Time  Nick       Message
02:44 paramat    okay hmmmm mgflat has docs and is ready for review when you have time https://github.com/minetest/minetest/pull/3313 in hidden form for now
02:56 VanessaE   paramat: you'll need to make sure to include a way to turn off trees and grasses.
02:56 VanessaE   most uses of flat mapgen seem to be for creative servers where wide-open space is desired.
02:57 paramat    yeah we're considering adding global flags to toggle ores, biomes and decorations
02:57 paramat    or at least decorations
02:57 VanessaE   decorations and biomes would probably be enough, I don't think server owners care if there are ores being generated
03:00 paramat    mg flags are intended to be updated, with 'trees' and 'flat' moved into mgv6 spflags as those only affect mgv6
04:29 dandelion  Created a pull request. Adds a ring buffer container to utils/container.h. https://github.com/minetest/minetest/pull/3354
08:58 celeron55  by the way, my conclusion from yesterday is that both libstdc++ 4.9's and libstdc++ 5.1's streams are broken
08:58 celeron55  but i cannot prove this because i don't have an easy way to test with older ones
08:58 celeron55  it's feasible because probably almost nobody in this world use streams like minetest's logging does
12:05 celeron55  /* set version information */
12:05 celeron55  void setVersionInfo(u8 major, u8 minor, u8 patch, std::string full) {
12:05 celeron55  just WHY is this comment here
12:07 celeron55  the answer: for no reason
12:09 celeron55  okay; well, actually it's there because the corresponding getter methods are named poorly and need a title comment to indicate what they even are, and then this is there for symmetry
12:09 celeron55  but why are they named so poorly
12:09 celeron55  there's no answer to that
12:12 celeron55  title comments are quite cancerous anyway; they will up all the code and in the end you can't find the actually useful comments
12:12 celeron55  fill up*
12:13 Calinou    aren't we supposed to use variables/functions like_this and not likeThis?
12:16 gregorycu  What's happening nigs
12:16 celeron55  Calinou: not for class member functions
12:17 gregorycu  Yeah, I hate comments like that
12:18 gregorycu  It's why I wish doxygen, or the ms doc tool supported documentation in different files, and code analysis annotations in separate files
12:19 celeron55  you could add a manual preprocessing step that copies the code to a temporary directory and adds comments from a custom format
12:19 celeron55  8)
12:19 celeron55  i would be actually very open to that
12:20 celeron55  or, i mean not manual but like something that doxygen doesn't expect
12:22 gregorycu  PilzAdam: I don't understand this issue: https://github.com/minetest/minetest/issues/3330
12:23 gregorycu  "Both get enabled. Result and canceling when entering in game : "
12:23 gregorycu  Do you know what was intended there?
12:23 celeron55  i think that issue is very clear
12:23 gregorycu  Result and canceling?
12:23 celeron55  instead of allowing that to happen we probably should just modify the mod spec to say that if two mods soft-depend on each other, the load order is random
12:24 celeron55  it's working as designed but the design isn't very useful
12:24 gregorycu  What is soft depends?
12:24 celeron55  it's the question mark
12:24 gregorycu  Basically means it isn't a strict dependency?
12:24 gregorycu  (It can load without it?)
12:25 celeron55  it means that if the questionmarked mod exists, it is be loaded before the mod that questionmarkdepends on it
12:25 celeron55  to be*
12:25 gregorycu  What is a hard dependency?
12:26 gregorycu  Well, I can guess
12:26 gregorycu  So, none of the two mods gets loaded?
12:33 gregorycu  Seems like very crap behaviour, to be honest, adding a soft dep causes a mod not to be loaded. Anyway
12:33 gregorycu  Doesn't matter
12:34 celeron55  it's simple logic
12:35 celeron55  nobody did it explicitly that way
13:02 rubenwardy To me it makes more sense to abort loading if two mods soft depend on each other
13:02 rubenwardy to me, soft depend means "load this mod first if it exists, otherwise don't throw an error"
13:03 celeron55  but they are 99% of the time perfectly capable of operating in any load order; how do you expect a player to solve that if it doesn't solve itself automatically?
13:04 rubenwardy What if they're not capable of operating in any order?
13:04 rubenwardy Isn't the whole point of optionally depending to make it load first?
13:04 rubenwardy Why optionally depend if you don't need it to load first?
13:04 celeron55  then you just lose some feature
13:05 celeron55  i challenge you to find a pair of mods that fail to work reasonably in that situation
13:05 celeron55  it's hard enough to even find a pair that soft-depend on each other 8)
13:05 rubenwardy food and the items it supports won't work if the items are loaded afterwards
13:06 rubenwardy soft depend loops happened
13:06 rubenwardy *happens
13:07 rubenwardy a while back, I ended up with a situation that food -> hunger, hunger -> mtfoods, mtfoods -> food.
13:07 celeron55  how did you fix it?
13:08 rubenwardy I can't remember
13:08 celeron55  if food items depend on food, aren't the items then themselves creating recipes from themselves into food?
13:08 celeron55  why would they soft-depend on food otherwise
13:09 rubenwardy https://github.com/rubenwardy/food/commit/fee1ca867079ba05484a89a82a8e101cecd50bb9
13:09 rubenwardy and there was a commit in mtfoods as well
13:09 celeron55  the result is just that it's defined essentially arbitrarily whether the recipes from food or the items take effec
13:09 rubenwardy but I gtg
13:09 celeron55  +t
13:09 rubenwardy I didn't know that hunger -> mtfoods
13:09 rubenwardy mtfoods -> food because mtfood wanted to use the powerful recipe library(tm) food has
13:10 rubenwardy food -> all farming/ingredient mods, as it needs to add groups to items
13:11 celeron55  that kind of sounds like mods should have separate content and library parts
13:11 rubenwardy I think my fix involved seperating food into food and food_basic: library and content