Time Nick Message 14:33 * Krock waves 14:59 * ANAND waves back 15:11 Krock VanessaE: for pipeworks: https://krock-works.uk.to/u/patches/0001-Fix-wrong-inventory_image-and-wield_image-types.patch 15:18 sfan5 perhaps the engine should've printed a warning instead 15:18 sfan5 oh well 15:32 Krock a warning that will be kept forever 15:37 sfan5 there is some concern with compatibility though: 15:37 sfan5 if the type of a field is ever changed, you can't just run a mod that has the new type on the old engine 15:38 sfan5 even though that would normally probably work even if the field is missing 15:38 Krock that's because the old engine could not handle it the way you wanted 15:39 sfan5 you could argue ignoring it is better than throwing an error 15:39 Krock indeed 15:40 Krock so what about mods that expect a certain type, like builtin? 9659 15:40 Krock #9659 15:40 ShadowBot https://github.com/minetest/minetest/issues/9659 -- Builtin should sanity-check some properties on node registration 15:41 sfan5 I'm not saying just changing it back is a solution 15:41 Krock warn whenever an unexpected data type was found? 15:41 sfan5 builtin could also do the "if it's not what I expect, ignore the field" thing, but it's annoying to do that in every place 15:43 Krock after all it would also be possible to add feature flags when a certain previously type-checked field gets more value possibilities 15:43 Krock it's not so nice, but would ensure that modders get what they intended 15:44 sfan5 yes 16:02 VanessaE um 16:02 VanessaE Krock: https://gitlab.com/VanessaE/pipeworks/-/issues/28 16:02 VanessaE your commit broke..something 16:02 VanessaE but I can't be sure what 16:09 Krock VanessaE: did you apply the patch I sent you twice= 16:09 Krock *? 16:10 VanessaE I honestly have no idea 16:10 VanessaE there's nothing within the last year in pipeworks' merged PR's nor anything that's open 16:10 Krock http://irc.minetest.net/minetest-hub/2020-05-07#i_5682052 16:10 VanessaE (from you) 16:11 Krock also earlier: http://irc.minetest.net/minetest-hub/2020-05-02#i_5679866 16:11 VanessaE oh, you only JUST said it also :) 16:11 VanessaE I probably shoulda scrolled up :P 16:11 Krock should've better used /query 16:12 Krock note to myself in the future 16:12 VanessaE yeah, I don't generally review old chat. always too much to go through 16:12 Krock I see 16:13 VanessaE I'll apply your patch 16:13 VanessaE thanks :D 16:14 Krock thanks. I hope the inventory image looks as intended (previously it used some default value) 16:14 VanessaE I would never have spotted that error 16:15 Krock strong type checks are a two-sided sharp sword :3 16:15 VanessaE hm, notify_hud_provider broke too.. dunno whose mod that isz 16:15 Krock similar issue? 16:16 VanessaE similar yeah. 16:16 VanessaE origin https://git.gryphno.de/neplevitan/minetest-mod-notify_hud_provider.git (fetch) 16:16 VanessaE ok your patch fixed pipeworks :) 16:18 Krock !next 16:18 MinetestBot Another satisfied customer. Next! 16:18 Krock oh, I can notify the author of this mod 16:18 VanessaE thanks 16:19 VanessaE come to think of it, I don't even remember why Dreambuilder even HAS that mod.. 16:19 Krock if you already have the error message - would you please be so nice and paste it somewhere? 16:19 VanessaE moment 16:19 VanessaE https://pastebin.com/emZSGZkf 16:19 Krock no worries. dreambuilder has so many mods it's hard to keep track of them :P 16:19 Krock thank you 16:20 VanessaE everything else in DB seems to be clean. 16:21 VanessaE I really need to keep up better with MT current events :P 16:21 VanessaE oh btw, 16:21 VanessaE 2020-05-07 12:16:10: ERROR[Main]: generateImage(): Could not load image "xpanes_space.png" while building texture; Creating a dummy image 16:21 VanessaE not sure where that's coming from 16:22 Krock VanessaE: if you want to be 100% sure, also run the chatcommand in #9748 so that entities are properly registered 16:22 ShadowBot https://github.com/minetest/minetest/issues/9748 -- Script: Enforce type checks if not nil by SmallJoker 16:22 Krock checking 16:22 Krock xdecor/src/nodes.lua: textures = {"xdecor_" .. name .. ".png", "xdecor_" .. name .. ".png", "xpanes_space.png"}, 16:22 Krock since 2019-07-23 14:03:20 16:22 VanessaE wtf? I don't even use xdecor 16:23 VanessaE AH 16:23 VanessaE my_mods/dreambuilder_modpack/castle_gates/doors.lua: textures = {"castle_jailbars.png", "castle_jailbars.png", "xpanes_space.png"}, 16:23 Krock no other matches within 360 mods 16:24 VanessaE Philipbenrs_mods/castle-modpack/castle_gates/doors.lua: textures = {"castle_jailbars.png", "castle_jailbars.png", "xpanes_space.png"}, 16:24 Krock strange enough. I get the same error using xdecor 16:24 Krock maybe somebody removed the texture 16:24 VanessaE possbly 16:24 * VanessaE goes off to file an issue 16:25 VanessaE oh, someone beat me to it. 16:25 VanessaE https://github.com/minetest-mods/castle_gates/issues/14 16:25 Krock should it be a blank pixel? 16:26 VanessaE I'm not sure 16:26 Krock hmm. it's the sides of the jailbars so I'd assume it is 16:26 Krock can write a quick fix for this one 16:26 VanessaE maybe those were replaced with a more "solid" image. 16:27 VanessaE (else you'd have bars that look paper thin) 16:28 VanessaE hm, notify_hud_provider does not seem to be used or depended on by any other mod. delete! 16:53 VanessaE there we go.. everything up-to-date and working properly now, new DB release issued 16:53 VanessaE *looks at the entity test thing* 16:54 VanessaE ehh, errr... not sure I want to know if anything is broken there :P 16:54 Krock it's for your safety 16:55 VanessaE well let's see.. 16:56 VanessaE bleh 16:56 VanessaE 2020-05-07 12:56:00: ERROR[Main]: ServerError: AsyncErr: ServerThread::run Lua: Runtime error from mod 'worldedit_commands' in callback on_chat_message(): /home/vanessa/.minetest/mods/entity/init.lua:5: bad argument #1 to 'assert' (value expected) 16:56 ShadowBot https://github.com/minetest/minetest/issues/1 -- GlowStone code by anonymousAwesome 16:56 VanessaE shaddup, bot :P 16:56 Krock https://github.com/minetest/minetest_game/pull/2614 removed the texture FYI 16:57 VanessaE that ^ error is thrown when I run /entities 16:57 Krock huh 16:57 sfan5 well it certainly isn't worldedit_commands fault 16:58 Krock assert() so.. add_entity returned no value 16:58 sfan5 an object can remove itself in on_activate so this is not a bug 16:58 VanessaE worldedit does use entities for its markers though 16:58 Krock oh! 16:58 Krock updated the code 16:58 VanessaE ok 16:59 * VanessaE updates and tries again... 17:01 VanessaE you're not gonna like the result :) 17:01 VanessaE I guess the ones it spawns are what it considers to be in error? 17:02 sfan5 it just supposed to spawn all entities that exist without causing an engine error 17:02 sfan5 it's* 17:02 VanessaE oh ok. then it's probably fine. 17:02 VanessaE no errors in the console 17:02 VanessaE just a lot of "spawned ..." messages 17:02 Krock no errors = pretty much safe 17:03 VanessaE fwiw, https://pastebin.com/1AWJEstu 17:03 VanessaE (idgaf about those two warnings :) ) 17:04 sfan5 VanessaE: unrelated question: what happens when signs_lib entities somehow disappear during server runtime? signs stay empty? 17:04 VanessaE an LBM restores them 17:05 VanessaE or you can /regen_signs if needed 17:05 sfan5 I mean while the block is still loaded 17:05 VanessaE yeah, for that the sign just stays empty until /regen_signs, or if you punch it 17:05 sfan5 good to know, thanks 17:08 VanessaE hm. dare I update the servers' engine/game builds? :) 17:08 VanessaE (they're both at 5.1.1) 17:09 sfan5 didn't some mod still crash last time you did that? 17:09 VanessaE a couple of mods did but I *THINK* I worked around it. now I don't recall :P 17:09 VanessaE though DB works fine with MT/mtg at git HEAD 17:10 VanessaE (updated an hour or so ago) 17:10 sfan5 HEAD has some issues that are still in 5.2.0 fixed after all 17:10 Krock documenting try/error results is what seems too much work at first, but would pay out after a week when everything's forgotten 17:11 Krock if there are mods overwriting the builtin entities (dropped item/falling node) then they might need a little adjustment 17:26 VanessaE wtf, I'll give it a go 17:28 VanessaE minetest@daconcepts:~$ /home/minetest/Scripts/build-minetestserver.sh *gently mashes "Enter"* 17:29 VanessaE (see, if I smack Enter firmly, I'm signaling to the machine that I'm confident in the decision, which in turn will give it cause to fuck with me :) ) 17:37 VanessaE ok.. 17:37 sfan5 it was minetest_game wasn't it? 17:37 VanessaE what does this mean in english? :) https://pastebin.com/FHxfnHjf 17:38 sfan5 oh 17:38 rubenwardy you're not overriding item_Entity correctly 17:38 rubenwardy update the item_entity.lua file in MTG 17:38 rubenwardy I should update my PR for that 17:38 VanessaE update it? er, I'm not intentionally overriding it? 17:38 rubenwardy default/item_entity.lua 17:38 VanessaE oh is this that one for pipeworks? 17:38 rubenwardy it was a bug in MTG 17:38 rubenwardy it may also be in pipeworks 17:38 VanessaE oh 17:38 sfan5 s/you/minetest_game (which has already been fixed) or another mod/ 17:39 rubenwardy yeah 17:44 VanessaE well in all of /home/minetest/mods, pipeworks is the only thing that overrides item_entity.. 17:45 sfan5 https://github.com/minetest/minetest_game/commit/bd64188ee566bc31c1cd50d63f0da3328d42ab51 17:45 VanessaE https://gitlab.com/VanessaE/pipeworks/-/blob/master/item_transport.lua#L263 17:45 sfan5 here's the commit that fixed it in MTG 17:45 VanessaE what, literally just ", ..." ? 17:46 sfan5 yes 17:46 rubenwardy are you sure you've updated MTG? 17:46 VanessaE rubenwardy: yes. 17:46 VanessaE commit ef45a8e14855a5631db8eab3712b0ae594a977ad 17:47 Krock "..." takes all following parameters (if there are any). va_args, if you want so. 17:47 Krock this is necessary to pass the moveresult to the original function 17:48 VanessaE wait a minute here, pipeworks is not overriding it. that's just some table entry 17:50 Krock VanessaE: the hud notifier mod is fixed now 17:51 VanessaE I already got rid of it 17:51 VanessaE it's superfluous anyway 17:51 Krock ok. so fixed as well. 17:51 VanessaE but this crash I'm not understanding. 17:52 Krock the item_entity error posted above basically says that the third function argument of on_step was nil 17:53 VanessaE why does it even care now when it was working fine before? 17:53 Krock since 5.3.0-dev it needs this third parameter 17:53 Krock this means MTG and engine have to be more or less the same version 17:54 VanessaE ok, and why isn't it telling me which mod is giving it bad info? 17:54 sfan5 the bad mod will be in the backtrace 17:54 VanessaE https://pastebin.com/FHxfnHjf 17:54 Krock here it's pointing to minetest_game 17:54 VanessaE that's the whole backtrace. 17:54 Krock exactly. error in minetest_game 17:55 VanessaE on_step = function(self, dtime, ...) 17:55 VanessaE builtin_item.on_step(self, dtime, ...) 17:55 sfan5 right and if it's not minetest_game there's just one other possibility 17:55 VanessaE in my server's copt... 17:55 VanessaE copy* 17:55 sfan5 a mod is changing the properties of the the item entity and indirectly causing this error 17:55 Krock setmetatable 17:55 sfan5 e.g. if physical = false then moveresult is always nil 17:56 VanessaE yep, only two params there 17:56 sfan5 the builtin entity keeps track of this, but if another mod changes that behind its back it will error but not show up in any backtrace or logs 17:57 Krock what if moveresult would also return a table when physical is nil? 17:57 Krock {collides=false} constantly 17:57 Krock it would make the parameter consistent, since there are no "nil" values for pointed_thing either, for example. 17:58 sfan5 kind of a waste of memory, it would also "hide" this bug without fixing it 17:58 sfan5 VanessaE: which item-related mods do you have? 17:59 VanessaE pipeworks and homedecor's item frame/pedestel mods the only one I can think of 17:59 VanessaE and item_drop 17:59 Krock indeed, though it would give mods more control about the entities 17:59 sfan5 !mod item_drop 17:59 MinetestBot sfan5: Item Drop [item_drop] by Calinou - https://forum.minetest.net/viewtopic.php?t=9638 18:00 VanessaE not that one 18:00 VanessaE the minetest-mods one 18:00 VanessaE there may be more that I am not thinking of right now 18:00 sfan5 that's the one I'd hoped to see 18:00 Krock !mod item_drop 18:00 MinetestBot Krock: Item magnet and in-world node drops [item_drop] by texmex - https://forum.minetest.net/viewtopic.php?t=16913 - https://github.com/minetest-mods/item_drop 18:00 Krock works for me... 18:01 Krock but I cheated 18:01 sfan5 https://github.com/minetest-mods/item_drop/blob/master/init.lua#L204 18:01 sfan5 here is the problematic override 18:01 VanessaE I'll patch it locally and see what breaks 18:03 VanessaE nopers. 18:03 VanessaE that didn't fix it 18:04 sfan5 did you patch both line 215 and 206? 18:04 VanessaE oh oopsd 18:04 Krock trash_can/init.lua: 18:04 Krock that mod also overrides it 18:04 sfan5 Krock: please don't tell me the trash can needs to override ..... 18:04 sfan5 .... 18:04 Krock >:D 18:05 Krock https://github.com/minetest-mods/trash_can/blob/master/init.lua#L262 18:05 Krock it's so beautiful 18:05 Krock redmattermod/init.lua:redmattermod_ARROW_ENTITY.on_step 18:06 VanessaE 2020-05-07 18:05:30: ERROR[Main]: ..._World/worldmods/dreambuilder_modpack/item_drop/init.lua:206: cannot use '...' outside a vararg function near '...' 18:06 VanessaE oh wait, maybe... 18:06 Krock nvm, redmattermod is a false alert 18:08 VanessaE 2020-05-07 18:08:36: ERROR[Main]: /home/minetest/minetest_core/src/serverenvironment.cpp:ea: std::__cxx11::string LBMManager::createIntroductionTimesString(): A fatal error occurred: attempted to query on non fully set up LBMManager 18:08 VanessaE what the actual fuck? 18:09 sfan5 haha 18:09 VanessaE all but one server are up and running now, that ^ is from my DB Creative server 18:09 sfan5 does env_meta.txt look like it might be broken? 18:10 * VanessaE looks 18:10 VanessaE https://pastebin.com/hpmmyUcU 18:11 VanessaE (you may want to look at the raw form) 18:12 sfan5 that looks like the expected output 18:13 VanessaE I could just reset the LBM list 18:14 VanessaE basically none of that should be useful anymore 18:15 sfan5 you could try that but that might not be the actual issue 18:15 VanessaE nope, that didn't work 18:18 VanessaE oh, wait, there's a little more 18:18 VanessaE https://pastebin.com/G2Mg1HhC 18:19 sfan5 I was about to type "LBMs are only saved at shutdown so this issue happening should be indicative of startup failing somewhere" 18:19 sfan5 why does the code care that the file is empty lol? 18:19 * VanessaE tries deleting the "empty" image 18:20 VanessaE nope, that wasn't it. 18:20 sfan5 does the error look different now? 18:21 VanessaE nope. 18:21 VanessaE other than that image not being mentioned. 18:21 sfan5 can you run it with --verbose? 18:23 VanessaE hang on. 18:25 VanessaE https://pastebin.com/aMPW4DQU 18:25 sfan5 uh huh 18:26 sfan5 this is going to sound stupid, but are you sure it's not a mod initiating a shutdown? 18:26 VanessaE I don't have any mods installed that should be able to do that 18:28 Krock ServerEnv::saveMeta() is called before actually loading it 18:28 sfan5 at this point I don't see another way except attaching gdb and seeing who called shutdown 18:32 VanessaE ok, let's see if gdb cooperates then 18:33 VanessaE https://pastebin.com/YrgiGtPm 18:33 VanessaE that's not very useful I guess 18:34 * VanessaE rebuilds with debug info 18:34 sfan5 gdb might be able to set a breakpoint on the right function even without debug info 18:35 sfan5 the function is Server::requestShutdown() 18:35 sfan5 server.cpp:3769 currently 18:37 Krock or os.exit() ? 18:37 VanessaE https://github.com/minetest/minetest/blob/master/doc/lua_api.txt#L5164 18:37 VanessaE but nothing running on the affected server calls that 18:37 sfan5 hm 18:37 VanessaE https://pastebin.com/sH1qpCWy 18:38 VanessaE that should be `thread apply all bt full` 18:38 VanessaE if I remember my gdb config right. 18:39 sfan5 yeah that's just the old backtrace but more details 18:39 VanessaE yeah, just figured the extra detail may help.. 18:40 sfan5 hm there isnt really a good way to debug this 18:41 sfan5 try a breakpoint on main.cpp:895 18:41 sfan5 or :925 if you don't have curses 18:42 VanessaE how do I do that? 18:42 VanessaE this is running detached, reading some commands from a file 18:42 sfan5 the command is `b file:line` but you'll want to do this from an interactive session 18:46 VanessaE ok, b main.cpp:895 18:46 VanessaE Breakpoint 1 at 0x6ca562: file /home/minetest/minetest_core/src/main.cpp, line 895. 18:46 VanessaE run it? 18:47 sfan5 yes 18:48 sfan5 currently considering if patching lots of debug output in would be more useful than messing with gdb 18:52 VanessaE there we go, helps if I run it as the right user (I'm not used to doing this interactive) 18:52 VanessaE ok, there's the crash 18:52 VanessaE now what? :) 18:52 sfan5 is it still the same? 18:53 VanessaE it looks so 18:54 sfan5 here's a diff http://sprunge.us/B8jEVz 18:54 sfan5 that should narrow it down without any needing any gdb action 18:56 VanessaE ok, applied that, compiling. 19:01 * VanessaE runs it... 19:01 VanessaE ok, there's the crash 19:03 VanessaE odd, I don't see any new output 19:03 VanessaE (e.g. init ok, start ok...) 19:04 Krock the point of these log messages is to figure out how far it gets before the shutdown 19:04 Krock FYI 19:05 VanessaE so I assumed. 19:05 sfan5 no new output is absolutely impossible 19:05 sfan5 unless you're running a client in singleplayer ;) 19:05 VanessaE wtf 19:05 Krock or you're running the wrong binary 19:06 VanessaE it said it applied the patch but the code isn't there. 19:06 VanessaE oh there it is. 19:06 VanessaE ok so the patch applied just fine. 19:07 * VanessaE rebuilds one more time just to make sure. 19:09 VanessaE ok, running it (without gdb) 19:10 VanessaE nope. 19:10 VanessaE it ain't showing me the new output, though the crash is still there 19:11 VanessaE minetest@daconcepts:~/minetest_core$ strings /usr/local/bin/minetestserver|grep "init ok" 19:11 VanessaE init ok 19:11 VanessaE and I *am* running that exact binary. 19:12 sfan5 maybe the output only does into the logfile or whatever 19:12 VanessaE nope, not in the log file either. 19:12 VanessaE and it's logging actions, warnings, etc. 19:13 VanessaE (though not VERBOSE lines, those only show up on the terminal) 19:14 sfan5 maybe there's an exception in the constructor 19:14 sfan5 and it doesn't get to print that because the lbm error happens first 19:14 sfan5 aha 19:14 VanessaE ? 19:14 * VanessaE hopes "aha" translates to "I know what broke" 19:14 VanessaE :P 19:15 sfan5 hm this is annoying to fix 19:16 sfan5 VanessaE: please comment out the entire method body of Server::~Server() 19:16 sfan5 server.cpp line 275 onward 19:18 VanessaE just that one method? 19:18 sfan5 yes 19:18 VanessaE ok, done. 19:19 VanessaE next? 19:19 sfan5 recompile, run it again in gdb 19:19 VanessaE ok.. 19:20 VanessaE [xkcd: Compiling!] 19:21 VanessaE epp, that bombed. 19:21 VanessaE /home/minetest/minetest_core/src/main.cpp:889: undefined reference to `Server::~Server()' 19:22 VanessaE (and a few more at other places) 19:22 VanessaE oh wait 19:22 sfan5 just the body 19:22 VanessaE you said the content, not t--yeah 19:22 * VanessaE fixes. 19:22 sfan5 so that you have Server::~Server() { /* .... */ } 19:23 VanessaE right 19:24 VanessaE there we go, that built clean 19:24 VanessaE gdb open. 19:25 VanessaE just run it? 19:25 sfan5 sure 19:25 VanessaE ok, that's new 19:25 VanessaE 2020-05-07 19:25:31: VERBOSE[Main]: ScriptApiEnv: Environment initialized 19:25 VanessaE [Thread 0x7ffff1a3b700 (LWP 24419) exited] 19:25 VanessaE [Thread 0x7ffff123a700 (LWP 24420) exited] 19:25 VanessaE 2020-05-07 19:25:31: ERROR[Main]: ModError: Invalid field "interval2020-05-07 19:25:31: VERBOSE[Main]: ScriptApiEnv: Environment initialized 19:25 VanessaE [Thread 0x7ffff1a3b700 (LWP 24419) exited] 19:25 VanessaE [Thread 0x7ffff123a700 (LWP 24420) exited] 19:25 VanessaE 2020-05-07 19:25:31: ERROR[Main]: ModError: Invalid field "interval" (expected number got string). 19:25 VanessaE [Inferior 1 (process 24406) exited with code 01] 19:25 VanessaE " (expected number got string). 19:26 VanessaE [Inferior 1 (process 24406) exited with code 01] 19:26 VanessaE oops 19:26 rubenwardy uh oh 19:26 VanessaE :P 19:27 VanessaE what oh G*d's green earth though? 19:27 sfan5 my guess: a mod tried to register an abm with interval = "123" when it should have been interval = 123 19:30 * VanessaE searches 19:30 VanessaE I have a hunch what it might be 19:32 VanessaE hm. hunch denied :-/ 19:34 VanessaE I can't begin to track this down without some reference to the mod name 19:35 sfan5 http://sprunge.us/PWlDat?diff 19:35 VanessaE gah, I hate that pastebin, line numbers get copied with the selection 19:35 VanessaE one line, I"ll do it by hand 19:36 sfan5 you can remove the ?diff from the url 19:36 sfan5 ..and I expected that you'd do it by hand anyway 19:36 Krock curl -Ls http://sprunge.us/PWlDat | git am - 19:37 Krock do the same for register_lbm too 19:37 Krock oh wait 19:38 Krock interval is abm-specific 19:38 VanessaE which is odd because it was throwing an LBM error, not ABM. 19:38 Krock misleading 19:39 Krock it's a follow-up issue caused by an error within the ABM registration 19:41 VanessaE AH HAH 19:41 VanessaE crafting_bench/init.lua 19:42 VanessaE I believe that's part of castles? 19:42 VanessaE https://pastebin.com/e6iYsmzM 19:43 VanessaE I think it refers to this mod, https://github.com/minetest-mods/crafting_bench 19:44 VanessaE hm. 19:45 VanessaE crafting_bench_crafting_rate = 10000 19:45 VanessaE local crafting_rate = minetest.settings:get("crafting_bench_crafting_rate") 19:45 VanessaE this must be returning a string. 19:46 sfan5 it always does 19:46 VanessaE much be another instance of that strict type checking? 19:47 * VanessaE wraps it in a tonumber() 19:47 sfan5 yes 19:49 VanessaE and I'll reset the engine back to a clean state as well 19:54 VanessaE fek. 19:54 VanessaE nope, that was not enough. 19:54 VanessaE oih wait. 19:54 VanessaE forgot to sync 19:56 VanessaE YES 19:56 VanessaE the downed server is back online. 19:56 * VanessaE posts the fix. 19:57 VanessaE https://github.com/minetest-mods/crafting_bench/issues/3 19:57 VanessaE https://github.com/minetest-mods/item_drop/issues/32 19:58 VanessaE thanks for the help in finding the root cause, that one was a right pain in the ass 20:11 VanessaE I trust you'll be pushing some commits to the engine to make such errors less troubledsome to sort out in the future :) 20:12 sfan5 happened 9 minutes ago 20:14 Krock that assert() check within builtin might help in the future. noted in case there are more changes needed in the same area 20:15 VanessaE sfan5: \o/ :)