Time Nick Message 12:57 Sharpik rubenwardy: Hello Rubenwardy, I'd like to ask If You are going to release 5.4.x android test build at https://github.com/rubenwardy/minetest/releases/ in near few days. Thanks for response and Your work 13:29 rubenwardy I had no plans to do so 13:29 rubenwardy 5.4 won't be released until December 13:34 Sharpik Ok, thanks for response 16:14 lhofhansl hi all, I updated #10426 to only handle worldmods (world specific mods) and no longer handles an optional "game" folder insight the world, since that is too controversial. 16:14 ShadowBot https://github.com/minetest/minetest/issues/10426 -- Handle world mods in the config dialog. by lhofhansl 16:57 sfan5 There's a backport branch for 5.3.1 with some anticheat fixes 16:57 sfan5 in your repo? 16:58 rubenwardy In minetest 16:58 rubenwardy Unless I didn't push it 16:58 sfan5 https://github.com/minetest/minetest/tree/backport-5 oh 17:04 sfan5 I wonder if any server owners are running the attached freemove fix yet 17:04 sfan5 it could use some testing 19:04 Krock ok guys I fixed the chat inputs 19:12 rubenwardy Yay 19:18 Krock #10456 19:18 ShadowBot https://github.com/minetest/minetest/issues/10456 -- Fix key input issues by SmallJoker 19:20 Krock I can now enter {}][ and ! which were not possible previously. perhaps there's also improvement for other layouts 19:20 Krock using stock irrlicht 19:20 rubenwardy as key bindings or in text input? 19:21 MTDiscord I'll add to my client and see how it goes 19:24 sfan5 .git/rebase-apply/patch:113: trailing whitespace. 19:24 sfan5 switch (key_code) { 19:26 Krock rubenwardy: latter. I believe former already worked 19:30 sfan5 there's also an unit test failure 19:30 Krock >:( 19:30 Krock 0 / 37 failed modules (0 / 198 failed individual tests). 19:33 Krock I don't understand why clang fails for that one. though the test is also kinda stupid 19:34 MTDiscord Would there be any interest in changing the color for the minimap radar to something other than pure green? 19:34 MTDiscord At the moment, it feels more like a debug tool than an actual gameplay element 19:34 Krock radars are commonly green, though. 19:35 MTDiscord true, but it could be toned down a bit (currently it's pure #00FF00 green) 19:37 sfan5 unlike operator== the std::hash implementation prefers keycode over char 19:37 sfan5 could that be it? 19:37 Krock https://github.com/minetest/minetest/blob/master/src/client/minimap.cpp#L339-L341 19:39 Krock no, the hash function has to be like that 19:40 Krock also it would either use hash or equality signs 19:40 Krock also the hash function is only relevant to the input handler - not to KeyPress itself 19:41 sfan5 4. For two parameters k1 and k2 that are equal, std::hash()(k1) == std::hash()(k2). 19:42 sfan5 I don't see the function guaranteeing this 19:46 sfan5 and because the comparison can work on either keycode or char I can't come up with a hash function that does 19:46 Krock "Key" is bit-shifted to mostly ensure that the values don't overlap 19:47 Krock though it's a pity that wchar_t is the size of size_t 19:49 sfan5 the issue is that {KEY_KEY_G, '.'} hashes to (KEY_KEY_G << 24) and {, '.'} hashes to '.' but operator== would compare the two equal 19:50 Krock I see 19:52 sfan5 and in fact this was added before your PR but I didn't notice 19:53 sfan5 (maybe that's what caused the issues?) 19:57 sfan5 if you don't mind, check if reverting https://github.com/minetest/minetest/commit/787561b29afdbc78769f68c2f5c4f2cff1b32340 fixes key input 19:59 Krock it does 20:03 Krock okay. let's see whether this attempt works. == and hash will now produce the same outputs for the same KeyPress 20:05 sfan5 shouldn't that be if (valid_kcode(Key) || valid_kcode(o.Key)) ? 20:05 sfan5 otherwise it might happen that k1 == k2 but !(k2 == k1) 20:06 Krock right! 20:07 Krock ah shit. Sneak move broke 20:11 sfan5 :/ 20:13 Krock KEY_KEY_W must be taken for upper and lowercase W, whereas / is always in combination with shift 20:13 Krock I believe the reason why it worked previously is a matter of priority 20:14 sfan5 yeah probably 20:23 Krock fsck this. will probably just split the chat fixes into a new PR and revert the other one.