Time  Nick        Message
00:00 IhrFussel   Is there a way to dump an entire inventory so that I can simply use string.find() to look for certain stacks?
00:11 rubenwardy  to_table
00:12 rubenwardy  bad idea to do that though
00:12 rubenwardy  unless for debugging
00:26 Sokomine    IhrFussel: maybe contains_item can be of use for you?
00:56 IhrFussel   If I call contains-item() 5-10 times wouldn't it be much slower than calling string.find() 5-10 times?
00:56 IhrFussel   contains_item()
01:03 BillyS      IhrFussel, I assume so, but dumping the table to a string might not be that fast either.
01:25 rubenwardy  it would probably be slower
09:19 IcyDiamond  rubenwardy: woo!
09:39 rubenwardy  Fuck me https://forum.minetest.net/viewtopic.php?f=9&t=20621
09:41 xerox123_   Too right, I know what my next installed mod is
09:46 IcyDiamond  rubenwardy: https://i.lunasqu.ee/Screenshot_18-08-06-12:46:30.png thanks a lot <3
09:47 IcyDiamond  https://i.lunasqu.ee/Screenshot_18-08-06-12:46:58.png also fluids
10:18 rubenwardy  Nice
10:18 rubenwardy  IcyDiamond: Consider using minetest.colorize on them, could make them look sexy
10:18 rubenwardy  *even sexier
11:13 IcyDiamond  rubenwardy: yeah ill see about that :P
11:23 Krock       !tell IhrFussel contains_item is faster than dumping & string find
11:23 MinetestBot Krock: I'll pass that on when IhrFussel is around
11:26 rubenwardy  Definitely lol
11:27 Krock       just to make sure
14:24 ANAND       Is it possible to link more than two lists using listring[] to allow shift-clicking?
14:35 Krock       ANAND: yes but it's only one "list ring"
14:36 Krock       so you can allow to move multiple lists  into a main one, but only move from the main one to another list
14:36 ANAND       Ah
14:37 Krock       IIRC, it will take the next ring item (inventory list info) after the one which was triggered by the shift-click call
14:38 ANAND       So multiple listrings for the main list won't result in unexpected behaviour?
14:39 Krock       code to confirm my "IIRC" https://github.com/minetest/minetest/blob/master/src/gui/guiFormSpecMenu.cpp#L3640-L3652
14:40 Krock       to move something out of the main list, it will take the first occurrence of the main list and take the next item to move the items there
14:40 Krock       That behaviour is surely handled properly
14:41 ANAND       Alright, I'll give it a shot
14:41 Krock       but just calling listring[] probably won't do what you want, so specify parameters
14:42 ANAND       Thought so... :)
14:44 ANAND       Does the order of the listrings matter?
14:45 Krock       ? Yes. Yes, it does.
14:45 rubenwardy  Yes
14:45 ANAND       The first is the origin, second the destination?
14:46 Krock       yes and no
14:47 Krock       imagine this: A B C. Shift click in C, and it'll be moved to A
14:47 Krock       shift click in A and it'll go to B
14:47 Krock       ofc is A defined first
14:47 ANAND       ohh!
14:47 ANAND       right
14:47 ANAND       that makes it very clear
14:48 Krock       !next
14:48 MinetestBot Another satisfied customer. Next!
14:48 ANAND       Waaaaiiit! Not yet...
14:49 ANAND       Do I have to use one pair of listrings for AB, BC, and CA?
14:49 ANAND       i.e. 6 listrings in total?
14:50 ANAND       or just A->B->C (i.e. 3)?
14:50 MinetestBot IhrFussel: Aug-06 11:23 UTC <Krock> contains_item is faster than dumping & string find
14:50 ANAND       I'm guessing 6
14:50 Krock       ANAND: no. specify one by one
14:50 ANAND       right, thanks
14:50 Krock       don't use listring[] without arguments when you have more than two lists
14:50 Krock       unless you know what you do
14:50 IhrFussel   Krock, are you sure that it#s still faster when I call it quite a few times? Isn't contains_item() a lookup every time?
14:51 ANAND       Understood
14:51 Krock       IhrFussel: ofc it is. but the lookup happens in C++
14:51 Krock       so it's a simple function call, a hundred microseconds (roughly guessed) in C++ and then back to Lua
14:54 IhrFussel   Okay then^^ Also is this guy implying that banning users shouldn't be a thing? There are lots of cases where banning is necessary and there is no other solution than to ignore troublemakers and let them destroy your server https://github.com/minetest/minetest/issues/7625#issuecomment-410674013
14:55 Krock       No, they're suggesting that players must gain their interact&shout privs
14:55 IhrFussel   The question is if banning makes even sense in MT if all we have is the IP
14:55 Krock       so that the loss is higher
14:55 Krock       probably worth closing that issue. Privacy issue.
14:56 IhrFussel   You cannot make players gain their basic privs on every server... that depends fully on the gameplay you want to achieve
14:57 rubenwardy  You can usually protect against grieifing and such pretty easy
14:57 Krock       surely you can. It's as simple as an "I agree" button on join
14:57 rubenwardy  The problem is with chat offenses and other subjective interactions
14:57 rubenwardy  I agree doesn't mean they'll do it
14:58 Krock       that's for interact. you could let them unlock shout by buying it or registering an email
14:58 rubenwardy  Every user on CTF has to click agree too
14:58 IhrFussel   I also don't see it as a privacy problem if we can only detect the device they play on...their device doesn't belong to them in that sense, it's just one of billions
14:58 rubenwardy  It's also hashed
14:58 Krock       the issue is that a device is pretty much unique identifyable
14:58 rubenwardy  The issue is that it's a tracking token
14:58 rubenwardy  And yes that
14:59 rubenwardy  It would be better to combine this with a central login system anyway
14:59 rubenwardy  Well, or not do this
14:59 IhrFussel   Just because you can detect a certain device doesn't mean you have access to any data...privacy is only about data
14:59 Krock       IhrFussel: no but you can track it
15:00 IhrFussel   Tracking only makes sense if you can associate the device with a profile which we can't
15:00 Krock       "makes sense" doesn't matter in the privacy rules
15:01 Krock       you can associate the device with nicknames, maybe on multiple servers if you centralize it
15:01 Krock       nevertheless can all this data be faked
15:02 IhrFussel   Lots of games give each device an ID and they don't face any lawsuits
15:03 IhrFussel   And just because we are OSS we have to accept troublemakers ruining our servers
15:04 IhrFussel   Some are THAT dedicated to trolling MT servers...you can ban their IP multiple times and they won't leave ... these people are wasting staff time and can (partly) destroy the gameplay
15:05 Krock       how about if you make the join procedure take more time? More effort? Require an email verification?
15:05 Krock       Let the staff ban them instead of wasting time
15:06 Krock       Using autocomplete it takes a few seconds to ban someone
15:07 nerzhul     warn about email verification, it's a personal data, GPDR is strict
15:07 Krock       oh right
15:07 ANAND       How would central login be implemented? Using email verification?
15:08 Krock       Eye and teeth scan
15:08 Krock       Or by SQL injection
15:09 ANAND       .....
15:09 ANAND       Forum accounts?
15:14 nerzhul     google or FB authenticatio n? :D
15:17 ANAND       lol
15:17 rubenwardy  My design would use its own accounts
15:17 rubenwardy  Probably move from cdb to a central system idk
15:17 rubenwardy  It would be based on certificates and stapling
15:18 rubenwardy  You could have xban on the central system as well
15:18 rubenwardy  Ie: a server code say, don't let this account or any associated IPs join my server
15:18 rubenwardy  Not entirely sure on the details of that yet
15:19 rubenwardy  It would have to be more similar to a password manager anyway
15:19 rubenwardy  :(
15:19 IhrFussel   I'd argue that my server already punishes troublemakers...at least if they played actively before cause they lose their level, their unlocked features (based on level), their protections, their stuff, their friends (ingame friend list) and more ... the real problem is just with those who join once and never again
15:19 rubenwardy  Like, usernames on a server are owned by a central user of not necessarily the same name
15:21 nerzhul     rubenwardy when the postgresql backend will come it can be easy to modify the pg code to implement a such thing and have a central db
15:21 rubenwardy  That sounds like a bad idea
15:21 rubenwardy  Let's not do that
15:21 rubenwardy  This is central as in minetest-wide and official
15:21 rubenwardy  Not as in shared by a few users
15:21 nerzhul     central auth for users ?
15:22 nerzhul     in MT coredev servers ?
15:22 nerzhul     on*
15:22 rubenwardy  No, all serverd
15:22 nerzhul     like minecraft then ?
15:22 rubenwardy  No
15:22 rubenwardy  Because this would work without the central database online
15:22 Krock       I'd prefer to have multiple central databases, run by people who actually run the servers
15:22 nerzhul     then if it's centralized per server network a central pg db is sufficient to have common auth. Now we have the new auth backend we can have common auth
15:23 nerzhul     Krock i'd prefer too, i justwanted to be clear :)
15:23 rubenwardy  Fragmentation
15:23 Krock       ^
15:23 rubenwardy  Also doesn't solve the issue
15:23 rubenwardy  This is pointless if every server has their own
15:24 rubenwardy  With my design, you could just have multiple CAs
15:24 rubenwardy  Then it would be federatised
15:24 rubenwardy  Or whatever the term is
15:25 Calinou     I've seen other open source/freeware games add centralized auth systems, it didn't always turn out well
15:25 rubenwardy  The important thing is the central server gives some sort of token, ideally a certificate, which is then used to login with untrusted servers
15:25 Calinou     most games ended up not requiring authenticated accounts anyway
15:25 Calinou     (i.e. they became optional)
15:25 rubenwardy  This would be opt out
15:25 rubenwardy  You can just fall back to SRP
15:30 rubenwardy  I have bigger priorities though
15:30 rubenwardy  Like actually having some QA
15:30 rubenwardy  And sane input bindings
15:44 nerzhul     rubenwardy: a certificate is too high security for a game like MT
15:45 rubenwardy  You could use tokens instead
15:45 rubenwardy  But it's less reliable if the central server goes down
15:46 rubenwardy  Plus the difficulty of issuing new tokens
15:46 rubenwardy  Alternatively you could just not worry about it going doqn
15:46 rubenwardy  Make it redundant
15:48 rubenwardy  There's no such thing as too high security
15:48 rubenwardy  There's only over engineering
15:50 Krock       well, certificates are surely never wrong but mostly too much effrot
15:50 Krock       *effort
15:52 rubenwardy  Yeah
15:52 rubenwardy  This is low priority anyway
15:52 rubenwardy  Therefore probably not worth the effort
15:56 Calinou     Xonotic uses a public/private key system for authentication
15:56 Calinou     a lot of players end up losing their private key when moving over machines (they don't copy their private key files) :(
15:58 sfan5       Calinou: xonotic doesn't really tell you that it has just generated an important file that you should totally™ not lose
15:58 sfan5       neither does mumble IIRC
15:59 Calinou     Mumble does, but it's not prominent enough to my taste
16:00 rubenwardy  Doesn't matter if they lose there key
16:01 rubenwardy  The central server will issue them a new one
16:01 Calinou     yeah, but they will appear as separate players
16:01 rubenwardy  No
16:02 Calinou     http://stats.xonotic.org/search?fs=&nick=calinou&server_name=&map_name=&submit= once returned 5 entries, mind you :P
16:02 Calinou     I got all of them merged eventually
16:02 Calinou     and now I make sure to copy my key when installing a new PC
16:02 rubenwardy  The certificate just says that the central server vouches that they're actually foobar
16:02 rubenwardy  I'm talking about minetest
16:02 rubenwardy  Or my proposal
16:02 rubenwardy  The purpose of the certificate is to allow logging in when the central server isn't available
16:02 rubenwardy  Which is one of the big criticism
16:03 rubenwardy  Actually
16:03 rubenwardy  A much simpler way would be to generate a load of one time passwords to login with. Less secure, but no certs
21:27 Fixer       HEY, LITTLE DRUM!