Time  Nick       Message
01:30 Bobr-m     Hi
01:31 bigfoot547 Hi
04:39 _stephen_  Looks like the latest minetest or maybe minetest_game is broken.
04:39 _stephen_  I did a checkout from git and I can't build.
04:39 _stephen_  If I checkout an older version I can get it to compile, but the current master will not.
04:39 _stephen_  http://paste.debian.net/971087/
04:39 _stephen_  http://paste.debian.net/971088/
04:43 Hijiri     what tool does the travis build use to lint PRs?
04:48 Hijiri     _stephen_: did you try updating jsoncpp
04:48 _stephen_  haven't yet.  I can see how that could caus eit.
04:49 _stephen_  What version should I be using?
04:49 Hijiri     I don't know, just try the newest stable version
04:49 Hijiri     or whatever you can get easilyy
04:53 _stephen_  Oh, looks like I need 1.8.0 according to 2362d3f926e2702585f60011d4cea90b4faf4bd6
04:58 Bobr-m     Hmm
05:00 _stephen_  damnit, newest the repo has is 1.7
05:00 _stephen_  Guess I'll see if that works.
05:15 _stephen_  What distribution of Linux do you generally build on?
05:16 Hijiri     Debian
05:16 Hijiri     Debian Testing
05:17 _stephen_  and looks like 1.7 wont do.
05:17 _stephen_  I'm just going to have to build an older version for a while.
05:21 _stephen_  I can build the version from just before that checkin, so I'll just use that until my distro catches up.
05:22 _stephen_  Thanks for the help.
05:49 srifqi     Trying to compile Minetest using buildbot.sh in Ubuntu on Windows. Wish me luck.
05:49 srifqi     (Targeting: win32)
05:54 srifqi     No more mutex and thread compilation error! :D
05:55 srifqi     But I get a lot of strtok warning.
05:55 nerzhul    srifqi, there is compilation problem, the build is industrialized, just follow either travis-ci build or gitlab-ci
05:55 nerzhul    but it's not really a discussion for -dev, maybe for -hub
09:27 nerzhul    sfan5, why there should be discontinuity ?
09:44 sfan5      pch and mt return different random values even for the same seed because they are not the same
09:44 sfan5      the point of the seed in world generations is that it returns the same values
17:25 RobbieF    Hi Shara
19:41 red-001    ShadowNinja, is #5965 a good way to go about implementing #5958 or would it be better to create a new security system from scratch?
19:41 ShadowBot  https://github.com/minetest/minetest/issues/5965 -- Load client-side mods into memory before executing them. by red-001
19:41 ShadowBot  https://github.com/minetest/minetest/issues/5958 -- Proper CSM security module
20:26 paramat    shall we tag latest MTgame as 0.4.16.1 because game#1761 is fixed and only trivial PRs have been merged since release? and before any other potentially risky PRs are merged?
20:26 ShadowBot  https://github.com/minetest/minetest_game/issues/1761 -- Regression - overriding of default functions does not work anymore (lavacooling)
20:28 Fixer      add tin blocks stairs/slabs...
20:28 Fixer      cat_please.jpg
21:26 paramat    #5972 tested and ready
21:26 ShadowBot  https://github.com/minetest/minetest/issues/5972 -- Mgv6 mudflow: Avoid floating stacked decorations by paramat
21:28 nerzhul    wow indent is purely crazy, too long lines, please do functions
21:28 nerzhul    what a unmaintainable function omg
21:29 nerzhul    9 levels of indentations
21:38 paramat    yes mgv6 code is pretty bad, there are many other things to clean up, but, another PR
21:48 nerzhul    paramat, you already said another PR but you should fix the code style with your pr, like everybody does.
21:48 nerzhul    i'm waiting your 2-3 months old PR for a unittest about the range check you fixes twice whereas a unittest can handle the border cases.
21:54 kilbith    you're expecting paramat to write a unittest?
21:54 kilbith    lol
21:55 paramat    the cleanup of mgv6 required is so big it makes sense to correct this in a separate PR. this is not code style but major code restruscturing
21:55 nerzhul    kilbith, i except paramt will stop to open duplicate issues and will conform to our coding rules and fix his PR before merge, else i will finish by closing them or reverting until he codes correctly
21:56 nerzhul    no separate PR, move some indented code part to a function.
21:56 paramat    sorry i don't want to write a unittest for mapgen limit
21:56 nerzhul    lol paramat you are just stupid man, you fixed it twice in master because you didn't code the unittest which cover edge case and this will show you EXACTLY the problem before merge on a so sensible function.
21:56 nerzhul    also i take 5 mins to do it
21:57 nerzhul    and i don't want to merge your PR on unreadable unmaintainable function with 9 indent levels, 140 lines and with lines over 90 chars which is our limit.
21:57 paramat    nerzhul my PR is fine for merging, i already intend to clean up mgv6 in a separate PR. restructuring mgv6 code to avoid this amount of indentation is a big job that will mean rewriting the whole function, this is best doen separately
21:58 nerzhul    no it's not fine for merging.
21:58 nerzhul    restructure it
21:58 nerzhul    fix the code style.
21:58 nerzhul    you never do separate maintenance pr, it's just ... smoke
21:58 paramat    so you can write the unittest, dev work is voluntary
21:58 nerzhul    no.
21:58 nerzhul    you do
21:59 nerzhul    it's your PR
21:59 nerzhul    not mine
21:59 kilbith    he can't write one
21:59 nerzhul    why kilbith , tell me ?
22:01 paramat    i have done lots of maintenance PRs, look at my history
22:01 nerzhul    then do it with this PR
22:02 paramat    a unittest does not have to be written by the coder of the relevant code
22:02 nerzhul    lol
22:02 nerzhul    i will not fix your shit man
22:02 nerzhul    i don't write often unit tests, but when i ask it there is a good reason
22:02 paramat    ok i'll look into moving the specific code block to a new function
22:04 nerzhul    ty
22:04 nerzhul    this function is over complicated, reading is important for everybody who doesn't know very well the code
22:04 paramat    the code block i alter has corect code style, it just happens to be in a larger function with indentation issues, therefore a fix requires a big rewrite that is possibly best done later, i admit this is a borderline case
22:05 nerzhul    not the line length
22:06 paramat    ok someone else can write the unittest, you don't have to, but i don't have to either. "i will not fix your shit man" lack of a unittest is not a breakage, there's nothing to fix
22:07 paramat    stop being an ass
22:07 nerzhul    for this PR there is no unittest for me
22:08 nerzhul    i talk about the unittest for the static function who checks the map borders
22:08 nerzhul    a unittest permits to ensure in some useful case you correctly handle the limits of your function and if you change it you just need to launch the unit test to ensure all matchs like it does, less tests to do
22:10 paramat    yes we are talking about the unittest for the mapgen limit, i know
22:15 paramat    erm, no lines are over 82 columns
22:16 paramat    tab size 4 remember
22:19 paramat    please don't make threats of unilateral action, it takes several disapprovals to close a PR or to revert
22:19 nerzhul    jsut make this function maintainable
22:21 rubenwardy which PR is this?
22:21 rubenwardy #5972
22:21 ShadowBot  https://github.com/minetest/minetest/issues/5972 -- Mgv6 mudflow: Avoid floating stacked decorations by paramat
22:21 rubenwardy that's fine imo
22:22 nerzhul    9 levels of indents are fine to you ?
22:22 rubenwardy like, it seems to be a fix to a section of the code, and only touches one part
22:22 rubenwardy no, but it should be done in a refactor - before or after this PR
22:22 paramat    the question is, if a code block i work on is inside a large amount of indentations, do i have to rewrite the whole function to reduce the indents? i'm not sure
22:22 nerzhul    + breaks + do while + for + for + for + for + if + while + for
22:22 nerzhul    just lol
22:23 nerzhul    my suggestion for this pr
22:23 nerzhul    move code inside "if (!dropped_to_unknown) {" to a function
22:23 rubenwardy that's also fair
22:24 nerzhul    this will remove 2 indents level from the function
22:24 paramat    this doesn't make the commit itself unmergeable, because including the cleanup would confuse the review of the fix, however does it need to be another commit within this PR? i think a separate PR would be ok
22:25 nerzhul    will you really do it or this is just like all other case a pure smoke ?
22:26 srifqi     Can we just have two commits in one PR? First for logic changes and second for code style fix.
22:26 paramat    yes a split there seems good, i'll consider it
22:26 nerzhul    srifqi, sounds reasonable
22:27 rubenwardy in this case, the splitting won't make it less reviewable
22:27 rubenwardy as that changes are all in one 20 line section
22:27 nerzhul    yes paramat do 2 commits in same pr
22:27 rubenwardy in other cases, it would
22:27 nerzhul    reviewable in both cases
22:27 rubenwardy but 2 commits is good
22:27 nerzhul    and less time lost to review a code style :)
22:27 paramat    "just like all other case a pure smoke ?" like one case when i made it clear i probably will not want to write a unittest
22:28 nerzhul    and if you did it, you won't fixed master 2 times because calculation was false
22:28 nerzhul    because all cases are directly covered by CI and unittests, it's very useful in border cases on complex maths calculs don't forget it
22:28 paramat    i doubt it, the issues and bugs were quite subtle
22:28 nerzhul    there is no subtile in IT, there is only bugs
22:29 paramat    you don't understand the issues at world edge, i know them better than anyone
22:29 nerzhul    then you are the best to write the border test case
22:29 nerzhul    it permit to change the funciton and ensure it works as intended next time :)
22:30 paramat    i am best qualified yeah :]
22:30 nerzhul    then next time i see a unittest case, please just do it, it you die or disapear it permits to ensure code quality for less knowledged people
22:32 paramat    anyway, it seems i am not forced to cleanup in this PR, another dev has a different opinion, but i will still consider it as a separate commit in this PR
22:32 nerzhul    it's nice to collaborate, as i take many time to answer on CSM
22:34 rubenwardy more unit tests would be good
22:34 rubenwardy but Minetest isn't really designed in that way, in most places
22:34 paramat    you can probably tell i don't like writing unittests ;]
22:34 rubenwardy no one does :)
22:35 nerzhul    oh unittests are difficult on games
22:35 nerzhul    if you want unittests on environment class it's difficult
22:36 nerzhul    also i don't really like our unittest custom framework, it has some cons, i prefer the long time and robust CPPUnit library which works with macros and has some interesting features
22:39 paramat    ok, i'll add a commit for cleanup to my PR
22:40 nerzhul    ty