Time  Nick         Message
03:00 Megaf_       actually sapier's changes improved performance a lot
03:01 Megaf_       so much that I'm able to host 42 players spread on 3 server on my 100 KB/s upstream
03:01 Megaf_       on of those servers running on a Raspberry Pi
05:05 VanessaE     http://pastebin.ubuntu.com/7377483/
05:05 VanessaE     um, help?
05:05 VanessaE     I'm told that merely teleporting oneself to 0,31000,0 is sufficient to cause thi
05:05 VanessaE     this*
05:06 VanessaE     (which is beyond the legal limits of the map,  realize)
05:08 cg72         no vanessa 0 3100 0 didnt crash
05:08 VanessaE     no wait, ignore those coords :P
05:08 cg72         31000*
05:08 cg72         it just hung my client
05:09 cg72         the player to player did it
05:09 VanessaE     ok so bad block in the map obviously...  but how the hell did the map get corrupted?
05:09 cg72         (-645.87902832031, 21.5, 767.87200927734)
05:13 VanessaE     ok yeah
05:14 VanessaE     if there is a bad mapblock here I can't find it (ignore world load errors is set to true now, nothing looks broken)
05:14 VanessaE     so that error must be a false alarm
05:15 VanessaE     the real crash must be elsewhere, like you were thinking.
05:15 VanessaE     /teleport <one name> <another name>     crashes the server
05:15 VanessaE     but the error in the log is the above paste
05:16 VanessaE     and nothing in the logs indicates a bad mapblock now either (with ignore errors set to true)
05:16 cg72         i tp cool to recker 5 or 6 times each time crashed
05:17 VanessaE     what if you teleport one of them to some place really far away from that location, let's say over to some other place on the map?
05:17 VanessaE     and then  /teleport cool recker
05:17 VanessaE     does it still crash?
05:17 cg72         didnt try that this was at like 4-6 pmest
05:18 * VanessaE   sets ignore errors back to false and restarts the server.
05:36 hmmmm        hmmm
05:36 VanessaE     https://github.com/minetest/minetest/issues/1268
05:36 hmmmm        if I were to simply guess, I'd say that error would be caused because loadBlock() is being called directly from ServerThread, where the teleport is executed
05:37 hmmmm        whereas if it were encountered "naturally" it would've been queued in the EmergeThread and since it'd be out of bounds, it'd drop that block load request
05:39 hmmmm        to check if I'm correct, comment out "if (blockpos_over_limit_(p)) continue; in EmergeThread::Thread
05:39 VanessaE     what really confuses me is that the mapblock coords have zero relation to the target of the teleport
05:39 hmmmm        if it gives you world loading errors when merely "approaching" the limit then i am
05:41 VanessaE     map.cpp line 2539
05:42 VanessaE     lemme give it a shot
05:43 hmmmm        i doubt generateBlock is ever getting called...
05:43 hmmmm        only if that block didn't exist in the first place
06:00 VanessaE     hmmmm: that line on mine actually is a few lines, not just an if/continue pair ( http://pastebin.ubuntu.com/7377691/ ). Commenting out from if to the closing brace } is enough to stop the crash.
06:00 VanessaE     oops, copy&paste error.  forgot to widen my terminal.  meh, you get the point.
06:00 VanessaE     lines 2539 plus a few more in current HEAD.
06:01 VanessaE     the correct paste, just because I'm anal like that:  http://pastebin.ubuntu.com/7377712/
11:28 Megaf_       Good morning
11:43 RealBadAngel morning Megaf
11:52 Megaf        What do you mean john_minetest ?
13:01 sapier       any comments to #1267?
13:01 ShadowBot    https://github.com/minetest/minetest/issues/1267 -- Add hack to avoid 2s startup delay on local games by sapier
15:41 * VanessaE   pokes hmmmm in the -dev logs :)
15:55 Megaf        sapier: pulling latest commit right now
15:55 Megaf        * [new branch]      bluetooth  -> origin/bluetooth
15:55 Megaf        bluetooth branch?
15:55 Megaf        why?
15:55 sfan5        wait wait
15:56 ShadowNinja  Comments on #1258?
15:56 ShadowBot    https://github.com/minetest/minetest/issues/1258 -- Organize builtin into subdirectories and use "core" namespace by ShadowNinja
15:56 VanessaE     Megaf: wtf?
15:56 sfan5        o fuck
15:56 sfan5        Megaf: sorry my fault
15:56 sfan5        I probably wanted to push that to my fork
15:56 sfan5        ok, gone
15:57 Megaf        careful there sfan5, terrible things could happen
15:57 sfan5        it's only a branch
15:57 Megaf        now
15:57 BlockMen     ShadowNinja, nothing new. I still think that it should be engine, not core
15:57 sapier       ShadowNinja:  it's still 2:2 for or against renaming to "core"
15:57 * sfan5      is for "engine"
15:58 BlockMen     3:2
15:59 VanessaE     I'm in favor of "core" myself but I tend to use "engine" and "core" interchangeably(and most often lately, the two words together).
16:00 VanessaE     3:2.5   :)
16:00 ShadowNinja  Hmmm.  Why engine?  It's longer.  3:3 (VE's not technically a core developer, but she's at least a major contributor)
16:00 sapier       that's worst variant possible VanessaE cause this will result in having minetest. core. and engine. as possible variants in lua api
16:00 VanessaE     sapier: oh, no I wouldn't advocate to add both words.  no way :P
16:01 ShadowNinja  sapier: Only the menu uses engine, so modders only have to worry about minetest and core.
16:01 VanessaE     sapier: I just meant I tend to write "engine core" when describing the main engine versus some subgame
16:01 BlockMen     let the documentation at minetest for now
16:01 sapier       lets just keep it minetest, I don't believe we're gonna  get an agreement to any of those variants
16:02 sapier       and to be honest I believe we've got more pressing issues then renaming and causing lots of work for everyone
16:03 RealBadAngel i just thought the same, its gonna break everything out there...
16:03 sapier       no matter how you rename "minetest" that's gonna result in all mods to be fixed or be deprecated
16:04 sapier       and having minetest as legacy name isn't an excuse that's just delaying the work
16:04 ShadowNinja  Well, what about the first commit.  The second is mostly a cleanup.
16:05 sapier       can you cleanup that pull first? I don't wanna have broken comments in there as for last big commit
16:06 sapier       any comments on that ugly hackfix saving 2 seconds on local join?
16:07 ShadowNinja  sapier: Hmmm?  What's wrong with the first commit?
16:07 VanessaE     local core = engine or minetest ; function core.job_processor( ....
16:07 VanessaE     that's the first thing I noticed.
16:07 sapier       pull contains the core renaming ;-)
16:07 VanessaE     doesn't that be...
16:07 VanessaE     yeah
16:09 ShadowNinja  VanessaE: Yes, the other option was naming it "tbl" which is very generic.
16:09 VanessaE     quite so
16:10 sapier       btw how is serialization loaded for async env?
16:10 VanessaE     well in a situation like this, since the engine table already exists, I'd personally have used that.
16:11 VanessaE     (or at least, tried to)
16:14 sapier       ShadowNinja: ? misc_helpers.lua and serialize.lua ? wha aren't they (necessary) in your new version?
16:15 VanessaE     builtin/misc_helpers.lua → builtin/common/misc_helpers.lua
16:15 VanessaE     builtin/serialize.lua → builtin/common/serialize.lua
16:15 VanessaE     both renamed without changes it says here
16:15 sapier       they're not loaded at all
16:15 sapier       none of those options
16:16 VanessaE     shit
16:16 VanessaE     yes they are
16:17 sapier       where?
16:17 VanessaE     builtin/init.lua lines 22/23
16:17 VanessaE     https://github.com/ShadowNinja/minetest/commit/9c552429b28bf88d67f80551a74ae2d708783c1b#diff-82774cb1985a3f757730324fadbbb49fR22
16:19 sapier       I see ... lua self tells what to do from c++ ... well ... that's ugly ... but not ugly enough to block it
16:48 CraigyDavi   Hello. I seem to be getting a very annoying bug on my server. While it is happening this shows up in server window: ERROR[ServerThread]: Server::ProcessData(): Cancelling: peer # not found. The problem is, it lags out the server untill it is impossible to do anything, sometimes even join the server. It seems to be a recent bug, never noticed it in 0.4.9 stable.
16:49 CraigyDavi   Here's an example: https://gist.github.com/CraigyDavi/ffc94f7c0337cfc5d49c At the top the server is working perfectly fine. At the bottom I have to shutdown the server as there is so much lag I cannot even join let alone use /shutdown command.
16:49 VanessaE     Make sure you've fully updated your server to git HEAD
16:49 VanessaE     Megaf and I have also had this happen but this has theoretically already been fixed.
16:50 * VanessaE   pokes sapier
16:51 crazyR__     i might make a php based live formspec designer this week. :/
16:52 CraigyDavi   I'm using a build from April 19th. I don't think a very recent version may fix it but I'll try that.
17:05 rubenwardy   crazyR__: PHP doesn't do live :P I assume you mean JS
17:08 crazyR__     lol well to be honest, i didnt mean fully live, page refresh wich is much faster than designing on minetest atm. but yeah i could should probably use js you update it.
17:09 crazyR__     *js to update
17:11 sapier       CraigyDavi: what happened between 17:22:59 and 17:26:38?
17:14 sapier       CraigyDavi: ignore my last question and update to latest git it's already fixed there
17:16 sapier       you missed the fix by a couple of commits because it's been fixed at april 19th
17:24 CraigyDavi   Okay thanks. Ill update now.
17:46 Calinou      idea: change default font to Droid Sans (and bundle it)? it looks like that with font size 18 (which is better suited for large screens): https://cdn.mediacru.sh/wL5TpI1CHFqn.png
17:47 Calinou      this is the settings I used: http://0bin.net/paste/5O0h9hI1x3qOMd1R#+HYLFO6mUE2gGFPF0+aprhQ+cuDCu7eeOJZ2L/vZYuI=
17:49 PenguinDad   Calinou: Droid Sans is already bundled but it doesn't support äöü
17:49 Calinou      oh
17:49 Calinou      it seems to do, in the OS.
17:50 Calinou      indeed, only ASCII characters work
17:50 Calinou      this is strange because the main menu supports it
17:51 Calinou      oh, it also doesn't work in Liberation Sans anyway
17:51 Calinou      in this case, it shows something tab-like, unlike Droid Sans which shows nothin
17:56 Megaf        VanessaE: isnt head == master?
17:57 * Megaf      doesnt know what head means
17:57 Megaf        (I mean, in Git)
17:57 VanessaE     master refers to a particular branch where the code is stored
17:58 VanessaE     HEAD (all caps) refers to the very latest commit at any given moment in whatever branch you're looking at
17:58 VanessaE     upstream usually refers to the main minetest source code repo at https://github.com/minetest/minetest
17:58 VanessaE     (or whatever the project in question is, whosever repo is considered the main, lead repo)
17:59 Megaf        VanessaE: what the git pull command would look like then?
17:59 VanessaE     that's the simple view.  there's more to it that I'm not good at explaining.
18:00 sapier       as noone did complain I'm gonna push #1267
18:00 ShadowBot    https://github.com/minetest/minetest/issues/1267 -- Add hack to avoid 2s startup delay on local games by sapier
18:02 VanessaE     Megaf: well let's say you're on the master branch and you want to pull from the upstream master, then you can usually just `git pull` with no other arguments and it'll bring you up to HEAD on that branch.  If you have some patches applied to your local copy of master (say from someone else's repository), you can usually `git pull --rebase origin master` to pull from upstream, moving your local patches to the top of the list after the pull is
18:02 VanessaE     done
18:03 VanessaE     sapier: I have not tried it, but if that one-step delay cuts 2 seconds from the startup without breaking anything, and it can't be done better right now, I don't see anything wrong with the idea.
18:03 sapier       it's only for localhost
18:04 sapier       won't have more then 100ms effect on connecting to remote
18:34 CraigyDavi   Should there not be a comment above here which says "Timeout for server to remove unused map data from memory"? https://github.com/minetest/minetest/blob/master/minetest.conf.example#L321
19:51 PilzAdam     sapier, are you going to fix #1230 ?
19:51 ShadowBot    https://github.com/minetest/minetest/issues/1230 -- Incorrect handling of modpacks in mod configurator (per world)
20:14 sapier       I don't know a way to fix this without rewriting full mod handlin
20:15 sapier       those mods are disabled because of disabling a modpack named identical ... at least that's what I understood
20:15 sapier       or even worse because of another mod being called identical
20:16 sapier       by now modnames have to be unique, beeing part of another modpack doesn't remove that limitation
20:17 sapier       correct me if I did understand wrong
20:52 sapier       any comments to #1253? it's cleaning up game.cpp a little bit and adds support for polarization 3d screens (about factor3 faster then first try)
20:52 ShadowBot    https://github.com/minetest/minetest/issues/1253 -- Add support for line based polarized 3d screens by sapier
21:13 PilzAdam     sapier, no, the bug is that you just open the configure window and it disables a mod
21:14 PilzAdam     e.g. I have a modpack mesecons with the mod mesecons in it
21:14 PilzAdam     I click on the configure window, enable the whole mesecons mod, leave the configure window, and then reopen it
21:14 PilzAdam     without clicking anything else the "mesecon" mod gets disabled
21:15 PilzAdam     *the whole mesecons modpack
21:22 crazyR__     Is there any reason why the technic mod wont save the settings file.and also it wont read the default settings.
21:23 proller      sapier, can you move code in separate commits with code change?
21:24 sapier       no because that isn't actually a 1:1 code move but resorting
21:25 sapier       most time it's only a few lines between changes
21:26 crazyR__     anyone?
21:26 sapier       PilzAdam I'm gonna check this I've got an idea why this happens but I'm not sure if it's a quick fix for colliding names
21:26 sapier       crazyR__: I'm not sure but don't you have to explicitly write settings from within lua api to make them persistent?
21:27 PilzAdam     I think we should mark #1230 as blocker
21:27 ShadowBot    https://github.com/minetest/minetest/issues/1230 -- Incorrect handling of modpacks in mod configurator (per world)
21:28 sapier       ?
21:28 sapier       a blocker is something not beeing there in previous version
21:28 sapier       I'm not even sure if it's a bug modnames are not supposed to colide not to other mods and not to modpacks too
21:29 sapier       it's quite ugly that's true
21:29 PilzAdam     why is a blocker something that was not there in previous versions?
21:29 sapier       because it didn't block old version?
21:29 sapier       and it's not a crash or data corruption either
21:29 PilzAdam     I understand blocker as a major bug that has to be fixed before next release
21:30 sapier       I suggest not using "blocker" tag for any ugly issue around
21:30 PilzAdam     just because you don't find a bug fast enough doesn't make it less important
21:30 sapier       it's not about finding it's about is this really preventing a release ... and imho this isn't severe enough to block a release
21:31 PilzAdam     it makes the configure window unusable
21:31 PilzAdam     or, almost unusable
21:31 sapier       no it doesn't it's only unconveniant for modpacks with invalid mod naming
21:32 sapier       and that issue about two mods with same name is mentioned in there too, we can't fix this one without rewriting whole mod mechanism
21:32 PilzAdam     "unusable" and "inconvenient" are pretty close together in GUI
21:33 sapier       and I don't think we should fix this anyway as we've got plenty of things relaying on mods having unique names
21:33 sapier       e.g. textures, sounds, entity names
21:33 PilzAdam     yea, the mod naming isn't as important
21:34 sapier       not important? I'd not consider it an valid issue at all
21:34 sapier       we can discuss about modpack name, mod name collisions
21:35 PilzAdam     the important part is that things change without the user doing anything
21:36 sapier       I'm gonna check
21:39 sapier       seems to be a display issue
21:40 sapier       I'm gonna fix it in my mainmenu cleanup branch, I don't wanna work on code to be removed anyway
21:53 sapier       PilzAdam: check #1241 the modpack <-> mod name collision issue should be fixed ther
21:53 ShadowBot    https://github.com/minetest/minetest/issues/1241 -- Add formspec toolkit and refactor mainmenu to use it by sapier
22:32 crazyR__     is there a tile animation that rotates the texture?
22:37 CraigyDavi   crazyR__, yep: http://dev.minetest.net/texture
22:38 crazyR__     thanks
22:40 crazyR__     oh wait thats not it, thats just normal image manipulation, im after the animation side of it
22:45 CraigyDavi   Oh, I'm not sure then