Minetest logo

IRC log for #minetest-dev, 2014-03-22

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

All times shown according to UTC.

Time Nick Message
00:08 PilzAdam ~ tell sapier your new chat menu double-posts when clicking the button to send
00:08 ShadowBot PilzAdam: O.K.
00:11 VanessaE well that explains that.
00:24 iqualfragile joined #minetest-dev
00:41 Fresh_me_ joined #minetest-dev
01:36 ShadowNinja PilzAdam: Hey!  I found that bug!  ;-)  I told him but haven't received a response.
01:39 domtron joined #minetest-dev
02:25 spillz joined #minetest-dev
02:26 spillz !tell sapier placing blocks still unresponsive in latest apk (dig works)
02:26 ShadowBot spillz: O.K.
06:10 PenguinDad joined #minetest-dev
06:21 khonkhortisan joined #minetest-dev
06:24 proller joined #minetest-dev
06:35 grrk-bzzt joined #minetest-dev
06:42 diemartin joined #minetest-dev
06:56 alexxs joined #minetest-dev
07:03 smoke_fumus joined #minetest-dev
07:15 salamanderrake joined #minetest-dev
08:35 Anchakor_ joined #minetest-dev
08:55 Anchakor_ joined #minetest-dev
08:59 ImQ009 joined #minetest-dev
09:08 sapier joined #minetest-dev
09:28 Jordach joined #minetest-dev
09:37 sapier http://animalsmod.comuf.com/downl​oads/Minetest-debug_variant1.apk http://animalsmod.comuf.com/downl​oads/Minetest-debug_variant2.apk ... opinions about the crunch button? I prefere the variant where button is next to jump
09:44 CheapSeth joined #minetest-dev
10:33 ShadowBot` joined #minetest-dev
10:45 Calinou joined #minetest-dev
10:56 Saunterer joined #minetest-dev
11:57 Zeitgeist_ joined #minetest-dev
12:20 tomreyn joined #minetest-dev
12:34 RealBadAngel joined #minetest-dev
12:38 rsiska joined #minetest-dev
12:53 EvergreenTree joined #minetest-dev
12:56 nore joined #minetest-dev
13:24 PilzAdam joined #minetest-dev
14:01 lanxu joined #minetest-dev
14:01 celeron55 joined #minetest-dev
14:04 Jordach joined #minetest-dev
14:06 yno joined #minetest-dev
14:06 Jordach joined #minetest-dev
14:07 GhostDoge joined #minetest-dev
14:10 sapier Any opinion about adding a style rule member functions of a class only used by the class itself HAVE TO BE private? I thought this is common sense.
14:11 proller what if in next day you want to use this function outside?
14:11 proller same with variables
14:12 sapier I've never read a more silly question proller, if you want to use it outside you're changeing code anyway so you can make it public too
14:13 proller but its wrong with variables
14:13 sapier but if you wanna use a private function you should think about that twice as those functions shouldn't be touched from outside
14:13 proller you must delete m_ from name
14:13 sapier no it isn't it's very bad style to manipulate class internal variables without telling the class about it
14:13 proller or make stipid accessors
14:13 RealBadAngel hi
14:13 RealBadAngel http://pastebin.com/j13pWmCB
14:14 RealBadAngel ^^ united node shader thx to core changes i made
14:14 RealBadAngel vertex one ofc
14:15 sapier using class variables form outside is allways something to avoid
14:15 proller again your imagined world 8(
14:15 sapier if a class is a storage class only you're right proller but those should be pod stucts anyway
14:17 sapier e.g. hud ... hud class is responsible for drawing all the hud elements on screen, it's got lots of own code doing things with it's member variables ... the specialized draw functions shouldn't ever be called from outside
14:17 sapier as well as it's crap do calculate displaycenter and changed screen size from outside and then tell hud to check if the value you just suplied has changed
14:18 sapier a clean way would be telling hud about current screen size and let it decide itself if something has to be done
14:19 sapier why I'm at hud right now? because hud needs to deal with different screen dpi's on android ... which is impossible if everyone messes around with it's internal variables
14:19 RealBadAngel HUD shouldnt be even aware of current display size
14:20 sapier as I said it has to be aware to handle different dpi's
14:20 Exio4 looks like OOP problems
14:20 sapier unless you wanna define another class beeing aware of screen and dpi which is gonna tell hud about it's element size
14:20 RealBadAngel it should work on arbitral coords which should be recalculated to actual screen proportions
14:21 sapier and how do you wanna recalculate without knowing ? ;-)
14:22 RealBadAngel knowing in sense of makin something, anything inside of HUD code depending on resolution
14:23 sapier true, for time beeing I just wanna fix builtin hud elements, lua based ones can be handled afterwards
14:24 RealBadAngel hud coords to screen coords shall be last step and without any dirty tricks
14:24 sapier as far as I remember lua elements are specified with absolute coords ... which is silly but it's done this way right now
14:24 RealBadAngel so rescaling window shall effect with same positions for all the elements
14:24 RealBadAngel now its fucked up
14:25 hmmmm joined #minetest-dev
14:25 RealBadAngel hi hmmmm
14:25 RealBadAngel hmmmm, take a look: http://pastebin.com/j13pWmCB
14:25 sapier we cannot scale gui elements same way as images
14:25 sapier that's plain impossible in irrlicht
14:25 RealBadAngel we can scale bounding boxes
14:26 RealBadAngel and then fit elements in there
14:26 sapier it's not very helpfull if it's working for some elements but not for others
14:26 RealBadAngel it has to work for all of them
14:26 sapier and btw we don't even wanna scale in all situations
14:27 RealBadAngel 1:1 is also a scaling
14:27 sapier no it's not 1:1 in any case
14:28 sapier sometimes we wanna scale sometimes not
14:28 sapier it's dependent on dpi AND screensize
14:28 RealBadAngel lemme rephrase it. i want same coords system for all the elements
14:28 RealBadAngel now, even labels and textfields have different ones
14:28 RealBadAngel its insane
14:28 sapier you can add this in a second step as I don't wanna rewrite whole hud right now
14:29 sapier that's rewriting formspec too which I wont do for sure anytime soon
14:29 RealBadAngel SN wants
14:29 RealBadAngel with tables instead of strings
14:29 sapier it'd be great if this was true but it's not done this way and fixing it breaks anything
14:30 sapier -any + every
14:30 hmmmm that is one concern i have with my own universal formspec idea - it'd break compatibility with older clients
14:30 sapier you'll have to fix mainmenu ingame menus, any mod around
14:30 RealBadAngel sometimes you have to trash everything in order to get things done in a good way
14:31 hmmmm btw guys did i tell you about the universal formspec idea
14:31 sapier well you're free to do so but it's not what I'm gonna do
14:31 hmmmm formspecs are represented in memory by a set of objects, at the language level
14:31 hmmmm everything else - the stupid string format, lua tables, even json - are just serialized forms of this internal representation
14:31 RealBadAngel sapier, i wont touch that code, i have enough work on gfx ;)
14:32 hmmmm over the wire it'll get seralized to json probably
14:32 hmmmm yea?  no?
14:32 sapier where's the benefit hmmmm?
14:32 hmmmm the idea is that you should be able to specify formspec in any sort of key value structure
14:32 hmmmm sapier, extreme flexibility.
14:33 sapier for what I understood you just add another layer of indirection
14:33 hmmmm there are some new features we'd like in formspecs that you can't really encode in that dumb unflexible string format
14:33 sapier another proprietary gui specification language different but not better then formspec
14:34 hmmmm and you'd be able to make an "open" gui specification language frontend to this as well if you'd like
14:34 sapier why not implement a rendering engine for some already existing gui language?
14:34 hmmmm why implement a rendering engine in the first place
14:34 sapier yes that's a lot of work but it's gonna be same for your language if it's supposed to be as flexible as the existing ones
14:34 hmmmm this entire thing is so fubar
14:35 RealBadAngel guys, UnifiedInventory is pretty complex example of usage of formspec
14:35 sapier you're always free to do what you like hmmmm but I'm not sure if this is worth it ... my own opinion is it's not but this is something everyone has to decide himself
14:35 RealBadAngel even strings handled smart can do miracles
14:36 sapier yes formspec is already quite flexible, I've written mainmenu I trust me ;-)
14:36 hmmmm i disagree because my idea encompasses everybody elses'
14:36 hmmmm it's a strict enhancement
14:36 hmmmm people are free to use whichever frontend language they'd like to specify the gui
14:36 hmmmm sure it might be a little bit more effort but
14:36 sapier no hmmmm it's about half a year of work without guarantee to be better
14:36 hmmmm it's like mainmenu
14:36 hmmmm half a year !?  i don't estimate that long at all
14:37 RealBadAngel heheh
14:37 sapier :-) I thought mainmenu is just a few weeks too
14:37 RealBadAngel hmmm, just code it
14:37 hmmmm yeah but i' malready working on something completely different
14:37 RealBadAngel so make a pause to code it
14:37 hmmmm map overgeneration will solve more problems than just biomes
14:37 sapier so no need to discuss about things noone has time to implement ;)
14:38 hmmmm and it'll pave the way for finishing mapgen v7
14:38 RealBadAngel im doing same things with shaders, when i need a pause i code something else
14:38 RealBadAngel just to refresh my mind
14:39 sapier I wanna finish android port so I can push the final parts of my network fixes ... after this is done and (most) bugs are removed I'm gonna look for mobf again .. :-) then it's most likely end of year
14:39 RealBadAngel have you two read the vertex shader code i pastebined earlier?
14:40 hmmmm yeah
14:40 RealBadAngel i do have now complete control over the shaders :)
14:40 hmmmm it looks like shader code
14:41 hmmmm did you want me to look for anything in particular?
14:41 RealBadAngel you should notice one thing: its a node shader, one for all
14:42 RealBadAngel one code, many instaces
14:42 RealBadAngel no more copy pasta :)
14:43 sapier great! :-)
14:43 hmmmm i just had pasta last night
14:43 RealBadAngel remember that taste, it wont happen again ;)
14:44 hmmmm i thought it was pretty tasty
14:44 RealBadAngel hmmm, have you tried playin with generated normals?
14:44 hmmmm no..
14:44 hmmmm i never did any graphics related things aside from the alpha shader
14:45 RealBadAngel just try it
14:45 RealBadAngel game with even default 16x textures looks like from another century
14:46 RealBadAngel im fuckin proud of the final effect
15:01 Zeitgeist_ joined #minetest-dev
15:03 alexxs joined #minetest-dev
15:04 EvergreenTree joined #minetest-dev
16:32 NakedFury joined #minetest-dev
17:01 Calinou joined #minetest-dev
17:13 nore can I push that commit? https://github.com/Novatux/minetest/commit​/859fd3102f55bc7cf1d3fdc18d8c1977562af748
17:15 sapier seems to be fine ... nore can you try to fix the digging in my android branch? I don't have an idea how to adapt it to your offset changes
17:15 nore sapier, do you have a link to the relevant code please?
17:18 sapier https://github.com/sapier/minetest/b​lob/android_3/src/touchscreengui.cpp L 471 ... this did give a valid shootine before but now it doesn't seem to match the offseted one. It's used at game.cpp L 2715 ... the way I tried to fix it there doesn't work
17:24 nore how do you change a shootline?
17:25 nore btw: how do we do when there are license violations?
17:25 nore s/how/what
17:26 EvergreenTree joined #minetest-dev
17:26 PenguinDad joined #minetest-dev
17:28 nore sapier, could you try shootline.start += intToFloat(camara_offset, BS), and same with shootline.end?
18:07 celeron55 what about we add a page on either the main site or the wiki that very shortly describes the history of minetest and then explains all the versions out there
18:08 celeron55 it would be handy to link such a one to people that are wondering about updating a 0.3.1 or aren't satisfied with the 0.4 series or whatever
18:11 celeron55 like "development of 0.3 was ended on <date> and now we have 0.4 which your linux distro might not have in it's repository; you should do this and that to get it; alternatively there is minetest-classic and freeminer that attempt to do things X and Y"
18:11 celeron55 (with better structure and a bit more information than on that line)
18:12 celeron55 something that would answer and come up as the first result for the google search "minetest versions"
18:13 celeron55 i guess i'll do it if nobody is interested; i wonder on what site it should be in
18:22 e1z0 joined #minetest-dev
18:33 e1z0 joined #minetest-dev
18:53 rubenwardy joined #minetest-dev
19:56 werwerwer_ joined #minetest-dev
19:59 kaeza joined #minetest-dev
20:25 PenguinDad joined #minetest-dev
20:42 EvergreenTree joined #minetest-dev
20:54 grrk-bzzt joined #minetest-dev
21:00 domtron joined #minetest-dev
21:05 CheapSeth joined #minetest-dev
21:37 Amaz joined #minetest-dev
21:38 PenguinDad joined #minetest-dev
21:47 hax404 joined #minetest-dev
22:24 Amaz joined #minetest-dev
22:47 casimir joined #minetest-dev
23:08 EvergreenTree joined #minetest-dev
23:21 rsiska joined #minetest-dev
23:52 sapier joined #minetest-dev

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