Time  Nick         Message
10:04 sapier       Any comments on defining doxygen style function documentation to be the official way of documenting functions in headers? ... not as in "You have to document everything in doxygen" but more like "If you want to document do it in doxygen style"
10:56 RealBadAngel hi
10:56 RealBadAngel i was lookin for embeddable database for mods
10:57 RealBadAngel what do you guys think about it: http://sphia.org/
10:57 RealBadAngel https://github.com/mkottman/lua-sophia
10:59 RealBadAngel sapier, ?
11:20 sapier       well doesn't seem to be very marture
11:22 sapier       I suggest using some tested and proven code not bleeding edge implementations no one knows how much bugs are hidden or how long they will exist
13:45 dsimon       RealBadAngel, would probably be best to implement it based on whatever backend store minetest is already using
14:02 sfan5        RealBadAngel: https://github.com/sfan5/minetest/tree/sophia-db already last year, the sophia database corrupted itself after use, maybe it's fixed now
14:14 sapier       I wonder why everyone tries to improve performance by replacing database instead of looking for more hard to find issues burried within our own code ;-)
14:33 PilzAdam     ShadowNinja, tabs are 4-space width
15:54 sapier       https://forum.minetest.net/viewtopic.php?f=3&t=9385 Please join discussion about android port singleplayer game
16:41 sapier       I've got a crash on android port in blitBackAll ... "block" is 0
16:58 sapier       come on guys if you call getBlockNoCreateNoEx you have to either be sure it wont return 0 or check for 0
17:11 rubenwardy   Are all players currently loaded into ram at once? How hard would it be to only load the players that are online?
17:14 sapier       not a big deal but how big are players?
17:14 sapier       hmm ok palyer size isn't the issue but player count
17:18 sapier       I wonder what's the source bug behind block not beeing available within finishBlock...
17:24 sfan5        PilzAdam: tabs have no defined width in spaces
17:25 PilzAdam     sfan5, they have to if we have a char limit for lines; in the old style guidelines it was defined to be 4
17:25 sfan5        I mean generally
17:26 sfan5        rubenwardy: it is not proven that the "RPi" subgame is really faster
17:26 rubenwardy   Yeah, I realise that XD
17:26 PilzAdam     I was referring to this: https://github.com/minetest/minetest/pull/1341#issuecomment-44471878
17:26 sfan5        *sigh*
17:28 sfan5        I wish compiling for android without libcrystax would not break horribly
17:38 sapier       it doesn't
17:39 sapier       I'm compiling with default ndk
17:39 sfan5        same here
17:40 sapier       you only have to fix wchar conversions
17:40 sfan5        I'm compiling with the android support library that does this for me
17:40 sapier       why do you use it if it doesn't work?
17:41 sfan5        I'm not sure whether it is the support libraries fault
17:41 sapier       obviously
17:41 sfan5        let me compile without it
17:42 sapier       I intended to use it too but didn't manage to get it work by now .. I'd be glad to remove the hack I had to add ... but imho that could be done later too
17:43 sapier       I'm gonna prepare my android additions for merge this weekend. Hopefully that's gonna stop ppl from fixing same issues over and over again
17:45 sapier       that doesn't mean the android version is perfect, there's a lot left to improve
17:51 sfan5        sapier: now it crashes instantly at startup
17:51 sfan5        so it was not the support libraries fault
17:52 sapier       if you can tell me how to use it I can try it
17:53 sfan5        cd $NDK_DIR/sources/android/support && ln -s . jni && $NDK_DIR/ndk-build && echo "Done!"
17:54 sapier       and how to use it?
17:54 sfan5        then link to $NDK_DIR/sources/android/support/objs/armeabi/libandroid_support.a and add $NDK_DIR/sources/android/support/include/ to the include path
17:54 sfan5        there probably is an easier way to do this when you use the NDK makefiles
17:55 sapier       ok seems to be more complicated
17:55 sapier       I'll check once I've released my new version
17:55 sfan5        uhhh.. what?
17:56 sapier       using the support library
17:56 sfan5        s/what?/wtf?/
17:56 sfan5        if doing some command line stuff to use a library is too complicated for you, you should not be working on an android port at all
17:56 sapier       broken wchar support causes various strange effects as it messes up memory handling completely
17:57 sapier       spoken by someone whos port instantly crashes ;-P
17:57 sfan5        it's not my fault
17:57 sapier       lets stop beeing personal and get back to real work
17:57 sapier       adding a lib doesn't help at all if it's not used
17:58 sfan5        are you joking?
17:59 sapier       I was but still support lib didn't work for me on my last try and I didn't want to spend the time to figure out if it's been my fault or if it's broken at all
17:59 sapier       especially as I have a workaround
17:59 sfan5        android's support library provides the wchar methods and overrides the header files of the libc
18:00 sapier       which causes quite a lot of issues if it's not done for all libs too
18:00 sapier       and at least irrlicht uses wchar too
18:01 sfan5        it works fine for me if I compile irrlicht without the support lib
18:01 sfan5        which branch at sapier/minetest is the right android one?
18:02 sapier       you obviously hace a completely different build env so I don't know where exactly the fix is
18:02 sfan5        I was asking which one is the newest one. I was not asking you to fix my problem
18:02 sapier       android_merge_9 will be in a few minutes right now it's a preleminary state
18:03 sapier       android_merge_branch_9
18:03 sfan5        thanks
18:05 sapier       I'm going to split some non android fixes from it this weekend ... and create _10 supposed to be the first pull request
18:06 sapier       if you get support lib to fly I may add it too, I really hate the workaround for wchar
18:07 sapier       http://animalsmod.comuf.com/downloads/Minetest-release_20140529.apk Updated to latest master and fixed some bugs, please test
18:07 sapier       sfan5 _9 is up to date
18:07 sfan5        ok
18:07 sfan5        the workaround needs to be replaced before release
18:08 sfan5        otherwise anything that is not in your const wchar_t *wide_chars won't work
18:08 sapier       if you manage to find a working solution I'm with you
18:08 sapier       but I don't consider it to be a blocker
18:09 sfan5        almost no other language works
18:09 sfan5        no obvisouly not a blocker
18:09 sfan5        why would people care about languages?
18:09 sapier       I don't think they will work anyway
18:09 sfan5        </ironic>
18:09 sfan5        why not?
18:09 sapier       android doesn't have propper language support
18:09 sfan5        who cares? we use gettext
18:10 sfan5        it is enough if we can get the system language from android
18:10 sapier       doesn't help anything if encoding is broken
18:10 sfan5        <.<
18:10 sapier       that's exactly what I doubt ... getting correct language from android ... but I'd be glad if you prove me to be wrong
18:10 sfan5        the encoding obviously needs to be fixed
18:11 CraigyDavi   The too many users message seems to be wider than it should be: https://cdn.mediacru.sh/jhNzu-bO8COa.png
18:12 sfan5        CraigyDavi: no, that's normal
18:12 sapier       error messages always have same size
18:12 sfan5        perhaps "not working as commonly expected" but not a bug
18:12 sapier       yes that's something to improve on having too much time :-)
18:13 CraigyDavi   Ok, I just thought that the error messages should be the same width as the main menu
18:13 sapier       depends
18:20 scientistnik hello evetyone!)
18:52 CraigyDavi   hi scientistnik
18:55 sapier       btw sfan5 those chars in my workaround are ansi chars not those from my language
20:28 RealBadAngel sapier, im not tryin to improve performance of database backends already used
20:28 RealBadAngel im lookin for easy to use data storage for mods
20:28 RealBadAngel which is not implemented at all
20:28 RealBadAngel so where here is improving performance? first it have to exists ;)
20:31 sapier       RealBadAngel: I wasn't talking about you specifically but about all those ppl who regularly try to improve performance by switching database ... have been quite some the last months ;-)
20:32 RealBadAngel i know
20:32 VanessaE     greetz, all
20:32 LemonLake    hi
20:33 RealBadAngel lua-sophia is interesting because of that:
20:33 RealBadAngel lua-sophia is a Lua binding to Sophia embeddable key-value database.
20:33 RealBadAngel The database stores only Lua string keys and values, for storing more complex types use a serialization library, such as MessagePack.
20:33 sapier       I don't deny it's interesting but a project being alive for less then a year without backing of any organization can't be called marture
20:37 ShadowNinja  A sane per-player data storage system IS needed.  players/ , a big file, or a file for every player is not sane.
20:38 ShadowNinja  players/ is always fully loaded and has O(n^n) save time
20:38 sapier       we already have a database backend is it capable of creating a second db to store player data in it?
20:38 sapier       or use the map db file?
20:39 ShadowNinja  sapier: Yes.
20:39 sapier       I wander what stops everyone from implementing it?
20:39 sapier       -a+o
20:39 ShadowNinja  sapier: It isn't a generic backed though, you'll need a separate class that implements it.
20:40 sapier       more then a week of work?
20:40 ShadowNinja  sapier: But we should probably use something like luasql.sqlite3 instead of developing a DB API in-house.
20:41 sapier       I'm not a friend of doing everything different
20:41 ShadowNinja  sapier: With a full API more.
20:41 sapier       is it really realistic to switch everything?
20:41 ShadowNinja  sapier: Switch everything?  Switch what?
20:42 RealBadAngel maybe better use leveldb for lua?
20:42 ShadowNinja  LevelDB might work.
20:42 ShadowNinja  With "playername containername" keys.
20:44 RealBadAngel https://github.com/marcopompili/lua-leveldb
20:44 sapier       ShadowNinja: sorry misread ... I'm not against using luasql.sqlite3 or similar especially as we didn't have to support it
20:44 sapier       choosing leveldb will make a android mips port impossible
20:46 sapier       leveldb isn't mips compatible
20:46 ShadowNinja  Hmmm, player file saving could be O(n) instead of O(n^n).  /me checks the code.
20:47 RealBadAngel google's database project not aviable for android?
20:47 RealBadAngel sounds like a joke
20:48 sapier       nope leveldb doesn't build on android mips
20:48 sapier       most android devices are arm
20:48 sapier       so this might be an acceptable limitation
20:49 ShadowNinja  Alright, it's O(n).
20:50 RealBadAngel quick google says it can run on mips with a patch
20:51 RealBadAngel https://github.com/xyhl515/leveldb-android/tree/master/libs/mips
20:51 RealBadAngel and heres library for mips
20:52 sapier       well mine didn't build for android
20:52 RealBadAngel i can see a patch for mips but is is from 2001
20:52 RealBadAngel lol
20:52 sapier       I'm using the official google sources ;-)
20:53 sapier       that's really strange
20:53 sapier       2001? I doubt that's a correct time
20:53 RealBadAngel https://groups.google.com/forum/#!topic/leveldb/q_9JmYoa3Uc
20:54 RealBadAngel theres discussion about it, and propably a solution too
20:55 RealBadAngel https://code.google.com/p/leveldb/issues/attachmentText?id=109&aid=1090001000&name=0001-Fix-Android-MIPS-build.patch&token=keBDOnBNaMhjGwcXXp2ZV9JY81M%3A1360111755013
20:55 RealBadAngel date is definitely wrong, in the link before they do talk bout this patch
20:56 sapier       yes that could fix the issue
20:57 sapier       I wonder why google doesn't merge critical fixes like that one
20:57 sapier       doesn't really improve my opinion about leveldb ;-)
21:01 RealBadAngel but we do already have it implemented
21:02 RealBadAngel just to add lua bindings for it
21:05 sapier       what do you mean with "it" right now?
21:24 sapier       didn't someone have a x86 android device in here?
21:28 RealBadAngel i do
21:28 RealBadAngel kitkat
21:28 RealBadAngel just need to reboot ;)
21:30 sapier       ok I've built for x86 too but I haven't tested it at all so I'd not expect it to work https://forum.minetest.net/viewtopic.php?f=3&t=9389
21:31 RealBadAngel ok, i will try it