Time  Nick          Message
00:02 paramat       +1
00:10 paramat       #5289 luckily i know what i'm doing here
00:10 ShadowBot     https://github.com/minetest/minetest/issues/5289 -- Dungeongen: Add and improve parameters by paramat
00:45 zorman2000    paramat, just wondering about your PR #5289
00:45 ShadowBot     https://github.com/minetest/minetest/issues/5289 -- Dungeongen: Add and improve parameters by paramat
00:46 zorman2000    Is it configurable yet, or are you just adding the parameters for now?
00:46 paramat       it will not be configurable for players, only for those writing core mapgens
00:47 paramat       player-settable dungeon params is intended but hmmmm is inactive and i don't know how to do this
00:47 zorman2000    Hmmm I see... wouldn't it be similar to how the rest of the mapgen parameters are set?
00:48 paramat       it would, yes, i'll need some help though
00:49 zorman2000    Got it... thank you! I might look in the long term... but I'm not a good C++ dev myself :)
01:06 zorman2000    Minor bug #5296
01:06 ShadowBot     https://github.com/minetest/minetest/issues/5296 -- Android: Missing/wrong cart texture
02:31 kaeza         so, double-clicking the server list is kinda useless
02:31 kaeza         even if you fill the name and pass fields then dclick, the first click clears the password
04:02 VanessaE      kaeza: the password field also gets cleared when you change the window size.  irritating.
04:33 presstabstart Hi. I'm working on a bug to do with crashing when broken chests are right-clicked (#1462). I think I found a quick fix, but not sure if it's adequate. I still get warnings on producing the bug but no crash
06:02 presstabstart wait nevermind, it's just a really hard bug to reproduce. apologies.
08:26 Zeno`         the password getting cleared when server list finishes updating annoys the hell out of me
08:27 Zeno`         usually I'm about to press enter and all of a sudden it's gone :/
08:54 crazyR        i upgraded minetest a few hours ago to latest.... and since then technic drills will not change modes.... and r the replacer tool does work properly..  I havent upgraded technic so i know its not that. the only change made to the server was upgrading the engine
08:54 crazyR        anyone got any idea's where i can start looking to fix it?
09:08 blaaaaargh    Zeno`, the focus also resets to the name field it seems
09:09 blaaaaargh    think of e.g., start writing password -> list updates -> focus on name field -> type half your pass in the clear
09:09 blaaaaargh    may or may not be an issue
09:32 crazyR        anyone?
09:46 crazyR        i might be wrong but i think it might have something todo with this? https://github.com/minetest/minetest/commit/d988f9b7694de2155425fa109523bb6f44d643ea
09:47 crazyR        ^ @red-001 / @red-002
11:38 red-002       ?
11:39 red-002       That shouldn't change anything
11:39 red-002       is that the only new commit?
12:00 crazyR        red-002 its the only one that(forgive me if m wrong im no expert) messes with key strokes?
12:02 red-002       what's your last good commit?
12:08 crazyR        last good commit?
12:10 crazyR        red-002 that last update i did was on Feb 9, 2017 previous to the update i did yday
12:14 red-002       well I would suggest you try to use git biscet
12:15 presstabstart I've got a problem. What do I do if an issue was sent on the minetest_game repo that's an engine bug? Do I send the PR to the normal minetest repo?
12:16 red-002       close the one in mt_game and open one in the minetest repo?
12:17 red-002       or if you are fixing the issue just link to the one in mt_game
12:17 presstabstart okie dokie
12:17 red-002       github supports cross repo links
12:17 presstabstart how
12:18 red-002       minetest/minetest_game#issue_ref
12:18 red-002       https://github.com/blog/967-github-secrets
12:19 presstabstart Also should I be testing every small bugfix extensively? Because I don't have time for that
12:19 presstabstart don't mean to insult
12:19 red-002       well test it as well as you can
12:20 presstabstart because it's quite a hard bug to reproduce, has to be frame-perfect
12:20 red-002       whats the issue?
12:20 presstabstart #1462 in mt_game
12:20 ShadowBot     https://github.com/minetest/minetest/issues/1462 -- With no smooth lighting EW faces should be brighter than NS faces
12:20 presstabstart not that one
12:20 red-002       game#1462
12:20 ShadowBot     https://github.com/minetest/minetest_game/issues/1462 -- breaking then right-clicking a chest causes a crash
12:21 red-002       huh I would have though that will be a mt_game thing
12:22 presstabstart had to do with an invalid pointer in game.cpp
12:22 red-002       well as long as your fix doesn't have any bad side efffects I say go ahead and make a pr
12:22 presstabstart what do you mean by mt_game issue?
12:22 red-002       well I didn't read the issue
12:22 presstabstart I had to edit game.cpp to fix it, so I assume it's an engine issue
12:23 presstabstart and it was seg faulting
12:23 red-002       I assumed it meant lua crash my bad
12:23 presstabstart ok thanks
12:24 red-002       yeah go ahead and make a PR that way more people will see it
12:25 red-002       and be able to test/review it
12:26 crazyR        red-002 never used git bisect. how would i use it to find the bad commit?
12:31 red-002       do git bisect start
12:32 red-002       then git bisect good last_good_commit
12:32 red-002       and git biscet bad current_broken_commit
12:33 red-002       then test the commit it picks and do biscent good if it's working and bisect bad if it's not
12:34 red-002       contiune untill you find the bad commit
12:34 crazyR        so in my case the last_good_commit would be latest or the one before  your commit?
12:35 red-002       it would be the one you used for your previous build
12:35 red-002       the one that worked
12:35 crazyR        ok. thanks
12:38 crazyR        do git bisect start
12:38 crazyR        git bisect good 5707b739f38cc5cf651de5b69d91d4f46511dac0
12:38 crazyR        git bisect bad e2ad76f91087215309012300eb9f1ebdd7fa40ee
12:38 crazyR        gave me this error: Some good revs are not ancestor of the bad rev. git bisect cannot work properly in this case.
12:45 burli         hi
12:45 burli         Is it possible to update a fork on github to the newest version?
12:46 Calinou       burli: yes, if you kept your master branch intact, it's easier
12:46 Calinou       just do this in your fork:
12:46 burli         I have a fork from minetest with a branch, but this fork is from last year and I want to update the master for a new PR
12:46 Calinou       git pull https://github.com/minetest/minetest.git
12:47 burli         ok, will try, thx
12:53 burli         Can I have a repository in a repository? e.g. minetest game in the games folder from minetest?
12:58 nerzhul       you can use a git submodule
13:20 nerzhul       can someone review #5284 seems ready for a merge
13:20 ShadowBot     https://github.com/minetest/minetest/issues/5284 -- FormSpec: Add position and anchor elements by adelcoding1
14:29 Thomas-S      It seems like the node replacement tool from the replacer mod doesn't work properly in the latest minetest version. Maybe a backwards-incompatible metadata change?
14:43 crazyR        red-002 i owe you an appology. it wasnt your comit
14:44 crazyR        Thomas yes it related to the one im trying to find now.... it has the same affect with drills and i think most tools that use shift click
14:44 red-002       metadata got broken again?
14:45 Thomas-S      Thanks for your efforts!
14:45 Thomas-S      Actually, I think that this is the searched commit: 2d1fca51e975a01f84fa3fd9b266435316fbe548
14:45 Thomas-S      It worked before if I remember correctly
14:45 crazyR        was just looking at that one now
14:46 crazyR        il revert it on my fork and test it out
14:56 crazyR        thats the one... commit: 2d1fca51e975a01f84fa3fd9b266435316fbe548 is the cause
16:00 rubenwardy    what's the issue? It works for me
16:01 rubenwardy    test case: https://gist.github.com/rubenwardy/bf4b27cd14ba2f57de5a2e4432becf1c
16:11 VanessaE      question:  if I wanted to create an image with *more* than 256 pixels for a palette, just so I could get a good layout for the colors, could I still use it as long as the colors I want fit within the first 256 pixels?
16:13 sfan5         i'm pretty sure the code rejects bigger images
16:13 VanessaE      ok.
16:14 VanessaE      damn :P
16:14 sfan5         if you absolutely want to you could use a crop image modifier
16:15 VanessaE      sfan5: I was thinking, say, 24x11 pixels, but using only the first 10 rows
16:16 Thomas-S      rubenwardy: The problem is the following: When you shift-right-click with the node replacement tool, you set it to the pointed node. The message, that the saved node changed appears, but the metadata change does not happen.
16:16 VanessaE      the documentation suggests that if I supply a 24x10 image for 240 pixels, it would also be discarded
16:17 VanessaE      rejected*
16:17 sfan5         just try the crop modifier thing
16:17 crazyR        ruben.. theres a problem with commit: 2d1fca51e975a01f84fa3fd9b266435316fbe548  it causes tools that use shift click to not work correctly...  technic drill and replacer are examples..
16:19 VanessaE      sfan5: I don't quite understand - how can one crop an image to exactly 256 pixels if one dimension is not a power of 2?
16:19 sfan5         oh is the palette not allowed to have fewer pixels?
16:19 VanessaE      that's the impression I got from the lua_api
16:19 sfan5         if thats the case crop modifier + [combine
16:21 Sokomine      maybe it's more related to the way metadata is stored? after all the message still appears; thus, shift-rightclick as such ought to work. storing the metadata doesn't work anymore
16:23 crazyR        not sure... all i know is that i reverted that comit and it worked fine :)
16:23 Sokomine      rubenwardy: https://github.com/Sokomine/replacer/blob/master/init.lua about line 93ff. storing the new metadata does not seem to work anymore (or is retrieving it what fails?)
16:24 Sokomine      the way i store the information about the stored node is a string in the form "node:name param1 param2" (as a string)
16:24 Thomas-S      I think that storing (not retrieving) is the problem. My node replacement tool still places the node I stored yesterday
16:24 VanessaE      sfan5: actually looking at the source code, it may not matter how big the image actually is
16:25 Sokomine      yes, quite probably. same here
16:26 rubenwardy    putting print(dump()) statements claim that the metadata changed
16:28 rubenwardy    although it always sets to dirt_with_dry_grass, even when I'm pointing at stone
16:29 rubenwardy    ah, I know why
16:29 rubenwardy    it's because the [""] key in the meta table overrides the metadata key
16:30 rubenwardy    so [""] shouldn't be exported to the meta field
16:30 Sokomine      can you give me an example of how to do it better?
16:30 rubenwardy    although, side note - this is definitely the wrong way to do metadata
16:31 rubenwardy    you shouldn't use to_table and replace, you should just use set_metadata
16:31 Sokomine      due to the way it's implemented now?
16:31 sfan5         it may not be the most elegant way
16:31 sfan5         but it definitely isn't "wrong"
16:32 rubenwardy    this bug existed before 2d1fca, it was just hidden (as "" was renamed to itemstack:get_name())
16:34 VanessaE      sfan5: in case you're curious, it rejects the palette if it's not exactly right (either too large or too small)
16:34 sfan5         huh+
16:34 VanessaE      2017-02-24 11:33:33: WARNING[Main]: CNodeDefManager::getPalette(): the specified palette image "unifieddyes_temp_palette.png" does not contain exactly 8 pixels.
16:34 VanessaE      oh well.
16:35 ShadowNinja   celeron55, rubenwardy: If you're refering to the Irrlicht bug where its log messages sometimes aren't filtered, that's unrelated to my log changes, my changes just made the bug more visible.  There's an issue on GH with more details.
16:35 sfan5         wait is there no way crop images?
16:35 celeron55     ShadowNinja: i think the problem is that minetest's log messages aren't shown when they should be
16:35 VanessaE      sfan5: no.  the closest are the [verticalframe or [lowpart operations
16:36 celeron55     ShadowNinja: or was, when i was looking at it like a year ago
16:36 ShadowNinja   celeron55: Oh, never heard on that issue.
16:36 ShadowNinja   of*
16:36 sfan5         VanessaE: [combine:<crop W>x<crop H>:0,0=<image> would achieve it probably
16:37 VanessaE      sfan5: that would work, if the image had one dimension that was a power of 2. :)
16:37 sfan5         you can use that with combine again to create a suitable image
16:37 VanessaE      so you're suggesting something like cutting up the mis-sized image and re-arranging it, at load?
16:38 sfan5         yes
16:38 VanessaE      right.  that makes sense.
16:38 sfan5         like [combine:4x2:0,0=[combine\:3x2\:0,0=source.png
16:41 sfan5         hm but i'm sure we had some modifier to cut things
16:41 sfan5         the dungeon master code used it
16:43 sfan5         hm was that [sheet?
16:44 sfan5         nah thats recent
16:45 rubenwardy    #5302
16:45 ShadowBot     https://github.com/minetest/minetest/issues/5302 -- Store legacy metadata separate from new item meta data by rubenwardy
16:45 rubenwardy    also makes the json rework a little cleaner
16:46 sfan5         nvm the DM code used texture coords directly
17:44 Sokomine      Thomas-S: my replacer mod works again now thanks to rubenwardy
17:45 Thomas-S      Thank you very much, Sokomine and rubenwardy! I'll update my server now.
17:48 Sokomine      btw, i'm still having problems with mod_security. req_ie() seems to return a null value even though the mod is listed in trusted_mods
18:00 Sokomine      ah yes. and i ought to fix the problem with prices (or any other label data) in formspecs appearing behind/covered by the picture of the node. is there a way to fix that? to get label data to display on top/in front of other elements of formspecs?
18:09 rubenwardy    make the images backgrounds
18:09 VanessaE      sfan5: can [combine take negative coords?
18:09 sfan5         dunno try it
18:09 VanessaE      willdo.
18:10 VanessaE      I appreciate the help btw.
18:16 rubenwardy    sofar, est31: https://github.com/minetest/minetest/issues/4153#issuecomment-282361290
18:20 VanessaE      sfan5: negative coords appear to be valid.
18:20 VanessaE      can someone confirm that (negative target coords for [combine, that is) and update lua_api.txt ?
18:21 sfan5         shouldn't those cause some out-of-bounds writes though
18:21 VanessaE      maybe but the result works.
18:21 VanessaE      http://pastebin.ubuntu.com/24060130/
18:21 VanessaE      this produces a block textured as I was hoping for.
18:22 VanessaE      (of course that's not what I plan to do with the image :) )
18:24 sfan5         setPixel might be checking for out-of-bounds
18:27 Sokomine      rubenwardy: i'm using item_image_button for the images. the label part of that might do - i'll check...
18:32 red-002       is fire that's created by tnt not suppose to go out?
18:33 sofar         it's supposed to go out
19:11 octacian      sfan5: since you seem not to oppose a clearinv command, would you be willing to review #4994?
19:11 ShadowBot     https://github.com/minetest/minetest/issues/4994 -- Clear Inventory Chatcommand (/clearinv) by octacian
19:11 octacian      Also, just realized that the WIP label should be removed.
19:11 octacian      Zeno` or someone else ^
19:13 Shara         octacian: does it still clear crafting inventory too?
19:13 octacian      Shara: yes.
19:13 octacian      And yeah, IK it's not that nice.
19:14 Shara         Yuch. Was really hoping you'd chance that so this would be useful to me :P
19:14 Shara         change*
19:14 octacian      However, I was kinda left with two options if I didn't clear the crafting grid: introduce a second command, or use some extra parameter
19:15 presstabstart when i want to fix something trivial like indents in a PR i just make a new commit right?
19:15 octacian      ..so I'm not really sure at this point
19:15 octacian      presstabstart: yes, you can, though I'd recommend you amend the previous commit and force push
19:16 Shara         I'd amend for something tiny like that as well.
19:16 presstabstart I've amended before but I thought that was just for commit messages
19:16 octacian      As devs like paramat really don't like having excess code (e.g. another command), I thought I'd leave it as is.
19:16 presstabstart also can I "merge" commits together?
19:16 octacian      Maybe if enough people want it kinda separated I'd open a second PR, but I'm not sure ATM.
19:16 octacian      presstabstart: yes, you can. With amend or squash
19:16 presstabstart thx
19:17 octacian      Essentially, if you do git add then git commit --amend you can update the commit message and the commit files themselves (those staged with git add).
19:17 Shara         octacian: Maybe I'll just eventually try and do what I want as a mod instead. It's been on my list for an age, but never high enough to actually do.
19:18 octacian      Shara: well, with the API to override chatcommands you could just create a very small mod to override this and register a separate one (if needed)
19:18 Shara         I just have a server specific mod I drop things into, so doesn't even need a new mod :)
19:19 Shara         But it still would have been nice if this did what I hoped without extra fussing about from me :P
19:19 octacian      Well, get a second dev (aside from Zeno`) interested in having them separate, and I'll open a new PR that separates the commands unless someone else has a better idea
19:20 presstabstart so i squashed the commits, how do I make it show online?
19:20 octacian      the only reason why I feel it is important to be able to clear the craft grid too, is because in some places like an adventure map or a school setting, a player could stash something in the grid that they weren't supposed to have.
19:20 octacian      presstabstart: do git push origin +<branch name>
19:20 octacian      ..replacing origin with the remote nam
19:20 octacian      * name
19:21 Shara         Couldn't they just as easily stash something in a chest or on most servers, in armour slots and so on.
19:21 Shara         ?
19:21 presstabstart says everything's up to date
19:23 octacian      Shara: yes, they could. However, grid is builtin and there all the time, so to ignore it wouldn't make sense. Armour is registered by a different mod, that mod needs to cover clearing it with clearinv.
19:23 octacian      Adventure maps or school environments would be able to control whether a chest was available without any code.
19:23 Shara         Yea, I know it's a mod. Why I said most servers.
19:24 octacian      Yep
19:24 Shara         Either way, I don't chase devs to try and make them do what I'd like, so I'll just make something of my own when I've got time.
19:24 octacian      presstabstart: here's a good tutorial on changing previous commits in git, it might help. https://www.atlassian.com/git/tutorials/rewriting-history
19:25 octacian      lol, I understand. Still, if you find a second dev who thinks it should be separated, I'll open another PR.
19:42 sofar         git rebase -i is your friend
19:42 sofar         just make a small commit and then move it up and 'f'ix your older commit
19:46 octacian      sofar: I need to learn about rebasing...
19:47 presstabstart thanks, it worked
21:04 Sokomine      btw, do we have a freeze command by now? something like that might also be useful on servers from time to time. far more useful than clearing inventories
22:51 paramat       game#1552 any comments? might be merged soon
22:51 ShadowBot     https://github.com/minetest/minetest_game/issues/1552 -- Leafdecay: node timer based implementation, API. by sofar
23:04 paramat       it's had long enough, i'll test and merge it later, and add a news post in the forum