Time  Nick          Message
01:49 VanessaE      help!
01:49 VanessaE      random server crashes with no log output
01:50 VanessaE      http://pastebin.ubuntu.com/6554237/
01:51 VanessaE      no segfaults, nothing anywhere in the logs, no evidence of out-of-memory conditions
01:51 VanessaE      it's happened four times in a row now, each time it seems to be associated with digging a pipeworks node (but that code has been stable for weeks)
01:51 VanessaE      if it were a mod error, there would be a backtrace
01:54 VanessaE      wraith says he placed a tube just prior to the crash
01:54 VanessaE      so I wonder if the crash is happening inside the rollback manager?
01:57 VanessaE      I am seeing if I can get wraith to sign back on so I can get him to cause another crash but this time with the server running under gdb.
02:02 VanessaE      http://pastebin.ubuntu.com/6554267/
02:02 VanessaE      um, wut?
02:03 VanessaE      why does it show pipeworks as directly at fault now?
02:03 VanessaE      when it shows no error or backtrace when not running under gdb
02:04 VanessaE      local backdir = minetest.facedir_to_dir(node.param2)
02:04 VanessaE      um
02:05 VanessaE      what's facedir_to_dir() supposed to return if given nil as its input?
02:05 VanessaE      (I didn't write this code so I'm not yet sure how to properly correct it)
02:05 VanessaE      oops, that belongs in #minetest
02:07 ShadowNinja   VanessaE: It returns nil.
02:07 VanessaE      hm
02:07 VanessaE      seems an "or {}" added to the above should be sufficient then
02:14 VanessaE      ok, that was not enough to stop it and the crash is NOT being logged
02:14 VanessaE      unless I run it inside of gdb
02:14 ShadowNinja   That would be just as invalud as nil, and it makes it harder to detect error conditions.
02:14 VanessaE      it just...restarts
02:18 VanessaE      well THAT commit that just now went into pipeworks should be right but it doesn't fix the fundamental problem of the engine not producing a backtrace when it dies.
02:22 VanessaE      it looks like it's throwing an error okay, but it isn't *logging* the error is the problem
02:23 ShadowNinja   This adds support for protection to the screwdriver,in addition to minetest.swap_node and better code style: https://github.com/ShadowNinja/minetest_game/commit/a4e4950db4d0a769a3a82f055c2bc7a64c1f167c
02:23 VanessaE      I caught it outside of gdb, the backtrace is on the controlling term but it isn't in the log
02:23 ShadowNinja   Anyone to revie?w
02:23 VanessaE      yep, that's the problem
02:23 ShadowNinja   ? >> 1
02:23 VanessaE      the engine is not logging the crashes
02:28 VanessaE      ShadowNinja: revie[0x1F]w  ?
02:28 VanessaE      ;)
02:31 VanessaE      no, something's wrong here.
02:31 VanessaE      ShadowNinja: https://github.com/VanessaE/pipeworks/blob/master/autoplace_tubes.lua#L29
02:31 VanessaE      damn it, wrong channel again
02:32 VanessaE      -->#minetest
02:37 ShadowNinja   https://github.com/minetest/minetest_game/pull/228
02:41 hmmmm         convertFaceToIndex() should be a LUT, not a giant if-elseif statement
02:41 hmmmm         that's ridiculous
02:48 ShadowNinja   hmmmm: Done. Do you know of a way to generate a entry from faces_table mathematically?
02:48 hmmmm         nope
02:49 VanessaE      ok, so the actual *crash* was the result of an obscure mod bug, but the lack of error messages in the log is the engine failing to write said errors to the log (they appear on the controlling shell though).
02:56 hmmmm         VanessaE, I actually came across the same thing for a new set of API i'm working on
02:56 hmmmm         absolutely no error messages at all, program just terminates immediately
02:56 hmmmm         wonder what ShadowNinja has to say
02:57 hmmmm         (in particular, inside of an on_generate callback, a pretty typical undefined variable error)
03:20 ShadowNinja   hmmmm: Nothing in the console?
03:21 hmmmm         nothing
03:21 VanessaE      ok, just happened again, this time while someone was placing some mesecons wires.
03:21 ShadowNinja   Well Vanessa had a error in the console I beleive.
03:21 VanessaE      again, no logging of any kind, just a restart, and the user reports that the server just...died.  no "Server shutting down" message
03:22 VanessaE      yes, that's correct shadow
03:22 VanessaE      the error was in the console, but not in the logs
03:22 hmmmm         oh, hrm
03:22 hmmmm         maybe there is an error in the console but it doesn't show up until after the debugger is done handling the signal
03:24 VanessaE      ok, I have a session open and crashed right now
03:24 hmmmm         oh duh
03:25 VanessaE      a Lua error in another spot in pipeworks (I will fix in a sec).  it's sitting in gdb waiting for me to resume
03:25 hmmmm         I just figured out what my own problem was
03:25 VanessaE      console has the error, debug,.txt does not
03:25 VanessaE      -,
03:25 hmmmm         i.e. blindly writing Lua interpreter code while being half asleep
03:25 VanessaE      heh
03:25 VanessaE      btw
03:25 VanessaE      it throws a SIGABRT
03:25 VanessaE      of all things
03:25 hmmmm         hrmm
03:26 VanessaE      fuck
03:26 VanessaE      hrm
03:27 VanessaE      there's something subtly different between the facedir_to_dir() that got pushed to git and the version that was being used previously in pipeworks
03:30 VanessaE      https://github.com/VanessaE/pipeworks/commit/7289bfcfbab13044b83720ff4c675a356b8dedc5
03:31 VanessaE      there's where I phased it out in favor of the default one
03:33 ShadowNinja   VanessaE: Two numbers got swapped because they were backwards.
03:34 VanessaE      pardon?
03:36 ShadowNinja   VanessaE: https://github.com/minetest/minetest/commit/4e5760a5416cbca6945b1b4484cbd96bea7b250c
03:37 VanessaE      well that may be, but that's not why my code is suddenly crashing
03:37 VanessaE      minetet.facedir_to_dir() is interpreting its input differently from what the function it replaced was.
03:37 VanessaE      +s
03:44 ShadowNinja   VanessaE: It didn't replace anything.
03:45 VanessaE      ShadowNinja: in pipeworks it did.  there was a copy of a substantially similar function included therein.  similar to the deal with lib6d.
03:45 VanessaE      I phased out the copy in favor of the one that made it into 0.4.8
03:47 ShadowNinja   Yes, but they look identical.
03:54 VanessaE      well at any rate I seem to have fixed it (or at least kludged around it ;) )
03:59 hmmmm         oh here we go
03:59 hmmmm         "terminate called after throwing an instance of 'LuaError'
04:00 VanessaE      heh
04:00 VanessaE      so the server crashed while trying to handle a crash? :)
04:00 hmmmm         I suppose so
04:00 hmmmm         oh wait
04:00 hmmmm         hold on a minute
04:01 VanessaE      "404, file not fould [...] in addition, the server encountered a 404 while trying to display a custom error page"
04:01 hmmmm         when you get a lua error in on_step or whatever, what happens?  do you get an error like you're supposed to?
04:01 VanessaE      ehm?
04:01 VanessaE      oh, no clue
04:01 VanessaE      I never use on_step
04:01 hmmmm         because I think you're supposed to try to catch LuaError
04:01 VanessaE      at least not directly
04:01 hmmmm         and because this went uncaught it's doing this
04:01 hmmmm         let me check something real quick
04:01 VanessaE      though I think pipeworks uses it by way of Nore's item transport code
04:02 VanessaE      yep, on_step is used there, though I doubt it's relevent.
04:04 VanessaE      inb4 "oh this is just shit, what fucking moron code this trash?"
04:04 VanessaE      :D
04:05 VanessaE      +d
04:09 hmmmm         so I don't get it
04:10 VanessaE      ?
04:10 hmmmm         so you need to catch LuaErrors in an try catch statement when calling a callback?
04:11 hmmmm         when your mod does something wrong, the server is supposed to terminate and it goes back to the main menu
04:11 hmmmm         and then display an error message of some sort
04:11 hmmmm         but right now it doesn't do that for on_generate callbacks because there's no catching of LuaErrors whatsoever
04:12 hmmmm         correct..?
04:12 VanessaE      it's always been the case in the past that the server would throw an error, log it, and dump you into a half-broken error screen, buy you never had to do anything special to get there
04:12 hmmmm         right
04:12 hmmmm         except now when a LuaError is thrown, it just explodes with SIGABRT like you said earlier
04:12 hmmmm         which I experienced earlier too
04:13 VanessaE      I should thing that would continue to be the proper behavior rather than through a SIGABRT
04:13 hmmmm         the "terminate called after throwing instance of ..." is a standard libstdc++ error message
04:13 hmmmm         alright, at this point it's pretty clear what the problem is
04:13 VanessaE      think*
04:13 hmmmm         but now my question is what happened for this behavior to change, and who did it, and what was the rationale?
04:14 VanessaE      I wish I knew
04:14 hmmmm         it was either sapier or shadowninja
04:21 ShadowNinja   hmmmm: That would be me. Before Lua errors didn't throw a LuaError, but there was no errorhandler.
04:22 hmmmm         well
04:22 hmmmm         I think the only other place you'd need to add a LuaError catch is in EmergeThread
04:22 hmmmm         I'll add that
04:23 hmmmm         seems to work exactly as intended if it is caught
10:13 RealBadAngel  hi, i would like to add simple feature to shaders
10:13 RealBadAngel  https://github.com/RealBadAngel/minetest/commit/d56214cb097ff0297e479501dd21d6d86c573b6d
10:14 RealBadAngel  anybody against it?
10:14 RealBadAngel  the code is to allow one overriding normal map for all nodes
10:51 proller       https://play.google.com/store/apps/developer?id=Vicono+Inc.
10:52 proller       https://forum.minetest.net/viewtopic.php?pid=121596#p121596
10:56 proller       with lot of adv
12:01 RealBadAngel  hello earthlings ;) again
12:02 RealBadAngel  can i push my change>
12:02 RealBadAngel  ?
13:09 thexyz        I wonder if Android is the most popular platform Minetest runs on now
13:10 thexyz        according to that google play page it has 10k-50k installs
13:34 celeron55     most likely not
13:36 celeron55     my estimate of minetest's installation base (i.e. how many computers have minetest on them) is more like 100k than 10k
13:36 proller__     20-30k
13:38 proller__     buildcraft have 10k-50k installs per 12 days
13:38 proller__     first comment on market - 29.11.2013
15:35 BlockMen      is it intentional that lava is also waving with the new shaders?
15:52 PilzAdam      ShadowNinja, https://github.com/minetest/minetest_game/pull/228 why do you remove default from depends.txt?
16:07 celeron55     i guess because it's not really dependent on default
16:08 celeron55     there's just one crafting recipe
16:08 RealBadAngel  hi
16:08 RealBadAngel  hi celeron55
16:09 PilzAdam      celeron55, a crafting recipe is enough to make it depend on default, I guess
16:11 celeron55     it doesn't matter in reality anyway
16:11 celeron55     but putting it in that commit is random
16:23 RealBadAngel  celeron55, PilzAdam anything against it? https://github.com/RealBadAngel/minetest/commit/d56214cb097ff0297e479501dd21d6d86c573b6d
16:25 PilzAdam      RealBadAngel, whats a reasonable usecase for this?
16:26 RealBadAngel  case is to use one TP selected normal file for all the nodes
16:27 RealBadAngel  i do have one, really nice ready
16:30 PilzAdam      I wonder if it woulb be better to add a fallback texture
16:31 RealBadAngel  http://i.imgur.com/InX6xHT.png
16:31 RealBadAngel  no, that is not a good idea
16:32 PilzAdam      why not?
16:32 RealBadAngel  because if we do start support normalmaps we shall support them in one style
16:32 celeron55     i don't understand why anyone would use anything like that
16:32 RealBadAngel  ie texture pack style
16:33 RealBadAngel  celeron55, thats a matter of taste, for me personal it looks great
16:33 RealBadAngel  everything looks like made of lego bricks
16:35 RealBadAngel  and all just by supporting one file
16:36 RealBadAngel  if texture pack will support override_normal.png it will be used for everything
16:37 RealBadAngel  PilzAdam, about fallback normal, hmmm, have you seen Bumped, yes?
16:38 ShadowNinja   PilzAdam: c55's right. It doesn't depend on default, it just has a craft recipe. Therefore you can copy it to, eg, a game without default and add a craft recipe in a seperate mod.
16:38 RealBadAngel  PilzAdam, i do have almost ready Bumped as a feature. it bumps all 16x default textures
16:39 RealBadAngel  http://i.imgur.com/ntnJJWm.png
16:39 RealBadAngel  this is how it looks like atm
16:41 RealBadAngel  i am working on making shadows smoother
16:46 Exio4         how are you making those?
16:46 Exio4         (bumpmaps)
16:47 ShadowNinja   RealBadAngel: Isn't there a way to mathematically calculate the facedir of a node without using a giant table and loops on it?
16:48 RealBadAngel  ShadowNinja, what exactly do you need?
16:48 RealBadAngel  Exio4, last screenshots are from the engine, theyre not made
16:49 Exio4         wow
16:49 RealBadAngel  well, the one that causes world to look like lego is made
16:50 Exio4         yeah, but i was talking about the last ones :P
16:50 RealBadAngel  one file that serves as a normalmap for all the tiles
16:50 RealBadAngel  its made as 16x16 lego brick tile
16:50 RealBadAngel  to fit the resolution
16:51 RealBadAngel  normalmap is made using CrazyBump
16:51 RealBadAngel  ShadowNinja, look at the facedir table description in mapblock_mesh
16:52 ShadowNinja   The engine uses a table too? :-|
16:52 RealBadAngel  yes it does, it is way faster
16:53 PilzAdam      faster than... ?
16:53 RealBadAngel  since it is used for each and every tile, lookup is always faster than calculations
16:53 RealBadAngel  6 times per node
16:54 RealBadAngel  i made the code that calculated that but dropped it after trying
16:54 RealBadAngel  it was horribly slow
16:56 RealBadAngel  PilzAdam, now 6d is faster than old facedir if you would like to know
16:56 RealBadAngel  as for my box it was like 5fps gain
16:57 ShadowNinja   RealBadAngel: Perhaps, but the screwdriver doesn't run often, and Lua tables are slower.
16:58 RealBadAngel  screwdriver uses tables only in one mode
16:58 RealBadAngel  that for VanessaE insisted
16:59 PilzAdam      also it seems like some mod sets facedir > 23, you should double-check the screwdriver
16:59 RealBadAngel  i will, but i doubt thats screwdriver fault
17:00 RealBadAngel  there was a check in the engine that checked it
17:00 RealBadAngel  and then terminated
17:00 PilzAdam      there was an assert() that crashed clients! I already replaced it so it at least doesnt crash...
17:00 RealBadAngel  it was there before
17:01 RealBadAngel  it was workin the very same way for old facedir values
17:01 RealBadAngel  i havent put in there
17:02 RealBadAngel  if it was crashin it was the mods that set up illegal values
17:02 RealBadAngel  screwdriver alone cannot do that
17:03 RealBadAngel  :P thx to tables :P
17:03 PilzAdam      mods shouldnt be able to crash clients
17:04 RealBadAngel  huh, there are thousand of way to do that
17:05 RealBadAngel  *ways
17:05 RealBadAngel  insert random keys poking into script for example
17:06 RealBadAngel  youre still sayin that mods shouldnt crash server?
17:06 PilzAdam      you should consider to re-read what I said
17:06 RealBadAngel  or perhaps modders should write better mods?
17:07 RealBadAngel  ofc i did
17:08 RealBadAngel  if any code changes the values range, it should be wrote in a way to know what values are possible
17:09 RealBadAngel  and that applies to any code, not only mods
17:09 RealBadAngel  thats why for example light was fucked up in shaders for ages
17:13 RealBadAngel  and you cant make the code check it all, because then it will be slowed down as hell
17:16 RealBadAngel  and there are many asserts in the code
17:16 RealBadAngel  dont blame for one that i havent put there, but just was close enough to it
17:29 RealBadAngel  john_minetest, i keep adding new features
17:30 RealBadAngel  http://i.imgur.com/InX6xHT.png
17:30 RealBadAngel  this is example usage for one normalmap for everything
17:30 PilzAdam      how about fixing old ones first?
17:31 RealBadAngel  PilzAdam, what i should fix?
17:31 PilzAdam      you said water shader needs a total rewrite
17:31 RealBadAngel  yes
17:31 PilzAdam      and needs to be configureable in nodedef
17:31 RealBadAngel  and i will rewrite water
17:31 RealBadAngel  but thats not the SHADER case
17:32 RealBadAngel  water have to be changed in the engine
17:32 RealBadAngel  so it wont have 6 surfaces any longer
17:33 RealBadAngel  john_minetest, 2020 ;)
17:33 PilzAdam      as long as you dont break the water of Nodetopia
17:34 PilzAdam      (which you somewhat did with the water shaders)
17:35 RealBadAngel  PilzAdam, such change will be cosmetical for gameplay
17:35 RealBadAngel  but critical for shaders
17:35 RealBadAngel  i just want to know what tile i should shade, thats all
17:37 RealBadAngel  so, what about https://github.com/RealBadAngel/minetest/commit/d56214cb097ff0297e479501dd21d6d86c573b6d
17:37 RealBadAngel  can i merge it?
17:37 PilzAdam      Im still not convinced that its useful
17:38 RealBadAngel  you liked Bumped already
17:38 RealBadAngel  so why not Legoed for example? :)
17:39 RealBadAngel  not to mention that such effect is very lightweight
17:39 RealBadAngel  its just a texture pack that consist only of one file
17:40 RealBadAngel  i like lego, some1 can like something else
17:41 RealBadAngel  its a feature, not hardcoded stuff
17:42 RealBadAngel  i can think of many possible shapes that could be used
17:43 RealBadAngel  for example one similar to sphax texture
17:44 RealBadAngel  Bumped/and this feature can make game lookin interesting for folks that have problems with many textures
17:47 RealBadAngel  having an option to change the look by providing just one file isnt good at all?
17:54 RealBadAngel  it is a sample usage of the feature
17:55 RealBadAngel  if you dont know apply the patch and try for christ sake
17:56 RealBadAngel  all of you are talking about what you have not tried
17:56 RealBadAngel  theres one, of your choice
17:57 RealBadAngel  my choice was lego like
17:57 RealBadAngel  if you want it i can upload it
17:58 RealBadAngel  gimme a minute
18:01 RealBadAngel  http://www.sendspace.pl/file/1e7afafb6c681bb232d7606
18:01 RealBadAngel  put it into /textures/whatever/
18:01 RealBadAngel  i mean in a folder as for texture pack
18:01 RealBadAngel  with this single file in it
18:02 RealBadAngel  and then you will be able to select it via texture pack selection menu
18:03 RealBadAngel  ofc, shaders have to be enabled to see the change
18:14 RealBadAngel  PilzAdam, can you try it also?
18:18 thexyz        celeron55: https://support.google.com/legal/contact/lr_dmca?product=googleplay
18:23 rubenwardy    Are you taking down build craft?
18:23 rubenwardy    Can't believe it is getting 3 stars
18:23 thexyz        why not?
18:23 thexyz        it's running pretty OK
18:23 thexyz        (controls suck a bit though)
18:23 thexyz        it has more features than other Minecraft clones AFAIK
18:24 thexyz        multiplayer + modding API + great community with tons of mods (not like people use them much, but they play on servers with mods)
18:24 thexyz        at the same time it has weird UI
18:24 thexyz        and, well, it lags
18:25 thexyz        what do you think it should be rated instead?
18:26 celeron55     hmmh, well yeah, we kind of have to do something now
18:26 celeron55     even if i thought this was acceptable, many of minetest's past contributors wouldn't
18:27 rubenwardy    What happened to the thread?
18:27 celeron55     "Full legal name of the copyright holder you represent"
18:28 celeron55     rubenwardy: it was trashed because of obvious reasons
18:28 RealBadAngel  john_minetest, how? whats the problem?
18:29 celeron55     "Full legal name of the copyright holder you represent" <- i wonder what this is; i'm not too eager to put this under my name
18:29 celeron55     legally i should list every contributor's name in that field
18:30 Warr1024      only the ones you're sure you can speak for...
18:30 celeron55     or really, every contributor should fill this form for themselves for their part of the source code
18:31 thexyz        celeron55: yes, that's why I'm asking you; you have the biggest number of contributions I think, moreover, every source file has a comment with your name
18:32 Warr1024      where is this app anyway?
18:32 RealBadAngel  john_minetest, thats propably because of height, i should forbid the drawtype to use parallax
18:36 RealBadAngel  changes the way you see the texture
18:36 RealBadAngel  its a geometry stuff
18:37 RealBadAngel  hold on
18:37 RealBadAngel  http://en.wikipedia.org/wiki/Parallax_mapping
18:38 sapier        guy's you don't know what you're missing eclipse lua debugger is phantastic ;-)
18:38 sapier        -ph+f
18:38 sapier        +,
18:40 sapier        true eclipse is slow but having full debug and breakpoint support within lua is a feature I was missing eversince
18:40 sapier        and I just found out it's allways been there
18:40 PilzAdam      john_minetest, eh, what?
18:41 ShadowNinja   -> #minetest
18:42 sapier        as far as I know builtin contains lua files to so IMHO usefull debugging environment for it is core relevant information too
18:43 thexyz        so can it debug embedded Lua?
18:43 sapier        exactly
18:45 sapier        wait I'm uploading a screenshot
18:47 sapier        it doesn't catch assertions for now but you can just replace assert by your own fct and place a bp in there so this is a minor limitation
18:48 sapier        http://imgur.com/F57R6x4
18:49 sapier        I guess it'd be even possible to run c++ and lua debugger at same time
18:50 sapier        well they wouldn't know about each other so this might not be that usefull
18:52 thexyz        pretty cool
18:53 sapier        yes and all you have to do is adding two lines to builtin
18:54 sapier        it's exactly opposite to c++ debugging, in lua the application connects to debugger (at least in embedded case)
18:54 sapier        local initconnection = require("debugger")
18:54 sapier        initconnection("localhost", nil, nil)
18:55 sapier        debugger is a file that's provided with koneki(eclipse lua plugin) you only have to copy it to your lua search path
18:55 sapier        that's all ... quite simple
18:55 thexyz        I think you could document it somewhere (wiki) if you have time
18:56 sapier        It's as simple It may not even need documentation, koneki already contains this documentation you just have to know it's really working
18:57 sapier        but of course you don't get debugging for nothing it's significant slower
18:58 sapier        subjective feeling something round about fator 5-10
18:58 sapier        +c
19:00 thexyz        yes, no need to document, okay
19:01 sapier        I'll write a short paragraph in modding section so others realize it's possible ... modders may not read this log ... and I guess most core devs aren't willing to switch their favourite editor just for lua debugging .... especialy as eclipse is a real pain in the ass for beginners
19:28 sapier        http://dev.minetest.net/Intro thexyz the "IDE" subsection
19:30 thexyz        thanks
19:32 ShadowNinja   How do #221, #228, and #229 look? (Related)
19:36 sapier        I hope on_place isn't called if a node is e.g. added by a entity
19:37 sapier        if "user" is used in that function it's always checked first to be not nil
19:47 sapier        shadow I added some comments to those three it's only minor changes I believe to be done
19:48 sapier        -done + necessary
19:48 celeron55     on_place is supposed to be called for any entity that places a node
19:49 sapier        then my comment to this commit is even more correct
19:49 celeron55     if you make a monster that builds stuff, it should call it
19:49 celeron55     (there are api functions for placing a node for that, instead of setting a node)
19:50 sapier        I don't have monsters that do place nodes for now but good to know
19:51 ShadowNinja   sapier: I don't think nil placers/users are valid there.
19:53 sapier        read seven lines up
19:54 sapier        btw others within that function check for it so if you're really 100% sure it doesn't happen you should remove the check there too
19:57 celeron55     could people please ALWAYS explicitly mention the full URL when discussing about something?
19:57 celeron55     eg. now i have no idea what people are talking about
19:57 celeron55     and i'm not going to manyally check three different pull requests based on numbers to find it
19:57 celeron55     manually*
20:09 rubenwardy    sapier, you should mention what IDE means
20:09 rubenwardy    gtg
20:09 sapier        was rubenwardy serious?
20:10 Calinou       probably not
20:11 sapier        I tend to assume someone who doesn't know what an ide is doesn't actually need it
20:17 Warr1024      IDE = precursor to SATA
20:17 sapier        no thats PATA
20:18 sapier        :-)
20:18 Warr1024      PATA basically = IDE renamed
20:18 thexyz        sapier: so do you use eclipse to develop minetest?
20:18 sapier        yes actually name PATA was defined along with SATA so it isn't a real predecessor
20:19 sapier        yes thexyz for core as well as mods ... if you wanna try avoid juno it was crap
20:19 thexyz        lol
20:19 thexyz        I'm still waiting & hoping for http://www.jetbrains.com/objc/features/cpp.html
20:20 sapier        yea juno was first 4.x release ... obviously they forgot to do at least a single performance test
20:20 sapier        any lua support in there?
20:20 thexyz        I doubt it, tbh; but I hope they'll make a good C++ ide for linux
20:21 thexyz        knowing their other products the probability of this is quite high
20:21 sapier        objective c .... isn't that apple eccosystem own language?
20:21 thexyz        yes, originall this is an objective c ide but they're working on a new c++ one
20:21 Calinou       did you try CodeBlocks and Geany?
20:22 sapier        why can't those big companys take a standard language and reinvent their own everytime just to make themselfs different from each other
20:22 Calinou       sapier, that's what they are doing
20:22 sapier        imho it just tells they can't compete with quality so they need some crap difference
20:23 thexyz        Calinou: we want some quality VisualStudio alternative
20:23 thexyz        you suggest geany ;_;
20:23 Calinou       and?
20:23 thexyz        and you still don't get it?
20:23 sapier        yes ... eclipse seemed to become that one but after ibm left development seems to have slowed down dramaticly
20:24 sapier        codeblocks was once my prefered ide but they stopped development
20:24 Sokomine      back in the turbo pascal days i did have an ide. now...it's mostly editor with syntax highlighting and command line tools. eclipse...seems a bit...big...and quite diffrent in handling
20:25 sapier        yes biggest problem with eclipse is it's number of features
20:25 sapier        but same thing is eclipses best feature
20:26 sapier        once you managed to configure eclipse to your needs it's great
20:27 Calinou       just keep a terminal at hand to do "make"
20:27 Calinou       or use a crontab to run it every minute
20:27 * Calinou     runs
20:27 sapier        I start to understand why debug build uses -O1 ;-P
20:27 Calinou       it removes bugs
20:28 thexyz        or perhaps i should stop being lazy and google for vim plugins and build my very own ide
20:28 Sokomine      thexyz: sounds good. once you're done, please tell me :-)
20:28 thexyz        lol
20:28 Sokomine      (perhaps someone has done so already. who knows)
20:28 sapier        I guess compared to that task even configuring eclipse isn't "hard"
20:30 thexyz        nah, it's not that hard
20:31 sapier        imho there's only one reason for using vim ... you're a old coder waiting for retirement and don't want to learn that modern stuff ;-) *just kidding*
20:31 celeron55     vim plugins tend to be a bit half-assed so it's kind of nontrivial to make proper use of them
20:31 celeron55     but... i guess that's what plugins always are
20:32 celeron55     1) replace your init system with emacs, 2) ???, 3) PROFIT
20:32 Sokomine      john_minetest: :-)
20:34 Sokomine      still, it's odd that ides where more important in the past. but with mt taking several minutes to compile anyway, that part of the ide may not be particulary relevant
20:38 sapier        imho most important part of an ide is it's debugging support ... and sorry vim and emacs aren't quite good there
20:40 RealBadAngel  i do have a question
20:40 RealBadAngel  who wants to work with me together on the shaders?
20:40 Calinou       Sokomine, if you only change a few files it may be only a few seconds
20:41 sapier        sorry rba I don't have any idea how to write shaders
20:41 sapier        but your latest work is amazing rba
20:41 RealBadAngel  celeron55, PilzAdam ?
20:41 RealBadAngel  kahrl? hmmm?
20:42 Warr1024      These shaders sound pretty cool, but I hope they remain a nice optional part of MT... I only have 1 machine that can barely support the old basic shaders as-is...
20:43 ShadowNinja   Segfault that I get whenever a mod crashes, I fix it, and I reenter the world: http://pastebin.ubuntu.com/6557905/
20:43 Warr1024      RealBadAngel: btw, do you have any place where you keep the latest screenshots?
20:43 RealBadAngel  if any of you wont rise a hand and join improving shaders  i can see no point in giving right to block the changes im doing to them
20:43 thexyz        sapier: I think there were some plugins to integrate gdb with vim
20:43 RealBadAngel  thats my point
20:44 thexyz        but it seems they all sucked
20:44 sapier        :-) yet a console based editor can't ever compete to a gui using full features of windowing concept ;-)
20:44 Warr1024      as for me, as long as there's still a "turn shaders off" checkbox and the game runs correctly without them (and essential gameplay elements aren't relying on shaders) then I have no complaints about anything you do with them :-)
20:45 sapier        except for very specific tasks maybe
20:45 RealBadAngel  all is optional here
20:45 RealBadAngel  without it we would have killed ourselves long time ago ;)
20:46 thexyz        sapier: vim/emacs army is coming for you right now
20:46 sapier        rba haven't your latest changes be merged? if you need someone to agree to them I'll check them
20:46 RealBadAngel  sapier, ive pulled a tiny bit of code today
20:47 RealBadAngel  last changes was about changes to my original code (which made shaders broken for me lol)
20:47 sapier        if you need someone to check just ask, I can't tell anything about the shaders themselfs but I'd not consider bugs in there to be critical, you're the only one to fix them either ;-)
20:48 Warr1024      Are there going to be different types/levels of shaders depending on capabilities, framerate, or user preferences?
20:48 RealBadAngel  i want you to download the file http://www.sendspace.pl/file/1e7afafb6c681bb232d7606
20:49 RealBadAngel  put it into textures folder, select it as you would select texture pack
20:49 RealBadAngel  https://github.com/RealBadAngel/minetest/commit/d56214cb097ff0297e479501dd21d6d86c573b6d
20:49 RealBadAngel  and apply this patch
20:49 sapier        :-) uagh ... that's nothing meant to be viewd directly :-)
20:50 RealBadAngel  of fucking course
20:50 RealBadAngel  but no one wants to check it ingame
20:51 RealBadAngel  especially celeron55
20:51 RealBadAngel  celeron55, have you ever turned shaders on? once?
20:52 RealBadAngel  it is ridiculous we have to fight to get our stuff merged
20:53 RealBadAngel  we are doing it for free, and feeling like  doing hard job
20:53 sapier        rba how shall i name that file?
20:53 RealBadAngel  override_normal.png
20:54 sapier        guess I should enable shaders too
20:55 RealBadAngel  lol, yes
20:56 RealBadAngel  bumpmapping shall be enabled
20:57 sapier        what am I supposed to see?
20:59 sapier        lego mode
21:00 sapier        can you stop it from beeing aplied to all faces?
21:00 Warr1024      yeah, it should only be the tops, right? :-)
21:01 RealBadAngel  sapier, thats the point, with one texture make it look one style
21:02 RealBadAngel  no matter how many textures are out there
21:02 sapier        not sure if that's really desired :-/
21:03 RealBadAngel  thats a texture pack option
21:03 RealBadAngel  why do you all are  blind?
21:03 RealBadAngel  i do want the feature
21:04 RealBadAngel  because all the default textures are ugly
21:04 sapier1       you could use a little bit more diplomatic words rba ;-)
21:05 RealBadAngel  yeah? after months of askin "please do accept what i was working for so long?"
21:06 RealBadAngel  i am doing good code.
21:06 RealBadAngel  and all the time im forced to beg for it to join
21:06 RealBadAngel  this is fucking sick
21:07 RealBadAngel  g'nite for today
21:07 RealBadAngel  tired of begging for today (UDP timed out)
21:08 sapier1       guess we need to work on our review process :-/
21:09 sapier        what about some rule if you wanna add a feature first review to pulls :-)
21:09 Warr1024      does the problem involve people's changes being rejected, or just being ignored for a long time?
21:09 sapier        -to + two
21:09 Warr1024      sounds good.
21:10 Warr1024      or that status as a contributor in credits includes certain responsibiliites to review pulls...
21:10 sapier        main problem is them beeing ignored ... if theres some critizism you can fix it but if noone tells anythin you sit around and wait
21:11 ShadowNinja   sapier: #228 is fixed.
21:12 Warr1024      Sadly, the method I've had the most luck with is just being annoying and catching one of the core devs on IRC...
21:12 Warr1024      ...though I guess RBA's in a unique position, being on the cutting edge with shader tech.
21:12 sapier        what did celeron tell about numbers shadow ;-)
21:12 Warr1024      I'd love to review the stuff (even if it's just a crashes vs. works for me) but I don't even have hardware capable of it.
21:13 PilzAdam      its not like RBA gets ignored, I talked to him in PM today (a lot)
21:14 sapier        it's not only rba having that feeling it's getting more and more annoying to fulfull the 2 agreements rule as ppl just don't agree ... they don't deny either
21:14 sapier        some vague saying "seems to be fine" can mean anything
21:15 sapier        -a+e
21:15 sapier        I'm writing to fast, need to improve my spelling
21:15 PilzAdam      oO sapier using , ?
21:16 sapier        I try :-)
21:18 Warr1024      once something gets one OK (and no not-OK's) it seems like it should be a priority to have someone else review it soon.
21:19 Calinou       we could have more reviewers?
21:19 ShadowNinja   sapier: Hmmm, what did he?
21:20 Warr1024      does review need to be done by a core dev who's familiar with MT architecture, or is it enough for someone who knows how to pull the correct source and/or apply patches automatically to indicate that it works as described?
21:20 Warr1024      maybe a call to action from the moderately-techinical portions of the community, along with a how-to guide, could open up a pool of potential reviewers...
21:20 sapier        ShadowNinja Celeron requested full urls instead of numbers only ... that was primary adressed to you ;-P
21:21 PilzAdam      Warr1024, testing is just one side, looking at and judging code is the other one
21:21 sapier        Warr1024 any review is welcome, in best case write a comment to that pull request and describe your findings
21:22 sapier        the more information is already there, the less needs to be done by core devs, the more efficient merge process can be done
21:23 ShadowNinja   Ah, O.K. https://github.com/minetest/minetest_game/pull/228
21:23 Zeitgeist_    can some one look into https://github.com/minetest/minetest/pull/1042 its related to this bugreport https://bugs.gentoo.org/show_bug.cgi?id=493650
21:23 sapier        as hmmmm agreed to this one yesterday I'm gonna push it in 15 minutes if noone complains https://github.com/minetest/minetest/pull/1030
21:23 Warr1024      I'm just thinking that if pull reqs undergo a more rigorous pre-screening, then core devs may feel that the remaining pull reqs are more worth their time.
21:26 sapier        guess I should test it prior final agreement :-)
21:28 sapier        ok shadow works for me merge it
21:29 ShadowNinja   Alright, pushing...
21:29 Warr1024      shadow?
21:29 Warr1024      oh
21:30 Warr1024      man, I thought you were talking about adding shadows behind the text in the chat interface :-)
21:30 sapier        no, :-) shadowninja just took a way to long nickname
21:30 Warr1024      no tab autocomplete?
21:31 sapier        tab autocomplete? ... never tried that one in this window...
21:31 sapier        Warr1024: ShadowNinja ... wow it's working :-)
21:32 Warr1024      I was going to try to subclass the irrlicht static text control to create 9 copies and redraw the text 8 times in black and once in white to add a border contrast.  It's not a really optimal solution, but it would serve at least my need, and I suspect that the FPS hit would still be worth being able to read chat text...
21:33 Warr1024      more ideal would be if I could render the text somehow to a texture or buffer, then either apply a transform of some kind, or just redraw the pre-rendered buffer.
21:33 Warr1024      but that's way over my head at this point.
21:33 sapier        can you fix that ugly chat overlaying formspec menus issue too?
21:34 Warr1024      oddly enough I don't think I've ever witnessed that in person before.
21:34 sapier        install mobf do some talking open menu try to switch tabs
21:35 Warr1024      it might also be nice to add some kind of font scaling to make chat more easily readable at high res.
21:35 sapier        ... wont work until chat text is timed out
21:35 ShadowNinja   Warr1024: OR: https://github.com/minetest/minetest/pull/608
21:35 Warr1024      ah, cool, it's already been done :-)
21:35 Warr1024      the best kind of code is the kind I don't have to write :-D
21:35 ShadowNinja   Warr1024: See the font_size setting in Freetype builds.
21:36 Warr1024      I was thinking it would be cool to try to add an outline around just the text instead of a semi-transparent block behind it, though the block shadow IS much cheaper.
21:36 Warr1024      like the way a good captioner works.
21:37 Warr1024      ~.
21:37 ShadowNinja   Yes, but you're not going to be able to get the text lined up just right that way, and it will never be accepted due to how ugly it is.
21:37 Warr1024      what do you mean?
21:38 Warr1024      for(dx = -1; dx <= 1; dx++) for(dy = -1; dy <=1; dy++) if((dx != 0) || (dy != 0)) rendertext(x + dx, y + dy, black, string); rendertext(x, y, white, string);
21:39 Warr1024      it's expensive (if font rendering is slow) but works well, assuming that the font renders the same each time, which SHOULD be true as all of the coordinates are integral.
21:40 Warr1024      unless, I guess, it's a proportional font, it's wrapped, and the bottom and right edges of the bounding box aren't also adjsuted accordingly.
21:41 Warr1024      if font rendering is slow, the text can be rendered to a buffer, followed by either a matrix convolution and compositing, or a color matrix and a handful of composites.
21:41 ShadowNinja   Well, anyone to agree on that shaded text? And should it adjust? (Polls say yes)
21:41 ShadowNinja   sapier: What about the other pull? https://github.com/minetest/minetest_game/pull/229
21:42 Warr1024      adjust?
21:42 Warr1024      i.e. make a user setting for transparency or something?
21:43 kaeza         Warr1024, it'd prolly be less expensive only drawing (x=0,y=-1), (x=0,y=1), (x=-1,y=0), (x=1,y=0), and still gets a nice border
21:43 kaeza         can you check?
21:44 ShadowNinja   Warr1024: Adjust to the length of the longest chat line.
21:45 Warr1024      kaeza: yeah, though there may be a bias of horizontal/vertical vs. diagonal lines, depending also on how the text is antialiased.
21:46 Warr1024      ShadowNinja: personally I would probably prefer non-adjusting, but either way it's a significant improvement.
21:46 sapier        I can't read that from docs ShadowNinja
21:47 ShadowNinja   sapier: https://raw.github.com/minetest/minetest/master/doc/lua_api.txt  --> Ctrl-F --> on_receive_fields.  No mention of sender being optional.
21:49 celeron55     lol @ shaded chat text
21:49 sapier        there's nothing written about pos either but pos is absolutely optional
21:49 celeron55     minetest had that originally but it was removed for some reason
21:49 celeron55     i don't recall by whom or why
21:50 Warr1024      would you be opposed to adding it back in?
21:51 Warr1024      I don't see any dissenting comments on the pull req...
21:51 ShadowNinja   sapier: So are we to assume that the node argument to set_node is optional? Unles specified as optional all arguments are mandatory.  If a optional argument isn't written as such then it is a documentation error. And it's silly to call on_receive_fields with a nil player.
21:52 celeron55     Warr1024: it's fine to me with PilzAdam's addition
21:54 Warr1024      awesome.  Sounds like between you and PilzAdam, ShadowNinja has the mandatory 2 review ok's to push that commit in...?
21:56 sapier        on_receive_fields is called uppon ANY formspec event
21:56 pitriss       And will be possible to disable that shading?
21:56 sapier        e.g. mod generated non node bound formspecs too
21:56 sapier        wait
22:05 ShadowNinja   pitriss: No, why would you want that?
22:05 * ShadowNinja recompiles from scratch after rebasing that.
22:06 pitriss       ShadowNinja: because it is ugly.. I don't like that.. I want to see that chat "moves" but when I want to read I can press F10
22:06 Warr1024      is there currently any way to change the persistence time of chat messages?
22:07 ShadowNinja   pitriss: But it's unreadable as it is, and it isn't ugly.
22:08 ShadowNinja   Warr1024: There's a setting hardcoded...
22:08 Warr1024      between adding a knob to turn on/off or change the shading vs. a knob to control how long chat text lasts on screen, I think we'd get more milage out of the latter...
22:09 Warr1024      of course, that'd be a subject for a separate pullreq.
22:09 pitriss       ShadowNinja: But I don't need to be bothered by gray square on screen.. For me is not wanted that.. For me is enough to see that chat moves.,. (so hide chat is not an option for me), when i want to read what happened I don't have problem with opening console.
22:10 Warr1024      realistically we're not going to be able to make minetest appear pretty to everyone.
22:11 Warr1024      the loss of aesthetics in exchange for making the scrolling chat interface more useful than as a "you've got chat messages, F10 to view" indicator is definitely worth it for me.
22:11 pitriss       yes I know, but this is really worst solution for me..
22:11 pitriss       omg I'm using bigger font so it is barelly readable as is.. I don't need that gray s**t under chat
22:12 Warr1024      well, we had ideas earlier about a more attractive looking outline shadow.
22:12 pitriss       instead this for me is better solution to have outlined characters
22:13 Warr1024      please do then, and send us a pull req.
22:14 Warr1024      I'd love to have those too.
22:14 ShadowNinja   Rebased and ready to push.
22:14 Warr1024      btw, does font_size control the fonts across the whole UI?
22:14 ShadowNinja   Warr1024: Yes.
22:14 pitriss       I'm not C coder,, but this solution is a crap.. I think is more important to see what you building than some spammers
22:14 Warr1024      is there any way to change font size just for specific UI elements like chat?
22:15 sapier        I don't think so
22:15 ShadowNinja   pitriss: The chat may be useless to you, but most people use it, so it should be usable.
22:15 pitriss       gray square will make worse visibility in night
22:16 sapier        what are you going to push?
22:16 Warr1024      pitriss: sounds like the problem has less to do with the UI and more to do with the servers you hang out on :-)
22:16 sapier        shadow?
22:16 pitriss       so it should be possible to disable that
22:17 pitriss       I don't think it is unusable as is.. In console it is pretty readable.. and when console is closed, then players probably build more than read chat
22:18 Warr1024      I like to be able to read chat while I'm building without having to stop and hit F10 every few seconds to see if I lost something important in the clouds.
22:18 Warr1024      ideally, it would be nice to be able to turn this on/off.
22:18 Warr1024      except what would the code maintenance costs be, and would it involve kicking the pull req back to the start of the review process again?
22:18 pitriss       heh so make that "shadow" toggled by keypress
22:19 Warr1024      I believe it's already togglable with F2 or something :-)
22:19 Ritchie       hello, if you cannot read chat due to bad background, change angle of your view (yaw for example)
22:19 Ritchie       or look into sky
22:19 pitriss       Warr1024: no F2 toggle whole chat
22:20 Warr1024      pitriss: yeah, hence the :-)
22:20 Warr1024      sorry, was being facetious.
22:20 ShadowNinja   Warr1024: Because of the way Irrlicht does fonts you need to make your own font for each text size that you want to draw.
22:20 ShadowNinja   sapier: Shaded chat text.  I won't push it just yet though.  Do you agree with it?
22:20 ShadowNinja   pitriss: The chat is white.
22:20 pitriss       Warr1024: but I want to "see" chat without being disturbed by gray square on screen
22:21 sapier        do you have a number?
22:21 Warr1024      anyway, I can think of like a hundred different UI tweaks that I'd love to see added in, myself, but I'd rather the core devs be working on bigger things...
22:21 pitriss       ShadowNinja: Yes, and I'm not looking into sky most of time.. :) So It is not problem for me.. Grass is green, dirt is brown..
22:22 pitriss       there should be not so huge problem make that possible to toggle it by some key..
22:33 pitriss       I think best for both sides is to bind toggle of this feature to (for example) F8 (and maybe configuration directive of default state (to be not needed to switch off that everytime i start client)).. then it will be great and not disturbing for ppl who doesn't like/need this feature..
22:35 Warr1024      I'm not so sure that it wouldn't be a signficant change to add a toggle to this thing.
22:35 Warr1024      though I guess in the end ShadowNinja would have to be the one to weigh in on that.
22:36 pitriss       I know lot ppl who like it as is.. so, toggle key is really good idea..
22:38 kaeza         http://xkcd.com/1172/
22:39 Warr1024      nice
22:39 pitriss       and BTW: I tested it right now.. with my setting (I have dejavu sans bold as default font in game) it is readable as is.. even with looking on default light gray clouds:)
22:40 Warr1024      I'm not opposed to making this togglable in some way, but if it's not worth anyone's time to add whatever switches and knobs are necessary for that, I think the benefits outweigh the costs for most of the users.
22:42 pitriss       Erm but this is game not resources heavy irc client IMHO.. So building is priority, chat should not disturb gameplay when anyone preffer building..:)
22:43 pitriss       So for me, toggle button is almost must have on this feature..
22:43 Warr1024      so you'd stop playing minetest completely if a gray shadow were added below you chat text?
22:45 pitriss       No, but why I should be disturbed and with eyesore just because of something is highlighted even if i don't need it?
22:46 pitriss       And yes.. if you highlight chat, then it will attract your eyes more => eyes will be tired quicker..
22:47 pitriss       in case you want to build.. Now it is accurate for me.. I can notice that something was changed.. but I'm not forced to move eyes there to look what happened there