Time Nick Message 00:22 MTDiscord Restyling unit tests sucks. I can't wait until I'm done. 14:06 Sokomine i'm having trouble with the texture "combine" command. the second picture doesn't show up: default_sandstone.png^[combine:32x32:0,0=default_sapling.png:0,16:default_apple.png 14:07 Sokomine in the end, four textures ought to be placed on top of one. is that acceptable regarding server and client performance? if perhaps a dozen textures are combined that way? (provided it works) 14:07 sfan5 should be 0,16=default_apple.png 14:09 MTDiscord Yep. Plenty of syntax errors in textures unfortunately fail silently. 14:13 erlehmann implying there is no input validation going on 14:14 erlehmann luatic, funny, i wonder what afl-fuzz thinks of texture combinators 14:14 MTDiscord As I already stated in an issue, texture modifiers use parsing from hell. 14:16 Sokomine excellent! thanks a lot. with a 32x32 texture, it now works fine :-) 14:38 Sokomine what still gives me trouble is automaticly creating textures for doors: [combine:16x32:0,0=default_wood.png:0,16=default_wood.png works just fine - except that such doors look odd. the wooden texture needs to be rotated by 90 degree - which ^[transformR90 does nicely - for an individual texture 14:39 MTDiscord so then use it on the individual texture 14:39 MTDiscord note you will need to \ before : or ^ 14:40 MTDiscord its documented in lua_api.txt 14:40 Sokomine oh? i might have missed that 14:41 MTDiscord actually 2 of them, one for minetest, one for lua 14:41 Sokomine aah. truely :/ i was blind. sorry! 14:42 MTDiscord also this seems to be #minetest questions 14:44 MTDiscord heh, this https://github.com/minetest/minetest/blob/master/doc/lua_api.txt#L450 is blatantly wrong 14:45 MTDiscord needs to \ 14:46 MTDiscord hmm, was mentioned https://github.com/minetest/minetest/issues/9958 14:47 MTDiscord i would think you would want the actual way to use it in lua, but idk 14:48 Sokomine yes, that'd be nice 14:50 MTDiscord sfan5, thoughts since you commented on that issue? 14:51 Sokomine can't say that the result is as expected: [combine:16x32:0,0=(default_acacia_wood.png\\^\\[transformR90):0,16=(default_wood.png\\^\\[transformR90) the textures are rotated, right, but also streched... 14:52 MTDiscord the \ are not need in front of [ 14:53 MTDiscord nor are the () needed 14:53 MTDiscord [combine:16x32:0,0=default_acacia_wood.png\\^[transformR90:0,16=default_wood.png\\^[transformR90 14:55 Sokomine Jonathon: thanks a lot! but sadly that doesn't seem to work either 15:05 MTDiscord are you trying to apply your non square texture to a square? 15:05 MTDiscord >16x32 15:20 Sokomine no, not really. trying to apply it to a door 15:21 Sokomine standard doors in mt are a bit lacking. good doors are important and can improve otherwise medicore buildings. i'd like to add pretty simple wooden doors in all the types wood comes with 15:23 MTDiscord no idea how the door is uv mapped 15:24 MTDiscord https://github.com/minetest/minetest_game/blob/master/mods/doors/textures/doors_door_wood.png it unwraps the whole door to one texture 15:25 MTDiscord so probably 16+16+some edge pixels x 32 15:26 Sokomine without the rotation element it does work as expected. just as soon as i get the rotation into the mix the texture is rotated - but also stretched 15:26 erlehmann As I already stated in an issue, texture modifiers use parsing from hell. 15:26 erlehmann luatic which issue? 15:28 MTDiscord probably one of https://github.com/minetest/minetest/issues?q=is%3Aissue+is%3Aopen+author%3Aappgurueu+texture+modifiers+ 15:28 MTDiscord @Luatic 15:34 MTDiscord @wsor : No, it's not even necessarily a double backslash in Lua. You can use long strings [[some\content]] in order to not escape stuff. 15:34 MTDiscord in the given example it is 15:35 MTDiscord No, the example doesn't use Lua quotes. 15:35 MTDiscord The first issue (https://github.com/minetest/minetest/issues/11588) BTW 15:58 MTDiscord Minetest aggressively updates worlds, which is quite a pain if you enter a testing world using a newer Minetest. 16:11 erlehmann luatic, i had the same issue. sfan5 what workaround would you accept? i'd honestly like a setting for this. 16:11 erlehmann or which would you find acceptable, i realize you won't want to make promises for unwritten code 16:53 Sokomine which kind of updates do you mean in particular? 16:54 erlehmann zstd 16:54 erlehmann you can't go home again 16:54 erlehmann once you loaded your home with newer version 16:55 erlehmann older version will reject 16:59 MTDiscord it won't even reject it gracefully - it usually crashes (window closes) for me 17:00 MTDiscord due to #11698 17:00 ShadowBot https://github.com/minetest/minetest/issues/11698 -- Async "mapblock version mismatch error" seems to be stuck in deadlock sometimes 17:00 Sokomine ah. thanks 17:32 sfan5 erlehmann: I can think of several which don't require dev work 1) run a 5.4 server with your world and only upgrade the client or 2) patch minetest to write the older version to disk or 3) or both! (if you need 5.5 features on the server) 17:32 sfan5 @Jonathon the docs could mention it but it should continue to say \ and not \\ (which would be wrong) 17:34 erlehmann sfan5 oh, i was not aware that a 5.4 server will solve it! 17:34 sfan5 running 5.4 will not magically fix your "broken" world but I'm sure you're aware 17:34 erlehmann uh, so stupid question – does a 5.5 server work with 5.4 clients? 17:35 sfan5 yes? 17:36 erlehmann oh, that means the server tells the client what format it wants! i have to research this, so i can figure out how to make an option by which the 5.5 client tells the server that it can not do zstd ig 17:37 sfan5 how would that matter? 17:37 erlehmann i have to figure it out 17:37 erlehmann i bet my mental model is wrong 17:38 sfan5 what the client says it supports will not influence what the server saves the map as 17:38 erlehmann ah i am looking for client side saving though 17:38 erlehmann or is that handled by the same code? 17:39 erlehmann if map format version >= 29 17:39 erlehmann found it 17:40 sfan5 if you're considering client-side saving you're on the right path 17:40 erlehmann path to what? 17:41 sfan5 path to getting a map that is readable by 5.4 17:41 erlehmann client side saving is the only reason i have a backup of some servers that were shut down 17:41 erlehmann ah! 17:48 sfan5 (note that downgrading the protocol version will also kill other 5.5 features you may care about, a workaround is left as an exercise to the reader) 18:16 sfan5 merging #11662, #11705, #11725 in 10m 18:16 ShadowBot https://github.com/minetest/minetest/issues/11662 -- Fix item duplication if player dies during interact callback (alternative) by sfan5 18:16 ShadowBot https://github.com/minetest/minetest/issues/11705 -- Limit stepheight smoothing to the stepheight and stop smoothing during jumps by TurkeyMcMac 18:16 ShadowBot https://github.com/minetest/minetest/issues/11725 -- Fixes around emerge handling by sfan5 18:57 sfan5 rubenwardy: gplay says there is 743,9 % more crashes in the new release but that seems like rubbish 18:59 rubenwardy I looked at the crashes and found nothing suspicious 18:59 rubenwardy I suspect it's a sample error 18:59 rubenwardy The errors are freezes, segfaults, and aborts. No Java crashes 19:00 rubenwardy The segfaults or aborts could be errors with assets but I doubt that 19:00 rubenwardy Or networking crashes I guess 19:03 sfan5 debug symbols would be really helpful 19:13 rubenwardy https://rwdy.uk/HRlGg.png 19:13 rubenwardy all the .so files 19:13 rubenwardy I think you said obj/local, which would be the first 19:31 sfan5 for another project I upload the contents of obj/local but with the objs folders deleted 19:32 sfan5 and that works 19:46 rubenwardy well, no errors 19:46 rubenwardy segfault is still ugly, maybe it only applies to new errors 19:47 sfan5 you can plug the addresses into addr2line manually to check what it should output 19:51 rubenwardy first one https://gist.github.com/rubenwardy/75691d39ce063273131993b896012de7 19:51 rubenwardy it's a destructor called on exit() 19:55 rubenwardy added a second one 19:55 rubenwardy second one is modalMenu.cpp:227 m_hovered->OnEvent(gui_event); 19:55 rubenwardy presumably m_hovered is null 19:56 rubenwardy hm no, there's a sanity_check 19:56 rubenwardy oh nevermind 20:00 rubenwardy modalMenu.cpp:300 gets an element from a position, but never checks whether it's null 20:02 rubenwardy I guess the root gui element should be returned instead of null 20:07 sfan5 free involuntary bug reports \o/ 20:07 rubenwardy lol 20:07 rubenwardy I'm guessing there's something weird about this device that's letting the user click outside the window or something 20:08 rubenwardy like, it would make sense for it to return null if it's outside of the root element 20:08 rubenwardy without being able to reproduce it it's hard to verify 20:09 rubenwardy hopefully with the debug symbols Play might group crashes better, and the biggest impacting ones can be determined 20:10 rubenwardy oh great. So, the abort is due to a non-0 exit code in android_main 20:11 rubenwardy the program does print the problem but we don't seem to have access to that 20:30 rubenwardy looked through a few of the app not responding reports, and they appear to be media loading related. Which makes sense 20:40 rubenwardy !title https://github.com/rubenwardy/minetest/commit/8dfeba02b9f084ddd52090ccd906534200f468b3 20:40 ShadowBot rubenwardy: Fix crash on hypertext[] with not enough parts · rubenwardy/minetest@8dfeba0 · GitHub 20:41 sfan5 lgtm 20:48 rubenwardy pushing...