Minetest logo

IRC log for #minetest-dev, 2014-01-22

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

All times shown according to UTC.

Time Nick Message
00:23 diemartin what's your opinion on adding a "crouch" button?
00:24 diemartin basically, it only moves the camera to half the height, and reduces speed a bit
00:24 diemartin I have a simple (WIP) implementation: https://github.com/kaeza/minetest/tree/crouching
00:25 diemartin obviously, I'd also like to get a proper "crouching" animation for the player
00:26 diemartin any comments?
00:44 iqualfragile diemartin: there is sneaking allready, do you add the crouching to the sneaking function or create a new thing?
00:44 diemartin it's a new thing
00:44 iqualfragile why?
00:46 diemartin beeecause I like it that way?
00:46 diemartin sometimes you want to walk slowly without crouching (sneak), while sometimes you may want to crouch
00:47 diemartin (as a side effect it gives one more button for mods to do special things)
00:47 iqualfragile thats confusing for users and in my opinion stupid, as those functionalities are quite similar
00:48 iqualfragile just camera lowering for sneaking would be too litle?
00:48 VanessaE I always viewed "sneak" as "slow", not anything at all like crouch
00:49 diemartin iqualfragile, [21:46:50] <diemartin> sometimes you want to walk slowly without crouching ...
00:49 VanessaE having two separate controls could be useful
00:49 iqualfragile i misstranslated it a bit, sorry
00:50 pitriss joined #minetest-dev
00:50 iqualfragile adding another key would make things a bit too complicated i thing but why not make double-sneak into crouching?
01:02 VanessaE G*d no!
01:02 VanessaE none of this double-tap shit1
01:02 VanessaE does no one realize how much of a pain in the ass that makes the control system, in practice?
01:02 VanessaE (e.g. the double-tap-space-to-fly proposal from a while back, for example)
01:03 Weedy_lappy joined #minetest-dev
01:03 Weedy_lappy joined #minetest-dev
01:31 RealBadAngel about "meters" problem... what if i add a field with conversion ratio and a suffix ? mods could define their own systems
01:32 RealBadAngel based on language or whatever
01:38 salamanderrake joined #minetest-dev
01:55 us_0gb joined #minetest-dev
02:02 VanessaE meh
02:02 VanessaE just keep it as meters
02:03 VanessaE everyone in the US understands this as ~3 feet
02:03 VanessaE (yes yes it's a bit over, but that doesn't matter here)
02:05 us_0gb THe world is on the metric system. We United Statesians need to get over it already and make the switch.
02:06 us_0gb Metric in a much cleaner system anyway.
02:07 us_0gb Also, in the case of Minetest, having it in meters really tells us how many nodes away something is without any math. Having it in feet would be far less useful.
02:15 paramat joined #minetest-dev
02:19 diemartin wasn't that lighting issue with caves/floating islands fixed ages ago?
02:21 paramat Concerning waypoints in meters, see what hmmmm said recently, nodes are only extremely roughly 1m, there should be no unit because the unit is obvious and there is less visual clutter, all you need is distance in nodes stated as an integer
02:23 paramat which lighting issue?
02:24 VanessaE paramat: random mapblock-sized sqares of darkness appearing in already-generated land
02:25 VanessaE e.g. places that were normally lit suddenly and inexplicably going dark
02:25 VanessaE as in pitch black, no light at all
02:25 VanessaE diemartin and I just fended off about a dozen such black spots on a brand new (as in less than 2 days old) map that runs vanilla minetest_game.
02:26 paramat aha ...
02:26 VanessaE and these are not those thin, spidery dim spots, but square, pitch black spots.
02:27 us_0gb I wasn't here, I didn't see what was said. I'm not sure how I feel about measurements without units. "A distance of five (full stop)" doesn't make sense. "Five units", "five nodes", "five meters", all would be equally fine.  But with or without units or measurement, it's an awesome feature.
02:28 VanessaE "five clicks"
02:28 VanessaE but that usually means something akin to km or miles or something
02:28 VanessaE "steps" maybe?
02:29 us_0gb Clicks are fine. Kilometers are fine. I guess miles are sort of okay. Steps might work. Just .... any sensical unit of measure.
02:30 paramat blocks or nodes, not metres, MT is not a strict representation of a real world, its an abstract virtual world, therefore scaleless
02:35 us_0gb That's a very good point, paramat! Maybe a mixed notation would be good later on too. A block is sixteen nodes, so maybe a distance might look like "1533 blocks 7 nodes". Then again, maybe such a representation is only present in the broken US system. Come to think of it, I haven't seen it anywhere else ...
02:38 paramat oh i forgot about mapblocks, personally i would have 'WP3', and below distance as an integer plus unit 'n' for nodes. spelling out 'Waypoint 3' is not needed and just clutter
02:41 zat joined #minetest-dev
02:44 us_0gb I prefer to have things spelled out, personally. In another channel, someone just mentioned "FSM". I thought they meant "finite state machines", but they meant "flying spaghetti monster". WAY different concepts. But the waypoint name seems to be set  by the user anyway, so it can be set to whatever we like.
02:47 deltib joined #minetest-dev
02:47 diemartin left #minetest-dev
02:48 diemartin joined #minetest-dev
02:50 SmugLeaf joined #minetest-dev
02:50 paramat ah i misunderstood, its text set by the user thats cool :)
02:50 SmugLeaf joined #minetest-dev
02:57 RealBadAngel so we stay with meters or what?
02:58 ShadowNinja Just the distance without any designation or "nodes" I think.
02:59 RealBadAngel that would feel stupid imho
02:59 ShadowNinja The units are rather obvious if they figre out that it's the distance.
02:59 ShadowNinja +u
02:59 RealBadAngel i have another idea
03:00 RealBadAngel move waypoint name to "name" field and fill "text" field with unit
03:00 ShadowNinja Pushing soon: http://ix.io/a2v
03:00 RealBadAngel so modder will put there any unit he wants to
03:01 ShadowNinja That could be done, with a default of course.  But then you could have multiple wayboints using different words for the same thing, which would be even more confusing.
03:02 RealBadAngel wonder if user would have multiple waypoints mods
03:02 RealBadAngel propably one of aviable, no all of them
03:02 RealBadAngel so he will have one system
03:03 ShadowNinja There's nothing preventing it.  And the admin might install, eg, a spawn waypoint adder in adition to a personal waypoint mod.
03:03 RealBadAngel i am doing now waypoints for UI, im almost done with it
03:04 RealBadAngel ofc, like theres nothing preventing mods adding zilions of picaxes
03:04 NakedFury just call them nodes instead of meters or feet or miles or whatever
03:05 RealBadAngel that will rise question again "why the fuck not meters or clicks??"
03:05 RealBadAngel let the modder decide is best solution imho
03:05 NakedFury well then in crude and rude terms fuck everyone and call it meters.
03:06 NakedFury EVERYONE understands 1 node is ameter. its expected
03:06 RealBadAngel you know whats funny in this?
03:06 NakedFury you went never cared to rename it from meters?
03:06 RealBadAngel "meters" problem was rised by europeans, americans seems to be ok with it
03:07 RealBadAngel rotfl
03:07 NakedFury true
03:08 RealBadAngel another solution could be to do it another way, hold on for a second
03:09 ShadowNinja Nodes could mean something significantly different than a meter in the future.  For example if someone made a mod that made the player bigger and made the scale of the mapgen bigger so that you could build things more precisely, nodes would seem smaller.
03:09 paramat yes, people understand the node as being roughly the universal human unit: the stride, yard, metre
03:10 bas080 joined #minetest-dev
03:11 NakedFury for the record meter is the sort of universal used distance unit for games.
03:11 ShadowNinja diemartin: I would lake if sneaking lowered the camera.  Although a sit option would be nice, IIRC the model already has a animation for it.
03:11 ShadowNinja like it if*
03:12 RealBadAngel http://i.imgur.com/IFej7fF.png
03:12 RealBadAngel im not sure if thats better, too much unnecesary text
03:13 ShadowNinja Seems good.  Does it have TTF support?
03:13 diemartin ShadowNinja, I'll take a look at it tomorrow. I'm off to sleep.
03:13 diemartin night
03:13 RealBadAngel with many waypoints set, and lookin in their direction this will cause mess
03:13 RealBadAngel m is short and good
03:14 ShadowNinja G'Night.
03:14 RealBadAngel night
03:15 RealBadAngel ShadowNinja, sneakin would require another animation, like crouching
03:15 RealBadAngel and then lower player size to fit 1 node
03:16 RealBadAngel and lower the camera then
03:17 RealBadAngel but hey, that would be crouching not sneakin in fact :)
03:17 RealBadAngel hmm, i like that
03:41 Miner_48er joined #minetest-dev
03:51 BrandonReese joined #minetest-dev
04:11 darkrose joined #minetest-dev
04:22 paramat left #minetest-dev
05:16 Zeitgeist_ joined #minetest-dev
05:16 Zeitgeist_ joined #minetest-dev
05:24 Zeitgeist_ joined #minetest-dev
05:24 Zeitgeist_ joined #minetest-dev
05:45 Gethiox joined #minetest-dev
06:19 damiel joined #minetest-dev
06:19 damiel joined #minetest-dev
06:52 ImQ009 joined #minetest-dev
07:59 Anchakor_ joined #minetest-dev
08:12 werwerwer_ joined #minetest-dev
08:34 OldCoder joined #minetest-dev
09:25 mrtux joined #minetest-dev
09:29 ImQ009 joined #minetest-dev
09:33 rsiska joined #minetest-dev
09:44 proller joined #minetest-dev
09:55 xyz joined #minetest-dev
09:55 xyz|imaginary joined #minetest-dev
10:21 troller joined #minetest-dev
11:07 RealBadAngel2 joined #minetest-dev
13:08 NakedFury joined #minetest-dev
13:12 zat joined #minetest-dev
13:21 hmmmm joined #minetest-dev
13:33 iqualfragile joined #minetest-dev
13:36 ImQ009 joined #minetest-dev
13:43 b4n joined #minetest-dev
13:44 b4n left #minetest-dev
13:50 proller joined #minetest-dev
13:55 ImQ009 joined #minetest-dev
14:30 VanessaE christ in a cartoon, CAN WE PLEASE get a proper /kick command?
14:30 VanessaE this ban/unban shit is getting old
14:40 PilzAdam joined #minetest-dev
14:47 xyz did you open an issue?
15:02 VanessaE I should need to open an issue for THAT?
15:06 xyz yes
15:06 xyz or do you expect everyone to read all irc logs?
15:09 VanessaE um...
15:10 VanessaE isn't that the standard expectation here?  and isn't that the sort of command that pretty much anyone who runs a server eventually finds that they need?  and that virtually every one of us here has used at least once on IRC?
15:11 PilzAdam VanessaE, you should write all that in the issue
15:13 bas080 joined #minetest-dev
15:16 Yepoleb joined #minetest-dev
15:19 xyz yes, what's obvious to you is not obvious to others
15:19 xyz * could be
15:22 nore joined #minetest-dev
15:55 Jordach joined #minetest-dev
16:24 VanessaE joined #minetest-dev
16:41 kaeza joined #minetest-dev
16:43 Gethiox joined #minetest-dev
16:58 sapier joined #minetest-dev
17:12 Calinou joined #minetest-dev
17:23 rubenwardy joined #minetest-dev
17:29 rsiska joined #minetest-dev
17:33 zat joined #minetest-dev
17:38 20WAAX6E3 joined #minetest-dev
17:40 nore joined #minetest-dev
17:40 20WAAX6FZ joined #minetest-dev
17:41 Evolykane joined #minetest-dev
18:08 VanessaE joined #minetest-dev
18:38 grrk-bzzt joined #minetest-dev
20:24 Evolykane joined #minetest-dev
20:32 EvergreenTree joined #minetest-dev
20:32 EvergreenTree joined #minetest-dev
20:45 sapier1 joined #minetest-dev
20:50 proller joined #minetest-dev
20:53 RealBadAngel https://github.com/minetest/minetest/pull/1112
20:53 RealBadAngel i have updated waypoints with customizable suffixes. now it can be anything or even blank.
20:55 sapier1 I like the waypoints, but why does this have to be a hud element?
20:55 RealBadAngel because IT IS generic HUD element?
20:56 sapier1 let me be more precise, e.g. if I have a path with lets say 5 waypoints, and I want to show this to 5 players, a mod needs to handle 25 different waypoints
20:56 sapier1 make it 2 paths and 10 players and it's 200
20:56 RealBadAngel you dont get the idea of them, better try it
20:57 IceCraft joined #minetest-dev
20:57 sapier1 I use waypoints in mobf for guards, and usually I have about 10-20 of waypoints for a single path
20:57 RealBadAngel we are talkin bout totally different things...
20:58 sapier1 do we? I have implemented a way to show those waypoints in map ... not with this cool distance feature
20:58 RealBadAngel my idea is to show on screen distance to defined world position
20:58 RealBadAngel position on screen will change
20:59 xyz lol @ customizable suffix
20:59 sapier1 don't get me wrong I like those waypoints they're quit usefull for some cases.
20:59 RealBadAngel so you have to look in waypoint direction to see it
20:59 RealBadAngel and you can easily go into its direction
21:00 sapier1 but when I think about managing those waypoints for 10-20 players each showing a 20 point path it's a nightmare
21:00 IceCraft joined #minetest-dev
21:00 IceCraft joined #minetest-dev
21:00 RealBadAngel xyz, i had to. europeans were screaming that americans would be confused with "meters" ;)
21:01 RealBadAngel sapier1, those waypoint do not form any path
21:01 xyz haha
21:01 sapier1 of course not
21:01 sapier1 but It'd be cool to be able to use that distance feature
21:01 RealBadAngel an theyre handled only client side
21:01 RealBadAngel server just sends definition
21:02 sapier1 yes and server has to store it to be able to remove them or do I have to remove all at once?
21:02 ShadowNinja sapier1: The waypoints are per-player intentionally.  A mod will add personal waypoints like "Home".
21:02 RealBadAngel im adding now waypoints tab to Unified Inventory
21:02 ShadowNinja And storing a few hundred doubles isn't that bad.
21:02 xyz RealBadAngel: i think you could've used getScreenCoordinatesFrom3DPosition instead of math
21:02 RealBadAngel you will be able to see how it works
21:02 sapier1 it is
21:03 RealBadAngel xyz, ive tried but i took its code instead and applied some changes to it
21:03 sapier1 ok I'll have a look at it by time
21:03 RealBadAngel most of the code comes from that function
21:03 xyz ah okay then
21:04 RealBadAngel first of all it returned absolute screen coords
21:04 RealBadAngel and i need at some point 0-1 range
21:04 RealBadAngel not yet, but in the future, for minimap
21:05 sapier1 I'd have prefered this to be a Show text above entitiy but I guess this is a completely different usecase
21:06 RealBadAngel what for an entity??
21:06 xyz well i'm not following you now but i guess it's fine
21:06 sapier1 did you ever see the mob lifebars?
21:07 sapier1 they're quite hacky but something like that beeing feature of an entity itself would be cool
21:07 RealBadAngel sapier, first of all, if player were inside a small cave
21:07 RealBadAngel lets say 1x2 nodes
21:07 RealBadAngel how would you show entities with waypoints to him?
21:07 sapier1 yes the entity based things wouldn't help as they'd been hidden
21:09 RealBadAngel exactly
21:09 RealBadAngel and on the other hand we can allow waypoints to have also images in future
21:09 sapier1 the other way round the bound to screen waypoints don't help to show complex paths as screen will be full ... so both are usefull for different things
21:10 RealBadAngel and for example scale the image when player gets closer to a waypoint
21:10 sapier1 I'd consider some usecases for this waipoints to be "cheating" ;-)
21:11 xyz it's time to say something about client-side Lua api!
21:12 RealBadAngel i can say something
21:12 RealBadAngel it would be great to have it
21:13 sapier1 You're welcome to do it xyz, right now I've got enough about beeing criticized for doing work. And client side lua api will be an issue to stand a lot of critizism no matter how you do it.
21:13 RealBadAngel btw, somebody mentioned i should use there try-catch for new hud parameter. its not needed at all.
21:14 xyz i bet you'll be the first one to criticize it for being insecure
21:14 xyz i'm currently busy with reworking protocol
21:14 sapier1 of course if you don't implement it in a ave way
21:14 xyz indeed implementing it in unsafe way would be awful
21:15 sapier1 I'd even stop contributing to minetest on unsave implementation as I'm not willing to take the risk to download viruses from mientest servers
21:15 xyz but in the end it's the definition of "unsafe" that matters
21:16 RealBadAngel simply deny client side mods do anything system related
21:16 sapier1 beeing unsave is for me that there are known intentionally unfixed issues which are able to destroy modify or transmit non minetest related data from client
21:16 Miner_48er joined #minetest-dev
21:17 sapier1 xyz please don't forget to write a enet dissector for enet
21:17 xyz why?
21:17 sapier1 I assume your fixes are based on enet arent they?
21:17 xyz well yeah, it includes switch to enet
21:17 sapier1 ok then I'd like to have a enet dissector
21:18 sapier1 either write it or find it, I haven't found one
21:18 xyz ugh? sorry, but it's you who wants it..
21:19 xyz so why should I write one? so confusing…
21:19 sapier1 good god you suggested enet because it's standard and you can find issues in communication with enet more simple... right now I can't find any issues because I have no way to look into network packets
21:20 sapier1 do you really believe we never ever will have to look at what's really sent from client to server?
21:20 sapier1 but I actually wanted to know something completely different
21:21 xyz dunno, I just don't care atm
21:21 sapier1 I may be able to remove the "conlock" but ... there are two issues
21:21 xyz for now it just works and I believe that guys who wrote enet know their stuff better than me anyway
21:21 sapier1 1) I'm not absolutely sure that none of those widespread server functions aren't called from emerge thread someway
21:22 sapier1 2) if I remove that lock we'll have to revert this in case we ever decide to do lua in a separate thread
21:23 sapier1 yxz have you tried to send lots of data through it while other clients are online?
21:23 sapier1 that's the one issue I haven't been able to test by now
21:24 sapier1 everything else in enet seems to be at least acceptable
21:24 sapier1 tcp is way faster of course but enet isn't as bad as I couldn't live with it
21:24 xyz not sure what you mean
21:25 sapier1 send >30mb of texture data to a connecting client and make sure this doesn't cause lag on already connected clients
21:26 sapier1 I don't know if enet can handle this task, I hope so
21:26 Gethiox2 joined #minetest-dev
21:28 sapier1 come on guys this is your chance to decide about removing a lock, yesterday all of you have been that passionate about to many locks?
21:28 xyz i think some of those stopped caring
21:28 xyz yea, I'll check enet with minetestambience
21:29 sapier1 great the easy way (re)starting flame ware every week and then running away when there's some reponsibility to take
21:29 sapier1 if you have a server I'll pleased to help
21:30 sapier1 do you replace packetformat too?
21:32 xyz yup
21:32 xyz it's work-in-progress and not really stable yet
21:32 sapier1 I know how much work this is ;-)
21:32 sapier1 that's why I didn't do it
21:34 sapier1 and no matter what packet format you implement if you do it right the low level packet format wont make a difference for it
21:35 xyz what do you mean?
21:36 sapier1 my legacy/tcp/enet versions don't have exactly same interface to server. Because of low level packet format doesn't interfere to application layer packet format. I assume your application layer format won't interfere to low level format too
21:37 sapier1 -don't + do
21:39 xyz yea just did a quick test and didn't notice any lag increase when connecting another client without cache to server with ~10 mb of data
21:40 sapier1 sounds good but 10mb on localnet wouldn't take more then a few seconds to transfer it'd be hard to recognize lag
21:40 xyz ah I think I got what you mean
21:41 xyz it's not on localhost
21:41 xyz server is in germany and localhost is in russia
21:41 sapier1 ok not a proof but a quit good hint
21:42 xyz oh well what was this for then?
21:43 sapier1 If enet couldn't have handled this I'd suggested stopping development ;-) ... hope you didn't transfere media per curl ;-)
21:46 xyz ah! i know!
21:46 xyz you're just too pessimistic about it for whatever reason
21:46 xyz probably because you spent a lot of time fixing the broken implementation and don't like your work being thrown away
21:46 sapier1 actually I never intended to merge it
21:46 proller revert it now!
21:47 xyz why did you code it then?
21:47 sapier1 celeron55 wanted it to be merged. My only intention was to proof you're wrong about the protocol itself, minetest protocol isn't bad it's just been a faulty implementation.
21:48 xyz oh?
21:48 xyz well whatever
21:48 sapier1 according to my mesurements enet is only about 10-20% faster, I haven't tried how much of this is caused due to 1500 byte packets instead of 512
21:48 xyz I didn't even look into it so I really can't discuss this subject
21:49 * proller faceplam
21:50 sapier1 Did you just say you didn't even look at it but suggest a switch because it's "bad"?
21:55 sapier1 proller you should learn to live with negaitve side effekts of forking
21:56 proller where is negative?
21:56 sapier1 considering your even more picky then normal behaviour you know yourself
21:57 proller doing something is more positive than doing nothing and speaking about nothing
21:57 proller your mobs crashing every day...
21:58 sapier1 you didn't post a crash issue by now
21:58 sapier1 tell me about it I'm gonna fix as soon as possible
21:58 proller and new bug appears
21:59 xyz sapier1: btw, in regards to https://github.com/minetest/minetes​t/issues/1106#issuecomment-32786161
22:00 xyz what's the problem with xp?
22:00 xyz and in any case, since irrlicht internally uses wchar-d strings we won't get true unicode support anyway it seems
22:01 xyz what we should do is make sure the protocol uses utf-8
22:01 sapier1 didn't know about irrlicht, so there's an additional issue with utf8 usage
22:01 xyz and then decode/encode from/to it whenever needed
22:01 sapier1 protocol?
22:01 xyz yes, protocol
22:01 sapier1 what level?
22:01 xyz what level?
22:01 sapier1 low level protocol doesn't care about what data is sent
22:02 xyz every string we send over net/store persistently shall be assumed to be utf-8
22:02 sapier1 imho this is a packet format issue
22:02 xyz then you see how your question is incorrect
22:02 sapier1 it's perfectly ok to define paket format to contain utf8 data only
22:02 xyz right
22:02 xyz that's what i was talking about
22:03 sapier1 yes on redesigning packetformat I'd encourage and support this completely
22:04 xyz relevant code https://github.com/xyzz/freeminer/​blob/utf8/src/util/string.cpp#L39
22:05 Neological joined #minetest-dev
22:06 sapier1 guess it's gonna be merged automaticaly with your new packet format so no need to do work right now
22:08 xyz do work?
22:09 sapier1 e.g. merge that code
22:10 xyz ah, i doubt it's coming into minetest anyway
22:42 sapier1 left #minetest-dev
23:01 jojoa1997|PC joined #minetest-dev
23:07 jojoa1997 left #minetest-dev
23:08 jojoa1997 joined #minetest-dev
23:57 EvergreenTree joined #minetest-dev
23:57 EvergreenTree joined #minetest-dev

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