Minetest logo

IRC log for #minetest-dev, 2018-06-30

| Channels | #minetest-dev index | Today | | Google Search | Plaintext

All times shown according to UTC.

Time Nick Message
00:26 ManElevation heya how do i use the bind_address =
00:26 ManElevation do i put my ip in it? works the same as server_address? i need to use bind_address = as i have a ipv6 router
00:57 ManElevation please help on https://forum.minetest.net/viewtopic.php?f=6&t=20319
01:16 paramat joined #minetest-dev
01:26 YuGiOhJCJ joined #minetest-dev
01:45 VargaD joined #minetest-dev
04:20 Darcidride_ joined #minetest-dev
04:21 Darcidride__ joined #minetest-dev
04:53 VargaD joined #minetest-dev
05:20 Foz joined #minetest-dev
05:39 twoelk|2 joined #minetest-dev
06:26 AntumD joined #minetest-dev
07:31 wowaname[m] joined #minetest-dev
07:37 Darcidride__ joined #minetest-dev
07:38 warr[m] joined #minetest-dev
07:58 proller joined #minetest-dev
08:05 proller joined #minetest-dev
08:12 nerzhul ManElevation: this is the devel channel, ask for your questions on minetest-hub please
08:27 nerzhul any other approval on #7503 ?
08:27 ShadowBot https://github.com/minetest/minetest/issues/7503 -- Log usage of /pulverize by ClobberXD
08:30 karamel joined #minetest-dev
08:32 nerzhul i push a tiny memleak fix
09:17 ANAND joined #minetest-dev
09:31 Krock joined #minetest-dev
09:41 Fixer joined #minetest-dev
09:54 Darcidride__ joined #minetest-dev
10:01 nerzhul merging #7506
10:01 ShadowBot https://github.com/minetest/minetest/issues/7506 -- Removed '$' symbols from the start of shell commands in the README file by RustyRaptor
10:12 red-001 shame c++11 doesn't have lambdas with templates
10:13 nerzhul yes we are deprecated, but C++17 is for later, C++14 too :)
10:14 nerzhul C++14 has nice constexpr functions it's very cool
10:14 nerzhul constexpr int myNiceCalculator() { return sqrt(4) + 587 + cos(34); }
10:14 nerzhul :D
10:15 nerzhul i don't know if we can use C++14 instead of C++11, it depends on Android and old Linux we support, i think it's not possible, jessie should be C++11 only, we should wait for Debian 10 heh
10:35 nerzhul rubenwardy is more incisive than me :)
10:47 Krock nerzhul, #7410 needs an entry for the android makefile
10:47 ShadowBot https://github.com/minetest/minetest/issues/7410 -- Modernize lua read (part 2): C++ templating assurance by nerzhul
10:48 Krock Will merge #7503 in 5 minutes
10:48 ShadowBot https://github.com/minetest/minetest/issues/7503 -- Log usage of /pulverize by ClobberXD
10:49 nerzhul Krock: exact :)
10:53 Krock merging..
10:53 nerzhul Krock is this okay if i add the missing makefile entry ?
10:54 Krock nerzhul, I haven't tested it but the code looks good overall
10:55 Krock Whether to merge or not.. I don't know. More dev opinions would've been great to have there
10:56 nerzhul yes but nobody commented while i reported it :(
10:57 Krock well then, if it's tested I guess it's acceptable..
11:02 nerzhul it's tested on my workspace without problem
11:03 nerzhul merging then
11:03 nerzhul Thanks Krock
11:04 nerzhul hmm there is the std::string reader implemented , re-testing to be sure
11:06 nerzhul hmmm it's not good, i should revert some readingn params heh, the std::string strenght make mt_game crash because some nil are not checked
11:08 nerzhul okay i found it
11:09 Krock nerzhul, consider enabling a vehicles mod, pipeworks and unified_inventory to have more functions called
11:09 nerzhul yeah i already get unified_inventory :)
11:09 red-001 does the new code error on nil?
11:10 nerzhul yes it was the regression cause on the two calls i found
11:10 red-001 nil is used for default arguments in a lot of places
11:10 nerzhul i reverted the change here
11:10 nerzhul not everywhere, but yes i changed it by accident in this two places
11:14 nerzhul more places :) thanks the diff heh
11:16 nerzhul it's okay now
11:16 nerzhul waiting for travis
11:16 nerzhul Krock can you take a look a last time ? red-001 don't hesitate too
11:16 Krock will test it
11:21 Gael-de-Sailly joined #minetest-dev
11:23 red-001 I imagine a lot of code might depend on minetest casting the value to the right type...
11:23 nerzhul it's not the case, hopefully
11:23 nerzhul not on strings or bool
11:24 nerzhul and other parts can be detected on migration, easily
11:24 Krock 2018-06-30 13:24:49: ERROR[Main]: ServerError: AsyncErr: ServerThread::run Lua: Argument -1 is not a bool
11:25 Krock punched the MTG boat
11:25 nerzhul which call ?
11:25 Krock I don't know which call. Backtrace doesn't exist
11:25 red-001 could this at least support nil as false/empty string/zero?
11:25 nerzhul i can change the behaviour to be less strict yes
11:26 red-001 arguable using nil as that is ok code style
11:26 red-001 unlike passing an int to bool which is clearly a mistake
11:26 nerzhul okay, being less string then
11:27 nerzhul we can be more stricter later
11:27 Krock nerzhul, probably InvRef::contains_item is the issue
11:27 Krock arg 3 is optional
11:28 nerzhul okay, i pushed a commit to be less strict
11:28 shacharr joined #minetest-dev
11:29 nerzhul please re-test :)
11:29 nerzhul at least being less strict will permit to re-use the std::string reader in some code parts
11:31 nerzhul yes i can revert my regression fixes commits :p
11:31 nerzhul all of them are useless if we are less strict and have nil default values
11:31 Krock err.. checkstring should stay
11:32 nerzhul checkstring causes the backtrace
11:32 Krock that's good..?
11:32 nerzhul i cannot have a default value with it
11:32 Krock if args are optional, they must be treated that way in the API functions
11:32 nerzhul we have two choices, either these calls are strict, either not
11:32 nerzhul just tell me what you want for this API
11:32 red-001 luaL_check* functions are just wrappers for lua_is* + lua_to* right?
11:33 Krock red-001, no, they also throw an error
11:33 nerzhul strictness or not strictness on readParam
11:33 nerzhul and i will adapt the PR
11:34 nerzhul we can have a readParamWithDefault permitting to have nil
11:34 nerzhul it permits to handle both cases
11:34 red-001 have a default argument with a default value?
11:35 nerzhul wrong, it will trigger the default value in all cases
11:35 Krock well great. applied the patch previously, built from scratch. now I checked out nerzhul's branch directly with two new commits and it builds from scratch. again. wtf gcc
11:35 nerzhul Krock, very strange
11:36 nerzhul i'm not working on header files: p
11:36 nerzhul maybe due to the templating ?
11:36 red-001 default value with nil, otherwise if it fails to convert/type is wrong error out?
11:36 Krock maybe because gcc is stupid at detecting which files were modified and which not
11:36 nerzhul red-001, readParamWithDefault<std::string>(L, 1, "test");
11:36 nerzhul ?
11:36 red-001 yes
11:36 nerzhul okay, i will rewrite the PR in that way
11:37 red-001 can templates have default arguments?
11:38 red-001 *default function arguments
11:38 nerzhul they can but this will not permit to have a strict and lazy function
11:39 red-001 the less strict one will still error is someone passes an string instead of userdata or bool?
11:43 Krock so readParam<std::string>(L, 1) and readParam<std::string>(L, 1, "default value") is not possible?
11:44 nerzhul oh
11:44 nerzhul sorry, yes it should be possible :)
11:44 nerzhul it's permitted
11:45 nerzhul i add luaL_checkboolean as it doesn't exists in the Lua API
11:48 nerzhul it's nice, thanks for your comments, i'm continuing the fixes and i will push it heh
11:50 nerzhul testing it now, and i will push
11:53 nerzhul okay it's pushed
11:54 nerzhul Krock, red-001 ^ don't hesitate to finish your tests & reviews, my mtg is okay
11:57 red-001 nerzhul, /time makes lua freak out
11:57 red-001 script/cpp_api/s_base.cpp:89: static int ScriptApiBase::luaPanic(lua_State*): A fatal error occured: LUA PANIC: unprotected error in call to Lua API (bad argument #-1 (boolean expected, got nil))
11:58 Krock (still building)
11:58 nerzhul red-001 with last commits ?
11:58 nerzhul can you tell me which call shows you that ?
11:59 rubenwardy Don't use checkboolean
11:59 red-001 minetest.set_timeofday I guess?
11:59 rubenwardy It means you can't pass the result of something like   a or b  into a function
11:59 nerzhul rubenwardy because we did shit with it in the whole code ? :p
12:00 red-001 ugh
12:00 Krock still punching that MTG boat. still an error.
12:00 rubenwardy As the type of a may not be boolean
12:00 Krock 2018-06-30 13:59:55: ERROR[Server]: In thread 7fffd29e3700:
12:00 Krock 2018-06-30 13:59:55: ERROR[Server]: /data/Minetest/minetest/src/script/cpp_api/s_base.cpp:89: static int ScriptApiBase::luaPanic(lua_State*): A fatal error occured: LUA PANIC: unprotected error in call to Lua API (bad argument #4 to '?' (boolean expected, got nil))
12:00 ShadowBot https://github.com/minetest/minetest/issues/4 -- cppcheck warnings
12:00 rubenwardy Same for b
12:00 rubenwardy So there's no point in checkbool
12:01 nerzhul why nil & false is the same, it's not javascript :p
12:01 rubenwardy They're not
12:01 red-001 nil is weird
12:01 Krock no, it's not the same
12:01 nerzhul then the strict helper should not be so strict on bool
12:01 red-001 it's not false, it's just not true
12:01 nerzhul we need to have false default value in it
12:02 nerzhul do you mean Lua does quantic things ?
12:02 nerzhul shroedinger bool
12:02 sfan5 won't lua_toboolean convert that for you?
12:02 red-001 no
12:02 red-001 nil is a type and a value
12:03 nerzhul const TValue *o = index2adr(L, idx);
12:03 nerzhul return !l_isfalse(o);
12:03 red-001 same as "hello" isn't false
12:03 nerzhul it's the tobool code un lua
12:03 nerzhul #define l_isfalse(o)(ttisnil(o) || (ttisboolean(o) && bvalue(o) == 0))
12:05 nerzhul then i re-used the  for booleans
12:05 nerzhul you can re-tet, i'm doing same here
12:08 nerzhul it sounds good to me
12:08 nerzhul waiting for reviews :)
12:19 Krock at least the boats do not crash anymore. works so far
12:30 Darcidride joined #minetest-dev
12:33 entuland joined #minetest-dev
13:02 nerzhul sfan5 i answer to you. I look at travis now
13:03 nerzhul linter fixed. ready for a merge when reviews are good :D
13:12 nerzhul Krock are you okay with #7500 now ?
13:12 ShadowBot https://github.com/minetest/minetest/issues/7500 -- Fix memory leaks in mod storage by red-001
13:12 Krock tested ? yes : no
13:13 nerzhul merging #7494
13:13 ShadowBot https://github.com/minetest/minetest/issues/7494 -- Fix world deletion by numberZero
13:14 nerzhul for me it's okay Krock
13:14 nerzhul i reviewed the code another time to ensure this object lifetile, and it's linked with the LuaRef, no other place
13:19 Krock nerzhul, but the point is on whether the garbage collector function is called in any case
13:19 nerzhul it's called when lua_state is destroyed, hopefully
13:19 nerzhul else we would have a big memleak issue when returning back to main menu :p
13:21 Krock we have a memory leak when returning to the menu. maybe it's a different part, but there's a memory leak.
13:23 ensonic joined #minetest-dev
13:24 nerzhul this fix fix one of them
13:45 nerzhul Krock: dv name fixed
13:46 Krock nice
13:47 nerzhul i hope it's okay now we got a design after merge i can continue to improve our lua API by implementing more readers :)
14:00 ensonic joined #minetest-dev
14:01 behalebabo joined #minetest-dev
14:19 red-001 shouldn't #7505 be closed?
14:19 ShadowBot https://github.com/minetest/minetest/issues/7505 -- Remove the '$' symbols in the bash snippets in the README file.
14:19 red-001 along with #7334 unless someone can reproduce is
14:19 ShadowBot https://github.com/minetest/minetest/issues/7334 -- Player backend migration skips some players.
14:19 red-001 it*
14:27 Krock there we are.. 777 issues again
14:30 red-001 some needs to open a PR
14:30 red-001 need those 111 PRs
14:31 red-001 so I guess  #7433 is a blocker now?
14:31 ShadowBot https://github.com/minetest/minetest/issues/7433 -- Really delete things in fs::RecursiveDelete by numberZero
14:32 red-001 can't delete worlds on windows anymore unless it gets merged
14:39 nerzhul are #7500 & #7410 okay for a merge now ? it seems all issues are solved, we are just waiting for one approval
14:39 ShadowBot https://github.com/minetest/minetest/issues/7500 -- Fix memory leaks in mod storage by red-001
14:39 ShadowBot https://github.com/minetest/minetest/issues/7410 -- Modernize lua read (part 2 & 3): C++ templating assurance by nerzhul
14:56 Krock nerzhul, former seems good. Latter.. duh. I cannot test all API functions whether they still work as before
14:56 nerzhul currently they work as before
14:57 nerzhul if there is a regression -hub will mention it fast i think and we can fix it very fast too, but i re-read the diff 3 times to see if there were regression and don't see anymore
15:01 Player-2 joined #minetest-dev
15:03 Krock looked it through too. looks ok. sfan5, any comments regarding #7410 ?
15:03 ShadowBot https://github.com/minetest/minetest/issues/7410 -- Modernize lua read (part 2 & 3): C++ templating assurance by nerzhul
15:04 Krock nerzhul, oh well. LINT wants you to change the template line
15:05 Krock trivial thing, can be done on-merge.
15:10 nerzhul yeah merging & fixing
15:10 nerzhul & merging #7500
15:11 nerzhul thanks for your time
15:11 ShadowBot https://github.com/minetest/minetest/issues/7500 -- Fix memory leaks in mod storage by red-001
15:11 Krock np. Other PRs are waiting aswell
15:13 nerzhul if you got time for #7433, i don't know who can test that ?
15:13 ShadowBot https://github.com/minetest/minetest/issues/7433 -- Really delete things in fs::RecursiveDelete by numberZero
15:14 Krock No, I could fire up the XP machine and backport it but.. meh.
15:15 paramat joined #minetest-dev
15:15 nerzhul xD
15:15 sfan5 Krock: no more comments from me
15:15 Krock good. thanks :)
15:55 ensonic joined #minetest-dev
16:16 nerzhul i ported one part of #7123 to #7513
16:16 ShadowBot https://github.com/minetest/minetest/issues/7123 -- Android: switch to cmake based standard builds by nerzhul
16:16 ShadowBot https://github.com/minetest/minetest/issues/7513 -- Backport Android cleanup parts from #7123 by nerzhul
16:17 nerzhul if a android lover can just look at the PR, this a trivial code move, i think it's better to port changes from my huge 7123 to master asap when they are ready instead of having a big bullshit pr like 7123
16:18 nerzhul after merging this PR i will start to move some download functions to gradle and then, the most boring part: the build :)
16:19 nerzhul note, i justed published #7511 & #7512 too
16:19 ShadowBot https://github.com/minetest/minetest/issues/7511 -- C converter cleanup by nerzhul
16:19 ShadowBot https://github.com/minetest/minetest/issues/7512 -- Lua templating reading (part 4): s16, v2s16, v2f by nerzhul
16:19 nerzhul sorry i was absent a moment but here is some PR to review heh
16:19 nerzhul and yes we have the 777 vs 111 problem :p
16:25 Krock *sigh*
16:32 nerzhul sorry :D
16:38 Krock Rebased a few PRs of mine. Up to date again and ready for review
17:19 Taoki joined #minetest-dev
17:50 entuland_ joined #minetest-dev
17:53 entuland_ joined #minetest-dev
18:02 opal joined #minetest-dev
18:23 Lunatrius joined #minetest-dev
18:46 YuGiOhJCJ joined #minetest-dev
18:57 Cornelia joined #minetest-dev
19:50 Gael-de-Sailly joined #minetest-dev
19:55 paramat joined #minetest-dev
21:16 paramat will merge #7498 in5 mins
21:16 ShadowBot https://github.com/minetest/minetest/issues/7498 -- Android: Resize jump and crunch buttons' touch target by srifqi
21:30 entuland joined #minetest-dev
21:31 paramat merging
21:33 paramat done
21:34 paramat wow lots of good PRs coming in, and few people to review :D
21:43 nerzhul we did some work today with krock
22:11 paramat yes i saw :) good to see

| Channels | #minetest-dev index | Today | | Google Search | Plaintext