Time  Nick        Message
10:15 fruitsnack  Hello, can somebody help me with building the android version?
10:16 fruitsnack  I'm trying to package 5.1.0 for fdroid, using the same build environment as for previous builds (0.4.16, 5.0.0, etc)
10:17 fruitsnack  The native part builds fine, however java part gives errors
10:17 fruitsnack  https://pastebin.com/FbPjtTG1
10:17 fruitsnack  Here's the build log
10:17 fruitsnack  I'm using repo version at 5.1.0 tag, both minetest and mtg
10:18 fruitsnack  (my connection might drop but I'm checking chat logs regularily)
10:19 sfan5       that's not your fault, the code is actually broken
10:20 sfan5       you'll need to cherry-pick the commit that fixed this
10:25 fruitsnack  https://github.com/minetest/minetest/commit/7f023de5cb2299f9ef9e1363afde414fe14ede51
10:25 fruitsnack  this one?
16:30 Andrey01    hello, it`s strange 'minetest.get_meta(clickednode_pos):get_string("origin") or minetest.pos_to_str(clickednode_pos)' expression results nil when first operand is actual nil and other is string returned from that function
16:32 Andrey01    string represents true value and hence the operation should return that string
16:34 Andrey01    minetest.pos_to_string* just
16:34 Andrey01    not minetest.pos_to_str
16:59 Ruslan1     ?
18:01 Andrey01    please maybe anyone knows a solution of my problem above?
18:03 rubenwardy  I don't understand your problem
18:06 Andrey01    as I already wrote a line with the logical expression 'minetest.get_meta(clickednode_pos):get_string("origin") or minetest.pos_to_string(clickednode_pos)' returns nil
18:06 Andrey01    but the first operand actually returns nil and other a string
18:07 Andrey01    and I don`t understand why that operation returns nil, if second operand is string, it should return string as a result, right?
18:09 Andrey01    'minetest.get_meta(clickednode_pos):get_string("origin")' returns nil and  'minetest.pos_to_string(clickednode_pos)' does string, but why does 'or' operation return nil eventually?
18:10 rubenwardy  so   `minetest.get_meta(nil):get_string("origin")`  ?
18:10 rubenwardy  there should be a nil access error on   `nil:get_string("origin")`
18:11 Andrey01    'clickednode_pos' in the case is a table with coords, bot nil
18:12 Andrey01    but*
18:13 MinetestBot 02[git] 04Montandalar -> 03minetest/minetest: Use appropriate LC_ locale macro for Windows (#9230) 134445889 https://git.io/JeFAW (152019-12-24T18:13:22Z)
18:13 MinetestBot 02[git] 04Wuzzy2 -> 03minetest/minetest: Fix item eat sound not played if last item (#9239) 13876a15b https://git.io/JeFAl (152019-12-24T18:12:46Z)
18:13 MinetestBot 02[git] 04Wuzzy2 -> 03minetest/minetest: Document permissible group ratings (#9236) 137199dc7 https://git.io/JeFA8 (152019-12-24T18:12:31Z)
18:17 Andrey01    local clickednode_pos = {x=pos.x, y=pos.y+1, z=pos.z}, so minetest.get_meta(clickednode_pos) returns meta and in my case get_string("origin") returns nil, because it doesn`t exist and therefore the first operand is nil, second is coords in string form
18:19 rubenwardy  ahh
18:19 rubenwardy  get_string() returns "" when it does not exist
18:20 rubenwardy  you want `get()`, which returns nil when it does not exist
18:22 Krock       get() returns the default value if it's a Minetest setting
18:22 Andrey01    ah, yes, I didn`t know there is so method, thanks!
18:25 rubenwardy  Krock: this ismeta
18:26 Krock       oh. context matters.
19:01 fluxflux_   when running a local server via the client GUI, would it be possible to not crash the entire client, when the server crashes due to a lua exception?
19:01 fluxflux_   i mean, if it crashes during startup, the client reports the error just fine. if it crashes after startup, the client just freezes and has to be killed.
19:21 nepugia     depends on the lua error
19:23 fluxflux_   it does? how so?
19:28 Krock       fluxflux_: well yes. error handling in callbacks isn't great - and there's already an issue about that
19:28 Krock       *a reported issue
19:28 fluxflux_   ok cool.
19:29 Krock       solution: don't crash the server
19:30 fluxflux_   yeah, not easy to do sometimes. it's only a minor annoyance, was curious if anyone else wished for better behavior as well.
19:32 nepugia     fluxflux_: i know severall failure modes that happen with lua errors, some crash the client, some make the screen go blank and some just kick you to the title screen (although argueably this last one can be correct)
19:33 fluxflux_   i mean, if it segfaults, it segfaults. i'm talking about when lua actually reports an error in the debug log.
19:33 fluxflux_   the last one is the behavior i wished for, but it seem to happen after the server has started.
19:43 nepugia     that is still a bug :), it certainly should not segfault or blank the client
19:43 nepugia     both are rather unhelpfull for debugging
21:14 Genshin     Hello