Time  Nick           Message
00:36 VanessaE       is it just me, or did the buttons in the Esc/pause menu get just a tad larger after these last few commits?
00:36 VanessaE       (as well as some fields in the main menu)
00:43 VanessaE       I will say this much for sure;  those latest patches definitely cut down the memory usage by a large margin.  and if they don't, then something else recently did.  9GB is MUCH better than 13 GB.
00:52 VanessaE       I can drop that to 7.2 GB or so by disabling bumpmaps and aniso/mipmap, but that still leaves about 2GB that's unaccounted for
01:15 Taoki          Heh, I can literally see the world loading faster in singleplayer with the latest improvements. Awesome work sapier
01:18 VanessaE       unfortunately the fps improvement doesn't hold for long when playing on my Creative server with HDX 512, but it's a definite improvement.
01:18 VanessaE       probably being limited by my GPU at this point - there's only so much memory shuffling it can do I guess.
10:24 Calinou        http://0bin.net/paste/xB+uCkOEFgMV-Co8#iBa31qFArNOOS1anyTOdMSNdXKR5o-OaR3300bZAD6x → Crashes in Carbone but not minetest_game, as of latest Git.
10:24 Calinou        also, smooth lighting is just broken.
10:27 Calinou        HEAD~5 works, smooth lighting is fixed.
12:50 sfan5          pushing in 5 minutes: http://sprunge.us/SLVf?diff
13:42 troller        sfan5, DONT!
13:42 troller        move it to porting.
13:43 sfan5          you're almost one hour too latew
13:43 sfan5          -wq
13:43 sfan5          s/.*//
13:43 sfan5          also cguittfont is not a minetest component, making it depend on porting is not a good idea
16:39 Calinou        I still segfault when loading Carbone :/
16:43 RealBadAngel   any logs?
16:45 Calinou        I posted a backtrace of gdb this morning
16:45 Calinou        http://0bin.net/paste/xB+uCkOEFgMV-Co8#iBa31qFArNOOS1anyTOdMSNdXKR5o-OaR3300bZAD6x
16:45 Calinou        minetest_game loads fine, but smooth lighting is quite broken
16:45 Calinou        I compiled HEAD~5 again, I still crash
16:46 Calinou        so, we can deduce that the commit that crashes is: “Don't keep temporary images in RAM”
16:46 Calinou        (probably)
16:47 RealBadAngel   here carbone loads without any problems
16:47 RealBadAngel   just downloaded it and tried
16:49 Calinou        trying HEAD~8… this morning, HEAD~5 worked
16:50 Calinou        that one works
16:50 Calinou        HEAD est maintenant à f8522d5 Fix server assert in case of invalid message from client, just kick that client.
16:50 Calinou        that one
16:52 RealBadAngel   only thing i noticed that it crashes when going back to main menu
16:53 RealBadAngel   not always, but it happens
16:53 Calinou        segfaults are so frequent when going back to main menu…
16:54 RealBadAngel   yup
17:07 RealBadAngel   looks like ive fixed extrusion code
17:07 RealBadAngel   it started to work with any resolution textures
17:21 sfan5          Calinou: without a debug build the backtrace is pretty useless and "bt full" is better than "bt" for backtraces
17:21 Calinou        ok, thanks
18:02 RealBadAngel   [  0%] [  0%] Built target cguittfont
18:02 RealBadAngel   mo update
18:02 RealBadAngel   hmmz, shoulnt it be "no update" ? :)
18:09 sapier         don't you know mo?
18:10 RealBadAngel   i just checked, mo is here ;)
18:12 RealBadAngel   btw, https://www.youtube.com/watch?v=C9M596KW1dw
18:12 RealBadAngel   fixed long time broken extrude code
18:13 sapier         what's been the bug?
18:13 RealBadAngel   you gonna rotfl
18:13 sapier         return prior doing it?
18:14 RealBadAngel   wait a sec, i will make PULL for it :)
18:19 RealBadAngel   https://github.com/RealBadAngel/minetest/commit/a8ffb487048a7fa5f3f07522414e37c8670fba0c
18:22 RealBadAngel   you should learn how to make proper commits full of  content, not like 1,5k of whitespaces ;)
18:22 sapier         well if you guys wouldn't initially add that much nonsense I'd not have to fix it
18:23 sapier         but as I'm annyoed about that discussion I'm just gonna tell my ide not to fix broken whitespaces any longer
18:26 RealBadAngel   just jokin, calm down :)
18:26 RealBadAngel   can i push that BIG fix o' mine?
18:27 proller        sapier, you can fix broken whitespaces
18:27 RealBadAngel   lol
18:27 proller        BUT MAKE IT IN SEPARATE COMMITS
18:27 proller        DO NOT MIX IT WITH CODE
18:28 proller        ^ very hard to get ?
18:28 RealBadAngel   since when do you care bout mt?
18:28 celeron55      i must say proller is actually right
18:28 proller        https://github.com/minetest/minetest/network
18:29 RealBadAngel   i dont say hes right or not, just asked what happened that he cares
18:29 proller        ^ answer in link
18:29 celeron55      proller merges stuff from minetest to freeminer; i guess that's why
18:29 proller        yes
18:30 sapier         I don't deny he's right ... yet the result is whitespaces wont be fixed at all
18:30 sapier         because noone has interest in doing a whitespace commit
18:31 celeron55      how about running a full trailing whitespace removal on the whole codebase at some point when there isn't much going on otherwise
18:31 kaeza          maybe do it all at once and simply put a hard rule saying no whites... that
18:31 RealBadAngel   good idea to do so after stable is released
18:31 sapier         by now general rule was no automatic codestyling to full codebyse
18:31 kaeza          and kill anyone that adds blank-lines-with-whitespace
18:31 RealBadAngel   as first thing to do
18:32 sapier         I'll just change behaviour of my ide and then I'm done with this issue too ... imho it's way to much discussion to be worth it
18:32 celeron55      i would hope github would have a way to set limits to commits so that nobody would have to review these manually
18:34 RealBadAngel   make a branch, try it, if worked do the same with master
18:35 RealBadAngel   https://github.com/RealBadAngel/minetest/commit/a8ffb487048a7fa5f3f07522414e37c8670fba0c
18:35 RealBadAngel   so can i push it? ^^
18:36 sapier         hmm can you explain why 128 and not 64 or 32? :)
18:37 celeron55      i'm wondering why not 254
18:38 RealBadAngel   because the other values not workin correctly?
18:38 RealBadAngel   ive tested it
18:38 RealBadAngel   half of the range works perfectly
18:38 celeron55      i mean, 255 actually
18:39 celeron55      well whatever, i guess it doesn't break anything
18:39 sapier         testing isn't a good way to decide for a value ... maybe it behaves different on other graphics cards
18:39 RealBadAngel   everything below and above the half makes parts of the interior invisible
18:39 celeron55      it's a stupid hack though
18:39 celeron55      it should work no matter what that number is (except 0)
18:39 RealBadAngel   oh, cmon, alpha cannot behave different
18:40 celeron55      and as it doesn't work, there has to be some other bug
18:40 RealBadAngel   kahrl wanted to detect if pixel is solid or zero
18:40 RealBadAngel   but textures have semi transparent pixels too
18:40 celeron55      yes; that doesn't mean that at any point there would be no edge
18:40 celeron55      the edge would be at alpha=1 then, or alpha=128 now, or alpha=whatever the number is
18:41 celeron55      and as the edge disappears in some cases, there is a bug
18:41 RealBadAngel   https://www.youtube.com/watch?v=C9M596KW1dw
18:41 RealBadAngel   take a look
18:42 RealBadAngel   this is how it works with the patch
18:46 celeron55      i'm not doubting whether it works
18:46 celeron55      hacks work sometimes
18:46 RealBadAngel   i wouldnt call it hack, but whatever
18:48 RealBadAngel   it works as it should work from the begining
18:48 RealBadAngel   otherwise its useless for my further plans
18:53 RealBadAngel   http://imgur.com/5Jiy6mb
18:54 RealBadAngel   you should get what i mean lookin at this image
18:54 Megaf_         sapier: Hi, can I suggest Minetest Android for tablets players on my server?
18:54 Megaf_         I will give them a link do download it
18:54 Megaf_         on my website
18:58 ShadowNinja    Have travis check commits for trailing whitespace and fail if it finds any.
18:59 ShadowNinja    It auto-runs on PRs, so no manual checks needed.
18:59 RealBadAngel   another example: http://i.imgur.com/QdPNoic.png
19:00 RealBadAngel   and explanation whats that: proper 3d torch model, rdy to use with torchlike drawtyp
19:01 RealBadAngel   which will work for any textures given, unlike one proposed by blockmen
19:01 Megaf_         sapier: whats your most updated android branch?
19:05 sfan5          ShadowNinja: or have a git commit hook that automatically removed whitespaces
19:06 ShadowNinja    sfan5: Nah,auto-commits are ugly.  And better to do it right than have a bot fix it for you.
19:07 sfan5          no not auto-commits
19:08 ShadowNinja    sfan5: Comments?  http://ix.io/d6U/diff Makes inv:set_lists({}) clear the inv instead of doing nothing.  Maybe I should use a temporary inv in case it fails though.
19:08 sfan5          a hook that runs before a commit is made and automatically removes any trailing whitespace
19:08 ShadowNinja    sfan5: Oh, that might be good, but is it possible?
19:08 sfan5          sure
19:09 sfan5          looks good
19:13 proller        astyle instead trailing whitespace
19:13 ShadowNinja    sfan5: Also, remove confusing extra block: http://ix.io/d6V
19:14 sfan5          proller: astyle does not fully support the style guidelines we have
19:14 proller        but support part of
19:14 sfan5          it will change correct code if it does not fully support it them
19:14 sfan5          s/them//
19:15 proller        https://github.com/minetest/minetest/pull/1390
19:16 sfan5          we won't remove something that doesn't have a replacement
19:16 sfan5          and the indev mapgen is actually useful and nice
19:16 proller        nobody use it
19:16 proller        ^ about math
19:16 sfan5          that is not a reason to remove stuff
19:17 proller        it useless and have 1% features from planned
19:17 proller        and nobody exept me uses it
19:17 proller        and used
19:17 sfan5          why isn't it removed in freeminer then?
19:17 proller        and nobody know how to correctly use it
19:18 proller        freeminer have fully changed version
19:18 proller        with lot of fractals
19:18 sfan5          why not integrate the good one into MT?
19:18 proller        all agains this way
19:18 proller        and it GPL
19:18 sfan5          mh
19:19 proller        http://forum.freeminer.org/threads/four-circles.105/ - freeminers ;)
19:19 proller        one from ~infinity possilble worlds
19:20 proller        it was in pull more than half year
19:22 ShadowNinja    sfan5: So, the second commit's fine too?
19:22 sfan5          *nods*
19:23 sapier         megaf I'm merging ShadowNinja's comments atm
19:24 RealBadAngel   sapier, ive answered to your comments in my pull
19:25 ShadowNinja    sfan5: Re-did the inventory patch: http://ix.io/d6W/diff
19:26 sfan5          still fine
19:29 ShadowNinja    Ugh, who touched database-sqlite3?...
19:30 proller        database-sqlite3.sapier.edition
19:31 ShadowNinja    Whitespace stripping, right.
19:31 proller        every piece touched by sapier turns to shit
19:32 RealBadAngel   proller, give it a break, ok?
19:33 RealBadAngel   you do have own project to care of, dont need to sit here and work as a claun :P
19:33 ShadowNinja    clown*
19:33 proller        i need to fiz sapierized code
19:33 proller        fix
19:34 sapier         fix bugs instead of trolling around ;-P
19:34 proller        you produce bugs faster than i can fix
19:36 RealBadAngel   claun, clown whatever, *funny* guy that goes back through window, being previously thrown through door, who is pissing everybody around.
19:41 RealBadAngel   sapier, theres no id=0 in game
19:42 sapier         if this works for game buttons I most likely have been wrong ... haven't tested it so I could've missed some reason why it works
19:46 RealBadAngel   go on test it with current Unified Inventory
19:46 RealBadAngel   you will see how it works
19:47 RealBadAngel   https://github.com/minetest-technic/unified_inventory
19:47 RealBadAngel   its designed to work almost exactly the same as tooltips in my Xubuntu
19:51 sapier         doesn't work on my tablet rba ;-)
19:52 Megaf          unified inventory worked fine on that build I tested sapier
19:52 Megaf          on my x86 android phone
19:52 sapier         well my screen is probably a little bit smaller then yours ;-)
19:55 RealBadAngel   what, UI?
19:55 sapier         ShadowNinja: I fixed most of your comments
19:58 ShadowNinja    sapier: Good.  :-)  Pushing 1346 (player file loading) in a few minutes...
19:58 sapier         did you replace the 1000 by a define? ;-)
19:59 ShadowNinja    BTW, ShadowBot's server was moved, the move should complete today, so she should be back by tomorrow.
19:59 ShadowNinja    sapier: Yep.
19:59 sapier         great :-)
20:00 ShadowNinja    Oh, and she should have a 1gb connection, so lag should be much lower.
20:01 sapier         megaf do you want x86 or mips build first
20:02 Megaf          sapier; Well, I want ARM for players on my server
20:05 sapier         as of 0.4.10 arm will be released while x86 and mips will be experimental
20:05 sapier         well arm will be as stable as a initial version can be :)
20:06 RealBadAngel   sapier, what does not work for you?
20:06 sapier         it's too big
20:07 RealBadAngel   ooops :) buy something bigger then ;) and meanwhile try it on real computer
20:08 sapier         what about a unified inventory "phone edition"? ;-)
20:09 Megaf          nah
20:09 RealBadAngel   it will become possible
20:09 Megaf          the default one works fine on phones
20:09 RealBadAngel   Zefram_fysh is working on scaling the whole interface
20:09 sapier         and don't forget about the "watch edition" too ;-)
20:10 RealBadAngel   but till then, 800x600 is absolute minimum
20:10 sapier         you know most tablets don't have 600?
20:10 RealBadAngel   those from latest century? propably yeah
20:10 sapier         no
20:11 sapier         1024x600 results in application size about 1024x570
20:11 RealBadAngel   my tablet has better resolution than my desktop monitor
20:11 sapier         I'm not taking about 500eur+ class tablets ;-P most ppl will have at max 200eur class tablets
20:11 sapier         and those are 1024x600 raw size
20:12 RealBadAngel   hmm, ive tried mt with UI on such tablet
20:12 RealBadAngel   and it worked
20:13 RealBadAngel   also, all folks connecting VanessaE's servers are using UI
20:13 sapier         ok :-)
20:13 RealBadAngel   and lots of them connect with buildcrap clients
20:14 RealBadAngel   but question was not how big screen you have but if patch is working :)
20:14 sapier         you haven't talked to martin_devil and his 128x128 mips watch
20:16 sapier         but that thing is as small we're gonna need another special menu, even simplified mainmenu is unusable there
20:18 Amaz           Is there any point in supporting it?
20:18 RealBadAngel   yes, we want him to be able to see logo at least ;)
20:39 sapier         we're already far more then that rba https://www.dropbox.com/s/x766pt2mpop3jke/VID_20140622_173023.mp4
20:41 VanessaE       protip:  mount the smartwatch on something and focus the G*d damned camera.
20:41 VanessaE       and use a stylus
20:41 sapier         that's not my video tell it to martin_devil
20:43 VanessaE       I know, I'm just saying.  I see that shit all the time.  Hey you!  Yeah, you out there, the idiot with the camera!  HOLD THE FUCKING THING STILL!
20:43 VanessaE       :)
20:44 VanessaE       (that one isn't aimed at martin)
20:44 sapier         but it's amazing I didn't expect it to run at all at devices like that
20:44 VanessaE       yeah I know.  that looked like it was actually performing better than my piece of crap sub-$100 tablet
20:46 sapier         lol he's not at your spawn
20:47 VanessaE       heh
20:59 Megaf          sapier; I'd like to try to compile your Android branch here.
21:01 sapier         you're welcome ... I suggest doing a debug build as you need certificates for release
21:02 sapier         you need android ndk as well as ndk somewhere on your disk
21:02 sapier         git svn and wget installed
21:02 sapier         and of course make
21:03 RealBadAngel   ShadowNinja, https://github.com/RealBadAngel/minetest/commit/4fbafdb62b6934681d579806a9a5363582b3ef8b
21:03 RealBadAngel   better?
21:04 sapier         goto ... bäääääääähhhhhh
21:04 sapier         :-)
21:04 RealBadAngel   SN insisted
21:04 RealBadAngel   for sake of one level of tabs ;)
21:04 VanessaE       haha
21:05 sapier         I'd even prefere the do {} while(false) way to this ;-)
21:05 sfan5          VanessaE: ffmpeg has a stabilize filter
21:05 VanessaE       sfan5: good luck getting people to use it.
21:05 * sfan5        meows
21:05 ShadowNinja    sapier: But that adds another level of indentation, and the point is to remove indentation.  ;-)
21:06 RealBadAngel   sapier, theres if else if, not loop of any kind
21:06 RealBadAngel   so break also doesnt apply
21:06 sapier         split to a separate function
21:06 RealBadAngel   lol
21:06 sapier         draw_tooltip ;-) isn't that function already to long? ;-)
21:07 RealBadAngel   cmon guys, youre lookin for inexistant problems here :P
21:07 sapier         we don't seem to have real ones left
21:07 RealBadAngel   i can give you very serious one if you wish
21:07 RealBadAngel   located at the very same file :P
21:08 RealBadAngel   rdy?
21:08 sapier         can someone explain to me why I can run arm and mips android emulator but not x86?
21:09 RealBadAngel   checking for optional parameters by examining size is wrong to the bone
21:09 RealBadAngel   it works only for client and server with same version
21:10 sapier         why?
21:10 RealBadAngel   now, if server uses extra fields, and there will come client that doesnt know such fields...
21:10 RealBadAngel   kaboom
21:10 sapier         no that formspec just wont be drawn
21:10 RealBadAngel   you will see bunch of errors and invisible buttons
21:11 VanessaE       http://www.zimagez.com/zimage/missinbuttonsonunifiedinventory.php
21:11 VanessaE       specifically, that.
21:11 sapier         of course as always if you add new features not supported by clients
21:11 RealBadAngel   thus, in 0.4.10 version bump is definitely needed
21:11 RealBadAngel   and the whole checkin system has to be redesigned
21:11 sapier         why?
21:12 RealBadAngel   to allow older clients to connect and use only known fields while ignoring extra ones
21:12 sapier         if you connect to a server using a old version you get told about features may not work
21:12 sapier         you can't
21:12 sapier         there's no way to be forward compatible in general
21:12 VanessaE       sapier:  look at the screenshot; unless you can think of a way to make a server that supports tooltips send data in such a way that old clients still work, old clients will finally have to be disconnected.
21:12 sapier         this works for some things but not for all
21:13 RealBadAngel   theres a way to avoid this in the future
21:13 RealBadAngel   but we have to check if field exists, not how many arguments are here
21:13 RealBadAngel   sadly we cannot modify old clients
21:13 sapier         you still wont have 100% compatibility
21:14 RealBadAngel   but with such design, we will avoid such situations in the future
21:14 sapier         no
21:14 RealBadAngel   ofc
21:14 sapier         you can reduce them but not avoid
21:14 RealBadAngel   for example tooltips just wont be shown
21:14 RealBadAngel   but button will be still visible and usable
21:14 sapier         old clients can't support new features by definition ;-)
21:15 RealBadAngel   yes
21:15 RealBadAngel   thats why bump is needed
21:15 RealBadAngel   and then redesign
21:15 sapier         then someone invents a cool_image_button[] voila that one wont be shown on old clients too
21:15 sapier         or a type=cool_image_button
21:15 sapier         same thing in green
21:16 RealBadAngel   we are talking about extending existing ones functionality
21:16 RealBadAngel   not completely new features
21:16 sapier         tooltips are completely new features
21:16 RealBadAngel   that from definition cannot be shown or interpreted by older client
21:16 RealBadAngel   theyre not
21:16 RealBadAngel   they are extending existing feature, the button for example
21:17 sapier         you could've implemented tooltips in a way not causing compatibilty issues ;-)
21:17 lemonlake      just make people update, simple as
21:17 RealBadAngel   its not me who invented that system
21:17 sapier         e.g. adding a new field tooltip[gui_element_name;text]
21:17 VanessaE       it's not too late to re-design how those tooltips are done.
21:17 RealBadAngel   now, sapier, lets watch you pointing yourself
21:18 RealBadAngel   and sayin, mea culpa :P
21:18 sapier         yes I didn't realize that issue yesterday too
21:18 RealBadAngel   me neither
21:18 sapier         but hey according to credit's we're about 10 core devs
21:18 RealBadAngel   megaf raised it today
21:19 sapier         yet as vanessaE said, it's painfull but not to late
21:19 RealBadAngel   so, since i gave you real problem, can you get off my code and let it merge? :)
21:19 VanessaE       if 0.4.10 goes out with this current way, or some other mod starts to use these new tooltips, then it'll be too late
21:20 RealBadAngel   that system is pre-tooltip era
21:20 sapier         I can't remember to have said you shouldn't merge it today ;-)
21:20 RealBadAngel   checkboxes are affected also
21:20 sapier         ok we shouldn't break all buttons and checkboxes
21:20 RealBadAngel   all elements with extended (optional) functionality
21:20 sapier         what about the tooltip[] idea, would fix the default issue too
21:20 VanessaE       that's a good idea.
21:21 VanessaE       I may suck at using formspecs but even I can appreciate that.
21:22 sapier         this way you could even add tooltips to different table lines ... no idea if someone would want to do this
21:22 RealBadAngel   if ((parts.size() >= 3) || (parts.size() <= 5))
21:22 sapier         or dropdown lines ... there it'd be better
21:22 RealBadAngel   anything coded like that will confuse older clients in case server sending 5 arguments
21:23 RealBadAngel   and wont let client use just 3 of them
21:23 sapier         true but that's a fact we can't change
21:23 RealBadAngel   and i repeat, its not just tooltips idea
21:23 ShadowNinja    sapier: I commented on your commit.  There are still some missing things too, like the enum thing.
21:24 RealBadAngel   i can change it, but only for 0.4.10
21:24 sapier         yes I didn't change that one because I don't really understand how this questionable style issue should be relevant
21:24 RealBadAngel   so client 0.4.10 will be still (propably) able to connect to newer than itself servers
21:25 RealBadAngel   until version bump comes in way
21:25 RealBadAngel   but at least adding new field to an element wont break anythig for client anymore
21:25 VanessaE       why the argument here?
21:25 sapier         first of all we need to fix the tooltip issue
21:25 VanessaE       just code it the way sapier suggested and be done with it
21:25 sapier         I'm open for better suggestions too
21:26 RealBadAngel   can you rephrase the idea? because i cannot see any
21:26 VanessaE       [06-23 17:18] <sapier> e.g. adding a new field tooltip[gui_element_name;text]
21:26 RealBadAngel   huh?
21:26 RealBadAngel   how could it work even?
21:26 sapier         a button has to have a name
21:27 VanessaE       instead of tacking the tooltip onto the end of the button field, put it into its own separate field, using the same element name to tie the two together.
21:27 sapier         on parsing you add all tooltips to a std::map<std::string,std::string>
21:27 sapier         first string is name, second tooltip
21:27 sapier         on adding the tooltip you don't lookup in spec but in this table
21:28 VanessaE       a client will just ignore formspec elements it doesn't understand I guess, while showing the buttons with the same old element format they had before.
21:28 RealBadAngel   lemme think a bit on this idea
21:29 RealBadAngel   first of all i will have to trash 2 days of work adding tooltips to UI, great
21:30 lemonlake      is it the client or server that errors on a malformed formspec element?
21:31 RealBadAngel   old client
21:31 RealBadAngel   not malformed
21:31 lemonlake      so yeah, it will ignore them if they dont exist
21:31 sapier         RealBadAngel: defacto for client it is malformed
21:31 ShadowNinja    https://github.com/minetest/minetest/issues/1362#issuecomment-46774287  Comments on removing those "#if 0"/"/*" blocks?
21:31 RealBadAngel   it shouldnt be
21:32 sapier         it should
21:32 RealBadAngel   also i cannot agree on that solution, hower only for tooltips it could be nice
21:32 RealBadAngel   *however
21:32 sapier         how are you supposed to find broken formspecs if it's not checked?
21:32 RealBadAngel   problem is in general way
21:32 RealBadAngel   any extra field confuses the client
21:33 sapier         imho that's by design
21:33 RealBadAngel   so defacto <optional> in description is bullshit
21:33 sapier         formspecs are simple
21:33 RealBadAngel   optional should work both sides
21:33 lemonlake      ^
21:33 sapier         no as long as client knows they're optional that's gonna work
21:33 RealBadAngel   not only the server that uses what it wants and pisses on client
21:34 sapier         we don't need to talk about formspec beeing suboptimal ;-)
21:34 RealBadAngel   i can change the code to be optional for both parties
21:34 sapier         that wont help for old clients
21:34 RealBadAngel   and table for tooltips wont be needed and problem solved
21:35 sapier         you can't change 0.4.9 code
21:35 RealBadAngel   0.4.10 will be old client in the future too
21:35 sapier         if you manage to do that you're great
21:35 sapier         still you'd break 0.4.9
21:35 RealBadAngel   no
21:35 sapier         explain
21:35 RealBadAngel   we will be exactly at same place
21:36 sapier         so what's your suggestion?
21:36 RealBadAngel   but we will avoid the situation to happen again, makin 0.4.10 deprecated as soon as some1 adds new field to whatever
21:36 sapier         no
21:36 sapier         we don't avoid this by just ignoring errors
21:36 RealBadAngel   damn
21:36 RealBadAngel   check if field exists
21:36 sapier         that's how bugs are hidden and never found
21:36 RealBadAngel   not how many arguments you have sent :P
21:37 RealBadAngel   is that hard to get?
21:37 sapier         and how to you wanna find a broken formspec with additional fields?
21:37 RealBadAngel   send 9 of them, client knows what to do with 5
21:37 RealBadAngel   let it use just 5
21:38 sapier         so you're just gonna ignore a button containing 1000 parameters
21:38 sapier         imho that's error hiding not fixing
21:39 VanessaE       but the client *doesn't* know what to do with 9, let alone 5 + "gibberish"
21:39 VanessaE       that's the problem
21:39 RealBadAngel   im gonna ignore not known parameters
21:39 RealBadAngel   thus, makin them really optional for both
21:39 sapier         but you don't know if those parameters are new or glibberish
21:40 sapier         noone ever will detect that class of formspec error
21:40 sapier         because server wont check too
21:40 RealBadAngel   only thing i know by now its broken
21:40 RealBadAngel   and thats the way to fix it
21:40 sapier         e.g. a missing ]
21:40 VanessaE       come on, just write it sapier's way, jeez
21:40 sapier         you'll not find it
21:40 RealBadAngel   no
21:40 sapier         because you don't even get a error message
21:40 VanessaE       and move to table-based formspecs for 0.4.11 and up
21:41 RealBadAngel   because it doesnt apply to tooltips only
21:41 sapier         RealBadAngel: you're right but what you suggest is replacing a variant detecting additional errors where none are by a variant just ignoring errors
21:41 sapier         both is wrong true
21:41 RealBadAngel   it applies to any parameter added before and in the future to original defined elements
21:41 sapier         but imho better detect errors then ignore them
21:42 RealBadAngel   it does detection only when client and server match the version
21:42 sapier         yes
21:42 RealBadAngel   otherwise its just bullshit
21:42 sapier         no it's not
21:42 RealBadAngel   ofc it is
21:42 sapier         your mod has to specify what client version is required
21:42 sapier         client doesn't send any fields
21:42 sapier         server
21:43 sapier         it's a mod
21:43 RealBadAngel   megaf  updated server, installed new UI
21:43 RealBadAngel   all the clients GUI were broken
21:43 sapier         if your mod isn't usable at client x.x.x but at x.x.y only it's fine
21:43 sapier         default is a mod too
21:43 RealBadAngel   now add tooltips for mainmenu
21:43 RealBadAngel   lol, thats gonna be a mess
21:43 sapier         that wont cause any damage
21:43 RealBadAngel   client wont be able to see any buttons in mainmenu :)
21:43 sapier         mainmenu is local only
21:44 RealBadAngel   ah, true
21:44 sapier         there'll never be a client getting mainmenu from server ;-)
21:44 RealBadAngel   so nvm ;)
21:44 RealBadAngel   but the picture stays here
21:44 VanessaE       [5 years later] <sapier> well that finishes sending the main menu from the server...
21:44 RealBadAngel   for any complex mod
21:44 sapier         yes it's ugly ..... but your suggestion is even more ugly ... sorry
21:45 RealBadAngel   its not ugly, its flexible
21:45 sapier         think about this button[field,field; field;}button[field; field; field] ... for what you suggest that'd be a perfectly fine button
21:45 RealBadAngel   when we will stay with current state, we will need version bumps for every new field added
21:46 RealBadAngel   to prevent older clients to join and see a mess
21:46 sapier         no
21:46 RealBadAngel   wanna bet?
21:46 sapier         no because there's been always problems with old clients and new mods
21:46 sapier         that's nothing new it has happened and will happen again
21:47 RealBadAngel   try to use in a mod imagebutton fields
21:47 sapier         if your mod doesn't work for client 4.4.9 don't use it
21:47 RealBadAngel   like those two booleans and pressed texture
21:47 RealBadAngel   boom, button invisible, error thrown
21:48 RealBadAngel   no tooltips used :P
21:48 sapier         I know about the issue but why is your mod different from others? you can't use mobf 2.3 with client 0.4.1 too
21:48 RealBadAngel   my mod?
21:48 sapier         unified inventory
21:48 RealBadAngel   i dont use above fields
21:49 RealBadAngel   i was forced to add ;;;; to my buttons
21:49 sapier         yes I know it's not tooltips only ... but we could fix it for tooltips adding additional value ;-)
21:49 RealBadAngel   and gues what, older clients are not happy with that
21:49 sapier         e.g. by providing individual tooltips for dropdown lines ;-)
21:49 RealBadAngel   so what if the tooltips will be fixed but problem stays?
21:50 VanessaE       fix the root problem separately then.
21:50 sapier         problem stays for sure but at least it'd not be broken for all buttons
21:50 RealBadAngel   i said i like the idea of such table
21:50 sapier         the only fix is adding a full featured gui engine
21:50 sapier         including api version
21:50 sapier         I'll add that to requirements list for formspec replacement
21:50 VanessaE       provide tooltips in a way that doesn't break older clients, fix the root of the issue for the future so it's less likely to happen again
21:51 VanessaE       sapier: inb4 linking to QT/GTK+ ;)
21:51 RealBadAngel   so instead of fixing it right now, right here
21:51 VanessaE       (or wx/tk/whatever else)
21:51 sapier         you can't fix it right now rba
21:52 RealBadAngel   i will have to wait for tooltips till christmas 2016?
21:52 VanessaE       *facepalm*
21:52 VanessaE       [06-23 17:18] <sapier> e.g. adding a new field tooltip[gui_element_name;text]
21:52 VanessaE       just.  do.  that.
21:52 VanessaE       jesus fucking christ.
21:53 RealBadAngel   i will have to revert lotsa code
21:53 VanessaE       I know, and it sucks having to do that.
21:53 sapier         either that way or live with any mod using that feature require client version 0.4.10
21:53 VanessaE       but mistakes happen
21:53 VanessaE       no one's perfect.
21:54 RealBadAngel   why im always forced to go hard way? :P
21:54 sapier         it's not you alone rba ;-)
21:54 VanessaE       RealBadAngel: because you don't run a server :P
21:54 sapier         I can't even tell how much code I usally have to rewrite ;-)
21:54 RealBadAngel   shit and i thought i will start coding real 3d torches tonight
21:55 RealBadAngel   i was so close...
21:55 * RealBadAngel sits in the corner and starts to cry...
21:55 VanessaE       aw
21:55 * VanessaE     hands RealBadAngel a kitten
21:56 RealBadAngel   btw, nobody commented out my torches idea
21:57 VanessaE       RealBadAngel: you don't wanna hear my opinion on it.
21:57 RealBadAngel   i do
21:57 RealBadAngel   http://i.imgur.com/QdPNoic.png
21:57 RealBadAngel   and with explanation why those torches looks so bad
21:58 VanessaE       3d torches as you propose them are a stupid idea and you're an idiot for proposting to do them that way.
21:58 VanessaE       :)
21:58 RealBadAngel   ^^
21:58 RealBadAngel   comments on the picture plz :P
21:58 VanessaE       the picture looks fine but let's see a VIDEO of the flames burning.
21:59 RealBadAngel   for that i do have to extrude each frame of torch animation
21:59 VanessaE       and THAT will be hell.
21:59 RealBadAngel   not really
22:00 RealBadAngel   we are talkin bout one animation
22:00 VanessaE       three animations.
22:00 RealBadAngel   one :P
22:00 VanessaE       three.
22:00 RealBadAngel   mesh can be rotated :P
22:00 VanessaE       no.
22:00 VanessaE       you're getting into blockmen territory again.
22:00 RealBadAngel   huh, lemme make a quick video
22:00 VanessaE       flames always rise when they burn, and any idiot will design their torch animations to make the flames rise
22:01 VanessaE       that means making them rise from the ceiling torch as well as the wall torch
22:01 RealBadAngel   ah that
22:01 VanessaE       and if you rotate the mesh, they will seem to shoot flames out at angles.
22:01 RealBadAngel   somebody that thought that torch could hang down from the ceiling shall be shot in place
22:01 VanessaE       so you will have to extrude all three textures separately.
22:02 VanessaE       torches can hang from the ceiling in the form of a hanging oil lamp.
22:02 VanessaE       see Nostalgia game.
22:02 RealBadAngel   sooner the better, it will spare humanity more stupid ideas
22:02 RealBadAngel   so make a lamp node
22:02 RealBadAngel   we are talkin torch here :P
22:03 VanessaE       the only way this can work is if you use a new texture with a new filename and NO FLAME, extrude THAT, and use particles to produce the flame.
22:03 VanessaE       and give it a new drawtype.
22:03 VanessaE       anything that changes the behavior of the existing drawtype is dead deal.
22:04 VanessaE       doing that would provide for a very nice looking torch
22:05 VanessaE       10x nicer than blockmen's attempt and 100% compatible with texture packs.
22:05 RealBadAngel   huh, somebody has the link to the vid with particles that was posted in #minetest a few days ago?
22:05 VanessaE       https://www.youtube.com/watch?v=5WnmQYLJzlE&feature=youtu.be
22:05 VanessaE       this?
22:06 VanessaE       note:  although you can't see it clearly in the video, those particles are textured.
22:06 RealBadAngel   quite nice
22:07 VanessaE       the textures become evident at around 1:13
22:07 VanessaE       er not 1:13
22:07 RealBadAngel   yup, vid is 1:12 long ;)
22:07 VanessaE       0:57
22:07 VanessaE       thereabouts.
22:07 VanessaE       I always misread that.
22:07 VanessaE       anyway
22:08 VanessaE       irrlicht particles clearly work fine even when there are tens of thousands displayed
22:08 VanessaE       and jin_xi probably doesn't have a high-end gaming system does he?
22:09 RealBadAngel   i will be back in 20minutes, have to reboot and do somethin
22:10 VanessaE       anyways, I'll make a torch that's just the wooden torch with a blob of coal tar at the end and no flame if RealBadAngel will use proper flame-like particles
22:10 VanessaE       for HDX>
22:53 sapier         that video looks quite cool ... not really like flame particles but maybe it can be tweaked
22:53 VanessaE       well those are default irrlicht particles I think
22:54 VanessaE       I'm sure the textures thereof can be changed
22:54 VanessaE       (as well as the spread/distribution)
22:55 VanessaE       here's another example:  https://www.youtube.com/watch?v=-brLc8uUeQk
22:56 VanessaE       see 1:25 in the video
22:56 kahrl          anybody mind if I correct two typos in comments in guiTable.cpp?
22:56 VanessaE       flamethrower effect.
22:58 sapier         just do it
22:59 * VanessaE     wonders what RBA went off to do ...
23:00 kahrl          I guess I could fix #1187 while I'm at it
23:00 kahrl          and how about using formspec tooltips instead of irrlicht tooltips for tables?
23:01 kahrl          (not sure if that would collide with anything RBA is doing at the moment)
23:01 sapier         guess we should wait a little bit till we know how the final solution will look like
23:02 VanessaE       as long as it works and doesn't break old clients' rendering of fields they don't recognize :P
23:02 sapier         guess there the tooltips are already specified ;-)
23:04 kahrl          aah! I can only find one of the two typos now!
23:05 kahrl          I saw them like a week ago and didn't write down what they were
23:05 kahrl          oh well :)
23:05 kahrl          ah, there it is
23:08 sapier         http://i58.tinypic.com/2chatev.png I guess that screenshot is better for android port :-) ... looking at a tree wasn't quite good advertisement ;-)
23:10 VanessaE       sapier: is there a reason, btw, that the "climb down" button only looks like an arrow and not a ladder with an arrow?
23:10 VanessaE       (compare your screenshot to the actual texture file - it looks like your screenshot on my tablet too)
23:10 VanessaE       (except for the square hand)
23:11 sapier         hmm interesting I didn't even realize that by now :-)
23:11 VanessaE       also, shouldn't you move the chat button to somewhere that isn't always obscured by the build name/version ?
23:11 sapier         most likely ... suggestions? right corner?
23:11 VanessaE       bottom, to the right of the D-pad.
23:11 VanessaE       to complement the inventory button.
23:12 sapier         no that's gonna be an issue for small screens where inventory slots tend to move there
23:12 VanessaE       hm
23:12 VanessaE       hm
23:12 VanessaE       same could be said for the lower right corner where the jump/climb buttons are.
23:13 sapier         jess but making same mistake twice doesn't make it better
23:14 VanessaE       true.
23:14 VanessaE       well you could put it up in the upper right corner and give it an opaque background
23:14 sapier         I had to put those buttons there to get a smooth control ... at least to me ... as I always say controls are beta only if someone provides better ones I'll not stop him (at least after initial version is released) ... it's time to get that out right now
23:15 VanessaE       or put it in the control bar at the bottom?
23:15 VanessaE       (the menu bar, whatever Android calls it)
23:15 sapier         I don't have access to it
23:15 VanessaE       damn
23:15 sapier         that's android only
23:15 VanessaE       put it in the "..." menu then
23:15 sapier         ... menu?
23:16 VanessaE       the three dots button between the task switch and screen shot buttons in your screenshot.
23:16 kahrl          https://github.com/minetest/minetest/blob/master/src/guiTable.cpp#L209
23:16 VanessaE       I call that the "..." button
23:16 sapier         hmm on ogles2 items are inverted again
23:16 kahrl          how would I say "if (columns.empty()) columns.append(a text column)" here?
23:16 sapier         if you tell me how to do so?  I don't know android very well
23:16 kahrl          I can't do that because columns is const, and anything else I can think of is ugly
23:18 VanessaE       sapier: no idea, but imho inventory and chat buttons could be placed there since they're not quick-access in the same way as the regular controls are.
23:19 sapier         if this ain't something to be done fast I'd delay it to next version as our target date for 0.4.10 is end of this week (feature freeze)
23:20 sapier         btw that all dark issue with bumpmapping enabled is there on android too
23:20 sapier         may this be an issue if graphics card doesn't support it?
23:21 VanessaE       could be
23:21 VanessaE       if preload visuals is turned on, this sometimes happens on some GPUs
23:21 VanessaE       I've seen it happen before
23:22 VanessaE       particularly with irrlicht 1.8
23:22 VanessaE       even if the GPU supports bumpmaps, in fact.
23:22 sapier         hmm ahh I replaced that H arrow button by the arrow one on changing those buttons to half sized ones
23:23 sapier         btw with ogles2 it's the plain inventory images which are broken not the inventorycubes ;-)
23:24 sapier         good night
23:24 VanessaE       ni9ght
23:24 VanessaE       night*
23:27 kahrl          https://github.com/kahrl/minetest/commit/0f646721f16d64c6745b6e886ff0f2efa56998a8
23:27 kahrl          ^ fix for #1187
23:28 kahrl          any comments? if not, I'll merge it in an hour or so
23:54 ShadowNinja    kahrl: looks fine to me.
23:55 kahrl          ok, thanks, I'll merge it then