Minetest logo

IRC log for #minetest-dev, 2018-05-12

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

All times shown according to UTC.

Time Nick Message
00:17 proller__ joined #minetest-dev
00:49 vexyl joined #minetest-dev
01:08 AntumD joined #minetest-dev
02:10 AndroBuilder_ joined #minetest-dev
02:59 Gael-de-Sailly joined #minetest-dev
03:03 CBugDCoder joined #minetest-dev
03:07 YuGiOhJCJ joined #minetest-dev
03:27 proller joined #minetest-dev
04:18 Sokomine joined #minetest-dev
04:29 ssieb joined #minetest-dev
05:23 proller joined #minetest-dev
06:02 ANAND The fact that setting_get()  / _set() has been deprecated in favour of setting:get() / :set() needs to be updated in the dev wiki
06:03 ANAND There's literally no mention of the new methods in the whole wiki
06:17 ANAND Actually, I've got the method names wrong myself!
06:43 fwhcat joined #minetest-dev
09:39 Krock joined #minetest-dev
09:54 Krock Fix for not run callbacks on player leave: https://pastebin.com/raw/BkbkRiR7
09:54 Krock brb. Will announce the push later
10:16 sfan5 ANAND: generally the reference you'll want to use is lua_api.txt
10:42 ANAND sfan5, ok
10:44 proller joined #minetest-dev
10:50 Fixer joined #minetest-dev
11:33 proller joined #minetest-dev
11:34 ensonic joined #minetest-dev
11:41 proller joined #minetest-dev
11:42 fwhcat joined #minetest-dev
11:51 proller joined #minetest-dev
11:55 proller joined #minetest-dev
11:58 proller joined #minetest-dev
11:58 Wuzzy joined #minetest-dev
12:09 proller joined #minetest-dev
12:13 Krock Will push https://pastebin.com/raw/BkbkRiR7 in 15 minutes
12:13 Krock and afterwards merging #7319
12:13 ShadowBot https://github.com/minetest/minetest/issues/7319 -- Camera: Improve subpixel movement by SmallJoker
12:28 proller joined #minetest-dev
12:30 Krock pushing
12:53 nerzhul thanks sfan5, your comment is just real !
13:27 fwhcat joined #minetest-dev
13:40 wowaname[m] joined #minetest-dev
13:42 Beton joined #minetest-dev
13:50 ensonic joined #minetest-dev
13:54 warr[m] joined #minetest-dev
14:31 Anchakor joined #minetest-dev
14:37 paramat joined #minetest-dev
14:42 paramat when's a good time for talking about 0.4.17? problem is we tend to lose nerzhul at the usual meeting time, so how about in 20mins or 1h20mins? i'm here for the rest of today. Krock nerzhul nore rubenwardy sfan5 Shara sofar
14:42 sfan5 I don't mind doing it earlier
14:42 rubenwardy I'm free
14:43 Shara Here
14:51 Krock Here
14:51 Krock well, this will only make sense when nerzhul is online and has got time too :)
14:54 Krock however, I won't be here in 1h20min but it could be held then surely too
14:57 paramat our usual meeting time seems not good somehow, people tend to be quiet, maybe we're more awake in the daytime :)
14:58 paramat 1 more commit to backport in #6746 please can anyone do that? sfan5 or anyone?
14:59 ShadowBot https://github.com/minetest/minetest/issues/6746 -- Backport 0.4 by sfan5
14:59 sfan5 I can do that later
14:59 paramat MTG backports is ready
14:59 paramat ok thanks
15:00 paramat i think we should freeze today if possible
15:01 Krock sounds good
15:02 paramat hopefully nerz will appear in a while
15:05 Shara I may or may not be here in an hour from now, but it's not too important.
15:06 Shara I'm just happy if release can be soon :)
15:15 BakerPrime joined #minetest-dev
15:30 nerzhul i'm there but i will do some things outside
15:31 nerzhul go ahead and release 0.4.17 please
15:31 nerzhul i will provide the android builds after that
15:31 nerzhul we cannot update translations for this release as they are based on master
15:42 paramat ok
15:43 Krock i.e. release today?
15:45 rubenwardy I'm having crashes on my server
15:48 rubenwardy ERROR[Main]: ServerError: AsyncErr: ServerThread::run Lua: Runtime error from mod 'default' in callback luaentity_Step(): ...test/minetest_stable/bin/../builtin/game/item_entity.lua:184: attempt to index local 'node' (a nil value)
15:49 Krock backport branch?
15:49 rubenwardy yeah
15:50 Krock exists since https://github.com/minetest/minetest/commit/a59df5ef5
15:51 Krock the area is not loaded yet.. and that builtin item code looks like it was never really touched by any cleanup commit
15:52 Krock local nn = node and node.name or "ignore"   should solve the issue for you
15:52 sfan5 isn't the point of get_node_or_nil to return nil when in ignore?
15:53 Krock no, or_nil only returns nil when the node isn't loaded yet
15:53 sfan5 oh
15:53 paramat yeah it depends on unloaded/loaded, not 'ignore'
15:54 Krock this commit worked fine in 0.5.x because there was a builtin item rewrite
15:55 Krock so we should fallback either to "ignore", remove the item removal commit or backport the cleanups
15:56 Krock brb
15:58 paramat oops my error then
15:58 sfan5 no that's my mistake
15:58 sfan5 i backported these changes, I need to ensure they work with the older variant of the code
15:59 sfan5 the question is what is supposed to happen to items in unloaded area?
16:04 paramat this commit has to be in backports as it's part of a bunch of commits that allow world outside 'mapgen limit' to be playable if generated, this is necessary for server owners as a defence against attacks and to limit world size
16:04 Shara It's basically the thing I've been waiting for 0.4.17 for...
16:06 paramat in unloaded an item should have it's velocity and acceleration set to 0 as before, then i guess set 'node' to 'ignore'
16:06 proller joined #minetest-dev
16:07 paramat ugh i wish posts were not deleted in backport threads, that explained why those commits are needed
16:08 sfan5 setting it to 'ignore' would delete the item though
16:08 sfan5 well it wasn't me who deleted them
16:10 paramat oh i see, that code is ok in 0.5. i wondered why i had been sloppy :)
16:10 proller joined #minetest-dev
16:21 ensonic joined #minetest-dev
16:32 paramat i mean, set node to ignore at line 184, as Krock suggested
16:33 paramat local nn = node and node.name or "ignore"
16:34 paramat bu then, it might be ok to set vel/acc to zero and 'return'
16:36 paramat yes the code previously 'return'ed, see the commit
16:37 paramat and see 0.4.16 stable https://github.com/minetest/minetest/blob/80dc961d24e1964e25d57039ddb2ba639f9f4d22/builtin/game/item_entity.lua#L178
16:41 paramat if node == nil then (vel/acc = 0 etc. return) elseif node.name == "ignore" then (remove object etc. return)
16:44 paramat oh, the posts weren't deleted, github just hid them, https://github.com/minetest/minetest/pull/6746#issuecomment-368584307
16:44 paramat github is so slow at page loading ..
17:32 AntumDeluge joined #minetest-dev
17:48 nerzhul Krock if i remember there were some C++03 compilation issues on 0.4.17 backport branch. Is this fixed ?
17:53 Krock not yet. It was an issue with the Android build. The error log can be found in the logs to write a patch for it
17:55 sfan5 nerzhul: do we want to fix this?
17:56 sfan5 what i mean is, it would be lots of work to rewrite the code into c++03
17:56 sfan5 the easier way would be to just require c++11 and add -std=c++11 to android's makefile
18:00 proller joined #minetest-dev
18:23 reductum joined #minetest-dev
18:24 proller joined #minetest-dev
18:43 paramat is it important seeing how bad android is currently? if the new controls are not in backports then i can't see good reason to bother with android for 0.4.17
18:47 sfan5 would be nice to have it compile correctly
18:47 sfan5 especially when fixinig it is fairly easy
18:48 paramat oh ok
18:51 proller joined #minetest-dev
18:59 ensonic joined #minetest-dev
19:02 Krock Was delayed by a bit. Are still people here to discuss the meeting points?
19:03 sfan5 i'm still here
19:04 Krock nerzhul, nore, paramat, rubenwardy, Shara, sofar - would you have some time to discuss the points listed here? https://github.com/orgs/minetest/teams/engine/discussions/10
19:06 Krock sfan5, so we should push a single commit to tell the Android build setup to use C++11?
19:07 sfan5 yes
19:07 Krock paramat, we MUST bother with Android. at least 70% of the daily players use it
19:07 sfan5 paramat's (or our?) plan with android is to make it better for 0.5 and then move players to it
19:08 sfan5 so don't /need/ to bother with 0.4.17 android
19:08 Krock so if the players look for updates, they'll find them at the app store from another developer - expecting the issues are gone with the new 0.4.17 version
19:10 sfan5 there is nothing players should "expect" from third-party apps
19:10 Krock if there's a simple solution to fix compiling, then we should include that into the backports
19:10 sfan5 and if they don't provide something it's not our fault
19:10 Krock right
19:10 sfan5 (i'm not arguing against including a compile fix, but e.g. backporting controls would be wasted time)
19:10 paramat yes here. and yes i'm neutral about android, was jsut a random thought :)
19:10 paramat *just
19:11 Krock yes, the controls definitely belong into 0.5.x
19:11 Shara controls change is more a feature than fix, even if addressing a long term problem
19:11 paramat in fact, sorry for my comment, i feel we should bother with android for 0.4.17
19:14 nerzhul i don't see why we should push the android C++11 commit it's not the contract for 0.4.17 series
19:14 nerzhul and for other users
19:14 nerzhul porting C++11 to C++03 is not difficult on the problematic parts
19:15 Krock in this case we'll end up initializing the variables noted in https://pastebin.com/raw/j4mjniaq in the constructor
19:15 lumberJ joined #minetest-dev
19:16 Krock does anyone have a working android build system to make and test the patch?
19:17 Krock the errors seem quite obvious. I could do a patch but not test it until I've set up the android build on this system
19:26 proller joined #minetest-dev
19:27 Wayward_One I do
19:28 Fixer i feel sorry for android users in some sense
19:31 Krock Wayward_One, got a patch for you to test
19:31 Wayward_One Okay
19:32 Krock *diff. https://pastebin.com/raw/UA5a8Y3m
19:32 Krock my compiler is still running. It's possible that I've made a typo somewhere
19:34 Krock paramat, would you mind testing the suggested change for item_entity.lua so we can ensure its' working alright?
19:35 paramat hmm yes ok
19:35 Krock or is rubenwardy using a patch on his server already to confirm it?
19:37 rubenwardy not yet
19:37 rubenwardy been doing coursework
19:37 rubenwardy which is also the Minetest contentdb
19:39 Krock Wayward_One, the diff works here. Any results yet?
19:39 Krock sorry -forgot to ask. Would you have some time to test? :)
19:46 sfan5 paramat: "And this proposal is unlikely to happen or be accepted." how do you come to this conclusion?
19:46 Krock personal option formulated as general statement for the entire dev team
19:46 sfan5 I'm seeing one coredev clearly against and one tending to "against", rest neutral or no opinion
19:47 sfan5 Krock: yes, I often have this impression when I read paramat's comments
19:49 Krock I'm not a fan of this style either but unlike other contributors I interpret them as personal opinion
19:53 paramat ? where's that quote from
19:53 paramat happy to retract if i was wrong
19:55 paramat and no, unless i make a mistake i only write that when i know many are against something
19:55 paramat can you link to where i wrote that?
19:55 Wayward_One Krock: yes, i was just testing it :) ran into a small problem though, most likely on my end, so i'll have to troubleshoot when i get home
19:56 Krock thanks a lot for testing. I'd be glad if you could provide more detailed information about the problem as soon you've got time :)
19:57 sfan5 paramat: #7303 your post 17m ago
19:57 ShadowBot https://github.com/minetest/minetest/issues/7303 -- WIP: Begin discord rich presence by Dumbeldor
19:57 paramat btw Krock are you -1 for discord? your comment seems possibly to be disapproval
19:57 Krock I didn't state any direct disapproval and only threw in some ideas about it
19:58 Krock if it were a disapproval I'd :-1: or state it as "I disapprove"
19:58 Krock I'm not against it as long the implementation is KISS
19:59 paramat ah found it
19:59 Krock whereas it can't get much simpler than how it currently is if I look at the Discord API correctly
20:00 paramat hmm well i wrote that because it seems unlikely, whenever have we added a feature agreeing that core dev won't work on it? we're obliged to work on everything, otherwise the feature would be poorly maintained
20:02 sfan5 see my proposal, it's pretty handy that this feature is optional
20:02 sfan5 so if you don't enable it, no related code will even be compiled
20:04 paramat ok. but is it acceptable to add code that we are stating we (core devs) will never maintain? seems unacceptable
20:05 paramat and of course, if a feature is that low priority it's probably not worth adding
20:08 paramat ok updated my post to add (in my opinon...)
20:12 Krock paramat, a lot stuff is unmaintained by us *looks at the rendering part* and we actually rely largely on PRs which solve these issues
20:13 Krock so I don't see a problem with adding this Discord PR, which can be opt-out anyway easily
20:14 Krock .. or even remove it from the code if it stays broken for months
20:16 paramat we still spend time on rendering stuff
20:17 paramat there's more to the discord PR than dev time to maintain, there are other reasons to avoid it, such as it being completely useless code bloat
20:24 sfan5 that's highly subjective
20:25 paramat yes
20:25 Krock these few kilobytes which aren't even compiled without the CMake setting set to "true" does not hurt anyone. We've also got redis and postgresql support - "only server will need it". Same for the discord PR: "only discord users will need it"
20:25 Krock but still we've got redis and postgresql support
20:26 paramat 'does not hurt anyone' is not enough reason to add a feature, there has to be some benefit :)
20:26 Krock neither is the "code bloat" coming from that particular file a major issue
20:26 paramat and it will hurt devs by taking up time and maintenance ;)
20:27 paramat the feature doesn't justify a single line in my opinion, it's a stupid social network gimmick
20:29 Krock "time and maintenance" in reality this would rather look like the following situation: 1) a new issue regarding feature XY not working 2) time passes 3) someone writes the PR for it to get it working again
20:29 Krock whereas "someone" is usually a person who is affected by that problem
20:29 paramat well, it seems not worth it to me
20:30 Krock okay
20:31 paramat we can see that most core devs are really 'meh' about this and see it as near-useless, in these situations the right thing to do is -1, for simplicity and prioritisation. devs are being too soft :]
20:53 paramat btw, i've marked some issues with no core dev interest for 2 years as 'possible close', please can core devs go through these and add their opinions?
20:53 paramat i'm sure many were just missed by core devs so i bumped them
21:00 paramat these will be closed eventually if there is no core dev interest
21:39 YuGiOhJCJ joined #minetest-dev
22:30 Taoki joined #minetest-dev
22:41 ircSparky joined #minetest-dev
23:07 paramat joined #minetest-dev
23:09 paramat rubenwardy do you have a way to reproduce that item_entity crash? so i can test my suggested code?
23:09 rubenwardy set the limit really low
23:09 rubenwardy drop an entitiy
23:10 rubenwardy that's how it happens on CTF
23:10 rubenwardy maybe
23:10 paramat 'limit'?
23:10 rubenwardy mapgen limit
23:11 rubenwardy happens every few hours on my server
23:11 paramat ok, hmm
23:12 paramat interesting, if you drop an item into the ignore at world edge it should be deleted in the ignore before getting to 'unloaded'
23:12 paramat i might try dropping an item at high altitude above an unloaded mapchunk
23:13 paramat anyway will test that code
23:14 paramat will do what you suggest too
23:32 Tmanyo joined #minetest-dev

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