Time  Nick          Message
00:20 est31         celeron55: reminds me of the fun I and a friend had with the windows ABI for 128 bit integers, and what LLVM thinks it was: https://github.com/rust-lang/rust/pull/38482
00:20 est31         and its still ongoing.
00:23 est31         (the thing these two projects have in common: both involve looking at assembly)
00:23 est31         although I have to admit that I didnt do the look at assembly part, this is what my buddy did
00:23 est31         (not fluent in assembly xD)
00:23 est31         either way, I guess I shut up, its offtopic to minetest dev :p
10:18 rubenwardy    seriously, who names a field `typ` over `type`?
10:25 Zeno`         lol
10:25 Zeno`         gotta keep those variable names short!
10:25 Zeno`         </sarcasm>
10:28 rubenwardy    updated # #4966
10:28 ShadowBot     https://github.com/minetest/minetest/issues/4966 -- Use tree to list mods rather than textlist by rubenwardy
10:35 rubenwardy    installmod, parse_register_line, and isValidModname are all defined methods in modmgr.lua
11:39 rubenwardy    so, about #4806
11:39 ShadowBot     https://github.com/minetest/minetest/issues/4806 -- Add search to advanced settings by rubenwardy
11:39 rubenwardy    currently enter is used to select a setting and either toggle or open the edit dialog
11:40 rubenwardy    but with that PR enter is used to search and then iterate through results
12:11 thatgraemeguy anyone smarter than me know what I'm missing and/or doing wrong? http://pastebin.com/raw/gyzDqkBs
12:12 thatgraemeguy same build script I've always used suddenly broken today :-/
12:15 AcidNinjaFWHR hi
12:16 AcidNinjaFWHR anyone can tell me how to migrate a sqlite db to redis ?
12:18 sfan5         ./minetestserver --world ./path/to/your/world --migrate redis
12:18 sfan5         try reading the docs next time
12:18 AcidNinjaFWHR Sorry.
12:18 AcidNinjaFWHR and thanks.
12:19 sfan5         thatgraemeguy: complain to your distro, this is not minetests fault
12:19 sfan5         also why are you linking to irrlicht statically?
12:20 thatgraemeguy I am not intentionally doing anything of the sort, I have no idea what that means. I'm just using the same cmake/make commands as always
12:20 thatgraemeguy I did go from ubuntu 16.04 to 16.10 last week though, so I guess I'll try and figure it out
12:43 nrzkt         AcidNinjaFWHR, if you want performance and you have a db size > available mem, i suggest you to use pgsql
12:59 emptty        Hello,
13:00 emptty        are the the changes of 0.4.15 documented somewhere?
13:01 nrzkt         git log 0.4.14~0.4.15
13:03 emptty        and in a compact and user-readable way ? :)
13:03 nrzkt         generally it's posted on minetest wiki by Calinou
13:04 emptty        yeah, but this time, http://dev.minetest.net/Changelog is empty
13:05 juhdanad      https://forum.minetest.net/viewtopic.php?p=243949#p243949
13:05 juhdanad      It is Fixer's chancelog
13:05 juhdanad      *changelog
13:14 rubenwardy    #4969
13:14 ShadowBot     https://github.com/minetest/minetest/issues/4969 -- Document registered_chatcommands, add def:run() by rubenwardy
13:20 Zeno`         rubenwardy, I don't understand what your PR does?
13:21 Zeno`         are all those items expanded underneath the mod's node selectable?
13:26 rubenwardy    yes
13:26 rubenwardy    it basically means you can collapse a modpack using the [+] rather than a check box
13:27 rubenwardy    obviously a textlist was originally used as the treeview hadn't been made yet
13:27 Zeno`         hmm ok. cool
13:57 Fixer         rubenwardy: ugh... useful feature... but modpacks are so broken :/
15:22 Calinou       nrzkt: sorry, I don't have anything written yet :(
18:34 Krock         btw, devs should have access to a pull's branch, for the case like #4963
18:34 ShadowBot     https://github.com/minetest/minetest/issues/4963 -- Dont compare short with bool by adrido
18:34 Krock         to repair that branch.. but sadly this won't affect master :/
22:12 sofar         fyi, have you guys seen the two_steps posts on the forums? it seems we're allowing the server to receive negative damage values from the client
22:13 VanessaE      the server should automatically translate those to positive values to punish the cheaters :P
22:17 sfan5         no that's impossible
22:17 sfan5         the damage packet contains an u8
22:17 sfan5         unless there is an integer overflow somewhere it's not possible to restore damage using that packet
22:17 sofar         is HP also an u8?
22:18 sfan5         i dunno
22:19 sfan5         when i made the "various anticheat improvements" pr i actually check that
22:19 sfan5         and i stopped thinking about it once i saw that the pkt uses an u8
22:26 sofar         I can never find my way from networkprotocol.h to src/*.cpp
22:26 sofar         the abstraction always gets me lost
22:28 sfan5         src/network/*packethandler.cpp is where you want to look
22:29 sofar         src/content_sao.cpp:s16 LuaEntitySAO::getHP()
22:30 sofar         src/remoteplayer.cpp:	args.setS32("hp", m_sao->getHP());
22:32 VanessaE      so it must be overflowing?
22:33 sfan5         underflowing to be precise, yes
22:33 VanessaE      er yeah
22:33 sofar         I'm thinking
22:34 sofar         implicit casting is the problem
22:34 sofar         PoC code:
22:35 sofar         https://gist.github.com/sofar/2268a96080dd7e7d7b14685ad771ed39
22:35 sofar         run that
22:35 sofar         oh, wait
22:35 sofar         nvm, that's b0rk
22:36 sfan5         http://sprunge.us/OdbA
22:36 sfan5         run that with 255 as an arg
22:36 sfan5         well actually that doesn't show the implicit casting problem
22:36 sfan5         but i believe it's happening here
22:36 sofar         I'm not sure, that's why I tried to code it up too, lol
22:37 sofar         #include <stdlib.h>
22:37 sofar         $ ./OdbA
22:37 sofar         Segmentation fault (core dumped)
22:37 sofar         lel
22:37 sofar         yeah
22:38 sofar         dmg=255
22:38 sofar         hp_new=21
22:39 sofar         yeah, this is wrong for sure
22:39 sofar         sfan5: I assume you're patching already?
22:40 sfan5         well first i'm verifying that it actually worsk
22:40 sfan5         works*
22:40 sofar         if (damage > hp) { } ... ?
22:42 sfan5         no it doesn't work
22:42 sfan5         if you send TOSERVER_DAMAGE with 255 you just die
22:47 sofar         maybe through TOSERVER_INTERACT? I see code there that sets the player HP
22:48 sofar         hmmm no, it just gets the hp
22:50 sfan5         ./content_sao.cpp:		setHP(getHP() - result.damage);
22:50 sfan5         ./content_sao.cpp:		setHP(getHP() - hitparams.hp);
22:50 sfan5         that are the only candidates left
22:50 sfan5         however setHP checks for < 0
22:51 sfan5         so unless result.damage/hitparams.hp is > 65535 nothing breaks
22:51 sofar         s16 hp
22:52 octacian      sofar, sfan5: let me know what you figure out. I must say though, the guy seems pretty knowledgeable.
22:53 sofar         neh, most of these exploits were well known already, and easy to use
22:53 octacian      true.
22:54 sfan5         ./content_sao.cpp:  setHP(getHP() - hitparams.hp);
22:54 sfan5         only that one
22:54 sfan5         the other is for luaentities
22:55 octacian      the thing he says the client can do that I don't get, is the anti-ban stuff. I have some ideas as to how that might work, but I'm not familiar with MT's core.
22:56 octacian      huh, he's supposedly added some new stuff. killaura, /spam (spams chat), and /die
22:56 sfan5         to me that sounds like the client is just ignoring that the connection ended, so you can keep walking around
22:56 sfan5         chat actually has flood protection built-in
22:57 octacian      that's prob what he's doing. wait, how does the client handle being banned? more specifically, which stops sending and receiving packets; client or server? I'd assume server, but
22:58 sfan5         the server caps the connection
22:58 octacian      good.
22:59 octacian      in that case, the anti-ban stuff though it might allow you to continue moving around, you wouldn't be able to interact.
23:00 sfan5         sofar: i don't see how negative damage would be possible via the TOSERVER_INTERACT -> punch way
23:01 sofar         no, I think that's not it either
23:01 sfan5         tooldefs and armor groups are not client controlled
23:01 sfan5         well what is it then?
23:01 VanessaE      what about fall damage?
23:01 sfan5         that's the TOSERVER_DAMAGE thing (which isn't it)
23:01 VanessaE      ok.
23:02 octacian      Maybe somebody should buy this and look from there, cause even though I'm a bit less skeptical now, for all we know, the things he's done may not even be real.
23:03 sfan5         buying it is not needed
23:04 sfan5         a packet trace would suffice to find out what's going on
23:04 octacian      not needed, but could make it easier.
23:04 octacian      but, since I myself wouldn't buy it, I'm gonna shutup about that :P
23:46 MillersMan    I'm currently working on #4839 and close to a first prototype for the API. Should I create a PR even though I only have an work in progress proposal of how it might work?
23:46 ShadowBot     https://github.com/minetest/minetest/issues/4839 -- Support for recursive refining map generators