Minetest logo

IRC log for #minetest-dev, 2015-06-30

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

All times shown according to UTC.

Time Nick Message
00:52 Wayward_Tab joined #minetest-dev
00:56 Wayward_Tab The minimap doesn't seem to work for android atm. I just git pulled and built, and added the respective lines to minetest.conf, to no avail.
01:11 OldCoder joined #minetest-dev
01:11 OldCoder joined #minetest-dev
01:37 sloantothebone Hello, is there any way to improve the android interface? I dont like it
01:37 sloantothebone You cant shoot bows frequently, or anything else
02:18 VanessaE hmmmm: as long as you're on a bugfix crusade, can you PLEASE fix the player-gets-deleted bug?
02:20 VanessaE my creative server is getting increasingly hard to sign onto because of it
03:18 paramat joined #minetest-dev
04:40 paramat making ignore walkable does not help issue #2702 so we are still looking for a fix for item-entities and falling-node-entities falling throough world base and becoming stuck as unretrievable entities
04:40 ShadowBot https://github.com/minetest/minetest/issues/2702 -- Bottom of the world weirdness
05:13 hmmmm hey, est didn't get a second +1 for the last commit
05:15 Puma_rc joined #minetest-dev
05:16 paramat savanna is added to biomesdev. next i will add
05:16 paramat oops..
05:16 paramat .. a 'polar icecap' glacier biome
05:18 paramat due to more biomes i will increase heat and humidity spreads to 1024, as of course a hacky temporary thing
05:19 paramat and allow rainforest trees to spawn in shallow water to create swampland
05:22 hmmmm in fact you know what, I'm really not happy with est's latest commit
05:23 hmmmm there could be a very harmful effect of removing the sleep() from the mesh generation thread
05:24 paramat naughty est!
05:25 hmmmm although mesh generation would become faster overall, this has the bad effect of adding much more latency to other threads
05:26 hmmmm i wouldn't be surprised if master HEAD has a lag regression problem (as if the most recent rendering regression wasn't enough)
05:30 hmmmm https://github.com/minetest/minetest/commit/29dda9f356042c403b3b7da1d717d32b45c9b6de#diff-7f3a3ff3a2c9b903da668a4598f5f8bbR260
05:30 hmmmm this could lead to consumer starvation
05:30 hmmmm not to mention numerous code style violations
05:39 paramat will push later #2858
05:39 ShadowBot https://github.com/minetest/minetest/issues/2858 -- Biome API: Increase heat and humidity noise spreads to 1000 by paramat
05:42 RealBadAngel joined #minetest-dev
05:45 Hunterz joined #minetest-dev
06:02 paramat now pushing 2858
06:02 Krock joined #minetest-dev
06:06 paramat complete
06:28 paramat left #minetest-dev
06:52 kilbith joined #minetest-dev
07:07 Darcidride joined #minetest-dev
07:25 crazyR_mobile joined #minetest-dev
07:43 RealBadAngel 0.4.13 is coming... https://xkcd.com/1532/
07:43 VanessaE hah
07:46 Krock xD
07:46 Darcidride joined #minetest-dev
07:55 kilbith joined #minetest-dev
08:00 Yepoleb_ joined #minetest-dev
08:01 jin_xi joined #minetest-dev
08:23 nrzkt joined #minetest-dev
08:26 Calinou joined #minetest-dev
08:27 Darcidride_ joined #minetest-dev
08:33 nrzkt Hello, RealBadAngel: great work on minimap, and thanks for fixing the Clang issue before merging :)
08:49 rubenwardy joined #minetest-dev
09:30 Darcidride joined #minetest-dev
10:06 nrzkt joined #minetest-dev
10:22 rubenwardy joined #minetest-dev
10:30 RealBadAngel joined #minetest-dev
10:50 proller joined #minetest-dev
10:52 kilbith sfan5, if you have a moment : https://github.com/minetest/minetest_game/pull/548
11:06 Darcidride_ joined #minetest-dev
11:06 julienrat_ joined #minetest-dev
11:42 rubenwardy joined #minetest-dev
12:13 RealBadAngel Krock, here?
12:13 Krock afaik, yes
12:13 RealBadAngel hehe, have some time? i would like to fix yours shaders issue
12:14 Krock sure
12:15 Taoki RealBadAngel: Any news on the VBO pull request?
12:15 proller joined #minetest-dev
12:16 Krock RealBadAngel, how exactly can I help you?
12:16 RealBadAngel Taoki, will make a PR candidate today propably
12:16 Taoki Awesome :)
12:17 nrzkt Taoki: What is VBO ?
12:17 kilbith nrzkt, https://fr.wikipedia.org/wiki/Vertex_Buffer_Object
12:17 Taoki nrzkt: A technique to cache geometry on the video card memory, to avoid the CPU streaming it each frame. It improves FPS a lot!
12:18 Taoki BTW: There are two other similar techniques... I'm not sure if we could use these as well. VAO (Vertex Array Objects) and FBO (Frame Buffer Objects)
12:18 kilbith Taoki, it doesn't cache only the geometry
12:18 Taoki Does Irrlicht have these? Do they work alright with Minetest?
12:19 Lunatrius joined #minetest-dev
12:19 RealBadAngel VBO stores vertices and indices of the object
12:20 * Taoki nods
12:22 nrzkt And irrlicht doesn't do it natively ? :(
12:22 nrzkt Thanks for the explanation
12:22 Taoki nrzkt: It does. It just needs to be enabled. And for Minetest that had some issues
12:23 Taoki Namely high memory usage. Which was difficult to fix, hence why it took so long.
12:23 nrzkt oh i see, what are the issues ?
12:23 Taoki ^
12:23 nrzkt okay !
12:23 Taoki :)
12:24 Taoki Seems RBA found that magical cure to it at last. Very curious to test it, hope there will be a PR tonight
12:24 Taoki And that it won't take ages to accept, heh
12:38 Lunatrius joined #minetest-dev
12:45 Darcidride__ joined #minetest-dev
12:50 Darcidride joined #minetest-dev
12:57 nrzkt i will review it faster if it's a good improvement :)
13:04 RealBadAngel improvement is good, but we dont know yet what is the reason for memory leaks
13:31 blaise joined #minetest-dev
13:35 Darcidride joined #minetest-dev
13:58 nrzkt do you use valgrind ?
14:01 RealBadAngel im not familiar with it
14:01 blaise valgrind is nice
14:14 SopaXorzTaker joined #minetest-dev
14:18 nrzkt launch minetest with valgrind --tool=callgrind in debug mode
14:18 nrzkt then use kcachegrind to read valgrind logs with a great QT interface :)
14:19 nrzkt callgrind is for calls sorry, you need memory checks. Valgrind talk about it directly but it can trace in logs too
15:07 hmmmm joined #minetest-dev
15:31 CraigyDavi joined #minetest-dev
15:35 est31 joined #minetest-dev
15:41 proller joined #minetest-dev
15:43 rubenwardy joined #minetest-dev
15:59 nrzkt est31: sorry cannot i'm at work :)
16:00 est31 ok
16:00 est31 you +1ed that PR nrzkt ?
16:01 est31 last time I understood you wrongly
16:01 est31 therefore I ask now
16:06 nrzkt yes :p
16:23 Player_2 joined #minetest-dev
16:25 kilbith joined #minetest-dev
16:27 est31 seems somebody is abusing our name: http://www.1mobile.com/developer/minetest-mobile-llc.html
16:28 est31 but as its no trademark, we cant do anything
16:28 Calinou ask celeron55 about it
16:28 Calinou he can probably DMCA due to lack of source code
16:29 kilbith nrzkt can do it too
16:30 est31 they say its 0.4.11-dev
16:31 est31 so depending on whether nrz has copyright on that yet or not he can
16:32 est31 nrzkt, did you upload the minetest apk to 1mobile market too?
16:32 est31 http://www.1mobile.com/minetest-2671770.html
16:32 est31 or was it them, scraping play store
16:32 est31 most likely them: "Minetest is now available officially on the Google Play Store"
16:33 est31 man this is scary
16:33 est31 lets hope there is no adware included
16:37 est31 you know like that story with sf.net and gimp
16:39 CraigyDavi joined #minetest-dev
16:48 kilbith should the PRs updating language translations (eg. japanese, turkish) that nobody understands here be merged exceptionnally after one +1 ?
16:51 Robert_Zenz joined #minetest-dev
16:57 proller joined #minetest-dev
17:02 MinetestForFun joined #minetest-dev
17:07 selat joined #minetest-dev
17:15 VanessaE kilbith: sounds reasonable to me.
17:16 VanessaE if there's something wrong, literally no one will be able to see it until someone who speaks that language shows up.
17:22 kilbith someone can adds insanities furtively to the game as well :/
17:23 kilbith we can't check that
17:23 Calinou remember in Minecraft, the "You are a NIGGER" singleplayer button in Afrikaans? :D
17:23 VanessaE good point.
17:23 Calinou such things happened very often, like in every snapshot
17:23 Calinou I think they eventually stopped crowdsourcing translations due to that
17:31 VanessaE nrzkt: that patch you and ets just approved didn't fix any errors, it rather seems to have introduced new ones.
17:32 VanessaE est*
17:32 kilbith wat
17:32 VanessaE http://pastebin.com/jxhkDEWc
17:32 VanessaE all I did was single-click on a world in my main, singleplayer menu.
17:32 VanessaE minetest_game was selected.
17:33 VanessaE sorry, kilbith. :)
17:34 kilbith indeed, can reproduce it
17:35 VanessaE kilbith: hours wasted? ;)
17:35 kilbith likely a silly mistake
17:37 rubenwardy joined #minetest-dev
17:42 est31 joined #minetest-dev
17:42 est31 hmmmm, you there?
17:42 est31 whats exactly your problem with the thread commit?
17:42 VanessaE est31: see logs also.  that last menu commit broke.
17:42 VanessaE http://pastebin.com/jxhkDEWc
17:43 est31 I've got two +1 (except me) one from nrz one from rba
17:43 kilbith she talks about the latest commit
17:44 est31 yes and I talk about my thread commit
17:44 est31 http://irc.minetest.ru/minetest-dev/2015-06-30#i_4303764
17:45 est31 so kilbith you have a fix?
17:45 kilbith not yet, you should revert it
17:46 kilbith it's probably silly error
17:46 est31 ok then I have a look at it
17:46 kilbith weird, it worked fine why i tested earlier
17:46 kilbith i could entering in the world and so on
17:47 est31 but that error screen is cool at least
17:47 est31 yea in fact it was intentional
17:47 est31 only to showcase which cool errorscreen we have
17:47 est31 lol jk#
17:48 kilbith i could fix it now but in deleting the loop
17:48 kilbith but it's more redundant
17:48 est31 erm ok it makes sense it crashes
17:49 est31 also the names of the functions are a bit bad for understanding what they do
17:49 est31 what should menu_worldmt_legacy do
17:49 kilbith i'd prefer any help than rushing on a revert, though
17:50 kilbith you have a better idea for the naming ?
17:50 est31 what is the .. "" for?
17:50 kilbith it needs to sends "creative_mode", not creative_mode
17:51 kilbith with the commas
17:51 kilbith ah got it, fixed
17:52 est31 ermm
17:52 est31 "" stands for "\""
17:53 est31 ?
17:53 kilbith a missing param in the function that needed to be sent
17:53 kilbith no
17:53 est31 thats a very weird thing
17:54 kilbith it works like that
17:54 est31 thats a very weird thing
17:54 est31 err
17:54 est31 sorry
17:54 kilbith i'm sending the patch
17:58 RealBadAngel joined #minetest-dev
17:58 est31 I have a fix too, we can compare them then https://github.com/est31/minetest/commit/ca7b79258411fe2db5276c564b0d8732e95e158f
17:59 TeTpaAka joined #minetest-dev
18:00 kilbith est31 : https://github.com/minetest/minetest/pull/2860
18:00 est31 kilbith, there are other problems too.
18:01 blaise joined #minetest-dev
18:02 kilbith which ones ?
18:02 est31 for example, you do local mode =
18:02 est31 and so on
18:02 kilbith i confirm, no need those ..""
18:02 est31 it should be renamed
18:02 kilbith your partch is better
18:02 kilbith patch
18:02 est31 then those ..""
18:02 est31 and you do if mode then
18:02 est31 what about mode = false?
18:02 kilbith wait, i close my PR then
18:03 est31 then I push that fix in 10 minutes
18:03 crazyR_mobile joined #minetest-dev
18:03 kilbith ok thanks
18:03 est31 looking at other parts of the code to improve
18:04 kilbith i don't understand mode = false
18:04 kilbith mode is just an iterator to the table
18:04 est31 in menu_worldmt()
18:04 est31 you do return world_conf:get(setting)
18:04 est31 so when that happens, everything is finr
18:05 kilbith there's no `mode` in that function
18:05 est31 and you will have to do core.setting_set(mode_name, mode_val)
18:05 est31 but imagine if world_conf:get(setting) returns false
18:05 kilbith ok, so it's just miss a statement with return nil then
18:05 est31 then you do local mode_val = menu_worldmt(selected, mode_name)
18:05 est31 and then mode_val == false
18:05 est31 then, in if mode_val then
18:05 est31 then you will chose the second part
18:06 est31 menu_worldmt(selected, mode_name, core.setting_get(mode_name))
18:06 est31 which is wrong
18:06 est31 therefore the check has to be if mode_val ~= nil then
18:07 kilbith if mode_val then  is as correct as  if mode_val ~= nil then
18:08 blaise joined #minetest-dev
18:08 kilbith same statement
18:08 est31 no
18:08 est31 what about mode_val == false?
18:09 Hunterz joined #minetest-dev
18:09 est31 newest version now https://github.com/est31/minetest/commit/ca7b79258411fe2db5276c564b0d8732e95e158f
18:10 kilbith dunno, i never used the ~= nil in my statements
18:10 est31 err sorry
18:10 est31 https://github.com/est31/minetest/commit/29b0dc89e6789bce8663b74feeb32f1961a784d7
18:10 est31 ^thats it
18:10 est31 forgot push
18:12 kilbith lgtm
18:12 sloantothebone joined #minetest-dev
18:15 est31 hmmmm, would have been nice if you gave examples for the style violations
18:15 est31 of course I did while (statement) do_something:
18:15 est31 I admit to be guilty of that crime
18:15 est31 also for the if later on
18:17 est31 but otherwise, I see nothing
18:23 est31 but perhaps I should have said "set the semaphore back to 0" instead of this comment: https://github.com/minetest/minetest/commit/29dda9f356042c403b3b7da1d717d32b45c9b6de#diff-7f3a3ff3a2c9b903da668a4598f5f8bbR259
18:24 est31 "Set the semaphore counter to 0, just in case doUpdate() is expensive when nothing has to be done"
18:25 est31 I have changed both use cases to live with the "update" model
18:25 est31 here, every iteration of the outer loop is just one "update"
18:25 est31 so that its not so close to the queue model the threads embrace
18:26 est31 just to avoid confusion
18:32 boonscape3 joined #minetest-dev
18:38 TeTpaAka joined #minetest-dev
18:41 boonscape3 joined #minetest-dev
19:05 est31 pushing fix for #2836 https://github.com/est31/minetest/commit/842579eb8c31e6a5d4fb3a56413380b11b7252b7
19:05 ShadowBot https://github.com/minetest/minetest/issues/2836 -- [Esc] does not close the in-game keybinding menu
19:05 est31 in 5 minutes
19:10 MinetestForFun joined #minetest-dev
19:20 est31 5 minutes till merge https://github.com/est31/minetest/commit/370a33686ac6ce731c989bcd513aa351666ab5ae
19:21 est31 after kilbith has brought up the "single core dev agree" point on translation PRs, I have got an idea for how to do it
19:22 RealBadAngel you will send the translation to proper embassy for verification?
19:22 est31 what ahout doing a "maintainer" based system, where people can volunteer to be maintainers for one or multiple languages. If two of the maintainers agree (excluding the original one :p)
19:23 est31 and one core dev has checked for abuse with google translate
19:23 est31 then it can be merged
19:23 est31 we could track the maintainers with github teams
19:23 kilbith that'll be hard to gather minimum 2 maintainers on some languages...
19:23 kilbith but ok for german, spanish and french
19:23 est31 ok two or all of the maintainers
19:24 est31 so they can be pinged e.g. by @minetest/i10n-de
19:24 est31 @minetest/i10n-cn
19:25 kilbith nobody joins too specialized channels
19:25 est31 no irc channels
19:25 est31 github teams
19:25 est31 https://github.com/orgs/minetest/teams
19:25 kilbith ah, yes
19:25 est31 perhaps we should make a new group account
19:26 est31 not minetest, but minetest-i10n
19:26 est31 what do people like celeron55  or others think?
19:29 est31 requirement for maintainer are github account and that you vote or comment on PRs within 7 days
19:29 est31 within 7 days after being pinged
19:29 kilbith and have >16 years old if it was possible to check
19:30 est31 yea they should know the language to a well enough level
19:30 RealBadAngel est31, m_blocks_cache is no longer public
19:30 RealBadAngel you can update issue
19:31 est31 which commit RealBadAngel ?
19:32 est31 RealBadAngel, its still not private
19:32 RealBadAngel damn, i moved the code that flushes it to proper place, but forgot just to change variable to private
19:32 RealBadAngel lmao
19:32 est31 heh
19:33 est31 then move it, push it (its trivial fix) and I update
19:33 est31 the issue
19:33 RealBadAngel i do have now little time to do so
19:34 RealBadAngel i could do that tommorow, if you can do it today, just do it
19:34 est31 ok
19:35 est31 https://github.com/est31/minetest/commit/9e3e6dc5ec8ea63d65c635fcac0c80e0cd0e666d
19:35 est31 RealBadAngel, ^
19:36 RealBadAngel yup
19:37 RealBadAngel btw, i was fighting with shaders code and opengl 2.1
19:37 RealBadAngel i just learned that there are instruction limits for shaders code
19:37 RealBadAngel and ALL the loops are uncoditinally unrolled by compiler
19:37 RealBadAngel pretty funny, eh?
19:37 est31 lol
19:38 RealBadAngel thats why parallax wont work on opengl 2.1
19:38 RealBadAngel at least relief version
19:38 RealBadAngel max nr of iterations i could get was 4
19:39 RealBadAngel why circa 20 is needed for good lookin
19:39 RealBadAngel *ehen
19:39 RealBadAngel when
19:39 crazyR_mobile joined #minetest-dev
19:40 est31 so RealBadAngel what do you think of that proposal
19:41 RealBadAngel idea is ok, but imho no need for channel, opened issue each time new translation goes in should be enough
19:42 est31 no channel but github teams https://github.com/orgs/minetest/teams
19:42 RealBadAngel yeah yeah, i mean no group needed
19:42 est31 so you mean every language should have a github issue?
19:43 RealBadAngel same as in case of every other PR
19:43 RealBadAngel once new translation gets merged you close it
19:43 est31 yea but how to ping all maintainer
19:43 est31 s
19:44 RealBadAngel no idea
19:44 est31 they shouldnt have to monitor the open prs
19:44 kilbith e-mail ?
19:44 est31 when we have teams, they have no further access
19:44 kilbith *mailing list
19:45 kahrl my opinion is that we should bring back weblate
19:47 est31 why was it canceled?
19:47 kahrl xyz stopped maintaining it
19:49 Calinou if you do, install the latest version, it comes with a nicer interface
19:51 est31 kahrl, thats an even better idea
19:52 est31 btw if nobody wants to host it, weblate.org has free hosting for open source projects
19:52 Calinou it seems they take a lot of time to reply to requests
19:54 kahrl yeah I sent a request at some point, but I didn't get a response (although it might have gotten eaten by the spam filter)
19:54 est31 where you know from Calinou
19:54 Calinou I just heard that somewhere
19:55 kahrl perhaps it'd work if we follow up the request with bugging nijel in #weblate :)
19:55 est31 hehe
19:58 Wayward_Tab joined #minetest-dev
19:59 est31 should I send a new request?
20:00 kahrl sure
20:03 est31 ok, sent
20:38 Robert_Zenz joined #minetest-dev
21:26 sloantothebone_ joined #minetest-dev
21:34 MinetestForFun_ joined #minetest-dev
21:37 sloantothebone joined #minetest-dev
21:48 proller joined #minetest-dev

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