Minetest logo

IRC log for #minetest-dev, 2014-03-02

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

All times shown according to UTC.

Time Nick Message
00:23 Miner_48er joined #minetest-dev
00:36 eeew joined #minetest-dev
00:41 sapier joined #minetest-dev
00:54 domtron joined #minetest-dev
01:17 domtron joined #minetest-dev
01:46 werwerwer_ joined #minetest-dev
02:48 domtron joined #minetest-dev
03:09 Miner_48er joined #minetest-dev
04:25 ImQ009 joined #minetest-dev
04:42 SmugLeaf joined #minetest-dev
05:26 robmyers joined #minetest-dev
05:59 nore joined #minetest-dev
06:17 salamanderrake joined #minetest-dev
07:07 PenguinDad joined #minetest-dev
07:16 troller joined #minetest-dev
07:43 proller__ joined #minetest-dev
07:46 Calinou joined #minetest-dev
07:48 domtron joined #minetest-dev
08:06 khonkhortisan joined #minetest-dev
08:14 domtron joined #minetest-dev
08:48 khonkhortisan joined #minetest-dev
08:50 domtron joined #minetest-dev
09:08 Selat joined #minetest-dev
09:10 Selat Can you just add Belarusian translation? https://github.com/minetest/minetest/pull/1161.
09:51 RealBadAngel joined #minetest-dev
09:59 Calinou joined #minetest-dev
10:11 xyz ~tell sapier address sanitizer will only work on full-eng android builds
10:11 ShadowBot xyz: O.K.
10:20 jin_xi joined #minetest-dev
10:24 Taoki joined #minetest-dev
10:25 Selat So what about translation?
10:25 pitriss Selat: someone told me that translations are possible only by website and translation pull request won't be accepted when i asked about czech one.
10:26 Selat To add translation to the web site I have to send pull request.
10:26 Selat It' new translation.
10:26 Selat Not modification of existed one.
10:26 pitriss Selat: no.. they told me that i need to translate it using weblate..
10:26 Selat *existing
10:27 Selat It just adds empty file for translation.
10:27 Selat I already sent two requests by weblate.
10:28 pitriss This can be also done by weblate then.. no need for pull request..
10:29 Selat xyz: So still I need to send a pull-requestt?
10:30 pitriss Ask someone here to create it and wait. Just be patient... Devs are busy often..:)
10:39 Selat Can anybody add file po/by/minetest.po or merge this pull https://github.com/minetest/minetest/pull/1161?
10:42 xyz pitriss: no this can't please don't tell people stuff you're not sure about
10:42 pitriss xyz: I am sure what someone here told me:)
10:42 xyz pitriss: I'm now telling you you can't do that from weblate interface
10:43 pitriss This was main reason why i give up with trying to translate:)
10:43 xyz pitriss: you can translate but you can't create a new language, what's not clear here?
10:44 pitriss xyz: I see that I can't do.. This was did by some mod here..
10:44 xyz what do you mean?
10:44 xyz nobody can create new translation from weblate interface
10:44 xyz you need to do that manually, then import them (by updating weblate's git clone)
10:44 pitriss I mean that empty languages was everytime I mentiooned that by some staff here
10:45 pitriss *was created
10:45 xyz meh, I'm too lazy to explain this once again; just don't tell people stuff you know nothing about, okay?
10:45 xyz one time more, you can't create new languages from weblate interface
10:47 pitriss I understand.. But ypi don't understand fact that someone here told me that *NO* pull request from ppl will be applied..
10:47 pitriss *you
10:47 pitriss if it is about translation
10:47 pitriss So i only told what i read here when I'm asked.
10:47 pitriss *eh I was asking..
10:48 xyz they probably meant that about pull request that modify existing translation, not add new
10:48 pitriss My english skills sucks today, sorry
10:49 xyz also I think it's bad that they just close them without, i.e. adding those translations to weblate; but that's another question
10:49 pitriss Maybe.. But they told me general *No* to translation pull requests.. There was no exception for empty ones.
11:15 Jordach joined #minetest-dev
11:52 tomreyn joined #minetest-dev
12:09 PilzAdam joined #minetest-dev
12:15 cj joined #minetest-dev
12:25 rsiska joined #minetest-dev
12:27 e1z0 joined #minetest-dev
12:35 iqualfragile joined #minetest-dev
12:44 salamanderrake joined #minetest-dev
13:03 __1__0 joined #minetest-dev
13:03 __1__0 hello
13:14 __1__0 there are ?
13:17 __1__0 i would like to reoganise the gui in the game, i must modify source code or create a mode ?
13:21 xyz what gui? main menu or?
13:22 __1__0 backpack,selection of items
13:23 __1__0 table's crafting
13:28 xyz that you can modify with mods
13:29 __1__0 =) in lua =)
13:30 __1__0 i love the idea
13:51 __1__0 left #minetest-dev
14:00 PenguinDad joined #minetest-dev
14:19 domtron_ joined #minetest-dev
14:20 sapier joined #minetest-dev
14:20 sapier xyz isn't emulator version a full-eng build?
14:21 xyz dunno but if you're running it on emulator you have a more important problem
14:22 sapier I try both ... emulator doesn't handle texture correct, no idea why
14:22 xyz because it's a fucking emulator
14:22 xyz I'm surprised you expect it to work
14:23 xyz even if it does it's damn slow
14:23 sapier for what I know emulators usually work ... while beeing extremely slow but thats not a functional problem
14:24 xyz no they usually don't
14:25 sapier well then you use wrong emulator ;-)
14:25 xyz based on my experience with virtualbox, vmware, qemu, bochs and maybe something more
14:25 sapier or you use them for something emulators are not suitable for
14:25 xyz see, that's what you're doing
14:26 xyz nice you figured it out yourself
14:26 xyz but you really aren't supposed to run a 3d app on android emulator
14:26 sapier maybe but ogles is quite limited so I expected it to be translatable or even doable in software
14:26 xyz and then complain that textures aren't working or whatever
14:27 xyz what's the problem then?
14:27 sapier none I just wonder if there are devices out there having same issues as emulator
14:27 xyz just build full eng rom for your device if you want to use clang address sanitizer (why, again?)
14:27 sapier e.g. like the npot problem
14:27 xyz there surely will be, considering how many devices are out there
14:27 xyz it's not a problem
14:28 xyz it's something you just fucking know, ogles1 can't handle npot textures
14:28 sapier I don't root this devices as it'd be useless for it's main purpose, test what user will experience
14:28 xyz it's written in docs everywhere
14:28 domtron joined #minetest-dev
14:28 xyz and you don't even need to know anything about it since Irrlicht's supposed to handle it for you
14:29 sapier xyz you're wrong about it, it's not related to ogles but to the device there are devices out there who can handle npot textures in ogles1 mode
14:29 sapier well I guess there ain't anything like ogles1 mode
14:29 xyz sapier: yes but you shouldn't rely on it
14:29 xyz sapier: where?
14:30 sapier my dad's huawei device supports npot texutes out of box
14:30 xyz that's nice
14:30 xyz now what?
14:30 sapier that's why it took some time to find out what's wrong ;-)
14:30 xyz what's wrong where?
14:31 sapier minetest icon is missing in credits ;-)
14:31 sapier it's 100x100
14:32 xyz ah it's bad
14:33 xyz you can run some opengles tracer and maybe it'll report those issues
14:33 sapier ok there's only one major problem on my list ... performance on vanessaE's spawn, I fixed the menu problem by replacing pause as well as chat menu by formspec based ones. this even removes a lot of code
14:34 xyz so where's your code?
14:34 sapier the android_cherry_picks branch on my repo
14:34 sapier still not completed, I plan to extract the non android related fixes and create pull requests for them first
14:36 sapier e.g. pause/chat menu replacement
14:49 xyz ugh
14:50 xyz what NDK version do you use btw?
14:50 sapier latest
14:50 xyz r9c?
14:50 sapier yes
14:50 xyz strange
14:51 sapier what exactly are you talking about ? the wchar issue?
14:51 xyz yes, maybe
14:53 sapier well that one is really strange for what I read latest android versions support wchars ... maybe it's related to the target version
14:54 xyz wtf is CALCDELTA by the way?
14:54 xyz can't say I love those changes to touchscreengui.cpp
14:55 sapier calculate delta handle overflows
14:55 sapier I'm not surprised ;-)
14:56 xyz yes, why's that a macro?
14:56 xyz and why aren't you surprised?
14:56 sapier I guess it's more a rewrite then a change
14:57 xyz yeah and I think it got worse but whatever
14:57 sapier bad habit of using macros for some recurrent things
14:57 xyz a rewrite which made it 3x the size, well
14:57 sapier worse in what direction?
14:57 sapier on the other hand I don't need that much changes in irrlicht
14:58 xyz there's now too much code
14:58 xyz what changes are you talking about?
14:58 sapier and I do handle digg and place events without using special buttons in gui
14:58 rubenwardy joined #minetest-dev
14:58 xyz I do too, now what?
14:58 sapier it's 1000 lines 300 dead ... so it's not mutch
14:59 xyz ok twice as much, but what's the benefit?
14:59 sapier it fit's minetests core architecture
14:59 xyz you skipped my question
15:00 sapier but what's the benefit --> it fit's minetests core architecture
15:00 xyz also you should probably remove the copyright
15:00 xyz yes but you skipped another one
15:00 xyz not the first time though
15:00 rubenwardy Let's troll freeminer by changing Minetest upstream so much, they can not auto merge!
15:00 sapier oh true ... as I said it's work in progress ;-) coypright notices are very last to fix ;-)
15:01 sapier rubenwardy I wont stop developing minetest in order to make live of freeminer easy ;-P
15:01 xyz rubenwardy: they've done that already
15:01 xyz sapier: please do
15:01 xyz also please answer my question
15:01 xyz <sapier> on the other hand I don't need that much changes in irrlicht
15:01 xyz <xyz> what changes are you talking about?
15:01 rubenwardy sapier, I wasn't meaning what you were doing
15:02 sapier last time I did look at irrlicht patch it contained some more code then the one I use now ... of course that's been 2 weeks ago
15:03 xyz this doesn't explain anything
15:03 sapier and rubenwardy I wont change code for reason of trolling for sure ... sound crazy to risk bugs for something like that
15:03 rubenwardy :P
15:03 xyz what code did you remove?
15:03 xyz from the patch
15:04 sapier xyz basicaly the touchscreengui has grown out of effort to merge stu's and yours, while i tried to fix them I realized it doesn't fit minetest architecture
15:05 sapier and it's not yet complete so maybe there's some useless code in there too
15:06 sapier e.g. those hundreds of lines defining names for debug output
15:06 xyz we'll see
15:07 sapier I guess in the end this will be about 500 lines of code
15:11 Taoki[laptop] joined #minetest-dev
15:11 xyz this sadly still doesn't mean much
15:11 xyz what if it's 500 lines of those CALCDELTA-tier stuff
15:13 sapier lol I know why I didn't want to push the wip code ;-P
15:13 hmmmm joined #minetest-dev
15:13 xyz no worries, I'll find things to complain about either way
15:14 sapier I know :-) and some of them are valuable too .. that's what review is for
15:14 sapier I'll replace the calcdelta by a inline fct
15:15 sapier or do you know about a existing function doing same?
15:16 xyz of course, just subtract them
15:16 xyz converted to int
15:16 harrison CYGNUS INTER ANATES
15:18 sapier sure this will work on 32 as well as 64 bit and 128 bit too?
15:19 xyz lel
15:19 xyz sure you need that much precision when comparing double clicks and stuff
15:19 xyz wait, it's not precision but you get the point
15:20 sapier at least getTimeMs is u32 input so this may already fail on 64 bit
15:20 proller joined #minetest-dev
15:21 sapier overflow wont cause precision failure but detection just wont work
15:21 sapier effectivly loosing event
15:21 sapier s
15:21 xyz perhaps
15:22 sapier of course that way is the "I just want to be really really sure" version ;-)
15:22 xyz too bad it still won't work in certain cases
15:23 sapier can you be more precise about the "certain cases"?
15:23 xyz getTimeMs overflows too
15:23 xyz your way of calculating delta is strange
15:24 xyz surely you know that one event happened before another yet you don't take this into account
15:25 sapier true ... ok delta calc is something I often get wrong .. true
15:25 xyz so why even do it like this if it won't work correctly anyway if event that happened later has overflown u32
15:25 sapier so fi c you know how to fix it your help is welcome
15:26 sapier -c -fi +if
15:26 xyz I don't think there's a real need to fix anything, just replace this macro with "delta = getTimeMs() - whatever"
15:26 sapier well I know there IS a way to handle "single" overflows
15:27 xyz of course
15:27 xyz and there should be a function written which computes difference between two moments in time
15:27 xyz returned by getTimeMs
15:28 sapier well imho this function should handle overflows correct too ... there's various cases where you need the delta between two time ms values
15:29 sapier guess it should be in porting so different 32/64/128 bit can be handled correct
15:34 Selat ыфзшукЖ црн нщг гыу ьуьыуе,
15:34 Selat sapier: why you use memset?
15:36 sapier to ensure there ain't garbish in there
15:37 sapier I know I could skip the explicit setting of some fields done later
15:37 xyz surely c++ is a hard language
15:38 Selat And why constructor is not suitable for such stuff?
15:38 sapier because it's a struct
15:38 Selat mm... struct = class with public by default.
15:39 sapier well ask irrlicht why they don't clean their fields
15:39 sapier I just know they don't because I had quite some issues with uninitialiued memory in that struct
15:39 xyz if it's pod then just do SEvent something = {0}
15:40 sapier and where's the difference?
15:40 xyz if it's not pod and you're doing memset ;_;
15:40 xyz 1 line of code
15:40 sapier you may have realized it's allocated with new
15:41 xyz uh uh?
15:41 xyz what do you mean?
15:41 Selat But it deleted in the same block.
15:41 Selat Why you don't want to use stack variables?
15:41 sapier you know we're quite low on stack memory on android?
15:42 sapier and pthread just ignores your request for more memory
15:42 hmmmm but how do you *know* that
15:42 hmmmm you should check with getrlimit and be sure
15:42 sapier because I tested
15:43 Selat Than maybe it's better to use static variable?
15:43 sapier and allocating won't fail if you have more stack so why use two different sets of code dependent on available stack memory
15:43 hmmmm because one is typically faster than the other
15:43 sapier might cause issues if you happen to call that function recursive
15:43 xyz surely stack will be brutalized by 1 SEvent variable
15:43 hmmmm well SEvent might be big
15:45 Selat And of course there is a big chance of using recursion in isHUDButton or other events.
15:45 sapier I know allocating isn't fastest thing to happen but it's done on user interaction so we're in slow code anyway
15:45 sapier I guess worst recursion level is 2 but SEvent isn't a small thing
15:46 sapier and 2 is enough to make a static version fail
15:46 Selat But you are not going to use recursion in such simple methods. So static will be faster.
15:46 hmmmm isn't the rule of thumb to not go more than 5 levels of recursion on embedded platforms?
15:46 sapier well I use recursion there as it's passing that event back to it's caller
15:46 xyz android is now embedded platform
15:47 sapier as I said I expect worst to be 2 but that's enough to render static variant useless
15:47 sapier https://gist.github.com/sapier/9308353 is this variant of calculating 32bit difference correct?
15:48 hmmmm int foo(blah, int recursion_level) { if (recursion_level > MAX_RECURSION) { dstream << blah blah blah ; return 0; } ... foo(blah->whatever, recursion_level + 1); }
15:48 hmmmm you should use UINT32_MAX
15:48 sapier I can't change the function signature
15:48 Selat sapier: in what function you use recursion?
15:49 sapier OnEvent -> translate (using the SEvent) -> OnEvent ... right now there's no situation (I know about) where a translated event has to be translated again ... but I can't guarantee
15:49 ImQ009 joined #minetest-dev
15:50 hmmmm anyway yeah getDeltaMs looks good in that paste
15:50 sapier And I don't think using static variables as stack variable replacement is good style
15:50 hmmmm or, of course, you could just make a getDeltaMs64 version
15:51 hmmmm for Windows they have GetTickCount64()
15:51 sapier argh ... I didn't expect to open up a issue that big for time calculation ... I was wrong
15:51 hmmmm but i think that might be >= vista only
15:52 hmmmm yeah, it is.... nevermind
15:54 Selat sapier: I can't find call to translateEvent in TouchScreenGUI::translateEvent.
15:56 Selat ok, I understand
16:03 zat joined #minetest-dev
16:31 sapier great ... u32 is typedefd to unsigned it
16:32 sapier why does anyone create a bitwidth datatype if it's just ignored at all
16:36 xyz lol
16:36 Calinou joined #minetest-dev
16:36 hmmmm i suppose the reason is for compatibility with platforms that lack POSIX conformance, and the idea is to replace that platform specific header file with the one for the appropriate arch
16:37 xyz cstdint to the rescue
16:38 sapier possible but results in effect you can't rely on overflow behaving as overflow
16:39 hmmmm it is pretty bad that they provide "uNN" types but they don't provide uNN_MAX macros
16:39 xyz so, since we're already using c++11 anyway
16:39 xyz why not use types from cstdint?
16:39 hmmmm we're not using C++11
16:39 xyz woah woah woah look
16:39 xyz I was quite sure the other day I saw something from c++11
16:39 sapier msvc isn't c++11 capable
16:40 hmmmm the last attempt to push C++11 that i'm aware of failed
16:40 xyz sapier: we don't have to use all c++11 features, just those that work reliably
16:41 xyz what if I told you snprintf is C++11?
16:42 sapier true but do you wanna discuss each time someone uses a feature again about this feature is supported? You know how eager everyone is about testin on all supported platforms
16:42 xyz yet no one seems to discuss anything
16:42 hmmmm snprintf() is supported by implementations earlier than C++11 though
16:42 sapier yes because right now we have rule no c++11
16:43 xyz sapier: smart people have already discussed many features and made charts showing what we can use and what we can't
16:43 sapier at least as long as we support msvc
16:43 hmmmm and older gccs
16:43 xyz sapier: if we have no-c++11 rule please replace snprintf with something
16:43 hmmmm if it doesn't compile on my machine, i'd say that's a big problem
16:43 xyz otherwise it's double thinking
16:43 xyz hmmmm: what's your gcc version?
16:43 hmmmm 4.2.2
16:43 xyz and why's that a big problem?
16:44 xyz it's your problem, don't you think?
16:44 hmmmm it doesn't support C++11
16:44 xyz yes
16:44 xyz still, it shouldn't be our problem
16:44 hmmmm ....
16:44 hmmmm so are you also going to claim that supporting MSVC isn't "our" problem either?
16:44 hmmmm who is "your" anyway?  freeminer?
16:44 xyz alright, alright
16:45 sapier what's benefit of switching to c++11 anyway?
16:45 hmmmm look if you want to go break shit, go do it in your own project, not minetest
16:45 xyz well, I see
16:45 sapier is there anything which can be done WAY BETTER in c++11?
16:45 hmmmm sapier:  not really.  i suppose iterator assignments.
16:45 hmmmm they repurposed the auto keyword for C++11
16:46 xyz and you'll try to stop c++11 just because it doesn't work with your 5-years old compiler on your PC, fine
16:46 xyz that's reasonable
16:47 sapier xyz I don't recall the reason for switching to c++11 can you tell me?
16:47 hmmmm why don't you just go use GetTickCount64() instead of detecting overflow in GetDeltaMs and drop Windows XP support because nobody uses that anyway
16:47 hmmmm dropping support is just wrong
16:48 xyz no, people do use winxp
16:48 hmmmm switch to C++11, and you'll break support for FreeBSD earlier than 10
16:49 xyz oh that's so bad
16:49 hmmmm xyz, what's your platform?
16:50 xyz hmmmm: my platform contains a compiler that's not 5 years old
16:50 xyz it's good enough
16:50 hmmmm yeah well your platform is shit.  we'll drop support for it.
16:50 hmmmm my opinion > your opinion
16:50 proller freebsd 9 have gcc 4.9
16:50 xyz sapier: simple stuff, like auto keyword, range-based for loops, maybe more stuff later
16:50 hmmmm ....it doesn't.  you add in a huge 3rd party additional dependency
16:50 proller in ports, or clang
16:50 xyz just some basic convenience
16:51 sapier hmmmm our porting layer only provides the ms functions only and I don't wanna use os specific functions in non porting code
16:51 hmmmm none of those things are good enough reasons to break msvc and older gccs
16:51 proller most world will switch to freebsd 10 in next 6-9 months
16:51 xyz it doesn't break msvc though
16:51 hmmmm sapier, i was being sarcastic
16:52 sapier xyz I didn't miss any of those features yet, I know they're good for specific usecases but are they worth it (right now)?
16:52 hmmmm none of them are worth it..... we already had this discussion with hdastwb on the forum
16:52 hmmmm (i think that's his username)
16:53 sapier usually we get those issues on switching platforms e.g. android, sailfish or whatever ... those often don't support latest code
16:53 xyz sapier: as I said, that's just convenience
16:53 xyz same for fixed size int types
16:53 sapier btw using isdigit in settings is a (small) problem too ;-) irrlicht has a function named same
16:54 proller irrlicht HUGE depency too, and xorg, and..
16:54 sapier I'm absolutely positive about the int types but I don't wanna do the work of chaning everything ;-) any volonteers?
16:56 sapier SEvent translated               = {0}; doesn't work
16:57 xyz changing what?
16:57 sapier error: invalid conversion from 'int' to 'irr::EEVENT_TYPE' [-fpermissive]
16:57 hmmmm the idiomatic C++ way to do that is to call the constructor with all fields zero (yuck :\)
16:57 ShadowNinja One of C++11's main advantages is Rvalues.
16:57 sapier of course I'm gonna add a 500 lines initializer ;-)
16:58 ShadowNinja It allows you to move objects rather that copying them and destroying the old one.
16:58 ShadowNinja Which should make containers like std::map faster.
16:59 ShadowNinja I believe the server requires Irrlicht.  What does it need it for?
16:59 sapier I'm sure c++11 has some nice features ... it wouldn't have made sense to write it if it didn't ;-) but as long as we support platforms without c++11 support we can't switch
16:59 sapier primary datatypes ShadowNinja ... don't ask, historic reasons
17:00 ShadowNinja u8, s8, u16, etc can be replaced with uint8_t, int8_t, uint16_t, etc.  Or can be typedef'd.
17:00 sapier NO TYPEDEFING for things like that
17:00 ShadowNinja sapier: What?
17:00 xyz sapier: what are those platforms? hmmmm's pc?
17:00 sapier you'll end up in not knowing anything about the type you have right now
17:00 celeron55 ShadowNinja: the server uses irrlicht's types like vectors
17:01 ShadowNinja I meant "typedef uint8_t u8;" and the like.
17:01 sapier which have been replaced by std ones ... or did we miss some of them?
17:01 ShadowNinja celeron55: So, not all of the code uses STL containers?
17:01 sapier the clean way is greping and replacing all u8 by uint8_t
17:01 sapier this way you stop ppl from copy n pasting the wrong types too
17:02 ShadowNinja <cstdint> is apparently C++11-only.  So we'll have to use <stdint.h> for a while.
17:03 sapier xyz any platform a dev in here uses is supported ;-) ... at least that's what I have realized ;-)
17:03 celeron55 i like c++11; i use auto, the new for loop and lambas, initializer lists and non-static direct class member initializers a lot in new code i write
17:03 xyz and stdint.h is c99
17:03 sapier didn't they deprecate the non ".h" things ?
17:03 celeron55 it's like a whole new language with all that stuff
17:03 hmmmm instead of dropping support for old platforms we should be picking up support for as many things as possible
17:03 hmmmm in my code I try to support as many of the weird UNIXes as possible
17:03 Naked joined #minetest-dev
17:03 celeron55 (and a lot better language)
17:04 sapier celeron55 what about code readability on using all that shiny new stuff? ;-)
17:04 celeron55 sapier: better than before
17:04 celeron55 everything is better
17:05 sapier wow :-) excessive using of old c++ stuff usually resulted in code only the author was able to read
17:05 celeron55 all those things move C++ a whole lot in the don't-repeat-yourself direction
17:05 celeron55 which i like
17:06 sapier well that's good but most of the platforms I usually have to work for don't support it so I can't really talk about it
17:06 celeron55 gcc 4.6 supports most things, gcc 4.7 supports practically everything
17:06 sapier well :-) 4.2 doesn't
17:07 celeron55 newest msvc supports most things, and newest clang supports practically everything
17:07 sapier ans sometimes my code has to compile on ancient platforms too
17:07 ShadowNinja for (std::vector<ActiveBlockModifier>:const_iterator iter = abms.begin(); iter != abms.end() iter++) --> for (auto &abm : abms)  MUCH better.
17:08 sapier oldest maintained platform I work for was built in last millenium ;-)
17:08 celeron55 ShadowNinja: yes, it's like fucking magic
17:08 celeron55 nobody can ever argue against that
17:09 celeron55 (except with portability)
17:09 sapier true those for iterator cascades are creapy
17:10 xyz sapier: but you aren't going to try to run minetest on it anyway
17:10 sapier but c++11 is no exchange for looking at your code ... there are plenty of locations in code which are crap no matter what language you use
17:11 sapier btw that delta calculation has a small but relevant bug ;-)
17:12 celeron55 19:01:51 <+ShadowNinja> celeron55: So, not all of the code uses STL containers?
17:12 sapier result of equal times is 0xFFFFFFFF ... while 0 would be better
17:12 celeron55 ShadowNinja: ^ eh, the vectors you do calculations with
17:13 celeron55 and store positions in
17:14 ShadowNinja celeron55: Ah, as in v3s16?  We could make our own templated Vector class.
17:14 ShadowNinja Fix for #1162: http://ix.io/aTm
17:14 ShadowBot https://github.com/minetest/minetest/issues/1162 -- Item.lua copy_pointed_thing error
17:14 celeron55 ShadowNinja: also whatever rectangles and colors and maybe something else
17:15 ShadowNinja A mod must be calling minetest.registered_nodes[x].node_punch manually.
17:15 xyz https://gist.github.com/xyzz/e42d6718f3fb2e1c009b
17:15 ShadowNinja celeron55: So it should be doable...
17:15 celeron55 ShadowNinja: yes it is, but what is the goal here?
17:16 celeron55 also, why not just distribute a portion of irrlicht for building the server, if that is the issue
17:16 ShadowNinja xyz: https://github.com/yvt/openspades/commit/ef5ccc6b2ede3d85c76598d59c7ca6ebfb9e5644
17:16 celeron55 it's BSD code, no need to rewrite
17:17 ShadowNinja Yes, that could be done.
17:17 xyz why ganoo?
17:17 xyz ShadowNinja: ^
17:18 ShadowNinja xyz: ganoo?
17:18 xyz ugh, I mean
17:18 xyz it's called c++11, not gnu++11
17:19 ShadowNinja xyz: Yes.  But older versions use --std=c++0x I beleive.  --std=c++11 is the new flag.
17:20 ShadowNinja (gnu must be a synonym for c)
17:20 celeron55 gnu isn't a synonym for c, it's the standard with gnu extensions
17:20 xyz so just use c++0x then
17:21 celeron55 (= not a standard)
17:21 ShadowNinja xyz: The c++0x name is depreciated.
17:22 celeron55 c++0x is the only thing gcc 4.6 will accept, and x++0x will be probably forever a valid alias for c++11
17:22 celeron55 so there's absolutely zero reason to use anything else
17:22 celeron55 c++0x*
17:25 Anchakor_ joined #minetest-dev
17:26 xyz right
17:27 xyz it's the most "portable" one
17:44 kahrl joined #minetest-dev
17:51 grrk-bzzt joined #minetest-dev
17:54 Selat So what about adding belarusian translation? I asked already about 5 times.
17:56 sapier is there a pull request?
17:56 sapier I don't speak belarusian so I can't do it ;-)
17:56 Selat Yes.
17:57 Selat I just want you to add empty file, than it'll appear in weblate.
17:58 Selat *then
17:58 sapier hmm that pull request isn't really shaped for merge
17:58 sapier I'm gonna cherry pick it but next time please cleanup your pulls
17:59 Selat How?
17:59 sapier rebase to latest master
17:59 Selat I don't know why this merge commits are added to the pull request.
18:00 Selat *these
18:00 sapier I assume you didn't create a separat branch for this ... that delete seems familiar IRC it was already merged
18:04 Selat So it's better not to do like this https://help.github.com/articles/syncing-a-fork?
18:06 xyz yes minetest git guidelines require you to use rebase instead of merge
18:06 sapier true, way of mergeing is somewhat strange her in mt and discussed every now and then
18:08 Selat this uses merge http://dev.minetest.net/Git.
18:10 sapier thanks, i fixed it
18:24 rsiska joined #minetest-dev
18:26 Selat sapier: can you merge this again? I forgot to add name of language.
18:26 Selat https://github.com/minetest/minetest/pull/1163
18:28 EvergreenTree joined #minetest-dev
18:28 EvergreenTree joined #minetest-dev
18:28 Selat Damn.. He is offline. I'll have to wait about a day again.
18:29 sapier is it right now?
18:29 Selat Yes.
18:29 sapier done
18:29 Selat Why I can
18:30 sapier we really should switch to gpu lighting
18:30 sapier light calculation is one of the main operations done by minetest
18:30 Selat I have not seen you at the list because you were on the top.
18:31 ShadowNinja Yes, we should.
18:31 ShadowNinja But that's a bit GPU intensive.
18:31 sapier it's crazy to see getInteriorLight as top non 3d-engine operation in profiling graph
18:32 sapier 5% of total time is used for this
18:32 proller and now in very buggy place
18:32 proller its^
18:34 xyz sapier: what about server?
18:34 sapier server?
18:35 sapier sorry give me a hint what you're talking about right now
18:35 xyz yes, how would lighting be calculated on a server?
18:35 sapier puhh no ide
18:35 sapier a
18:35 xyz lol
18:36 sapier sounds interesting ... but will cause things like vanessae's multiserver machine quite useless as you're gonna need hell of a machine for a server
18:37 sapier and I assume graphics cards are way more suited to calc light values
18:37 proller sapier, you want to remix and broke all rendering too?
18:39 PilzAdam sapier, light is currently calculated by the server
18:40 sapier PilzAdam getInteriorLight is called on client so at least that part is not done at server
18:41 mrtux joined #minetest-dev
19:18 sapier https://github.com/minetest/minetest/pull/1164 that one is for those of you complaining about me adding too much code ;-P
19:19 sapier -450 lines of code ;-)
19:21 sapier if noone complains within the next hour I'm gonna merge it
19:23 celeron55 it's funny how even this special-purpose declarative language is better than a hardcoded alternative
19:26 sapier All the code is already in guiformspecmenu so why have it in there twice
19:28 sapier another positive effect is we only have to change guiformspec menu to fix all the gui at once e.g. in case of adding high dpi menus
19:30 proller now too hard to start game with 240x240 screen 8(
19:30 sapier I guess it'll be hard no matter what we do ;-)
19:32 sapier I'm gonna merge #1153 any complains?
19:32 ShadowBot https://github.com/minetest/minetest/issues/1153 -- Fix for unloaded but active block problem by CiaranG
19:35 sapier btw does cherry-picking to master comply to our merge rules?
19:37 celeron55 cherry-pick is basically a single-commit rebase
19:38 sapier ok so it's fine
19:41 BrandonReese_ joined #minetest-dev
19:42 xyz sapier: please don't indent with tabs
19:42 xyz I mean, don't "align"
19:42 sapier I don't do this on purpose where did I do it?
19:43 xyz https://github.com/minetest/minetest/pull/1164/files#diff-18513665750ef5adf42b5ec29e14162eR973
19:43 damiel joined #minetest-dev
19:43 damiel joined #minetest-dev
19:44 xyz it's usually ShadowNinja who complains though
19:44 sapier argh ... the old function parameter topic again ... I guess I'm gonna switch this completely ... will result in some more lines of code but noone can complain about tabs there any longer
19:45 sapier why can't you guys(including github) just use sane tab widths ... we don't have mechanic typing machines nowadays ;-)
19:47 sapier xyz better this way?
19:48 BrandonReese joined #minetest-dev
19:48 xyz why can't you guys(including sapier) just use tabs the way they're intended to be used ...
19:49 sapier because there ain't a real use for tabs nowadays ;-)
19:50 xyz you can't really align stuff with tabs because other people may have different setups
19:50 sapier that makes tabs useless and we could just remove them completely ;-)
19:51 xyz useless why?
19:51 sapier especially if you consider the amount of time wasted on discussion ;-)
19:51 xyz that's only because you can't get it right ;-(
19:51 xyz I honestly don't think it's that hard
19:52 sapier well 8 spaces tabs are wrong on almost any screen ... thats why some crazy guys invented mixed mode indentions ... level 1 4 spaces ... level 2 1 tab ... level 3 1 tab + 4 spaces
19:53 sapier luckyly there aren't much projects left using this crazy schme
19:54 sapier -much + many
19:56 sapier I'm gonna merge #973 next if there's no reason not to merge it?
19:56 ShadowBot https://github.com/minetest/minetest/issues/973 -- improved win32 file version information by sapier
19:56 celeron55 i neither think it's hard, but i guess it depends how braindead your editor is
19:57 khonkhortisan joined #minetest-dev
19:57 sapier imho tabwith 8 is just waste of screen space :-)
20:06 mkn joined #minetest-dev
20:07 mkn joined #minetest-dev
20:08 sfan5 sapier: you forgot a comma
20:08 sfan5 ""Minetest"" is property of Minetest community don't use the name -> ""Minetest"" is property of Minetest community, don't use the name
20:10 sfan5 also RUN_IN_PLACE=n would be better than '"run in place build' or 'in system install build' IMO
20:10 Zeitgeist_ joined #minetest-dev
20:11 sfan5 also for copyright use 2014 instead of 2013
20:11 sapier :-) maybe I should just write 2015 ;-)
20:12 sfan5 and this one: minetest engine -> Minetest Engine
20:12 sapier are you sure? thought capitals are used for names only?
20:12 sfan5 I am not sure at all
20:13 sfan5 but it should not cause any problems to capitalize it
20:13 sfan5 (look nicer IMO)
20:13 sapier hmm I'm exactly oposite opinion ... we need someone else to decide ;-)
20:14 sapier VanessaE ? you're native ;-)
20:16 Zeitgeist_ joined #minetest-dev
20:23 proller joined #minetest-dev
20:24 damiel joined #minetest-dev
20:24 damiel joined #minetest-dev
20:31 PilzAdam sapier, shouldnt you use wgettext() in #1164
20:31 ShadowBot https://github.com/minetest/minetest/issues/1164 -- Replace builtin pause and message menus by formspec based ones by sapier
20:31 PilzAdam +?
20:32 damiel joined #minetest-dev
20:32 sapier no formspec text itself is plain not wchar
20:32 damiel joined #minetest-dev
20:34 xyz Minetest™ is a property of the Minetest™ Community®
20:35 xyz Minetest™ Engine® Core Main Application©
20:35 sapier isn't TM only for registred trademarks ? ;-)
20:35 sapier and I won't use any special chars they just cause issues
20:35 xyz minetestâ„¢.exe
20:35 xyz nope?
20:36 xyz you're no fun :(
20:37 xyz actually, if you want to improve win32 builds, get a certificate and sign the binaries
20:37 xyz that way people can be sure the builds are from the Minetest™ Organization®
20:38 sapier well you can do this if you want to improve it even more then this already does
20:38 xyz you can get it for free, just apply some google
20:38 xyz that was just a suggestion
20:39 sapier thanks but I usually avoid things from google ;-)
20:39 xyz did you really have to state this? I don't think anybody cares
20:39 sapier this commit is about filling the existing fields only ... they are there so why not use them
20:54 damiel joined #minetest-dev
20:55 grrk-bzzt_ joined #minetest-dev
21:01 PilzAdam joined #minetest-dev
21:13 Zeitgeist_ joined #minetest-dev
21:13 Zeitgeist_ joined #minetest-dev
21:25 salamanderrake joined #minetest-dev
21:57 ShadowNinja sapier: "Minetest engine"
21:59 sapier ok
23:00 proller joined #minetest-dev
23:19 sapier left #minetest-dev
23:22 * VanessaE peeks in for a moment
23:23 VanessaE "tm" isn't "reserved" for anything, you can write "tm" after a trademark without said trademark being registered, normally only (R) is used for registered.
23:24 VanessaE it might be most accurate to write:   "Minetest", "The Minetest engine", and the Minetest logo are trademarks of the Minetestâ„¢ community.
23:24 VanessaE er, the TM would be used in all three places if you wanted to be anal about it
23:25 VanessaE er four.
23:34 VanessaE thing about trademarks though is that, at least under US law, if you have one that's important to you, you're expected to eventually register it and you're required to actively defend it at almost any cost (literally) if you want to keep it.
23:38 VanessaE and there's the secondary issue of what a certain other company might have to say about the MT community claiming a trademark.  there's a phrase in US law called "trade dress" that basically says that a product is defined by what it does and how it looks, and certain competitors might not look too kindly on the "Minetest" name if a trademark is claimed on it.
23:39 VanessaE (e.g. the name and look are so similar that it could get us into trouble)

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