Minetest logo

IRC log for #minetest-dev, 2017-11-19

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

All times shown according to UTC.

Time Nick Message
00:12 paramat will also merge #6633
00:12 ShadowBot https://github.com/minetest/minetest/issues/6633 -- Minetest dependencies for Linux in table by Reedych
00:18 paramat merging those 3 now then
00:31 paramat merged
00:31 paramat it would be nice to get below 120 PRs
01:07 paramat joined #minetest-dev
01:10 basicer #6570 seems like a good canidate for getting back to 120 :)
01:10 ShadowBot https://github.com/minetest/minetest/issues/6570 -- Allow enter to select items from combobox's list. by basicer
01:17 paramat =D
02:22 Tmanyo joined #minetest-dev
02:38 Megaf joined #minetest-dev
02:51 ThomasMonroe joined #minetest-dev
03:57 torgdor joined #minetest-dev
04:24 coyote joined #minetest-dev
06:20 coyote joined #minetest-dev
06:59 Fritigern joined #minetest-dev
07:33 Fritigern joined #minetest-dev
08:17 Fritigern joined #minetest-dev
08:22 ssieb joined #minetest-dev
08:55 Hunterz joined #minetest-dev
09:09 nerzhul joined #minetest-dev
09:33 Darcidride joined #minetest-dev
09:39 YuGiOhJCJ joined #minetest-dev
09:46 jcalve joined #minetest-dev
10:00 coyote1 joined #minetest-dev
10:16 davisonio Get this problem when someone joins the world on 0.5.0-dev: ServerError: AsyncErr: ServerThread::run Lua: OOM error from mod '*builtin*' in callback on_joinplayer(): not enough memory
10:17 davisonio even memory problems with builtin now :o
10:23 coyote1 $ free -h ?
10:23 coyote1 yes, it can be bug :P
10:27 sfan5 just because the OOM occurred inside builtin doesn't need to mean that builtin was the culprit
10:28 davisonio got enough free memory and some swap too
10:28 davisonio true
10:28 davisonio i might see if compiling without luajit does something
10:31 sfan5 when you have OOM issues it's usually luajit being stupid
10:50 Krock joined #minetest-dev
10:55 Kright joined #minetest-dev
11:13 nerzhul davisonio, Lua is limited to ~1GB per stack it can be a very bad mod (1GB lua state is crazy)
11:19 VargaD joined #minetest-dev
11:22 YuGiOhJCJ joined #minetest-dev
11:23 Kright Hello. I have an idea of new feature for graphics engine. In RL Earth is round, so distant objects will be slightly below horizon. To imitate this effect, one can approximate sphere with parabola. We only need to modify vertex shader. y -= square_distance_from_vertex_to_camera = dx*dx + dz*dz, where dx and dz is distance from camera to vertex in world coordinates (modified y in world coordinates too). So, vertex shader become more complex - instead of mul
11:34 proller__ joined #minetest-dev
11:35 sfan5 i'm not sure if that effect is so desirable
11:35 sfan5 open an issue for it so it won't be forgotten
11:37 Kright ok)
11:37 davisonio nerzhul: ok, I'll see what I can do probably luajit is better
11:40 Krock joined #minetest-dev
11:46 sfan5 nerzhul: i thought it was luajit who had the 1GB / 2GB memory problem (?)
11:49 nerzhul sfan5, it's possible as it's a JIT there is a max buffer pool size  yes
11:49 sfan5 i don't think it's the pool size, luajit uses 32-bit addresses with a few bits to signal something IIRC
11:50 sfan5 which naturally limits the amount of memory to 2 or 4 GB
11:50 sfan5 which is what gc64 is supposed to fix
12:28 Fixer joined #minetest-dev
12:53 Kright joined #minetest-dev
13:08 fwhcat joined #minetest-dev
13:18 Fixer seems like Texture Pack got changed to Texture Packs and I can't translate it for some reason
13:19 Shara If you mean the tab name, it was changed from "Texturepacks" since it's not a single word
13:34 coyote joined #minetest-dev
14:25 coyote joined #minetest-dev
15:19 rubenwardy I'd like to suggest switching to SemVar after 0.5.0
15:19 Krock straight to version 6.0?
15:20 rubenwardy and start releasing patch releases between full releases
15:24 Fixer Shara: it does but it was not in pot file iirc
15:25 Shara pot file?
15:25 rubenwardy I think that's the .po translation file
15:25 Fixer yeah
15:25 Krock translator's strings file, which needs manual updating
15:25 rubenwardy it needs to be generated from a script
15:25 rubenwardy ^
15:26 Fixer yes
15:26 sfan5 would anyone mind if i go ahead and create the suggested backports branch for minetest_game?
15:26 Fixer please update it
15:26 Krock go ahead
15:26 rubenwardy sure, sfan5
15:26 Shara Well, translation from "texturepacks" and "texture packs" shouldn't be different anyway.
15:26 Shara sfan5: would love that to be done :)
15:27 sfan5 rubenwardy: btw do you support doing the same for the engine repo?
15:27 rubenwardy yes
15:27 sfan5 ok
15:27 rubenwardy later this week I can finish up my PR
15:28 rubenwardy I forgot to change lua_api.txt and readme, which is why it's WIP
15:28 rubenwardy a backport branch should be directly based on stable-0.4
15:28 sfan5 yes that's what i planned
15:28 sfan5 `backport-0.4` or `backports-0.4` ?
15:29 rubenwardy hmmm
15:29 rubenwardy former sounds better
15:30 rubenwardy really doesn't matter though
15:30 sfan5 the smallest questions are the hardest in life
15:30 rubenwardy lol true
15:30 Krock trivialities must be discussed
15:30 rubenwardy to great length
15:30 rubenwardy 120 is too little a number of PRs
15:31 rubenwardy we must discuss these things until the count is tripled
15:31 * Shara will try and find some more silly typos to correct then!
15:31 Krock Shara, best start in the lua_api, there's most of them :P
15:32 rubenwardy Shara, lua_api rewrite pls
15:33 rubenwardy just don't convert it to markdown
15:33 sfan5 H T M L
15:33 rubenwardy <.<
15:36 Shara Haha
15:42 Roger9 >.>
16:02 Kright joined #minetest-dev
16:14 Shara Issue with the lua api is knowing everything well enough to know what needs changing. I've started looking at it more than once, but it usually makes me want to run away and hide.
16:14 Shara It's over 5k lines of stuff to check basically.
16:24 sfan5 Shara: https://github.com/minetest/min​etest_game/commits/backport-0.4
16:28 Shara thanks sfan5 :)
16:30 Shara Could we maybe get a news post or something on the forum to alert people that this exists and that they should switch to it if they want bug fixes?
16:31 rubenwardy I suggest doing that when we get some commits on it
16:31 rubenwardy if we haven't already
16:31 sfan5 uh
16:31 Shara It has them?
16:31 rubenwardy cool
16:31 rubenwardy it does
16:31 * rubenwardy should have checked
16:31 Shara :P
16:31 rubenwardy that was quick!
16:31 sfan5 i didn't take 1h just to create branch you know
16:31 rubenwardy thanks very much :D
16:32 Shara But yea, news post should be made, because people won't use it if they don't know it's an option now
16:32 sfan5 i should probably also create a backport-0.4 branch on the engine but that's much more work >.>
16:33 Shara I thought 0.4.17 wasn't going to be the bug fix release? Though in general I suppose a backport branch that gets updated with bugfixes would be better?
16:34 rubenwardy 0.4.17 is the bug fix release
16:35 sfan5 mostly bugfix*
16:35 sfan5 at least I don't see anything wrong with having minor features even in 0.4.17
16:35 Shara Some would be nice
16:36 Shara though it's mostly bugs I worry about
16:36 sfan5 also in case it isn't obvious: the backport-0.4 branches are basically 0.4.17-dev (= what is to become 0.4.17)
16:36 sfan5 ^ that should be noted in the news post
16:41 Shara Makes sense. Will you make the post?
16:42 sfan5 i'll do the backports for the engine now
16:42 sfan5 you can draft one if you want to
16:43 Shara I can't post there, so I just want to make sure someone does.
16:44 rubenwardy I can move a topic into there, or sfan5 or I can post it there
16:48 Shara I can write something if it helps, but it's really just a couple of lines :P
16:49 sfan5 would help yes
16:51 nerzhul rubenwardy, i'm fine with it but why not now ? 0.6.0 is not now :p
16:51 rubenwardy ?
16:51 nerzhul you talked about switching to semver
16:51 nerzhul (yeah i'm late)
16:52 rubenwardy well, we could just mark it as 1.0.0 or 5.0.0
16:52 rubenwardy I prefer the latter as 0.4.x should have be 4.x.0
16:53 nerzhul 1.0.0 can be nice to be proper for a new departure :p
16:59 Megaf joined #minetest-dev
17:00 rubenwardy is it agreed that we will start doing backporting releases after 0.5.0?
17:00 rubenwardy sfan5, nerzhul
17:00 sfan5 huh?
17:01 rubenwardy *continue doing
17:01 sfan5 i thought 0.4.17 was the only release with this speciality
17:01 rubenwardy I think it's worth doing in the future too
17:01 sfan5 that means 0.4.17 will turn unmaintained and dev will continue as uaual on 0.5.x
17:02 rubenwardy yeah, any further backports should be on 0.5.0
17:02 rubenwardy so between full feature releases we release patch versions with are the latest feature release (eg: 0.5.0) with bug fixes
17:02 sfan5 oh you mean just that we maintain the last version while the current ver is in dev?
17:03 rubenwardy yeah
17:03 sfan5 if we continue development as it was before that shouldn't be needed
17:03 nerzhul for me 0.4.17 would be released before 0.5.0 as bugfix release
17:03 sfan5 since 0.4.x was backwards compatible completely you could just run your server on git HEAD with no problems
17:03 Shara It would be nice if bug fixes are made to current stable releases in general. Not everyone wants to run dev version at all times
17:04 sfan5 this is how it should be done with 0.5.x too IMO, it wasn't really much of a problem in the past
17:04 Shara Then you get easy option to patch bugs but not to need to deal with new and untested features unless you want to
17:06 rubenwardy I think QA is something that we need to pick up the slack on, and this is a good start
17:06 rubenwardy we also need Lua unit tests
17:06 rubenwardy and I need less coursework so I can do all this
17:06 sfan5 this just creates more work for an already slowly moving development process
17:06 nerzhul rubenwardy, for good QA first we need more and more unittests
17:06 nerzhul especially in core engine
17:06 nerzhul but some are related to rendering and hardcore to code
17:07 nerzhul it's because Server and Client objects are too heavy
17:08 rubenwardy well, maybe leave this until after 0.5.0
17:09 rubenwardy (discussion of back porting)
17:10 Shara The current need for it is more than usual due to the compatibility breakage of current dev. So whether to continue it or not after 0.5 (or whatever you'll call it) can be decided again
17:13 coyote joined #minetest-dev
17:14 rubenwardy it's also worth noting that basically *no one* is testing 0.5
17:14 rubenwardy because it's network incompatible
17:14 Shara Yup.
17:14 rubenwardy so it's worth doing at least one quick release after 0.5
17:15 Shara From a user perspective, there is no reason to touch it, unlike normal releases where you can still join any servers
17:53 paramat joined #minetest-dev
17:55 paramat 0.4.17 seems probably a bad idea, and many people agree, see discussion in #6542
17:55 ShadowBot https://github.com/minetest/minetest/issues/6542 -- 0.4.17 release
17:56 paramat although it's probably ok for MTG
17:56 rubenwardy I personally think that it's a bad idea as it encourages users to stay on 0.4, and any backporting is better done for 0.5
17:56 rubenwardy however users want it
17:57 rubenwardy and 0.5 will be almost a year after 0.4.16
17:57 Shara The only reason I wanted 0.4.17 is so that bug fixes would be available.
17:58 Shara It doesn't really matter what you call things, or whether there's a big official release, so long as there's a branch that has the bug fixes for those who aren't rushing to update
18:00 paramat rubenwardy then you should go with what you know is best
18:03 paramat it's good to build up a good appetite before the feast of 0.5, it encourages people to change to 0.5 which we know will help concerning the mobile app situation, as we can then move players to the official app
18:05 paramat i support moving to 5.0.0, just not 1.0.0 as that's too much associated with 'complete'
18:06 rubenwardy I agree with that
18:06 sfan5 <paramat> 0.4.17 seems probably a bad idea
18:07 sfan5 server owners may need time to move to 0.5 and denying them bugfixes in the meantime doesn't seem useful at all
18:07 sfan5 0.4.17 will not contain major features on purpose, to get people to move to 0.5
18:11 paramat well, as discussed, 0.4.17 will encourage servers to not move to 0.5.0, and this is a rare and valuable opportunity to move players to android, which will only be effective if servers move quickly to 0.5.0
18:12 paramat encouraging servers to stay on 0.5.0 will also cause a larger split in the community, more need for 2 versions of mods etc. surprisingly many people agree with me in the thread, wasn't expecting that =)
18:13 sfan5 so you want to throw away the work that goes towards maintain a backports branch once 0.5 is released?
18:14 sfan5 since that will exist anyway server could just stay on the backport branch instead of staying on the 0.4.17 release
18:14 sfan5 the end result is basically the same
18:15 nerzhul i don't see how this can block them to move . the only problem with this if they are released too close i think
18:15 rubenwardy I'm +1 for 0.4.17
18:15 sfan5 obviously whether it's an official release affects distro maintainers, but servers are what matters for moving userbase
18:15 rubenwardy I retract my comment at 15:56
18:15 rubenwardy urgh
18:15 rubenwardy I retract my comment at 17:56
18:16 sfan5 it's not like 0.5 won't have enough features that will make servers want to move anyway (csm flavor limits!!)
18:31 paramat " you want to throw away the work that goes towards maintain a backports branch once 0.5 is released?" no
18:33 rubenwardy reopened game#853
18:33 ShadowBot https://github.com/minetes​t/minetest_game/issues/853 -- Hunger
18:33 paramat ok, i have one request then, no features added for 0.4.17 however small
18:33 sfan5 too late i'm not going to rebase half of this again
18:33 paramat slippery slope
18:34 Shara Ability to cut colours out of chat should be included please.
18:34 paramat ^ except that
18:34 sfan5 read the list here https://github.com/minetest/minetest/projects and decide whether that's "too many features"
18:34 paramat =)
18:34 paramat ok
18:34 paramat coloured chat restrictions is justified
18:40 paramat wow there are far too many features added, where do i list the ones to leave out?
18:40 paramat remember we have neen consistently saying 'no features added to 0.4.17'
18:40 paramat *been
18:40 sfan5 i don't remember anyone saying that
18:40 paramat nerzhul and others
18:41 sfan5 there's barely any real "feature" in there anyway, but i'd like to know which ones you want to omit
18:41 paramat rubenwardy
18:42 paramat will do in 0.4.17 issue
18:46 paramat "Smooth lighting: Fix light leaking through edge-connected corners" should not as that is now causing severe lighting bugs
18:47 sfan5 oh didn't know
18:48 paramat honestly, all devs have been saying 'bugfixes only for 0.4.17' for months
18:48 sfan5 please repeat that another ten times instead of listing the commits to omit
18:49 paramat working on it
18:50 paramat also 'Allow zoom to actually show more data' needs a restriction which i am yet to work on
18:50 paramat https://github.com/minetest/minetest​/issues/6542#issuecomment-345540072
18:51 paramat New lighting curve (because it's not optimum yet and subtle shadows have disappeared)
18:52 sfan5 for the section that "need work", those need to be cherry-picked too because a possible later PR will be based on that code
18:52 sfan5 not doing that would make merging more difficult
18:52 sfan5 unless you mean that 0.4.17 should contains no new light curve, at all
18:55 paramat "0.4.17 should contains no new light curve, at all" correct, as current master needs fixing, i'm working on it
18:56 sfan5 those statements are contradictory
18:56 paramat where is the 'needs work' section?
18:57 sfan5 i was referring to the entries in what you linked where you said you would improve them
18:57 paramat ah ok
18:57 sfan5 either you 1) will adjust the current light curve and want that to be in 0.4.17 or 2) 0.4.17 will have the same light curve as 0.4.16
18:57 sfan5 which one is it
18:57 paramat i see
18:58 paramat same light curve as 0.4.16
18:58 sfan5 okay
18:59 paramat surely that's the same as not backporting 'New lighting curve'?
18:59 sfan5 it is
19:00 sfan5 but if you want to improve the current curve, i would need to backport the curve commit even if the current state is not satisfactory
19:00 sfan5 what about the zoom thing? "will improve" or "same as 0.4.16"?
19:01 paramat same as 0.4.16
19:01 sfan5 okay thanks
19:03 paramat in current master a client can choose zoomFOV 1 degree and load in blocks up to 4000 nodes away, not good
19:04 paramat my next PR will put some restriction on that
19:05 paramat i agree that 0.4.x will eventualy have to have all it's bugs fixed, to leave it in a good state
19:24 sfan5 paramat: finished with preparing a backport-0.4 branch with the state you wanted
19:25 sfan5 will update the list of commits, then you can take another look if you want
19:27 paramat thanks, appreciated
19:36 sfan5 paramat: updated, please take another look; also: https://github.com/minetest/minetest_game/projects
19:48 Kright joined #minetest-dev
20:03 paramat will do
20:08 ThomasMonroe joined #minetest-dev
20:13 CalebDavis joined #minetest-dev
20:15 torgdor joined #minetest-dev
20:22 sfan5 paramat: looked at it yet?
20:22 sfan5 since i'd like to push the creation of the backports branch
20:22 paramat ok, will do now, sorry, hub stuff
20:24 paramat fine for MT engine
20:25 sfan5 ok thanks
20:26 sfan5 Shara: same thing for the engine -> https://github.com/minetest/m​inetest/commits/backport-0.4
20:26 sfan5 if you're running the release versions of the engine you might want to upgrade to that
20:27 sfan5 rubenwardy: should the two version-touching commits be reverted on stable-0.4 then? (back to the "Merge 0.4.16 into stable-0.4" commit)
20:27 nerzhul sfan5, paramat nice job on backport branch
20:27 sfan5 thanks
20:27 nerzhul when do you want to release this branch ?
20:28 sfan5 merge it into stable-0.4 when 0.4.17 is released
20:28 nerzhul yes, but when do we want to release ?
20:28 sfan5 maybe at the start of feature freeze period of 0.5?
20:31 paramat well 0.5.0 is still far off, 0.4.17 is best released asap
20:32 paramat if they're close you reduce appetite for 0.5
20:33 sfan5 you can't fix ANY bugs for the stable version between the release of 0.4.17 and 0.5, that will make server owners angry
20:33 paramat oh hmm
20:33 sfan5 like i said: 0.5 has enough features already, we don't need to worry about 0.4.17 being "too good"
20:34 paramat lord fingle tells me he'll probably have time to work on android in January, so 0.5 should be 2-3 months into 2018
20:34 Shara sfan5: thanks
20:34 paramat otherwise we lose the chance to move players to the official app
20:35 sfan5 paramat: if you consider the state of the mtg backport branch fine we should be looking to make a news post on the forums
20:35 paramat android needs to be reasonable before 0.5 is released, rubenwardy agrees
20:35 Shara rubenwardy already made a post
20:36 paramat hm mtg backport also has lots of features
20:36 Shara https://forum.minetest.net/vi​ewtopic.php?f=18&amp;t=18850
20:36 sfan5 i expected you to say that :P
20:37 sfan5 oh ruben already made a post, thanks
20:38 sfan5 paramat: for mtg i considered api changes fine but applied the strict rule of having no new ingame content
20:38 Shara Worked on it with him earlier (though it's mostly his :P)
20:39 sfan5 since api changes are no user visible i don't think they discourage upgrading to 0.5
20:39 sfan5 not*
20:42 paramat they're modder visible =)
20:42 paramat https://github.com/minetest/minetest_g​ame/issues/1948#issuecomment-345547820
20:50 Icedream joined #minetest-dev
20:53 YuGiOhJCJ joined #minetest-dev
21:09 sfan5 will apply changes later
21:09 sfan5 any concerns about force-pushing the backports branch?
21:20 paramat hm no-one's around, i can't judge that, sorry
21:38 Raven262 joined #minetest-dev
21:45 AntumDeluge joined #minetest-dev
21:45 sfan5 not doing that would make applying your changes fun as i'd have to revert 20 or so commits just to delete one
21:47 sfan5 i guess the rules don't apply to "backport-0.4" similar to how they don't apply to minetest/{minetestmapper,master-server}
21:53 paramat ok, since it affects backport branch only, force push is ok
23:03 torgdor joined #minetest-dev
23:08 Megaf joined #minetest-dev
23:23 paramat joined #minetest-dev
23:39 Icedream joined #minetest-dev
23:40 torgdor joined #minetest-dev
23:58 AntumD joined #minetest-dev

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