that patch you proposed today morning makes no sense to me
17:02 sfan5       or is m_localserver automatically initialized when the class is created?
17:02 hmmmm       apart from the deadlock-on-systems-that-use-pthread issue they're been helpful
17:03 sfan5       i didn't even read the whole code because i had no time
17:03 sfan5       so i might be completly wrong
17:05 Zeno`       sfan5, localserver is allocated but as far as I could see not deleted
17:05 Zeno`       that's all the gist was aobut
17:05 Zeno`       about*
17:05 sfan5       allocated?
17:05 sfan5       where?
17:06 Zeno`       https://github.com/minetest/minetest/blob/master/src/client.cpp#L306
17:07 sfan5       where's the problem, it only gets allocated when local-map-saving is requested
17:07 Zeno`       but never gets deleted
17:08 Zeno`       well, unless I am missing something (which I could be)
17:08 sfan5       can you give me the link to your patch again?
17:08 sfan5       afk food
17:08 Zeno`       https://gist.github.com/Zeno-/b2e9c9f8b183ab38f74a
17:09 Zeno`       The gist (that one ^^) also removes line https://github.com/minetest/minetest/blob/master/src/client.cpp#L298
17:09 Zeno`       but I'm not sure if that's correct; certainly world_gameid is never used, but I didn't look at all the side-effects (which is why it's only a gist)
17:13 Routh       Hey folks. Need some insight. I'm new to LUA scripting, however a fairly competent coder. I'm attempting to debug an issue with Sokomine's mobf_trader mod. Currently I've determined that the crash is because the mobf_trader.MAX_TRADER_PER_PLAYER var is 'nil' - despite it's being defined at line 31 in init.lua - The declaration seems fine, yet a 'print' of the var value reveals it is nil. Has something changed with LUA var declaration?
17:14 est31       Routh: --> #minetest
17:15 hmmmm       Routh:  yes, setting and getting global variables has become disallowed
17:15 hmmmm       undeclared global variables that is
17:15 Routh       @est31 #minetest for a dev issue?
17:15 hmmmm       it's a mod issue
17:16 Routh       @hmmmm Ok, is there a good site for the LUA language documentation?
17:16 est31       Routh: yea strange...
17:16 est31       http://rubenwardy.com/minetest_doc/lua_api.html
17:16 hmmmm       what's wrong with the official lua website?
17:16 hmmmm       you mean lua language or the minetest API?
17:17 est31       The link I posted was minetest API
17:17 Routh       Both really, but that link that est31 gave just helped me find both, thanks.
17:17 Routh       Hopefully will have a pull request for Sokomine by end of day
17:18 Routh       Annd I'll move mod debug to #minetest
17:18 Routh       Thanks
17:18 hmmmm       np
17:18 est31       np :)
17:19 hmmmm       something about schematic specifiers is wrong in the documentation
17:20 hmmmm       did the person who wrote it marked up like that take the liberty to modify some wording?  it's (p / 255) * 100
17:20 hmmmm       and the condition should be 0 <= p <= 255
17:21 sfan5       Zeno`: ooh makes sense
17:23 sfan5       hmmmm: is this something that can be merged while in feature freeze? https://github.com/minetest/minetest/pull/2228
17:25 sfan5       Zeno`: https://github.com/sfan5/minetest/compare/minetest:master...sfan5:issue2024
17:25 Zeno`       IMO I think it can be merged
17:25 Zeno`       hmmmm?
17:26 sfan5       Zeno`: what about the bugfix?
17:26 Zeno`       well, it's related to resolve
17:26 Zeno`       Resolve()
17:27 Zeno`       I'd have to scrollback
17:27 sfan5       huh?
17:27 Zeno`       the bug is when server_name == ""
17:28 Zeno`       kind of
17:28 sfan5       <sfan5> Zeno`: https://github.com/sfan5/minetest/compare/minetest:master...sfan5:issue2024
17:28 sfan5       I'm talking about that
17:28 Zeno`       oh yeah, that's fine
17:29 Zeno`       it's just a bug fix
17:29 Zeno`       something trivially missed
17:29 Zeno`       s/missed/overlooked
17:29 sfan5       actually
17:29 sfan5       hm
17:29 sfan5       the issue is something different
17:29 T4im        hm… there should be some way added though to allow testing undeclared global variables for declaration (which is commonly done testing them for nil) or allow supressing the warnings in those cases where it is a intentional and good thing to do
17:29 Zeno`       yeah the issue is different
17:30 Zeno`       that's what I meant
17:30 sfan5       Zeno`: the problem is that address and remote_port are not set correctly when connecting to a local server
17:31 Krock       is the infinite loop on the speed test wanted or an issue?
17:31 Zeno`       in client.cpp (line 288?) the path name and of the local server are expected to be set in g_settings... yeah
17:31 sfan5       local map saving should probably be forced off when a local server is started
17:34 Zeno`       yeah, it's tricky though....
17:34 Zeno`       https://github.com/minetest/minetest/blob/master/src/game.cpp#L2127
17:35 Zeno`       the thing is std::string *address is not modified if *address is empty
17:35 Zeno`       and that led me to connect_address.Resolve(address->c_str());
17:35 Zeno`       connect_address is modified (kind of)
17:36 Zeno`       but *address is not
17:36 Zeno`       so by the time it gets to the local map saving g_setting(".............whatever............"); is still the value of the previous connection
17:36 sfan5       what happens when local map saving is enabled in singleplayer?
17:36 Zeno`       it breaks
17:37 Zeno`       that's why it's been disabled for simplesingleplayer
17:37 est31       hmmmm: the audio crash log with a debug build
17:37 est31       http://pastebin.com/Dm9S6Cu7
17:38 Zeno`       I only looked at it for a while by tracing it I can see exactly what's happening. I'm just not sure *where* it should be fixed
17:38 sfan5       it should just be disabled for simple singleplayer
17:38 Zeno`       sfan5, it is
17:38 sfan5       or address should be set to 127.0.0.1 for local server
17:38 sfan5       oh sorry
17:38 Zeno`       the problem is from the server tab
17:38 sfan5       Zeno`: s/simple singleplayer/local server/
17:41 Zeno`       sfan5, that's the obvious solution
17:41 Zeno`       https://github.com/minetest/minetest/blob/master/src/client.cpp#L288
17:41 Zeno`       which is why I asked why g_settings() was used
17:41 sfan5       what else am i supposed to use
17:42 celeron55   wtf is going on on the forum
17:43 Zeno`       sfan5, that's what I'm not sure about. Perhaps the server address needs to be passed to the client ctor?
17:44 Zeno`       sfan5, what I'm saying is that what's going on is easy to see, but how to solve it is... a bit more difficult
17:44 sfan5       ideally it would just set the correct address and remote_port for local servers
17:44 Zeno`       so to answer "what else am i supposed to use"... I really don't know yet :(
17:45 Zeno`       sfan5 then perhaps adding the address and port to Client::Client() may be the most straightforward solution
17:45 sfan5       why?
17:46 Zeno`       so g_settings can be avoided
17:46 sfan5       why do we want to avoid g_settings?
17:46 sfan5       g_settings is fine
17:46 Zeno`       because they're not updated until after a successful connection
17:47 Zeno`       there are lots of ways to solve this; which is what I meant by "I don't know what the best way is"
17:47 sfan5       how does this work then?
17:47 sfan5       if they are only updated after opening a map folder how does it work?
17:47 Zeno`       ok, do this:
17:48 Zeno`       a) connect to a remote server
17:48 Zeno`       b) quit to main menu
17:48 Zeno`       c) change to server tab
17:48 Zeno`       d) create a new local server
17:48 Zeno`       e) connect to that new local server (from the server tab)
17:49 Zeno`       and see where the local map is saved to...
17:49 Zeno`       it's not what you'd expect
17:50 sfan5       i know the bug
17:51 sfan5       i was asking how the current code works if the settings are only updated after a successful connection
17:53 Zeno`       because writeSettings() or whatever it's called is called as part of the destructor of Game (or something like that)
17:53 Zeno`       Shall I attempt to write a fix? I'm not worried if you don't want me to, heheh
17:54 sfan5       please do
17:54 Zeno`       ok, I'll do that tomorrow and make a gist
17:54 sfan5       where is remote_port even set
17:54 celeron55   here's my reply https://forum.minetest.net/viewtopic.php?f=7&t=5432&p=169057#p169057
17:54 sfan5       in builtin apparently
17:55 Zeno`       celeron55, none of us understand C++11 though :(
17:55 Zeno`       anyway, I must sleep!
17:55 sfan5       celeron55: ubuntu 12.04's gcc does not support c++11
17:55 sfan5       or at least there were some problems
17:55 celeron55   ubuntu 12.04 is practically obsolete
17:56 celeron55   stop using it
17:56 sfan5       tell that the people at travis
17:56 Zeno`       there are a lot of LTS linux versions that do not (properly) support C++11 (and, yes, including travis)
17:56 sfan5       the minetestmapper job only builds w/ CXX11=1 when using clang
17:56 celeron55   ubuntu 12.04 has gcc 4.6, it's new enough for many C++11 features
17:56 celeron55   it doesn't have some of them, but many
17:57 Krock       celeron55, IF C++11 is used, I would like to continue my weekly builds. This requires the implentation of those parameters. I don't know much about this, therefore I'm worried about the future
17:57 sfan5       i don't exactly remember what was wrong with CXX11, minetestmapper and gcc 4.6
17:57 celeron55   Krock: implementation of what?
17:57 sfan5       Krock: just update to a newer visual studio i guess
17:57 Krock       https://github.com/freeminer/freeminer/issues/157
17:58 Zeno`       if C++11 becomes accepted then where does that leave people who cannot compile c++11?
17:58 Krock       sfan5, I think the additional packages which come with VS matter, not the program itself
17:58 Zeno`       I'm not suggesting that c++11 should never be accepted, but... LTS, LTS, LTS
17:59 celeron55   microsoft messing up windows XP support is not going to stop me pushing C++11
17:59 celeron55   it will work with mingw anyway in any case
17:59 sfan5       Zeno`: merge? [y/n] http://sprunge.us/DUES?diff
17:59 celeron55   Zeno`: latest ubuntu LTS is 14.04
18:00 Zeno`       celeron55, not the entire world uses ubuntu
18:00 celeron55   if other distributions can't catch up to even something as old as 12.04's gcc 4.6, it's their problem, not ours
18:01 Zeno`       no, not all enterprise distributions support a version of gcc that is compatible
18:01 Zeno`       anyway, it will happen when it happens
18:01 sfan5       >enterprise
18:01 celeron55   it looks like it will never happen
18:01 celeron55   switch your distro today
18:02 sfan5       this reminds me of the guy who insisted that minetest was an enterprise app because it can be run on servers that cost upward 500$ a month
18:02 VanessaE    sfan5: I remember that :P
18:02 Zeno`       :/
18:02 Zeno`       that's not what I am saying
18:02 Zeno`       by anyway
18:02 Zeno`       s/by/but
18:02 sfan5       Zeno`: what about my patch
18:03 Zeno`       sfan5, I'm not sure... I think I'd have to test
18:03 sfan5       Zeno`: if my patch does not fix the issue something's seriously wrong with minetest or my patch
18:05 sfan5       Zeno`: would be nice if you could test it though
18:05 Zeno`       yep, I will for sure
18:06 celeron55   http://strawpoll.me/3514458
18:06 celeron55   here's my GCC version poll
18:07 celeron55   (i don't want to require forum login for a simple thing like this)
18:07 VanessaE    celeron55: my answer to that poll:  whatever is supplied by the distro whose "current LTS" is the most outdated. :)
18:07 celeron55   well go and answer that
18:07 celeron55   the answer to that is going to stay the same for half a year anyway
18:07 VanessaE    I've got 4.8.2 here, Xubuntu 14.04
18:07 sfan5       4.9.2 here, arch
18:08 celeron55   VanessaE: input it to strawpoll, otherwise it doesn't count
18:09 sfan5       (yes, i voted on strawpoll)
18:09 VanessaE    ubuntu 12.04 has gcc 4.4.7, fwiw
18:09 VanessaE    and yeah, I voted
18:10 VanessaE    hm, 4.7.2 on my server box, can't change my vote :-/
18:11 celeron55   ubuntu 12.04 has gcc 4.6
18:11 sfan5       VanessaE: whats your server os? my ubuntu has 4.8
18:11 Brains      Rolling release FTW.  And grey hair.
18:11 VanessaE    celeron55: oh yep, you're right.
18:11 VanessaE    sfan5: debian 7.4
18:13 T4im        debian: old-stable 4.4.5,  stable 4.7.2, testing 4.9.1, unstable 4.9.2
18:14 T4im        testing common for clients, stable common for servers (and if they are either badly maintained or have certain stability requirements old-stable)
18:15 VanessaE    I'd say that's accurate, or deb-stable for the server, and at least some ubuntu variant for the client anyway
18:15 VanessaE    (which is more or less the same thing)
18:15 Zeno`       what about RedHat?
18:16 celeron55   this is the time of debian's release schedule where debian stable is actually somewhat up-to-date as latest stable was released this month
18:17 est31       upstream bug for the audio problem: https://bugs.launchpad.net/ubuntu/+source/openal-soft/+bug/1416042
18:18 Calinou     http://packages.ubuntu.com/search?keywords=gcc
18:18 Calinou     4.6.3 in precise
18:18 VanessaE    yeah, I was looking in the wrong repo
18:19 VanessaE    4.6 there, 4.7 in debian-stable, 4.8 recent client OS's, seems to me 4.7 is probably a good target?
18:20 celeron55   (looks like i actually forgot 4.5 from that poll, but no distro is using it anyway)
18:20 celeron55   well, the poll is getting so few votes that it's kind of useless
18:21 celeron55   but what i'm proposing is to start requiring 4.6 so that we can use many (not all) C++11 things
18:22 Calinou     people should upgrade to 14.04 already :\
18:22 Brains      Maybe a revised requirement after a nice stable release, gives some time for the laggards to catch up / make arrangements / whatever.
18:22 celeron55   basically the only distributions that anyone uses which fail the requirement are redhat and freebsd
18:22 sapier      I don't like half grown solutions I'd vote for requireing full c++11 support if we wanna use it at all
18:22 VanessaE    centos7 has 4.8 available...
18:22 celeron55   sapier: nobody supports full C++11
18:22 VanessaE    fedora 21 does 4.9
18:22 sapier      clang does
18:23 VanessaE    not sure what's actually current in that distro though
18:23 sapier      gcc 4.8 too
18:23 celeron55   we can't require gcc 4.8, it's not in debian stable and that's quite a strict limitation that i won't go past
18:24 celeron55   debian stable has 4.7
18:24 Calinou     requiring what Debian stable requires sounds good.
18:24 kilbith     jessie will be released very soon
18:24 celeron55   4.7 basically has everything except the threading stuff
18:24 Calinou     but lots of users won't upgrade as usual
18:24 sapier      I don't wanna have to permanently look for errors happening because of incomplete c++11 support ... if we don't demand it it's more then likely we'll have to find workarounds for missing features anyway
18:25 celeron55   sapier: this page is quite accessible: https://gcc.gnu.org/gcc-4.7/cxx0x_status.html
18:25 sapier      celeron55: and thereading stuff is the part about c++11 I miss most
18:25 T4im        how do you know kilbith? :/ they always tell that a very short time before they actually do… and till then its just waiting that the bugcount goes down
18:25 VanessaE    RHEL6.x seems to have gcc 4.7
18:25 T4im        debians clang versions: old-stable 2.7, stable 3.0, testing and unstable 3.5
18:26 sapier      ok most important part (atomics) of threading is supported in 4.7 too ... yet it's quite ugly
18:26 VanessaE    gentoo has 4.8...
18:27 VanessaE    ok I guess that's enough :P
18:27 Brains      4.8.3 is stable on funtoo (and presumably gentoo)
18:27 celeron55   clang 3.0 is pretty bad
18:27 celeron55   3.5 is fine
18:27 sapier      well the main question is how much additional reports about compile errors do we wanna have
18:28 kilbith     T4im: it's frozen since november, so it's incoming
18:28 celeron55   many competent C++ programmers already hate C++03, how many lost programmers do we want to have?
18:28 est31       sapier: Isn't it possible to set up cmake to do version checking?
18:29 celeron55   doing that is very easy actually
18:29 est31       Then the build doesnt even start, and people know their compiler is not recent enough
18:29 sapier      and amount of compile errors depends on someone force pushing c++11 code the next months ... if we just continue development and use c++11 every now and then there most likely wont be a big issue but if someone does a "lets change everything to c++11" pull we'll have them
18:29 T4im        kilbith: november isn't that long ago… don't be surprised if they freeze for half or a full year
18:30 Calinou     last time they froze between june and may
18:30 rubenwardy  How about we use Boost as well as C++11?
18:30 Calinou     they are trying to make it shorter
18:30 celeron55   rubenwardy: what?!
18:30 Calinou     rubenwardy, nice joke :P
18:30 sapier      *throwing things at rubenwardy for silly jokes*
18:30 rubenwardy  XD. C++11 is cool, just don't let it make the code stupid.
18:31 sapier      c++11 has some nice features true, yet I don't think it's spread enough yet ... at least not till debian stable has full support :-)
18:33 Calinou     yes, let's wait for jessie release
18:33 celeron55   well here's a story: we wait until debian 8 is released, having gcc 4.9 with full C++11 support; everyone is happy for all the time?
18:33 sapier      of course that's  personal opinion only ;-)
18:33 celeron55   it was a short and happy story
18:34 celeron55   now
18:34 celeron55   except that then ubuntu 14.04 is not supported as it will only have 4.8
18:34 * Brains    wonders what the story is behind the person running minetest on a host stuck on 4.2 is...
18:34 celeron55   and the next ubuntu LTS is coming after 1.5 years from that point or something
18:34 sapier      doesn't 4.8 provide full c++11 support?
18:34 celeron55   does it?
18:35 Brains      4.8.1 provides almost all of C++11, IIRC.
18:35 sapier      according to gcc docs it's c++11 feature complete
18:35 celeron55   hmm actually it does
18:35 celeron55   Brains: freebsd has 4.2 and will never update
18:35 celeron55   they're taking clang into use... some day during the next century
18:36 sapier      haven't there been rumors about some bsd version switching to clang?
18:36 sapier      ok you heared same rumors
18:36 celeron55   freebsd is hopeless in my books, it's not worth looking at regarding to this issue
18:36 celeron55   i assume they'll update once minetest needs newer than 4.2
18:37 celeron55   otherwise they're not interested
18:37 celeron55   because minetest is obviously the killer game for freebsd
18:37 celeron55   they don't have other games!
18:37 sapier      hmm dropping bsd support doesn't sound right to me
18:37 sapier      but I don't see a better option too
18:38 rubenwardy  Can't you cross compile at all?
18:38 celeron55   http://www.freshports.org/lang/gcc48
18:38 celeron55   they can just install that
18:38 celeron55   it's just that the system doesn't come with anything newer due to licenses
18:38 est31       its not dropping support. You can still download the compiler, and let it compile minetest.
18:39 celeron55   wait what https://www.freebsd.org/releases/10.0R/announce.html
18:39 Calinou     there is Red Eclipse for FreeBSD, official builds :P
18:39 * T4im      wonders if the compiler can then be compiled
18:39 celeron55   does this mean that freebsd 10 has clang and it has been released?
18:39 est31       T4im: yes
18:39 sapier      so no issues with freebsd
18:40 celeron55   they have clang 3.3 as it seems
18:40 T4im        5 November 2012 Clang becomes the default compiler in FreeBSD 10.x
18:40 Calinou     which Clang versions fully support C++11?
18:40 celeron55   which has everything
18:40 celeron55   http://clang.llvm.org/cxx_status.html
18:41 celeron55   3.3 is the one that has everything
18:41 Calinou     3.3 apparently
18:41 est31       T4im: GCC can be compiled with every compiler supporting C++98 https://gcc.gnu.org/install/prerequisites.html
18:41 Brains      As an aside, anybody done any benchmarking of gcc vs clang builds?
18:41 T4im        ah, nice
18:41 Calinou     performance is probably very similar
18:42 Calinou     Phoronix frequently benchmarks GCC vs Clang, the conclusion is that performance is often the same.
18:42 Brains      Calinou: It should be...  The devil is in the details usually though.
18:42 T4im        performance might even vary between releases
18:42 T4im        performance difference*
18:42 sapier      so debian 8 release would be all major distrubutions having at least gcc 4.8 or clang 3.3
18:43 T4im        debian 8 would even come with both ;)
18:43 celeron55   and i declare windows to officially not be a problem at that point
18:43 celeron55   because it's not allowed to be
18:43 celeron55   if it is, it can go straight to hell
18:44 * Brains    is kinda impressed by how much VS2013 apparently doesn't support of C++11... https://msdn.microsoft.com/en-us/library/hh567368.aspx
18:44 celeron55   now everyone knows when MT will move to C++11 (at debian 8 release) and why (then everything supports full C++11 out of the box) and cannot complain, and everyone has plenty of time to get ready
18:46 celeron55   Brains: well, i hope the MSVC people are ready to continuously write pull requests for fixing problems related to those
18:46 VanessaE    celeron55: so, about 5 years from now :P
18:47 celeron55   VanessaE: factually probably about half a year
18:47 celeron55   or less
18:47 celeron55   or more
18:47 celeron55   that's what about means
18:47 VanessaE    I was kidding :)
18:48 Brains      celeron55: I kinda hope they aren't...  I suffered through their initial template implementation years and years ago.  I'd like them to suffer a bit in penance now...
18:48 rubenwardy  Many Minetest players use Windows. Mainly because they're forced to by parents or by game companies.
18:48 rubenwardy  Or don't care about FOSS and Linux is unfriendly to them
18:48 Brains      rubenwardy: MSVC isn't the only option for Windows folks AFAIK...  (Not having done it myself)
18:49 T4im        path dependency… "I need it for <x>" - "no you don't, you haven't even tried the open source equivalent yet"
18:50 T4im        Brains: don't they have mingw)
18:50 * Brains    dualboots, but that is more for games in the middle...  Older games (emulated) and newer games (native) work fine.
18:50 T4im        ?*
18:50 Brains      T4im: That is what I was thinking but the build instructions in the README are what I was going off of and they even say they are outdated.
18:50 Krock       hmm I wonder if it's possible to use MSVC libraries for GCC
18:51 celeron55   it seems that mingw might not support C++11 concurrency but mingw-w64 does
18:51 celeron55   mingw-w64 is anyway considered the better alternative so that's not an issue
18:51 celeron55   (and it builds 32-bit binaries regardless of the name)
19:38 rubenwardy  Two core devs have +1. This can be merged. https://github.com/minetest/minetest/pull/2185
19:38 rubenwardy  #2185
19:38 ShadowBot   https://github.com/minetest/minetest/issues/2185 -- Fix NDT_GLASSLIKE normals by kahrl
19:40 rubenwardy  It's a bug fix. Although maybe wait if the release is in progress
20:49 sfan5       Zeno`: ping
22:06 sofar       sfan5: out of curiosity - I consider my pull request rather mundane and low impact, so it's relatively easy to merge. Do you guys have something like a merge window where after release all the trivial changes get added? Or do pull requests get randomly picked up as needed at any time in a release?
22:06 sofar       moreover I'd like to know since I'm contemplating making some game-changing patches next
23:02 hmmmm       <sfan5> hmmmm: is this something that can be merged while in feature freeze? https://github.com/minetest/minetest/pull/2228
23:02 hmmmm       yes
23:03 hmmmm       are you guys okay with dropping windows 2000 support?
23:15 Brains      We went with C++11 in the forseeable future, I think Win2k doesn't have a chance.
23:15 jasonjayr   unsolicited 2 cents: Folks who are still on win2k should just hop on linux if they don't want to/can't afford to upgrade ...
23:31 paramat     hi hmmmm, pre-release bugfixes and corrections for mgv7 are ready for review/merge https://github.com/minetest/minetest/pull/2229 good screenshots too
23:35 hmmmm       i'll have to take a look
23:36 hmmmm       paramat:  i noticed that mapgen v5 generates terrain relative to water_level which defeats the purpose of the water_level variable.  is this intentional?
23:38 paramat     that's fixed in these 2 commits https://github.com/minetest/minetest/pull/2175 also waiting for review/merge
23:43 hmmmm       ahh ok
23:54 exio4       has anyone ever used mt in w2k?