Minetest logo

IRC log for #minetest-dev, 2022-01-06

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

All times shown according to UTC.

Time Nick Message
00:56 AliasAlreadyTake joined #minetest-dev
01:07 GreenXenith joined #minetest-dev
01:09 olliy joined #minetest-dev
01:28 tekakutli joined #minetest-dev
01:51 olliy1or joined #minetest-dev
02:17 v-rob joined #minetest-dev
03:03 v-rob joined #minetest-dev
03:14 olliy joined #minetest-dev
03:28 queria joined #minetest-dev
03:34 queria joined #minetest-dev
04:32 v-rob joined #minetest-dev
05:00 MTDiscord joined #minetest-dev
05:52 v-rob Man, EnrichedString is gross.
06:12 YuGiOhJCJ joined #minetest-dev
07:13 Yad joined #minetest-dev
07:48 tekakutli joined #minetest-dev
08:35 erlehmann joined #minetest-dev
08:44 calcul0n__ joined #minetest-dev
09:16 calcul0n_ joined #minetest-dev
10:14 Fleckenstein joined #minetest-dev
11:03 tech_exorcist joined #minetest-dev
11:58 appguru joined #minetest-dev
12:14 rubenwardy anyone know what's happening here?  https://forum.minetest.net/viewtopic.php?f=6&t=27657
12:14 rubenwardy !title
12:14 ShadowBot minetest not connecting to content db - Minetest Forums
12:19 celeron55 i think their router has been hacked or something
12:19 celeron55 Linuxdirk's first answer is spot-on
12:20 MTDiscord <luatic> Not necessarily hacked, sometimes "security" software will MITM your traffic to "scan" it. I have experienced this when using my school WiFi. Obviously, they must then send the wrong certificate. Websites with HSTS enabled won't accept that.
12:20 MTDiscord <luatic> My workaround is to just use a VPN.
12:47 appguru joined #minetest-dev
13:15 tech_exorcist joined #minetest-dev
13:16 tech_exorcist joined #minetest-dev
14:06 m42uko joined #minetest-dev
14:28 Fixer joined #minetest-dev
14:32 Taoki joined #minetest-dev
16:00 m42uko joined #minetest-dev
16:04 freshreplicant[m There's always this: https://www.grc.com/fingerprints.htm
16:06 rubenwardy but what if that page is compromised
16:07 Krock .. why does container[] reset the child element's origin, but not scroll_container[] ?
16:08 Krock ._.   scroll_container, not scrollbar_container. nvm
16:32 tekakutli joined #minetest-dev
16:54 rubenwardy trivial update: https://github.com/rubenwardy/minetest.github.io/commit/cc565858e6fddab024af0f2ab6d3a1a56d195e57
16:55 rubenwardy remember that the direction.md document has the links I removed
16:56 rubenwardy well, most of them
17:14 Krock > mostly iteratively
17:14 Krock what's this supposed to mean?
17:15 Krock is that equal to "spontaneous development"?
17:16 Krock and an "open issues" link would surely be nice to have (without the "high priority" filter)
17:20 rubenwardy iteratively is code for without much planning. You choose what to do next based on the current state
17:25 Krock ah that's how it's meant. yes, that makes sense.
17:31 perfomanceHigh joined #minetest-dev
17:36 automatedtester joined #minetest-dev
17:43 perfomanceHigh is there a document or page that outlines design decisions?
17:43 perfomanceHigh Lua is not that highly regarded anymore, so i'm trying to read why Lua was chosen over alternatives.
17:49 sfan5 the choice to use Lua was made 7+ years ago so how it's regarded today is not particularily relevant
17:50 sfan5 also I can't think of an alternative off the top of my head
17:50 sfan5 the important part here is that a language has to be embeddable
17:51 Krock wHy dO We nOT UsE jAvAscRipT
17:51 sfan5 javascript is well embeddable these days actually
17:52 rubenwardy Javascript is tempting :D
17:52 Krock mostly using CEF, which is a giant beast. slim javascript interpreters might work well for minetest's use-case tho
17:52 perfomanceHigh yeah, a percentage of devs would unironically use javascript over lua.
17:52 perfomanceHigh i personally am not advocating for or against anything, but i wish to read about the decision process/trade-off made
17:53 rubenwardy The main reason is that Lua is small, lightweight, and fast
17:53 rubenwardy it's a common game language as well
17:53 rubenwardy Javascript has gotten faster in recent years
17:53 sfan5 well I was thinking like https://mujs.com/
17:53 sfan5 but that's just an interpreter
17:54 sfan5 nobody wants to touch V8 for embedding
17:54 perfomanceHigh mmmh, i have heard arguments against it being fast. but ofcourse, that's just a matter of scale. therefore i need to know where Lua is used within the big picture and what the trade-offs are.
17:54 rubenwardy LuaJIT is fast
17:54 sfan5 Lua is mostly not used for performance critical parts
17:54 celeron55 duktape is another lua-like javascript interpreter
17:55 sfan5 e.g. no number crunching, buffer wrangling
17:55 celeron55 i've used it and it does work
17:55 celeron55 but whatever, this entire topic doesn't really make sens
17:55 celeron55 +e
17:55 celeron55 lua works fine and there's a lot of code written using it already that MT is compatible with
17:56 celeron55 and lua isn't a half bad language to learn if you're aiming for the game industry
17:57 perfomanceHigh so the argument is. "Lua is sufficient, already invested in and there was no reasonable alternative at the time the decision was made"?
17:58 MTDiscord <Jonathon> honestly even some of the most bloated servers still manage to keep there legs under them with luajit
17:59 celeron55 back then the only reasonable alternatives would really have been python, which is a pain to embed as it's designed as a host language, and something like guile, which is lisp so nobody wants to program that
18:00 celeron55 lua is also secure in addition to being small, lightweight and fast
18:00 celeron55 and extremely easy to embed - which is obvious today as many scripting languages copy lua's interface
18:02 rubenwardy If I were to add another API, I'd go for something like C# or a general interface like C ABI
18:02 rubenwardy both of those are much more complicated
18:02 MTDiscord <luatic> Yes. Lua is a great language. It can be sufficiently fast; highly performance-relevant code should ideally be part of the engine or Lua libraries written in C(++). Lua's popularity has been stable. It's frequently used in the games industry, as has already been pointed out. I think the burden of proof lies on you, performanceHigh: Why not Lua? Lua is maybe the only thing that MT didn't get horribly wrong.
18:03 ROllerozxa having javascript as an embedded language honestly sounds extremely unweildy even with a minimal library like duktape
18:03 sfan5 not like js has a larger stdlib than lua ;)
18:03 rubenwardy JS isn't that dissimilar from Lua
18:03 rubenwardy just bigger
18:03 rubenwardy JS is designed for embedding as well: web browsers
18:04 celeron55 also i think lua is a better language design than javascript
18:04 celeron55 javascript is kind of an "oops well let's leave it like that" design
18:04 rubenwardy there's less of it
18:04 MTDiscord <luatic> ^
18:04 celeron55 lua was designed by software reserachers
18:04 celeron55 who knew what they were doing
18:04 MTDiscord <luatic> JS is dirty as heck, Lua started out clean
18:04 MTDiscord <luatic> JS is slowly getting cleaner, but still is a mess; and a bloaty one, at that
18:04 rubenwardy JS is more useful
18:05 rubenwardy and has typescript
18:05 perfomanceHigh JS is truly the pop culture of languages. everyone loves to hate it, yet it keeps failing upwards.
18:05 MTDiscord <luatic> Teal (typed Lua) exists too
18:06 celeron55 JS - the client-side PHP
18:06 rubenwardy JS is so much better than PHP
18:06 MTDiscord <luatic> Not ES5
18:06 MTDiscord <luatic> I think a "niche" language is exactly the right choice for a "niche" voxel engine.
18:06 celeron55 (i actually like PHP, it suits the job 8))
18:06 MTDiscord <threehymns> @preformanceHigh, that sounds more like C++ to me LOL
18:08 perfomanceHigh old C++ devs hate new C++, but it is spoken well off by everyone else (from my experience)
18:08 rubenwardy and the rest of C++ devs hate old C++
18:08 rubenwardy meanwhile I hate all C++
18:09 MTDiscord <threehymns> I just gave that one a star ! 🙂
18:11 MTDiscord <luatic> (1) we can see that (2) the IRC people can't see that and now wonder what a "star" is
18:11 MTDiscord <Jonathon> meanwhile people in irc not caring because there stuck in the 1800s feature wise. or till ircv3? or whatever the latest standard was
18:11 MTDiscord <luatic> Libera doesn't even support a chat history
18:11 MTDiscord <threehymns> At least ruben can see it!
18:12 celeron55 at that this rate someone will soon kickban the discord bridge
18:12 rubenwardy 1800s would be telegraph
18:12 perfomanceHigh i use discord, but i find IRC is more productive.
18:12 rubenwardy pretty sure this isn't telegraph-based
18:12 celeron55 -that
18:12 perfomanceHigh chat history being wiped is nice as well
18:12 MTDiscord <Jonathon> except its not
18:12 MTDiscord <Jonathon> its all logged to irc.minetest.net
18:12 MTDiscord <threehymns> which is sadly a seperate website
18:13 MTDiscord <Benrob0329> What does this have to do with Minetest?
18:13 MTDiscord <Jonathon> nothing
18:13 MTDiscord <Jonathon> more #minetest content
18:13 MTDiscord <threehymns> #meta-discussion
18:14 perfomanceHigh yeah, but like, information theory wise, i know my messages are logged. it's just a difference of the conditions under which they are collected.
18:14 ROllerozxa an IRC client can be run on a crunched together ball of orange peels, try to do the same with discord without getting banned
18:14 perfomanceHigh irc just has this clean, decentralised feel i wish i used more
18:15 rubenwardy please    -> #minetest
18:15 perfomanceHigh discord, but you host it yourself rather than sell the data of yourself and all your users.
18:15 celeron55 the beauty if IRC is it's the bare minimum, and as a result, IRC hosting can be done for free
18:16 celeron55 of IRC*
18:16 celeron55 without corporate interests
18:16 MTDiscord <luatic> true
18:16 celeron55 ideologically it suits projects like MT very well
18:18 MTDiscord <threehymns> agreed, but discord is not explicitly against opensource, they just understandably dont want people making copies of their own software.
18:19 MTDiscord <Jonathon> wrong channel, please cease
18:19 MTDiscord <Benrob0329> Anyways, as a note once the docs project matures a little bit more (give us a couple weeks to get our templates and formatting under wraps) if any core devs want to review our information for accuracy that'd be great :D
18:19 tech_exorcist joined #minetest-dev
18:20 MTDiscord <Benrob0329> We do the best we can of course, but the actual devs are obviously the most familiar with the source code, especially when there is influence on an API from 3+ locations in the tree.
18:21 MTDiscord <luatic> I have already seen that Desour is lurking, sometimes dropping comments.
18:22 MTDiscord <Benrob0329> Indeed
18:27 rubenwardy Should Minetest only load from share/mods if the game is also in share? Or am I reading this wrong
18:27 rubenwardy because that seems wrong
18:28 celeron55 maybe
18:29 celeron55 but i won't pretend i know how users are actually using share/mods :D
18:30 rubenwardy subgames.cpp:110
18:32 rubenwardy I suspect this isn't tested much at all
18:32 rubenwardy the mainmenu doesn't seem to support share mods
18:34 celeron55 wait so there's already logic to avoid loading share/mods if the game is loaded from the user directory
18:34 celeron55 and user/mods is loaded only if the game is loaded from the user directory
18:35 celeron55 no
18:35 rubenwardy user/mods is always loaded - see user!=share bit
18:35 rubenwardy share/mods is only loaded if the game is in share/games
18:35 celeron55 user/mods is loaded for any game, yes
18:35 celeron55 i think these make sense
18:36 rubenwardy I don't think so, I think it should always support both. But that's something for another PR
18:36 rubenwardy I'd like to refactor and combine all this code. There's currently duplicates in C++ and Lua
18:36 rubenwardy context: #11784
18:36 ShadowBot https://github.com/minetest/minetest/issues/11784 -- Use virtual paths to specify exact mod to enable by rubenwardy
18:36 celeron55 well this logic does completely predate the main menu and probably any configurability in choosing mods in any way
18:37 celeron55 initially all mods were always loaded
18:37 rubenwardy yeah
19:17 olliy joined #minetest-dev
19:29 Fleckenstein joined #minetest-dev
20:06 Fleckenstein joined #minetest-dev
20:07 sfan5 merging #11872, #11935 in 7m
20:07 ShadowBot https://github.com/minetest/minetest/issues/11872 -- Fix damage wraparound if very high damage by Wuzzy2
20:07 ShadowBot https://github.com/minetest/minetest/issues/11935 -- Fix incorrect bit positions in paramtype documentation by aerkiaga
20:09 MTDiscord <luatic> sfan5: the damage wraparound might require more attention, thinking of exploitation attempts
20:10 MTDiscord <luatic> "This PR makes sure that the wraparound no longer occurs by switching to s32. While in theory that boundary can be exceeded as well, this scenario is too theoretical." Is it? What about cheating?
20:11 sfan5 the TOSERVER_DAMAGE takes an u8, so this was never an issue even with the old code
20:14 MTDiscord <luatic> NVM, there is a rangelim included. Merge it.
20:14 sfan5 the value can overflow before that
20:14 sfan5 but for that you need a tool with level=999999 or whatever stupid
20:15 MTDiscord <luatic> Are levels not s16?
20:15 sfan5 🤷
20:17 lhofhansl joined #minetest-dev
20:18 lhofhansl Just saw the share/games discussion in the logs. Wanted to mention my previous somewhat related efforts: #10426
20:18 ShadowBot https://github.com/minetest/minetest/issues/10426 -- Handle world mods and world game mods in the config dialog. by lhofhansl
20:18 Yad joined #minetest-dev
20:19 rubenwardy ah, interesting
20:19 v-rob joined #minetest-dev
20:20 rubenwardy this stuff should really use a tree rather than a bunch of random fields
20:20 lhofhansl Yes. In general per game mods are very powerful, IMHO
20:20 sfan5 https://github.com/minetest/minetest/pulls?q=is%3Apr+is%3Aopen+milestone%3A5.5.0 <<<< reviews please
20:21 sfan5 and/or testing
20:33 sfan5 for #11831 btw I plan to check if the implementation can't be made simpler but idk when I will have time do to that
20:33 ShadowBot https://github.com/minetest/minetest/issues/11831 -- Don't go out of the map during raycast by savilli
20:33 sfan5 also #11866, not essential but you'll have to live with erlehmann's complaints otherwise
20:33 ShadowBot https://github.com/minetest/minetest/issues/11866 -- Raise max mapgen limit constant to align with MapBlock by sfan5
20:48 v-rob Do we need this code anymore? https://github.com/minetest/minetest/blob/master/src/util/string.h#L382-L416
20:49 rubenwardy no, as that's in C++11
20:51 sfan5 definitely not but maybe hold off with a cleanup as it looks we're going c++14 soon-ish
20:51 rubenwardy :O
20:51 rubenwardy also, disregard my second comment on the create game PR
21:43 v-rob Left three comments on #11863, but otherwise looks good and works properly.
21:43 ShadowBot https://github.com/minetest/minetest/issues/11863 -- Fully remove bitmap font support by sfan5
21:51 olliy1or joined #minetest-dev
21:52 olliy joined #minetest-dev
22:00 m42uko joined #minetest-dev
22:06 sfan5 thanks, will check back tomorrow
23:32 v-rob joined #minetest-dev

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