Minetest logo

IRC log for #minetest-dev, 2018-01-31

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

All times shown according to UTC.

Time Nick Message
00:00 red-001 as in technic updating it's own nodes?
00:00 red-001 or technic changing nodes around it?
00:00 VanessaE yeah
00:00 VanessaE pipeworks too
00:00 VanessaE think nodebreakers, deployers.
00:01 red-001 wouldn't not recording that break rollback?
00:01 VanessaE it could.
00:01 ThomasMonroe red-001, thanks for the tips on #6965
00:01 ShadowBot https://github.com/minetest/minetest/issues/6965 -- [WIP] Basic pause menu settings by ThomasMonroe314
00:01 VanessaE but it would be worth it.
00:01 red-001 I suppose it could be a setting in technice
00:01 VanessaE rollback should only record the actions of actual players.
00:01 red-001 technic*
00:02 red-001 does rollback even record tnt?
00:04 VanessaE the most common use case would be a row of deployers, all pointing at empty spaces, placing something like wheat seeds, with a row of nodebreakers on the opposite side, waiting for a signal from a mesecons node detector, breaking all of the then-grown wheat, shuffling it off to a chest for staging, then off to furnaces, autocrafters, et al.
00:04 VanessaE so really 99% of the work is done by pipeworks, rather than technic.
00:04 VanessaE idk if it records tnt.
00:08 sofar tnt uses a vmanip
00:08 sofar at least mtg does
00:09 red-001 and that doesn't get recorded?
00:10 sofar I have nfc
00:11 sofar but, if anything bypasses rollback, I would think vmanips do
00:28 Fixer VanessaE: maybe database needs other layout?
00:29 EvergreenTree joined #minetest-dev
00:29 VanessaE I wouldn't know.
00:30 Fixer VanessaE: and it lags your server?
00:32 VanessaE only if I need to do a rollback; I've long since disabled it.  the rollback logs were so huge they were filling the disk from time to time (there's only about 117GB free right now)
00:39 twoelk joined #minetest-dev
00:53 paramat joined #minetest-dev
01:09 paramat luavoxelmanip actions are not recorded by rollback, that would slow it down too much
01:19 EvergreenTree joined #minetest-dev
01:27 paramat game#2016 now tested, ABM is 3 times less intensive but flora spread is faster
01:27 ShadowBot https://github.com/minetest​/minetest_game/issues/2016 -- Flower spread ABM: Optimise by paramat
01:28 Shara paramat: Can't decide what I think about quicker overall flower spreading
01:32 ThomasMonroe joined #minetest-dev
01:38 paramat it's 33% faster in test, but can be tuned to be equal speed
01:38 paramat if i slow it down a little more to be equivalent that makes the ABM even less intensive so happy to do so
01:39 Shara I suppose the speed difference will only be really noticeable in large open areas
01:39 paramat yes, when farming flowers in an empty area
01:40 Shara How big an area were you testing to get 33%?
01:40 paramat the intention was to make flower farming a little faster due to its slowness
01:41 Shara Due to how you are adapting this, on one level I'm worried it could spread too quickly through large areas, but then on the other hand it shouldn't end up slower in whatever the typical size area players use is
01:41 paramat the test was with 1 stationary player, so the area was the ABM active block range of 3 mapblocks
01:41 paramat radius 3 mapblocks
01:42 Shara Might be okay then.
01:42 paramat good point. anyway i can equalise speed if desired
01:42 Shara I do like the general idea, just a bit cautious.
01:44 paramat good
01:45 paramat i'll merge some game PRs tomorrow as there are now 2+
01:45 Shara Glad to see obsidian panes PR ready :)
01:47 pgimeno_ on the deletion of generated blocks discussed earlier, I have a concern, which is that many mods (including nyancat before it was removed) use random numbers without a seed based on location, math.random() in this case. Whatever is generated by these means, must be immediately saved because it can't be reproduced if not.
01:47 ThomasMonroe joined #minetest-dev
01:50 twoelk left #minetest-dev
01:52 turtleman joined #minetest-dev
02:01 paramat i guess in that case the chunk would be considered altered so would always be saved in the database
02:42 Tmanyo joined #minetest-dev
05:14 Darcidride joined #minetest-dev
06:47 torgdor joined #minetest-dev
07:38 nerzhul joined #minetest-dev
07:38 nerzhul !help
07:38 ShadowBot nerzhul: help [<plugin>] [<command>]
07:38 Karazhan Available commands: b64decode, b64encode, builder, crc32, help, md5, reverse, sha1, sha256, sha512, strlen, weather
07:42 Karazhan joined #minetest-dev
07:42 nerzhul !help git
07:42 ShadowBot nerzhul: Error: There is no command "git".
07:42 Karazhan Available commands: b64decode, b64encode, builder, crc32, help, md5, reverse, sha1, sha256, sha512, strlen, weather
07:44 Karazhan joined #minetest-dev
07:44 nerzhul !help me
07:44 ShadowBot nerzhul: Error: There is no command "me".
07:50 sofar problems?
07:52 Darcidride_ joined #minetest-dev
08:05 nerzhul sofar, no
08:39 Darcidride joined #minetest-dev
09:44 Fritigern joined #minetest-dev
10:29 Beton joined #minetest-dev
10:48 YuGiOhJCJ joined #minetest-dev
12:44 Fixer joined #minetest-dev
13:32 ThomasMonroe joined #minetest-dev
14:17 paramat joined #minetest-dev
15:14 paramat concerning server attacks, this issue is now important #6984 owners need to be able to reduce mapgen limit and have already generated world outside function normally
15:14 ShadowBot https://github.com/minetest/minetest/issues/6984 -- Allow entities to be active outside the set 'mapgen limit'
15:15 paramat the issue is, how to stop entities falling into ignore, i feel they should collide with ignore
15:15 paramat like players
15:16 paramat the code that removes entities outside 'mapgen limit' needs to be removed somehow
15:21 antims joined #minetest-dev
15:29 ThomasMonroe joined #minetest-dev
15:35 paramat anyway, i'll make a PR now that will hopefully allow owners to use 'mapgen limit' as a defence, it can then be backported to 0.4.17
15:37 ThomasMonroe joined #minetest-dev
15:40 pgimeno_ in the context of #6898, is forward compatibility a requisite to backport bugfixes?
15:40 ShadowBot https://github.com/minetest/minetest/issues/6898 -- Allow distinguishing mods by modpack by pgimeno
15:55 pgimeno_ I'm talking about this: https://github.com/minetest/minetest/​pull/6898#pullrequestreview-89875626
15:56 Fixer joined #minetest-dev
15:58 ThomasMonroe joined #minetest-dev
16:32 paramat erm
16:35 paramat the comment means that the bugfix breaks old/new client/server compat, so can only be applied to 0.5 and can't be backported to 0.4.17
16:36 nerzhul joined #minetest-dev
16:36 nerzhul yes
16:36 nerzhul you can't backport it
16:43 pgimeno_ I'm wondering how that happens. The patch affects decisions on which mods are loaded in the server, but not the protocol.
17:30 Darcidride_ joined #minetest-dev
17:47 Krock joined #minetest-dev
17:55 Jordach joined #minetest-dev
18:11 Fixer those bot-clients just motivating you to progress with anticheat, imposing mapgen limits is treating the symptoms
18:27 Gael-de-Sailly joined #minetest-dev
18:33 QwertyCool joined #minetest-dev
18:33 QwertyCool hola
18:34 sofar hi
18:36 QwertyCool hey sofar could you comment on https://github.com/minetest/minetest/issues/6977
18:36 QwertyCool or message me here about it
18:38 sofar just start with making the error message more informational
18:38 sofar e.g. print out what is being placed at what position exactly
18:38 sofar if that info is even available
18:39 QwertyCool problem is, that any unknown block is converted  to "ignore"
18:39 QwertyCool which is why i was suggesting maybe changing THAT functionality to increase the accuracy downstream in the error logging
18:39 sofar you may want to propagate the error down the stack to the caller, and have them print the error
18:39 sofar I wouldn't change the functionality for now
18:40 sofar not without knowing what is going on
18:44 QwertyCool great idea. would that be whever ::setnode is being sent from the minetest-game? or is that still part of engine?
18:47 QwertyCool ERROR[Main]: Map::setNode(): Not allowing to place CONTENT_IGNORE while trying to replace "air" ..
18:47 sofar well setNode() is being called likely from lua
18:47 QwertyCool ehh
18:47 sofar so if the function in which that happens returns an error to the caller
18:48 sofar you may be able to have more access to relevant data up the call stack
18:48 sofar a stack trace at that point will be handy
18:48 QwertyCool you're super right, but my experience with c++ is super limited my Lua is NIL
18:49 ThomasMonroe lol
18:49 QwertyCool what program do you use to stack trace?
18:50 ThomasMonroe src/map.cpp line 211: void Map::setNode(v3s16 p, MapNode & n)
18:51 red-001 run it under gdb putting a breakpoint on that print
18:57 paramat Fixer yes but serverside defence against hacked clients is difficult and will take time to implement
18:57 Fixer agree, it just motivates more ::)
19:04 nerzhul joined #minetest-dev
19:07 QwertyCool sorry if this is a dumb question but is GDB a build/make flag? or
19:08 QwertyCool y'all probably don't want someone this inexperienced messing around witht he code, but i'm trying my best to learn so one day i can contribute more meaningfully... sorry again for all the newb questions.
19:09 rubenwardy gdb is a program
19:09 QwertyCool https://www.gnu.org/software/gdb/
19:09 QwertyCool this one?
19:09 rubenwardy yeah
19:09 rubenwardy usually installed on most linux systems at the same time as c compilers
19:11 QwertyCool perfect ah
19:11 QwertyCool working from windows
19:11 QwertyCool my bad
19:14 sofar we could use more windows developers
19:14 sofar sorry, I said that wrong
19:14 sofar we could use windows developers
19:14 sofar I don't think anyone is doing any serious development on windows atm
19:14 red-001 QwertyCool, what IDE?
19:15 red-001 or are you using prebuild binaries like any sane person?
19:15 QwertyCool sublime for everything non cpp and MSVS for cpp
19:16 QwertyCool prebuilt binaries?
19:16 red-001 MSVC has a built in debugger
19:16 QwertyCool thats what i thought :$
19:17 Krock on the left of the line numbers you can simply click in the bar and a red dot will appear -> breakpoint
19:17 QwertyCool just gotta figure out how to compile mintest entirety in msvs
19:17 red-001 ^
19:17 QwertyCool do i have to re-compile it after that?
19:17 QwertyCool or
19:17 red-001 no
19:17 Krock QwertyCool, run cmake and open the sln file with it
19:17 red-001 Krock, setting up libs is a pain
19:17 Krock red-001, for 32-bit there are pre-built ones
19:18 Krock and addi (forums) had somewhere a script to set up the entire thing with a single script
19:23 QwertyCool what libraris are you reffering to?
19:23 Krock sqlite3/irrlicht/luajit/zlib/libogg/libvorbis and all the nice stuff around
19:28 QwertyCool uou guys are one step ahead of me haha, i now understnad what you mean
19:28 QwertyCool no libogg prebuilt for win 64
19:29 QwertyCool no i just have to use 32bit?
19:31 QwertyCool oh dope
19:31 QwertyCool https://github.com/iosdevzone/IDZPrecompiledOgg
19:33 Krock merging game#2035 and trivial game#2036 in ~5 minutes
19:33 ShadowBot https://github.com/minetest​/minetest_game/issues/2035 -- add obsidian_pane by TumeniNodes
19:33 ShadowBot https://github.com/minetest​/minetest_game/issues/2036 -- Stairs Unbind table groups. by Jat15
19:33 Beton joined #minetest-dev
19:48 Thomas-S Krock, wait with merging game#2035 please
19:48 ShadowBot https://github.com/minetest​/minetest_game/issues/2035 -- add obsidian_pane by TumeniNodes
19:49 Krock Thomas-S, timer stopped working. rescue the titanic. now.
19:49 Thomas-S There seems to be an indentation mistake in /mods/xpanes/init.lua lines 107/132
19:49 Thomas-S Spaces are used there instead of tabs
19:50 Krock aagh.. force pushing quickly
20:15 celeron55 i got an email asking for co-developers for an opengl 4.5 fork of irrlicht
20:16 celeron55 i think this one https://github.com/buildaworldnet/IrrlichtBAW
20:16 Megaf There's the fork we've been asking for!
20:16 Megaf :P
20:17 celeron55 it's from someone called "DevSH", seems to have a facebook page https://www.facebook.com/DeVshGraphicsProgramming/
20:19 celeron55 well i'll have to respond i don't know much about opengl 4.5 but i'm asking around
20:20 rubenwardy I know him
20:20 Krock just for opengl? what we need are keyboard fixes
20:21 rubenwardy He's tried to ask me to help on my uni's facebook group
20:21 rubenwardy he's the graphics programmer for buildaworld.net
20:22 celeron55 maybe it would be worth it to point this out to all of the few people who have come here to ask about helping improve graphics
20:22 celeron55 ...assuming someone can be bothered to find those from logs
20:23 celeron55 Krock: he says SDL2 in the e-mail; that could mean some input fixes also
20:24 celeron55 "We're looking to take the Engine a whole leap forward to expose full OpenGL 4.5 features, bring mutli-threaded asset streaming from assimp, GPU Driven Scene Graph, SDL2 Device, Android and Possibly Vulkan." "Are you interested?" "Regards, DevSH"
20:24 celeron55 that's the email in its entirety
20:26 Krock indeed, use of SDL2 could contain some fixes in the input system. An experimental branch can always be made if someone is keen to do it
20:28 celeron55 if someone is really interested, PM me and i'll give the email to respond to
20:40 jcalve joined #minetest-dev
21:08 Darcidride_ joined #minetest-dev
21:12 ThomasMonroe joined #minetest-dev
21:19 ThomasMonroe joined #minetest-dev
21:31 Lunatrius joined #minetest-dev
21:53 paramat joined #minetest-dev
22:00 Jordach joined #minetest-dev
22:56 rubenwardy left #minetest-dev
23:05 ThomasMonroe joined #minetest-dev
23:12 turtleman joined #minetest-dev
23:28 QwertyCool would it be a bad idea to download a newer version of irricht?

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