Time  Nick         Message
02:47 thexyz       not releasing until #910 is fixed
02:51 kahrl        any progress on that one?
02:57 VanessaE     imho the entity duplication and inventory crash bug (#831) need fixed before 0.4.8 also..
02:58 kahrl        those are not regressions though
02:58 kahrl        not the entity duplication anyway
02:58 kahrl        not sure about 831
02:59 VanessaE     no one said it had to be a regression :)
03:01 kahrl        so we should not release 0.4.8 at all? :)
03:03 kahrl        I don't foresee anyone successfully investigating the entity duplication in the near future
03:06 VanessaE     I only mention it because as more people begin to use them for cool shit, more problems show up
03:23 kahrl        does anyone have a windows build ready where #910 happens?
03:23 kahrl        if so could you try dumping out the contents of wlabel at guiFormSpecMenu.cpp:458?
03:24 kahrl        thexyz: ?
04:41 kahrl        guess if nobody is working on #910 (and I can't because I don't have windows), we can't release 0.4.8
05:42 celeron55    what does it do on wine?
05:44 kahrl        I'm too lazy to set that or mingw up
06:01 kahrl        ugh who is adding all those issues to the 0.4.8 milestone
06:01 sfan5        -7
06:02 sfan5        oops
06:02 kahrl        at 43 issues the milestone is practically useless now
06:05 VanessaE_[T] weren't there only like 15 or so originally?
06:06 kahrl        yeah
06:06 VanessaE_[T] jeez
06:06 * sfan5      didn't add any new issues
06:07 kahrl        don't bug trackers normally log when somebody changes labels/milestones for an issue? does github discard that data?
06:07 celeron55    the github issue tracker is quite barebones
06:07 celeron55    so that probably goes to /dev/null
06:07 VanessaE_[T] no idea but I would have thought so
06:09 kahrl        why is there [WIP] stuff in the milestone
06:09 kahrl        proller: ?
06:11 celeron55    just remove them, you obviously know much better than whoever added 30 more issues to the milestone
06:12 celeron55    (we do need to know who did that though)
06:12 sfan5        kahrl: "std::cout<<wlabel;" makes gcc throw some wonderful errors at me
06:12 kahrl        sfan5: because cout is a char stream and wlabel is wchar_t string
06:13 kahrl        I'd be interested in the binary contents of wlabel
06:13 sfan5        binary.. ok
06:14 kahrl        for (u32 i = 0; i < wlabel.size(); ++i) dstream<<std::hex<<(unsigned int)wlabel[i]<<" "; dstream<<"\n";
06:14 kahrl        or something like that
06:25 sfan5        kahrl: prints out "wlabel: " in wine with ja language setting
06:25 sfan5        I hope that was helpful because I need to go now
06:27 kahrl        I assume you printed "wlabel: " above that loop?
06:27 sfan5        yes
06:27 sfan5        now bye
06:27 kahrl        okay that was helpful, thanks
08:18 proller      kahrl, i dont know why it gets milestone flag
08:18 proller      maybe automaticaly?
08:25 kahrl        proller: that could be it
08:28 kahrl        although when I open a new issue, it shows "No milestone"
08:31 kahrl        maybe it's just pull requests?
08:45 proller      yes, all was opened as pulls
09:39 kahrl        anyone mind if I merge #927?
09:44 VanessaE_[T] that one doesn't affect me, so knock yourself out ;)
09:48 Calinou      same
09:50 proller      927 must be merged without asking
09:51 kahrl        why didn't you do it then :P
09:53 proller      because i saw it 2 min ago
09:53 kahrl        it's fine I was just teasing ;)
09:56 kahrl        okay I removed the milestone tag from some of the least essential issues (ymmv)
09:56 kahrl        now we're down to 1 page at least
09:58 kahrl        proller: do you agree that it should be removed from controversial things like wingsuit?
09:59 proller      no 8)
10:00 proller      most of wingsuit is bugfixes
10:00 proller      controversial only <10 lines
10:00 kahrl        can you make the bugfixes a separate pull request?
10:01 kahrl        otherwise this is just going to stay in limbo and block the release
10:02 proller      i think it must be commited, and than, if peoples do not like it - simply disabled
10:02 proller      but it too hidden feature and no one will notice
10:02 kahrl        but as the code is now it can't even be disabled
10:03 kahrl        not even from lua
10:03 proller      i can make setting for it
10:03 proller      it can be disabled with one /* ..  */
10:03 proller      bit it setting no one newer will change
10:04 proller      but
10:04 kahrl        it seems weird to allow a client setting to modify the physics on a server, though
10:05 proller      better to commit and than make poll in forum
10:06 kahrl        maybe
10:07 kahrl        you need to convince those who were against it, though
10:10 kahrl        anyway, let's try to deal with at least 1 issue on the milestone list per day
10:11 kahrl        how about https://github.com/minetest/minetest/pull/528
10:11 kahrl        I think at the current rate of approving code it should be postponed
10:13 VanessaE_[T] <PilzAdam> not voxely enough <PilzAdam>
10:13 VanessaE_[T] (I like the idea though)
10:13 proller      i merged it to -next, and it works
10:13 kahrl        how does it perform?
10:13 proller      i'm +1 for 528
10:14 VanessaE_[T] dammit, missed the / in my joke
10:14 kahrl        if we want to move carts according to the same rules this seems to be pretty expensive on both the server and client
10:15 VanessaE_[T] merge it as long as it won't slow down the fixing of something else, e.g. something "big"
10:15 kahrl        maybe rails should switch to storing their configuration in param2, so just 1 node needs to be examined
10:15 proller      kahrl, what ?
10:15 kahrl        that is actually how MC does it
10:16 Calinou      rails are configured server-side on minecraft, yes
10:16 Calinou      you see them configuring as you place them (bad when you lag :P)
10:16 kahrl        well we could have prediction :P
10:19 VanessaE_[T] I think #651 needs to go in with it, to allow carts to rotate smoothly
10:19 VanessaE_[T] but it's not critical
10:21 proller      why sand not falling
10:22 VanessaE_[T] but some way needs to be added to allow an entity to stop at a predesignated spot, because of the lag question
10:22 kahrl        VanessaE_[T]: isn't that what 651 does?
10:22 kahrl        oh you mean a (x,y,z) spot not a yaw
10:22 VanessaE_[T] kahrl, 651 is for rotation, not movement
10:23 VanessaE_[T] yeah
10:23 VanessaE_[T] an x/y/z spot, or distance and bearing
10:23 VanessaE_[T] whichever maks he most sense in practice.
10:23 VanessaE_[T] the*
10:24 kahrl        perhaps just an axis aligned plane that acts as an additional collision plane
10:25 kahrl        or arbitrary plane if you want to be fancy
10:26 VanessaE_[T] perhaps
10:38 proller      all pulls with ok was tested and works - https://github.com/proller/minetest/blob/next_tools/util/next/report.log
10:40 kahrl        proller: tested how much?
10:40 proller      3 weeks on sky server
10:40 proller      an i sometime play as client
10:43 kahrl        of course they still have to be looked at in detail
10:43 kahrl        for example you listed PilzAdam:vbo as "ok" although it has massive memleaks
10:44 proller      not very massive
10:44 kahrl        depends on how you play and your hardware I guess
10:44 proller      but it muuuuch faster
10:45 VanessaE_[T] the vbo patch isn't THAT bad
10:45 proller      i'm ready ro restart minetest every hour to get this speed
10:46 kahrl        it didn't have any effect on framerate for me
10:46 proller      try press r
10:46 VanessaE_[T] it's a great improvement for me as long as I restart every 30-60 mins.
10:46 proller      or increase view distance to 500
10:47 VanessaE_[T] (after a while, the speed benefit falls off)
10:48 kahrl        brb, lunch
12:10 thexyz       celeron55: it reproduces in wine
12:14 kahrl        I wonder at what point the string gets truncated
12:14 kahrl        fgettext?
12:16 kahrl        might be interesting to examine the formspec before it is passed to engine.update_formspec
17:20 rubenwardy_  What is the name of the technique used to create caves?
17:20 rubenwardy_  Node trees or something?
18:38 sfan5        sophia ( http://sphia.org/ ) DB backend pull request incoming!
18:42 PilzAdam     wouldnt it be better to wait for libmtmap?
18:44 sfan5        maybe
19:01 thexyz       hmm
19:01 thexyz       we should've used it instead of leveldb
19:01 thexyz       or that hyperleveldb
19:01 thexyz       PilzAdam: who's working on that thing?
19:07 thexyz       I think I should rephrase this. Anyone working on that thing? How will including another backend (I guess it's as simple as LevelDB to use) make it harder to implement libmtmap? Also, I'm not really sure what's it for.
19:08 PilzAdam     thexyz, well, better implement it once in libmtmap than twice
19:09 thexyz       what do you mean "twice"?
19:09 thexyz       and could you explain shortly what libmtmap is?
19:09 PilzAdam     libmtmap is a libary that handles the saving of the world
19:10 thexyz       why do we need it?
19:10 PilzAdam     so other programs can use that libary to do stuff with Minetest worlds
19:10 celeron55    which doesn't exist but what hmmmm wants to make
19:10 thexyz       I doubt there are any programs except minetest mapper (and it's in Python)
19:10 sfan5        wat
19:10 PilzAdam     with "implementing twice" I mean: we would now implement it in Minetest, and then have to do it again in libmtmap
19:11 PilzAdam     thexyz, python mapper doesnt work with leveldb
19:11 thexyz       so I don't think we should delay backend-related pull requests because of that
19:11 thexyz       yes, it doesn't
19:11 sfan5        somehow it manages to corrupt libsophia's crc value
19:11 sfan5        and it won't load the db
19:11 PilzAdam     to make it easy for future mappers to support all kind of backend its nice to have a lib for it
19:11 thexyz       it is nice
19:12 thexyz       are there any ETAs?
19:15 sfan5        (currently not working version of sophia backend: https://github.com/sfan5/minetest/tree/sophia-db )
19:51 kahrl        eh, since hmmmm is gone for a few days I guess I'll do the libmtmap
19:52 kahrl        question: should that library include mapblock deserialization
19:54 kahrl        on the one hand we should, because we don't want to reimplement that in every program that uses libmtmap
19:55 kahrl        on the other hand that pulls in so many things that it isn't even funny (e.g. IGameDef)
19:55 kahrl        also remember hmmmm said libmtmap should be written in C
22:13 kahrl        any comments re: mapblock de/serialization in libmtmap?
22:14 kahrl        without this discussed I can't design the libmtmap interface
22:14 PilzAdam     I think it should be in it
22:15 kahrl        how would a mapblock be returned?
22:16 PilzAdam     the name->id vector could be passed to the function getMapBlock()
22:16 PilzAdam     (Im not very familiar with mapblock serialization, though)
22:24 kahrl        I wonder if it's even possible to make a C interface as all the serialization stuff relies on iostreams
22:24 PilzAdam     why does hmmmm even want it to be in C?
22:24 kahrl        dunno
22:25 kahrl        perhaps it's needed to call it from python?
22:26 kahrl        of course a benefit would be that if you have a libmtmap.dll, you can use it from whatever compiler you like, and don't need to worry about name mangling
22:27 VanessaE     any chance of making sure this lib can also be called/used from within a regular old bash script?
22:27 kahrl        um
22:27 kahrl        how do you use libs from bash?
22:28 VanessaE     well you know, e.g. via some simple command-line util
22:28 kahrl        I guess on windows you could use rundll
22:28 kahrl        well anyone is free to write such a utility based on libmtmap
22:28 kahrl        but it's outside the scope of the lib itself
22:29 VanessaE     fair enough
22:39 kahrl        although if someone submits a patch that allows loading the library via bash's "enable -f" I'd consider it