Minetest logo

IRC log for #minetest-dev, 2013-12-04

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

All times shown according to UTC.

Time Nick Message
00:05 john_minetest It is Jeija's mod iirc
00:05 VanessaE working in it now.
00:05 john_minetest oh, okay
00:05 iqualfragile jeija disappeared
00:05 john_minetest Lol, right, I got that subdomain at mesecons. Just remembered.
00:10 VanessaE damn it
00:10 VanessaE iqualfragile: it set ME as the author.  this is not right.
00:10 iqualfragile i know
00:11 VanessaE also, "add a picture" and "add a version" should allow URLs
00:11 iqualfragile just write a notice in the description, i am quite sure jeija will understand
00:11 john_minetest I am considering making a minetest hipster swag yolo retro edition game (for mac ;) )   ->   http://pastebin.sabayon.org/pasties/store/15345/mi​netest_hipster_swag_yolo_retro_edition_for_mac.png
00:11 VanessaE if there's a file on the web somewhere, I should be able to point to that.
00:11 iqualfragile but i think about adding a field for maintainer
00:11 john_minetest Who made the mmdb web interface?
00:11 iqualfragile nah, thats some kind of schema: no external dependencies
00:12 iqualfragile john_minetest: jeija
00:12 john_minetest Last post: 2013-07-03 12:01:56
00:12 john_minetest strange...
00:13 VanessaE iqualfragile: just change the "Author" field to "Uploader"
00:13 VanessaE and then add separate "Author" and "Maintainer" fields.
00:13 john_minetest And only allow new mods when download link is provided.
00:13 VanessaE that way the database doesn't get screwed up
00:14 iqualfragile VanessaE: yeah, i think i can manage that :D
00:15 iqualfragile john_minetest: why?
00:16 sapier VanessaE if mmdb would support files this would break minetest download
00:17 sapier and of course it'd break trust in signed off versions
00:17 VanessaE sapier: why?  just have mmdb check that the supplied URL points to a downloadable file, and have it actually perform the download and store the file.
00:17 VanessaE in other words, the mmdb should automatically do precisely what the uploader has to do manually
00:17 sapier ok misunderstood this
00:18 sapier as long as the file is stored in mmdb everything is fine
00:18 VanessaE yues
00:18 VanessaE yes, that's the theory.
00:18 sapier ok good night ;-)
00:18 sapier left #minetest-dev
00:19 john_minetest iqualfragile: Because without a downnload link it isn't useful for anything.
00:19 EvergreenTree joined #minetest-dev
00:19 john_minetest The mmdb is there to provide info and especially a download link platform for mods.
00:19 iqualfragile john_minetest: not realy, still contains i.e. a link to the github repos
00:19 iqualfragile mods without downloads are hidden in the client
00:20 john_minetest iqualfragile: They just clutter the database.
00:21 john_minetest It doesn't make sense having mods without download there.
00:21 iqualfragile john_minetest: nah, its allright, they can be deleted if the actually turn out to be unused
00:22 john_minetest Why posting a mod without providing a download link?
00:23 john_minetest Without a link that database entry is worthless. If they don't have a link yet, they should wait until posting.
00:31 Neological joined #minetest-dev
00:35 VanessaE frak.
00:38 Anchakor_ joined #minetest-dev
00:38 john_minetest hmmm?
00:39 VanessaE RBA ping'd out.  #224 against minetest_game is waiting on him.
00:55 VanessaE make that #225...  forgot to put the changes in their own branch..... https://github.com/minetest/minetest_game/pull/225
00:57 RealBadAngel joined #minetest-dev
01:06 RealBadAngel joined #minetest-dev
01:24 RealBadAngel https://github.com/minetest/minetest/pull/1037
01:24 RealBadAngel thats a little helper function done by PilzAdam
01:25 RealBadAngel can we have this merged?
01:25 VanessaE minetest-game #225 depends on that ^^^
02:20 werwerwer joined #minetest-dev
02:35 OldCoder joined #minetest-dev
03:03 werwerwer_ joined #minetest-dev
03:43 pandemonium joined #minetest-dev
04:00 OldCoder what version of irrlicht is presently recommended?
04:10 djdduty joined #minetest-dev
04:10 djdduty joined #minetest-dev
04:48 thexyz OldCoder: 1.8 I think, it doesn't have issues with white inputs
04:49 OldCoder thanks
05:20 kahrl_ joined #minetest-dev
05:53 Miner_48er joined #minetest-dev
06:10 mrtux joined #minetest-dev
06:14 Miner_48er joined #minetest-dev
07:32 iqualfragile joined #minetest-dev
08:13 darkrose joined #minetest-dev
08:24 salamanderrake joined #minetest-dev
09:07 dzho_ joined #minetest-dev
09:10 pitriss` joined #minetest-dev
09:17 khonkhortisan joined #minetest-dev
09:32 PandemoniuM joined #minetest-dev
09:42 proller joined #minetest-dev
10:22 iqualfragile joined #minetest-dev
11:51 proller joined #minetest-dev
11:58 sapier joined #minetest-dev
12:30 john_minetest joined #minetest-dev
12:33 john_minetest Let's see what I promised to do yesterday when I was half asleep...
12:34 john_minetest minetest hipster swag yolo retro edition game (for mac)  and that mapgen chooser that work
12:34 john_minetest s
12:34 john_minetest * shows more info
13:21 ImQ009 joined #minetest-dev
13:27 hmmmm joined #minetest-dev
13:37 john_minetest I tested Zeg9's wieldlight branch. Really nice.
13:38 john_minetest bbl
13:56 Gethiox joined #minetest-dev
14:10 ShadowNinja joined #minetest-dev
14:12 zat joined #minetest-dev
14:28 thk joined #minetest-dev
15:54 john_minetest joined #minetest-dev
16:08 Jordach joined #minetest-dev
16:20 NakedFury joined #minetest-dev
16:26 Gethiox2 joined #minetest-dev
16:38 VanessaE joined #minetest-dev
16:49 Akien joined #minetest-dev
16:57 sapier joined #minetest-dev
17:00 sapier https://github.com/minetest/minetest/pull/1030 any core dev to agree to this one?
17:01 RealBadAngel joined #minetest-dev
17:03 sapier rba you know more about opengl then me could mipmapping anisotrophic bilinear and trilinear be done as one or mor dropdowns?
17:04 john_minetest sapier: I will test it.
17:04 john_minetest the search
17:04 sapier I thought so john :-)
17:05 john_minetest It is great to have a fast computer. compiling is fast... :)
17:10 VanessaE sapier: bi/tri could perhaps be made into a dropdown or radio buttons but the other two are theoretically independent.
17:11 sapier saving a single element doesn't seem very helpfull to me
17:12 john_minetest sapier: It runs nice in "normal" mode.
17:12 Calinou joined #minetest-dev
17:12 john_minetest going to use the force.
17:14 sapier can anisotrophic and bi/tri be used same time?
17:15 Calinou probably?
17:15 Calinou mipmap = far away textures don't look noisy
17:15 sapier thought there can only be a single filter active same time?
17:15 Calinou mipmap+aniso = same, but far away textures do not look blurred or plain
17:16 Calinou mipmap+ainso+bilinear = nearby textures are smoothed
17:16 VanessaE yes they can.
17:16 VanessaE bi/tri are mostly for up-close stuff
17:16 Calinou nearmap+aniso+trilinear = see above? which difference does it have?
17:16 Calinou maybe trilinear is better smoothing of nearby textures
17:17 sapier so anisotrophic is for far away things and bi/tri for close ones?
17:17 VanessaE I think bi- is texture-based and tri- is texture+distance(angle) based
17:18 VanessaE sapier: I regularly use mip+aniso+tri together.
17:19 sapier I still wonder if this is of any use ;-)
17:19 VanessaE the only thing we don't have in that regard is some sort of proper scaling algorithm for distant objects with transparency
17:19 sapier ok yes it is usefull
17:19 VanessaE (e.g. glass, papyrus still look like shit in the distance)
17:20 Calinou by the way, how do you change the anisotropic filtering level? how do you know it, too?
17:20 Calinou by the way 2, it would be nice if there was a checkbox for anti-aliasing in the options menu
17:20 VanessaE no clue.
17:20 sapier but I'm gonna make bi and tri xor
17:21 VanessaE sapier: yes.
17:21 Calinou or just make trilinear available
17:21 Calinou I don't see any use for bilinear filtering today
17:21 sapier if I understood correct anisotrophic uses bi or tri
17:21 sapier no idea what happens if you don't specify anything
17:21 VanessaE Calinou: some hardware may not perform well with trilinear, that's why I included both.
17:22 Calinou very, very old hardware :P
17:22 Calinou you can use aniso without bi/trilinear filtering, that's what I do
17:22 Calinou no noise on far away textures, but they still look detailed.
17:22 john_minetest sapier: I checked the menu with search. Found only one unrelated bug.
17:23 john_minetest Minetest crashes when trying to install the "test" mod.
17:23 john_minetest mmdb should check if the attatchment is a zip file.
17:25 sapier what's the exact error message?
17:25 john_minetest l_mainmenu.cpp:807: static int ModApiMainMenu::l_extract_zip(lua_State*): Assertion 'fs->getFileArchiveCount() > 0' failed.
17:28 john_minetest Since the test download is a .Ink (windows symlink) file I guess that this might be the problem.
17:28 john_minetest But it is an error on the mmdb side, not on the minetest client side.
17:29 sapier no application should crash on broken remote data
17:29 john_minetest no remote data should be broken
17:29 sapier isn't an excuse for unsafe code
17:29 john_minetest I think it would be enough to throw an error in that case.
17:30 john_minetest Let me try to fix that.
17:30 sapier I'll fix it ... it's my own code so noone misinterpret this as me accusing anyone else ;-)
17:30 john_minetest okay. if you want...
17:30 john_minetest :)
17:31 sapier a application shall never ever crash on invalid data ... it may show an error but not crash
17:33 john_minetest you shall not crash! fly you bools!
17:52 ShadowNinja Hmmm, any thoughts on the "Stack is over 30" crash? It seems to happen in a call to perlin:get2D. See https://forum.minetest.net/vie​wtopic.php?pid=120604#p120604 (Only the --info log has this bug and is usefull)
17:52 sapier did you check stack beeing popped correct?
17:53 ShadowNinja sapier: I haven't touched that code, but my table creation commit adds feilds to table, which become one entry on the stack by the time that function returns.
17:53 ShadowNinja tables*
17:55 ShadowNinja My errfunc commit is more likely to cause this, but I puped more things from the stack, not less.
17:55 ShadowNinja poped*
17:55 sapier I thing so but maybe something was missed
17:57 sapier john new version shouldn't crash and provide error message
17:58 sapier +instead
18:01 PilzAdam joined #minetest-dev
18:02 sapier I'd like to merge that one if there's an agreement and no further issues with it
18:05 ShadowNinja sapier: Could the async API cause this?
18:05 sapier if it happens in game then no async is menu only by now
18:07 sapier are you guys really sure irrlicht is capable of applying bi/tri and anisotrophic same time?
18:07 VanessaE sapier: positive.
18:07 VanessaE I use all three at once, and each has a different effect on my hardware
18:08 john_minetest lol
18:08 PilzAdam if bi- and trilinear filter are enabled irrlicht uses the trilinear one only
18:08 sapier SMaterialLayer.h irrlicht header explains usage so yes both can be used same time
18:08 VanessaE PilzAdam: he means using aniso + either of the two filters
18:09 john_minetest VanessaE: caught between a rock and a hard place :D
18:16 Taoki joined #minetest-dev
18:17 VanessaE meanwhile:  minetest #1037 and minetest_game #225 need attention.
18:22 ShadowNinja VanessaE: Better: http://paste.progval.net/show/677/
18:23 ShadowNinja That way doesn't copy the whole nodedef. You just have to rmove a metatable.
18:23 ShadowNinja (Or use setmetatable)
18:24 PilzAdam Id prefere ShadowNinja's way
18:24 * VanessaE shrugs - if it's better, okay
18:24 bas080 joined #minetest-dev
18:25 VanessaE PilzAdam: kinda ironic since you mostly initiated the pull-requested method :)
18:26 PilzAdam its the simplest way from a mod point of view
18:28 ShadowNinja https://github.com/minetest/minetest/blo​b/master/builtin/misc_register.lua#L153 That just has to be removed, but why was it added?
18:29 PilzAdam this dissallows mods to accidentally modify the nodedefs at runtime
18:29 VanessaE better?
18:29 sapier it's overriding the __newindex method ... maybe it shouldn't be {} but nil
18:31 kaeza it should be a function that calls error()
18:32 VanessaE is #225 suitable now?
18:32 kaeza setting it to {} makes the code edit a table without raising an error, and setting to nil makes it edit the original table
18:33 ShadowNinja PilzAdam: Actually, it only prevents adding new fields, not modifying existing ones.
18:33 kaeza ^ also that
18:33 PilzAdam thats a bug ;-)
18:33 VanessaE in which case, #1037 is still needed.
18:34 * VanessaE puts #225 back the way it was.
18:35 bas080 joined #minetest-dev
18:36 ShadowNinja And it creates a new metatable for each node. How can I chek if the game has started and the initialization phaze where node edits are allowed is over?
18:36 PilzAdam minetest.after(0) is the first env step
18:36 celeron55 the modstore should *really* support pageup, pagedown, scrollwheel and dragging of the scrollbar for scrolling
18:37 celeron55 some of those is probably not possible but at least most should work
18:37 VanessaE celeron55: +1000
18:37 celeron55 now it seems only the minisculous buttons work
18:37 celeron55 nobody wants to press buttons for scrolling since 1980
18:38 Zeitgeist_ joined #minetest-dev
18:38 sapier there are two ways to make it work
18:39 VanessaE ShadowNinja: the clone node method is needed then, since those grasses don't have a drops field defined.
18:39 sapier new gui toolkit supporting containers
18:39 PilzAdam VanessaE, then we just remove that metatable field
18:40 PilzAdam it should work how ShadowNinja suggested it, no matter what needs to be changed to make it like this
18:40 VanessaE PilzAdam: and watch as it breaks 100 other things? :)
18:40 sapier generate mousewheel events as button clicks to create something that is similar to mousewheel
18:41 celeron55 by the way is the modstore supposed to work?
18:41 sapier yes
18:41 celeron55 for me it has the slight problem that it doesn't work
18:41 sapier only if curl is enabled of course
18:42 celeron55 it appears to work, but no mods get installed while it says they're being installed
18:42 sapier modpack or mod?
18:43 celeron55 umm... actually, it works
18:43 sapier really not installed or only button in modstore not changed? presence detection for modpacks is wrong
18:43 sapier ok so the modpack issue
18:43 celeron55 it's just hard to find a mod because the store page doesn't say the modname but the fancy name
18:43 celeron55 then you go to the configuration page and be like "wtf is the name of whatever i just installed"
18:44 sapier hmm it's diffucult to show all that information in that compact view
18:44 celeron55 there's plenty of space above rating
18:44 sapier not in search tab variant
18:44 sapier https://github.com/minetest/minetest/pull/1030
18:45 ShadowNinja Should indexing a non-existant field be redirected to minetest.(node,tool,item)def_default?
18:45 celeron55 sapier: maybe show it in the dialog that says it's installed?
18:45 sapier that's possible yes
18:45 PilzAdam ShadowNinja, no
18:45 celeron55 "short name: <whatever>"
18:46 sapier I'm adding it
18:48 ShadowNinja PilzAdam: minetest.after/on_globalstep isn't really reliable though, as you can still modify the itemdef if your code runs before the game_started one. But I guess things in builtin will run first...
18:52 ShadowNinja PilzAdam: Does this look good? http://ix.io/9fN
18:53 ShadowNinja PilzAdam: Also this: http://ix.io/9fO
18:53 PilzAdam dunno, Im not familiar with metatables
18:53 VanessaE bah
18:54 VanessaE way to overcomplicate it
18:54 john_minetest Isn't https://github.com/minetest/minetest/pull/1025 merged yet? Or is that a new one?
18:55 john_minetest https://github.com/minetest/minetest/pull/1024 <- duplicate
19:00 salamanderrake joined #minetest-dev
19:06 ShadowNinja kaeza: What do you think of the above patches?
19:06 kaeza ShadowNinja, look fine
19:07 kaeza no wait
19:07 kaeza the second one is plain wrong
19:08 kaeza `table' is not on scope in alias_metatable.__index (actually, it refers to the `table' global, the table library)
19:10 kaeza wait..
19:12 ShadowNinja minetest.after isn't available yet either, and neither is register_globalstep, and I can't rearange the order. :-(
19:13 ShadowNinja kaeza: Right, fixed.
19:13 kaeza also, __index should be function(self, name)
19:14 ShadowNinja Yes, I forgot the table argument.
19:16 jojoa1997 joined #minetest-dev
19:16 kaeza ShadowNinja, I think the second patch shouldn't be applied
19:17 ShadowNinja kaeza: I don't know how I can do it, I need to set game_started without globalstep.
19:18 kaeza no that first one is OK and should do what's intended
19:19 kaeza by the time the after() callback is called, all mods should have been loaded
19:20 kaeza so setting game_started there is the correct way
19:20 kaeza the second one may cause problems, I'm looking into it
19:25 ShadowNinja kaeza: No, after is in misc.lua, which depends on misc_register.lua. Cyclic dependencies.
19:26 ShadowNinja kaeza: Second meaning the alias_metatable one?
19:26 kaeza yes
19:26 proller joined #minetest-dev
19:26 ShadowNinja That one works correctly.
19:27 kaeza oh well then
19:28 RealBadAngel PilzAdam, https://github.com/minetest/minetest/pull/1037
19:28 VanessaE RealBadAngel: too late.
19:28 VanessaE they're already discussing an alternative.
19:28 RealBadAngel what alternative?
19:29 kaeza ShadowNinja, to solve the cyclic dependency, you could move misc.lua sooner in the load order (in builtin.lua), but dunno if anything else may depend on it
19:29 kaeza oh right... register
19:30 kaeza in any case, what is the point of this patch again?
19:30 kaeza is it worth the effort?
19:32 ShadowNinja kaeza: Easy redefining of items.
19:36 Miner_48er joined #minetest-dev
19:38 [PavelS] joined #minetest-dev
19:38 Jordach hello [PavelS]
19:38 [PavelS] Jordach, Hello
19:42 Taoki joined #minetest-dev
19:45 emptty joined #minetest-dev
19:54 Miner_48er joined #minetest-dev
19:57 VanessaE my/RBA's patches would have taken far less time to put into practice.....
19:57 emptty I have the feeling that mods placed under CUSTOM_SHAREDIR are not found at runtime
19:58 emptty I'm digging to see where the list of mods is parsed during startup. Could someone help me to get oriented in the code?
19:59 sapier thegame
19:59 sapier emptty do you do a run in place build?
20:00 sapier if you do custom shareddir is ignored
20:00 emptty nope, I try to make animal_pack working with the debian package :)
20:00 sapier good idea :-)
20:01 emptty the mobf package works and everything, but the minetest package does not see it for some reason
20:03 sapier strange
20:04 emptty the mods are installed under /usr/share/games/minetest/mods/minetest/mob_guard and such
20:04 emptty I have                   -DCUSTOM_SHAREDIR=/usr/share/games/minetest \
20:05 sapier strange path
20:05 emptty yeah, that's because when the FHS standard was created, back in the eighties, there were no games on unix machines
20:05 VanessaE um...wut?  [12-04 15:04] <VanessaE> position = (-2.14748e+06,-1.63842e+06,-4712)
20:05 VanessaE when is this ever gonna be fixed?
20:06 VanessaE [12-04 15:06] <twoelk> no wonder no client worked
20:07 VanessaE he says he has no idea how he even got there in the first placde.
20:07 VanessaE place*
20:08 VanessaE the only thing I know that can send a person that far is a trampoline and he wasn't near one.
20:14 emptty1 joined #minetest-dev
20:17 VanessaE in other words, why aren't player coords being clamped to +/- 31k?
20:17 VanessaE and don't tell me its impossible.  if you can move the player, you can clamp his coords.
20:22 VanessaE and what about player files being re-saved all the fucking time?
20:22 sapier https://github.com/minetest/minetest/pull/1030 Celeron55 is this successfull dialog ok?
20:25 emptty I should increase debug_log_level to get more info? To what value to get it all?
20:26 sapier I guess only way to find out is using debugger and step
20:30 Calinou joined #minetest-dev
20:32 emptty k, thanks
20:39 EdB joined #minetest-dev
20:40 celeron55 22:07:44 < emptty> the mods are installed under  /usr/share/games/minetest/mods/minetest/mob_guard and such
20:40 celeron55 mods/gamename is not supported anymore, or is it?
20:40 celeron55 i mean, mods/gameid
20:41 VanessaE afaik, no.
20:43 emptty so that should be  /usr/share/games/minetest/mods/mob_guard and such (with  /usr/share/games/minetest/ being path_share) ?
20:45 VanessaE why are you putting mods in the game dir?
20:45 emptty because I dunno what I'm doing
20:46 VanessaE they belong in your homedir
20:46 VanessaE ~/.minetest/mods
20:46 emptty they are in a package, installed with the system
20:46 VanessaE ohh
20:47 emptty (I'm building the package, actually)
20:47 VanessaE well I guess in that case, it's right
20:47 PilzAdam <emptty> so that should be  /usr/share/games/minetest/mods/mob_guard and such (with  /usr/share/games/minetest/ being path_share) ? <- yes
20:47 jojoa1997 joined #minetest-dev
20:49 ShadowNinja emptty: mobf comes as a modpack, so it should be /usr/share/games/minetest/mods/mobf/mob_guard. With a modpack.txt in the mobf folder.
20:50 ShadowNinja That way things stay organized and you don't end up with 500 folders in your mods folder.
20:50 EdB left #minetest-dev
20:53 ShadowNinja What do you think of adding a mod_data dir or similar? It would be used for settings and the like that are not world-specific.
20:56 emptty ShadowNinja: what is the expected content of modpack.txt ?
20:56 PilzAdam nothing, but it shouldnt start with a #
20:56 PilzAdam *nothing special
20:58 OldCoder joined #minetest-dev
20:59 emptty an empty file is ok?
21:00 PilzAdam yep
21:04 sapier1 joined #minetest-dev
21:06 emptty It still does not work for me
21:07 emptty but as soon as I ` ln -s /usr/share/games/minetest/mods/mobf ~/.minetest/mods/ ` then it works
21:07 emptty so, someone is not exploring path_share when looking for the installed mods
21:13 emptty I think that src/script/lua_api/l_mainmenu.cpp line 681 is faulty
21:14 emptty When builtin/modmgr.lua asks for the modpath, the engine only returns the user path
21:15 sapier that's intended
21:15 emptty But that prevents the lua world from retrieving the mods that are in the share path, right?
21:16 sapier if I remember correct mods are read by a different mechanism
21:17 sapier modpath returns user modifyable modpath shared folder isn't modifyable by user
21:19 emptty that's true
21:20 sapier ok you're right we only read mods from that folder ... guess there's a inconsistency to fix
21:21 emptty So the bug is in lua?
21:21 emptty It should check for shared mods too?
21:21 sapier mod and moddependency within minetest are a real mess code paths and behaviour is spread throughout multiple files redundant and inconsistent
21:21 emptty somewhere in modmgr.preparemodlist
21:22 sapier imho minetest core should provide those information consistent for game startup as well as mainmenu but that's not a quick fix
21:24 EdB joined #minetest-dev
21:24 emptty I must confess that I'm bit lost. I couldn't even fill a useful bug here
21:25 sapier chances this will be fixed for 0.4.9 are minor but as I need that code for modstore dependency handling too it'll be most likely fixed in 0.4.10
21:26 emptty ok
21:26 emptty I'll try to think of a solution on my side
21:27 emptty I have the feeling that extending modmgr.preparemodlist with a call to get_mods("/usr/share/games/​minetest/mods",global_mods) could help, regardless of how ugly it is
21:28 sapier for your package maybe but a hardcoded path wont be added to minetest
21:28 emptty I'll try later, I must now sleep
21:28 emptty sure, that would be a local and temporary fix, waiting for  a better solution
21:29 zat1 joined #minetest-dev
21:29 emptty thanks for your help
21:29 john_minetest Wait... 0.4.10 comes after 0.4.9?
21:29 john_minetest duh.
21:29 sapier of course ;-) 9 10 11 12 .. ;-P
21:30 john_minetest Forgot that it was <never_changes>.<always_four>.<major>
21:30 sapier there's no rule version numbers are expected to be single digit
21:31 Calinou we'll forever be at 0.4 :/
21:31 Calinou we're increasing _patch_ versions with very visible changes... does not make sense
21:31 sapier no once we see no other way as breaking map format we will switch to 0.5 ;-)
21:32 sapier we're not breaking api
21:32 Calinou increase major version number for that...
21:32 Calinou it's not a shame to have a version number like "4.8"
21:33 sapier we're far from a serious 1.0.0 ;-)
21:33 Calinou not really?
21:33 sapier and imho majors are to be changed by celeron
21:34 sapier and switching just because some people are frightened about an additional digit isn't a reason ;-)
21:35 kaeza I can see a point in what Calinou said
21:35 celeron55 everyone has a different view on what the version number should reflect
21:35 Calinou all the core devs are on the same level
21:35 Calinou even celeron55 himself said that he is not the ultimate dev :P
21:35 kaeza the fact that it's mineTEST plus it being version < 1.0 makes it feel unpolished
21:36 Calinou it doesn't frighten me, but it does not make the most sense to increase patch versions for visible changes
21:36 celeron55 kaeza: it's unpolished even if you called it mineprofessional 47.0
21:36 kaeza lol yes
21:37 celeron55 i think there is no valid reason to raise the minor version until something meaningful is changed
21:37 celeron55 eg. a change of gui toolkit would be a valid reason
21:38 celeron55 or, really, we might have to do what linux did
21:38 john_minetest Actually 1.0 isn't that far away imo.
21:38 celeron55 just see that "oh we're just doing this incremental stuff without problems, call it x.0 then"
21:39 john_minetest As long as we don't do a 0.5 -> 0.9.9 jump and then do tons of "0.9.9.april_patchlevel_23" it should be okay.
21:40 john_minetest Version numbers are basically percentage of completeness between 0.0.1 and 1.0.0. Above that it is only an indicator "We are still working on it and there is a new version."
21:40 celeron55 some people are going to interpret 1.0 as "it's now finished to an extent that it can be critically reviewed"
21:40 john_minetest That should be the goal.
21:41 sapier minetest is very very far from finished
21:41 john_minetest 1.0 should be stable and feature everything that is needed for the game. Further features can come later.
21:41 sapier we're basicaly rewriting core parts atm
21:41 PilzAdam john_minetest, *for the engine
21:42 john_minetest PilzAdam: Yeah. When the engine has all needed features, the game can be developed with minetest_game updates. The engine will evolve, but not critical.
21:43 celeron55 um
21:44 celeron55 saying anything like that is just overly idealistic
21:44 OldCoder joined #minetest-dev
21:44 john_minetest Look at linux kernel.
21:44 sapier imho as long as we talking about critical design issues within core there's no chance for 1.0
21:44 john_minetest 1.0 was a release when X11 worked quite well and it was so advanced that it could  be used for productive work.
21:44 PilzAdam there is no "Minetest" game, there is only Minetest the engine, and then that thing "minetest_game"
21:45 celeron55 getting the minetest engine to "feature completeness" will take like 1000 years with current resources
21:45 EvergreenTree joined #minetest-dev
21:46 celeron55 john_minetest: if you think of it like that, then minetest should have been 1.0 some three years ago
21:46 sapier imho once we managed to remove env/maplock and implemented client side lua we're close to a first 1.0
21:46 john_minetest When the engine works and the code is nice and the features are so complete that there could be new ones added without changing the whole engine, it should be time for 1.0
21:46 Calinou Minetest will never be finished, so that's normal
21:47 kaeza is software in general ever finished? that's the greatest fallacy in the whole world
21:47 celeron55 kaeza: it is if you never change the goals
21:47 PilzAdam kaeza, my HelloWorld.java is finished
21:47 celeron55 but the goals of minetest are not fixed
21:47 john_minetest Linux kernel will never be finished. 1.0 was the release where it was really usable with sockets and everything. Now we have 3.9.12 or something.
21:47 celeron55 and probably never will be
21:48 sapier official linux kernel version is 3.12
21:49 sapier there's no third digit in "calling name" of kernel
21:49 john_minetest When we have a new <always_four> version every year, we would have 5 years of development until 1.0
21:49 celeron55 john_minetest: it's called the minor version
21:49 sapier 0.10
21:49 sapier 0.11
21:50 sapier 0.12
21:50 celeron55 lol
21:50 sapier there are a lot of always four numbers :-)
21:50 celeron55 minetest 0.210.4
21:50 kaeza 0.1337
21:50 celeron55 the minor never ends!
21:50 kaeza that will be teh release
21:51 celeron55 but really instead of 0.210.4 it's probably going to be like 0.4.1890
21:51 john_minetest In 40 years we will have 0.4.489
21:51 kaeza maybe in like two universe timespans or so
21:53 celeron55 anyway; it's stupid to set goals like "client-side scripting = 0.5" when nobody is even working towards that
21:53 john_minetest * Assuming we put a new "major- minetest style" up every month
21:53 john_minetest Goals shouldn't be that fixed.
21:54 celeron55 john_minetest: they should, because otherwise you don't know when to change the version
21:54 john_minetest Goals should be set on a shorter view.
21:54 john_minetest Like the milestones for 0.4.9 on github.
21:55 john_minetest https://github.com/minetest/minetest/is​sues?direction=desc&amp;milestone=2&amp​;page=1&amp;sort=created&amp;state=open
21:55 sapier waht about setting goals that will result in a minor increase ? this way it's not set that e.g. client side lua is 0.5 it could be 0.6 too
21:55 celeron55 sapier: so that adding any one of them will cause a minor increase?
21:56 sapier that's what I meant yes
21:56 celeron55 and then a new minor increase when some other of them is added later
21:56 john_minetest I think it would be more productive to say: "Hey, we need client side lua" and then wait. When somebody starts it it is put in one of the next releases.
21:56 celeron55 it would work
21:56 celeron55 you can make a suggested list of them
21:56 john_minetest okay, also would work that way.
21:57 celeron55 john_minetest: the issue in what you are saying is that minor versions are *never* scheduled
21:57 celeron55 only patch versions are scheduled like that
21:57 john_minetest My point is, that in an open source project, it is impossible to say: "0.5.0 is when we have client side lua". When nobody cares about it we would be stuck in 0.4.x forever.
21:58 sapier I assume finished mapgen v7 might be worth a minor increase too
21:58 john_minetest And when there is 0.9 and everything implemented for 1.0, there might be some feature that is not in that list but is actually another milestone.
21:59 celeron55 anyway, it makes sense to say "minor version will be increased in any release that adds any of these: client-side lua, completely new map format, new gui toolkit, replaced 3d engine"
21:59 sapier then there will be a 0.10 john_minetest
21:59 john_minetest What exactly is major and minor in our case?
21:59 NakedFury joined #minetest-dev
21:59 sapier that's what we seem to define right now
21:59 celeron55 also on that list: tcp protocol, voxel-area entities
22:00 sapier removed env lock
22:00 john_minetest I mean the names.
22:00 john_minetest <major>.<minor>.<subminor>?
22:00 sapier <major>.<minor>.<patch> ?
22:01 john_minetest oh, okay.
22:01 sapier or <major>.<alwaysfour>.<minor> ;-P
22:01 celeron55 it's major.minor.patch
22:01 john_minetest Okay.
22:02 celeron55 this explains how to use them best for public APIs: http://semver.org/
22:02 celeron55 not directly applicable to a program
22:02 john_minetest between 0.3 and 0.4 there are incompatibilities.
22:03 john_minetest anyways
22:03 john_minetest this cake is great. It's so delicious and moist.
22:03 sapier Guess we should add a wiki page with goals that will trigger a minor increase
22:04 john_minetest Imo it would be better to specify goals for the NEXT minor increase.
22:04 celeron55 john_minetest: no, because we don't know what people get inspired to do
22:05 celeron55 it's impossible to force work to be done on any single large feature
22:05 celeron55 people just do what they care about
22:05 sapier and someone may even write something major worth we haven't even thought about
22:05 john_minetest If we specify 90 minor release triggers, we would have to do stuff like 0.28.1 which would be confusing
22:05 celeron55 john_minetest: it's fine
22:05 sapier 0.28.1 is perfectly ok
22:06 celeron55 altough by that point i would have declared it to be 1.0 for sure
22:06 celeron55 i mean, i'd basically consider well-working voxel area entities to be worth 1.0
22:06 john_minetest I see those version numbers as percentage of work done. 0.5 means 50%. 0.28.1 might mean 28% done. Which is far less since it could become 1.0 every second.
22:07 celeron55 john_minetest: they don't mean that; stop thinking that
22:07 celeron55 john_minetest: it's ridiculous that i even have to say this to you
22:09 celeron55 versions are meant for keeping track of things; not for giving an impression of something to people
22:10 celeron55 have you noticed that when you buy a game in a game series like "LOL GAME 3", it's actually LOL GAME 3 version 1.0.3.a.asd_0vd9suv
22:10 sapier besides cacluclation of completion state is usually wrong to a degree it's not usefull at all
22:10 sapier calculation
22:13 john_minetest When there is "Version 0.1.48" people might think it is less advanced as a "Version 0.4.9" although it might be the same version with different numbering shemes.
22:15 sapier so because of firefox is version 102.1 we need to switch too?
22:15 john_minetest My intention is to make minetest versioning more like percentage+patchlevel.
22:16 john_minetest When using minor versions with care it should be possible without problems.
22:17 john_minetest People are used to decimal system. When the patchlevel is more than 10, ok. But when a version is 0.9.9 the next logical step would be 1.0.0 and not 0.10.0
22:17 john_minetest It is all very confusing.
22:17 celeron55 when you're thinking how people perceive your versions instead of doing things that people actually care about, you're doing it wrong
22:18 john_minetest So you see versions primarily as help for better package management and less to show: "We are about that far"?
22:18 Ritchie joined #minetest-dev
22:19 celeron55 john_minetest: we don't know how far we are
22:19 sapier john that's exactly what a version is ... we don't even know what is 100%
22:19 john_minetest A lot of projects don't know.
22:19 john_minetest :/
22:19 john_minetest Hmmm...
22:20 celeron55 if i knew how far we were, i'd version minetest as "minetest X%"
22:20 sapier I don't know how long wine was <1 ;-P
22:20 celeron55 but we don't, so i don't
22:21 iqualfragile joined #minetest-dev
22:21 john_minetest Well, you are the core devs, it's up to you. Anyways it would be great to have a wiki site with important things that have to be made.
22:22 john_minetest I am thinking about getting into coding again.
22:22 celeron55 http://www.nothings.org/stb_vorbis/#history
22:22 celeron55 there's some versioning for 'ya
22:23 john_minetest That is exactly what happened to linux.
22:23 celeron55 i think that's a cool way to do it
22:23 john_minetest They had sockets and with it X so they did a jump from 0.5.x to 0.9.x and failed.
22:24 john_minetest Yeah, when we do it like that we will be following a tradition of OSS projects. So let's just jump to the 0.9.0 right now ! ;)
22:29 john_minetest So much stuff in the todo-queue
22:30 sapier celeron did you have time to see the new "successfull dialog"?
22:35 deltib joined #minetest-dev
22:36 VanessaE ShadowNinja, kaeza, what is the disposition of #225/#1037?
22:38 kaeza VanessaE, node2 is global, a `local' should be added (first line of func)
22:38 kaeza (#1037)
22:39 NakedFury joined #minetest-dev
22:39 kaeza also `node'
22:39 VanessaE oh
22:39 VanessaE so you aren't gonna do the other method you and ShadowNinja were discussing?
22:41 ShadowNinja Well... It was impossible the way that I was doing it.
22:41 VanessaE ok
22:43 kaeza VanessaE, also, #225 has some issues with indentation after the clone_node() call
22:43 kaeza other than that I don't see anything wrong
22:43 VanessaE that indentation is intentional.
22:44 VanessaE I use the clone_node() call as the opening of the stanza, and the register_node() call as the closer, as in a begin/end
22:45 kaeza oh okay then :P
22:47 bas080 joined #minetest-dev
22:49 deltib joined #minetest-dev
22:56 Ritchie joined #minetest-dev
23:32 john_minetest I just had the best error message EVER: "00:32:40: ERROR[main]: Something failed"
23:32 john_minetest after about 1 hour of media loading, OC's server.
23:32 john_minetest cya
23:59 sapier left #minetest-dev

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