Minetest logo

IRC log for #minetest-dev, 2018-03-10

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

All times shown according to UTC.

Time Nick Message
01:29 Beton_ joined #minetest-dev
03:04 Cornelia joined #minetest-dev
03:31 paramat joined #minetest-dev
03:37 paramat merging #7104
03:37 ShadowBot https://github.com/minetest/minetest/issues/7104 -- Spawn level: Add 'get_spawn_level(x, z)' API by paramat
03:41 paramat merged
03:57 AndroBuilder joined #minetest-dev
04:33 paramat merging #7120
04:33 ShadowBot https://github.com/minetest/minetest/issues/7120 -- Lua_api.txt: Split long lines part 2 by paramat
04:40 paramat merged
05:32 Darcidride joined #minetest-dev
08:09 Andrej1 joined #minetest-dev
08:38 Krock joined #minetest-dev
08:40 Krock Will push this typo fix in 5 minutes: https://pastebin.com/raw/wzs61YbL (solves #7118)
08:40 ShadowBot https://github.com/minetest/minetest/issues/7118 -- Lighting glitch since "Variable name fix + structure creation unrolling in lighting code"
08:46 Krock pushing...
09:03 AntumD joined #minetest-dev
09:11 nerzhul ty Krock
09:11 Krock np
09:11 nerzhul it seems a forced push commit didn't pass on my end
09:11 Krock yes, I saw your comment :)
09:11 nerzhul if should be getNodeNoCheck(relpos, ...) :)
09:11 nerzhul oh yes, i'm not sure it passed by mail heh
09:12 nerzhul at least not squashing commits permits to bissect it easily :p
09:31 SilverLuke joined #minetest-dev
09:41 fwhcat joined #minetest-dev
09:52 fwhcat_ joined #minetest-dev
09:52 Krock Any demand on starting a new discussion for the meeting today to collect the discussion points?
09:56 fwhcat joined #minetest-dev
10:22 nerzhul i cannot be there, it's my son's birthday, but my point is to release 0.4.17 ASAP to push many fixes to 0.4.X branch, and if we need it do a 0.4.18 with 0.5. There are too many bugfixes waiting for 0.5 and we don't have a release date as some required features are not there
10:25 Krock nerzhul, happy birthday to him :) I hope other devs will have some time, otherwise this can be moved to next week. I don't have a 0.5 release in mind yet for exactly that reason of too many PRs/issues open
10:26 nerzhul Yeah i agree 0.5 is not important now, but 0.4.17 is important, there is a very huge patchset useful for players, also on android
10:35 Lymkwi joined #minetest-dev
11:09 ssieb joined #minetest-dev
11:10 Gael-de-Sailly joined #minetest-dev
11:13 Darcidride_ joined #minetest-dev
11:16 proller joined #minetest-dev
11:30 nerzhul is this normal that the voxel area doesn't reorder points ?
11:43 Fixer joined #minetest-dev
11:47 nerzhul Krock if you get time, i pushed a PR to add unittests on VoxelArea object: https://github.com/minetest/minetest/pull/7121
12:00 Wuzzy joined #minetest-dev
12:04 Megaf Are we good with web#121 now?
12:04 ShadowBot https://github.com/minetest/m​inetest.github.io/issues/121 -- Channel description improvements 2. by Megaf
12:08 Megaf sfan5: What's the issue on having a webchat for -hub?
12:08 sfan5 people can't talk
12:08 Megaf all other channels have
12:08 sfan5 well I guess that applies to -dev too
12:08 Megaf and some of mt members think it can have a link and some of mt members think it can't
12:09 sfan5 I haven't seen anyone advocating for a link which is why I asked
12:09 Megaf ¯\_(ツ)_/¯
12:10 Megaf Shara: rubenwardy: ping
12:10 Megaf webchat is muted on -dev as well
12:10 Megaf yet there's a link
12:11 Megaf sfan5: ^
12:12 sfan5 if you look two lines above you can see I already noticed
12:14 Megaf I'd say we remove webchat links for every channel
12:16 sfan5 you want people to figure out this irc thing on their own? nobody will ever join
12:16 sfan5 it's fine with webchat links for every channel
12:21 Megaf I first added webchat link for eveychannel because people can always try to join and try to send a private message to an op
12:24 MegafTestingWebC joined #minetest-dev
12:25 MegafTestingWebC joined #minetest-dev
12:25 MegafTestingWebC joined #minetest-dev
12:25 Megaf Ok, even authed to the services it's not possible to message here
12:25 MegafTestingWebC left #minetest-dev
12:31 Megaf sfan5: https://github.com/minetest/minetest.gi​thub.io/pull/121#issuecomment-372026266
13:18 Beton joined #minetest-dev
13:29 BakerPrime joined #minetest-dev
14:44 fwhcat joined #minetest-dev
15:37 Gundul joined #minetest-dev
15:37 Lone-Star joined #minetest-dev
15:40 CBugDCoder joined #minetest-dev
16:02 AntumD joined #minetest-dev
16:05 Antum joined #minetest-dev
16:39 Krock joined #minetest-dev
16:58 Krock nore, rubenwardy, sfan5, Shara, sofar: Would you have some time around 18:00 or 19:00 UTC ?
16:59 sfan5 fine with me
17:00 Krock I collected a few ideas/points to discuss in the meeting: https://github.com/orgs/minete​st/teams/engine/discussions/6 Feel free to extend it
17:01 Krock ^ addressed to everybody
17:05 sfan5 oh right UTC, 18:00 would be better in that case
17:22 AntumDeluge joined #minetest-dev
17:31 paramat joined #minetest-dev
17:31 paramat i can do meeting at 18 or 19 UTC
17:33 Krock that's great. for now it looks like 18:00 fits better
17:33 Gael-de-Sailly joined #minetest-dev
17:34 Shara I will be mostly here from now until I fall asleep, whatever time that happens to be. :)
17:42 paramat added some points to discussion doc
18:06 Krock Four devs - not too many, should we start anyway?
18:07 sfan5 better than none
18:07 Krock haha of course :)
18:09 paramat nore rubenwardy sofar ShadowNinja
18:10 YuGiOhJCJ joined #minetest-dev
18:10 Krock Okay, let's get started with those who are currently here. First topic: https://github.com/orgs/minete​st/teams/engine/discussions/5  Doxygen as a new PR rule. Opinions?
18:11 paramat i need to research the advantages because i don't see many at the moment, i'm sure there are more than i realise though
18:12 Krock It requires more time to document the functions but that's exactly what we need: documentation.
18:12 sfan5 i doubt a doxygen documentation is that useful
18:12 Krock for larger functions it can indeed provide helpful information about what the function does
18:13 Shara Documentation is definitely important, but whether that's the exact approach needed, I'm not sure.
18:13 sfan5 well, I consider it common sense to document large c9ode sections with overviews
18:13 paramat i agree with better comments, just not sure about doxygen or insisting on it, yet
18:13 sfan5 but I don't think it needs to be 1) in Doxygen format or 2) a rule
18:14 Krock indeed. in-line documentation does its job too
18:15 Shara We could just do with... more of it
18:15 Krock so http://doxy.minetest.net/ will just stay there as a product of trying new things?
18:15 paramat i'm not keen on anything that requires a certain program to view, clarity when viewing comments in code seems highest priority
18:15 sfan5 probably
18:16 paramat doxy format is a litte messy looking
18:17 davisonio joined #minetest-dev
18:17 Krock so, majority is opposed to using doxygen. rather do proper in-line commenting?
18:17 paramat i'm opposed for now
18:18 sfan5 yes
18:18 Shara Same
18:20 Krock Next idea: Shall we open new project to categorize long, medium and long-term issues?
18:20 sfan5 categorize in which ways?
18:21 Krock rubenwardy doesn't define that exactly in his post but I assume it's defined by the difficulty of the implementation to solve the issue
18:21 Krock i.e. PR size and amount of effort, I guess
18:22 Krock or even more likely how important they are, like we already have the label Low priority, Blocker and High priority
18:22 sfan5 sounds like more like what should be looked at right now, later or far in the future
18:23 Krock indeed
18:23 sfan5 which is pretty pointless to me, since usually devs work on what they want to work on
18:23 Krock I feel the same. Labels already exist for important issues
18:23 Shara I took that to mean more about roadmaps
18:24 Shara But I'm not really sure
18:25 paramat difficult to discuss without rubenwardy, discuss another time maybe
18:25 Krock Projects would however show them more easily and allow to move them to "WIP", "Needs review" and "Complete"
18:28 Krock Noted as "needs further discussion". Let's go on with the PR discussions: #6660 is waiting for quite a while, being rather trivial
18:28 ShadowBot https://github.com/minetest/minetest/issues/6660 -- Remove caching from getTextureDirs by numberZero
18:30 sfan5 my stance: " seems fine to me +1 "
18:30 Krock rubenwardy took back his review - but for now I think it's best to merge it
18:31 paramat i'm neutral
18:31 Krock the settings speed is another topic and can be solved in a later PR
18:32 Krock will merge as-is
18:33 Krock for the reason that it currently only worsens the player experience with the cached value
18:34 Krock #6951 CSM, security related
18:34 ShadowBot https://github.com/minetest/minetest/issues/6951 -- [CSM] Don't create the script environment if CSM is disabled by red-001
18:34 paramat i assume 6660 doesn't affect in-game performance?
18:35 Krock paramat, it only does on-load
18:35 paramat fine
18:35 Krock but even then it's a very small delay
18:36 Krock Regarding 6851: it moves the death formspec back to C++ to not load the environment without CSM enabled
18:37 Krock Shara, may I ask what exactly was the issue with it? "For example i believe at least one of Shara's servers are closed to the public until these go in." (paramat)
18:38 Shara I assume that comment was not about 6851
18:38 paramat ? that's another issue
18:38 Krock is it the aspect of being able to change builtin to include cheats?
18:38 Krock paramat, what issue did you mean then?
18:38 Shara My server is closed to new users until I can safely limit/prevent new mapgen
18:38 paramat for Shara's servers i'm referring to world edge changes that are todo for backport
18:39 Shara I think the needed PRs were merged now, but not yet in backport branch
18:39 paramat nothing to do with CSM builtin
18:39 Krock ah, so these are two separate issues. My bad for connecting them
18:39 Gael-de-Sailly joined #minetest-dev
18:39 Shara No problem :)
18:40 paramat 6951 seems a good idea for simplicity and avoiding the alternative, rubenwardy is strongly opposed to what happened to the death formspec
18:41 paramat considering the controversy of CSM i think we need to roll it back a little as 6951 does
18:41 Shara A lot of people wanted that changed.
18:43 paramat it seems a bad idea to start using CSM for builtin stuff, and is another method of cheating that needs securing
18:43 sfan5 well 6951 certainly can't be merged as-is
18:44 paramat red-3-2-2019
18:45 sfan5 the pr does not remove the need for hashing builtin btw
18:45 paramat oh ok :)
18:46 sfan5 if a servers wants to disable client-provided mods, but use its own csm mods you need to load builtin
18:46 sfan5 (I assume this is planned)
18:46 paramat rubenwardy's comment https://github.com/minetest/minetest​/issues/6736#issuecomment-358479221
18:46 Krock #5393 is about that exact issue
18:47 ShadowBot https://github.com/minetest/minetest/issues/5393 -- Server-provided client-side scripting
18:47 sfan5 otherwise the server has two options 1) don't allow use of CSM at all or 2) make us of CSM but allow the client to inject their own code
18:47 sfan5 +e
18:50 Krock so builtin should be secured on a way that at least minetest has to be recompiled to make builtin modification possible
18:50 Krock that's possibly the most we can do for this, as the client's data can not be trusted
18:51 nerzhul can someone take a look at https://github.com/minetest/minetest/pull/7121 ? :)
18:51 nerzhul #7121
18:51 ShadowBot https://github.com/minetest/minetest/issues/7121 -- Add Voxelarea unittests by nerzhul
18:52 Krock ^ our next discussion topic
18:52 Tmanyo joined #minetest-dev
18:53 paramat i think that should wait until after the meeting
18:53 paramat but will look
18:53 nerzhul oh it's the dev meeting :)
18:54 sfan5 concept sounds good to me
18:54 TC01 joined #minetest-dev
18:54 sfan5 haven't found any code problems from a quick look
18:55 Krock left two comments, otherwise looks ok
18:56 paramat lol
18:58 Krock There are still a few points on the list.. so let's proceed with #6900 - should we revert the previous feature PR to call after_place_node or merge this one?
18:58 ShadowBot https://github.com/minetest/minetest/issues/6900 -- core.rotate_node: Do not trigger after_place_node by SmallJoker
18:58 sfan5 this one IMO
18:58 Krock Maybe someone has got a better idea on how to work around this callback loop problem
19:01 Krock Namely https://github.com/minetest-mods/nixie_tubes/ seems to be broken due this issue, so I will test again the PR (last test was too long ago) and merge it if the loop is fixed. Is that okay?
19:03 paramat no opinion
19:04 Krock Fine. Does anyone have an android build env set up to test #6796 ?
19:04 ShadowBot https://github.com/minetest/minetest/issues/6796 -- Android build fixes for c++11 by stujones11
19:04 nerzhul Krock: i got it
19:05 paramat hm i'm still waiting for lord fingle to get involved in android dev as he said he might
19:05 sfan5 I don't but I could probably try a build of that tomorrow
19:05 Krock paramat, at the end we could wait forever
19:05 paramat yes, can't wait for him
19:05 Krock sfan5, that would be great if you could do that. Thanks.
19:05 paramat obviously
19:06 paramat i'll PM him on the forum again concerning controls
19:07 rubenwardy hi
19:07 paramat woo
19:07 nerzhul Krock: maybe i can test it this evening too
19:07 Krock hello rubenwardy
19:07 nerzhul and at least port android.mk fixes at a point
19:08 Krock nerzhul, no, you shall be at your son's birthday party ;)
19:08 nerzhul lol it was this afternoon :)
19:08 Krock nice. testing is welcome
19:09 Krock Automatically generated settings headers: #6728
19:09 ShadowBot https://github.com/minetest/minetest/issues/6728 -- Autogenerate settings header by numberZero
19:09 nerzhul i hope at a point we should switch to cmake based build, but i don't know if it's possible easily, using the cmake + ndk standards google added for those use case
19:09 sfan5 haven't reviewed, can't comment
19:10 Krock but concept-wise? Is generating the headers generally good?
19:11 Krock the description could also do well with a more detailed explication
19:13 sfan5 it has performance advantages for sure, but I'm not fully convinced yet
19:14 nerzhul if we use primitives types for our core settings the performance will be increased and we can remove all settings caches
19:18 nerzhul the idea is good, but it's not easy. I think at a point we should have a blazing fast settings object containing the good primitives
19:22 Krock Okay. The 0.4.17 release - what other fixes for the backport are we missing?
19:22 paramat rubenwardy see logs, any comments on discussion?
19:22 paramat world edge stuff
19:23 Krock If there are none, we need a volunteer to get the backport PR up-to-date so it can finally be released
19:24 Krock paramat, the "world edge stuff" is basically the 4 PRs here? https://github.com/minetest/minetes​t/pull/6746#issuecomment-368584307
19:24 paramat yes
19:26 Krock they were tested and approved. Sounds good...
19:26 Wayward_One Krock, I've already tested #6796 and it works
19:26 ShadowBot https://github.com/minetest/minetest/issues/6796 -- Android build fixes for c++11 by stujones11
19:28 sfan5 (I'll be gone now)
19:28 Krock Wayward_One, that's great to hear.
19:29 rubenwardy doxygen should be used if a function or class needs a comment. functions only need comments if their behaviour is not obvious, and non-trivial classes should always have comments to define scope
19:29 rubenwardy paramat: false
19:29 rubenwardy paramat i'm not keen on anything that requires a certain program to view, clarity when viewing comments in code seems highest priority
19:30 rubenwardy doxygen comments are literally just an extra slash
19:31 paramat yes not too bad, but @param @return too
19:31 rubenwardy you don't need those
19:31 paramat i don't mind a 3rd slash
19:31 rubenwardy and they're hardly less clear
19:31 paramat well i mean, clutter
19:32 rubenwardy the project things was supposed to be similar to   dev.minetest.net/TODO    but less pipe dreams
19:35 rubenwardy also doxygen doesn't require an extra program
19:35 Krock rubenwardy, it's difficult to tell when a function needs documentation and when it's superfluous.
19:35 rubenwardy just a web browser
19:36 Krock a web browser to read the data that was parsed by another program for doxy.minetest.net?
19:38 rubenwardy also, you can just view the source code
19:40 paramat a web browser is a separate program
19:40 rubenwardy do you really not have a web browser open to code?
19:40 rubenwardy also, not necessarily. You could be using an OS like emacs
19:40 Krock nobody gets around StackOverflow, paramat.
19:40 paramat most do, but i'm talking about the principle of it
19:41 rubenwardy in any case, there's the source code
19:41 rubenwardy I'd argue documenting every single method is worse than not using doxygen
19:41 rubenwardy +that
19:41 rubenwardy // Gets a texture
19:41 rubenwardy /// Gets a texture
19:41 rubenwardy ITexture *getTexture();
19:42 paramat viewing the comments in source code is obvious and irrelevant to the argument
19:42 rubenwardy how is it irrelevant?
19:43 paramat i'm not actually strongly opposed to doxygen comments, just not keen
19:45 rubenwardy anyway, never
19:45 rubenwardy anyway, nevermind
19:45 rubenwardy I feel like a rule is too restrictive anyway
19:45 paramat really, i'm almost neutral. just oppposed to insisting on specifically doxy right now, as are the other devs
19:47 nerzhul personnally i will try to implement doxygen comments on my future PR modified functions
19:47 nerzhul if some go to that way we can start to cover the code
19:47 paramat and cautious because md formatting for lua_api.txt has made a visual mess of the original and i'm not sure it was a good idea
19:47 nerzhul but yes, comments inside functions are always useful
19:48 paramat if doxygen only extracts comments i can't see much advantage, comments are best viewed in context
19:48 rubenwardy not really
19:48 rubenwardy doxygen has search
19:48 nerzhul not only comments, you have also relational data
19:48 nerzhul it's a standard in C++ industry
19:48 rubenwardy yes
19:49 paramat yes, there are probably advantages i am not aware of, need to research :)
19:49 rubenwardy it might as well be part of the C++ language for how ubiquitous it is
19:49 rubenwardy my favourite part is this: http://doxy.minetest.net/cl​assServer__coll__graph.svg
19:49 rubenwardy note how there's links
19:50 paramat ok how can that create that if we only add a 3rd slash?
19:50 nerzhul it's nice, but also shows that server class has too many direct relations
19:50 nerzhul paramat all IDE supports doxygen correctly, and for those comments there is only one norme
19:50 nerzhul norm*
19:51 rubenwardy it's good software practice, as long as you don't go overboard like my example
19:51 paramat it's actually possible i will support doxy comments after some research
19:52 paramat so, nuff said =)
19:53 paramat so there are some meeting topics still to do
19:54 paramat when to release 0.4.17 and 0.5.0
19:54 paramat how far do we allow server-sent CSM to delay 0.5.0?
19:55 paramat considering it will probably be 'compiled out' of 0.5.0 anyway
19:59 Krock basically only 0.4.17 is a thing. we can wait with 0.5.0 a bit
19:59 Krock no need to rush for two releases at once
20:02 Krock of course, we can delay this topic to next week 8)
20:04 paramat btw is it ok to encourage people to use backports and consider it somewhat 'stable'?
20:05 Shara Since it's meant to be bugfixes, I would hope so.
20:05 paramat i mean, if it's fairly stable then we may not have much need to make a 0.4.17 release soon
20:06 Krock rolling stable version :D
20:06 paramat yeah
20:06 Shara A release would also have to contain any features, and not just be bugfixes
20:06 paramat well, 0.4.17 will only be bugfixes
20:06 Krock no, releases are only new versions, which does not imply any changes
20:07 Shara Would be rather pointless with no changes
20:09 paramat bugfixes. also 0.4.17 is intended to be 'the final 0.4 stable version'
20:10 paramat ie wrap up 0.4 properly with a proper release process and feature freeze testing
20:11 paramat which it seems will happen at release of 0.5.0, or maybe just after
20:15 paramat anyway, meeting seems over
20:15 Krock yes.. thanks for your efforts :)
20:15 Krock efforts, time, opinion, whatever.
20:20 p_gimeno aww, I hoped #6898 would be discussed
20:20 ShadowBot https://github.com/minetest/minetest/issues/6898 -- Allow distinguishing mods by modpack by pgimeno
20:21 p_gimeno it's possible to make a backpatchable, but ugly version
20:22 p_gimeno but having worlds breaking just for adding stuff to the mods folder sounds unacceptable to me
20:27 paramat many devs may not have time for MTG, which is fine, but here are 2 PRs for consideration game#2062 game#2068
20:27 ShadowBot https://github.com/minetest​/minetest_game/issues/2062 -- Icesheet biome: Add 'default:cave_ice' node to enable caves in land ice by paramat
20:27 ShadowBot https://github.com/minetest​/minetest_game/issues/2068 -- Pine trees: Add small pine tree and mix into forests by paramat
20:29 paramat #7117 is simple and ready
20:30 ShadowBot https://github.com/minetest/minetest/issues/7117 -- Biome API: Add 'get_biome_name(biome_id)' API by paramat
20:30 Krock p_gimeno, yes but the meeting already took longer than expected with the other topics. That PR and 0.4.17 discussion have to be done next week, as too long meetings are quite tiring
20:33 Shara paramat: sure, will check pine trees
20:56 p_gimeno Krock: thanks
21:12 fwhcat joined #minetest-dev
22:13 DI3HARD139 joined #minetest-dev
22:43 Fixer_ joined #minetest-dev
22:52 Gael-de-Sailly joined #minetest-dev
23:28 Fixer joined #minetest-dev
23:35 paramat thanks
23:36 paramat will merge those 2 later then
23:37 Shara :)
23:38 paramat posted suggestion for new needles texture game#2075
23:38 ShadowBot https://github.com/minetest​/minetest_game/issues/2075 -- Thicker pine needles texture
23:40 Shara I find the opaque version shown there ugly... but then I always find opaque leaves ugly
23:42 paramat yes doesn't look great
23:43 paramat maybe i could try using bark colour for the opaque background colour
23:43 Shara That could make more sense
23:46 paramat maybe a dark grey-brown, black is a bit severe

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