Minetest logo

IRC log for #minetest-dev, 2015-07-08

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

All times shown according to UTC.

Time Nick Message
00:00 kahrl a wild bee hive appears!
00:00 kahrl skewing my FPS comparison ever so slightly :S
00:01 VanessaE lol
00:04 kahrl ok, it's hovering around 20 FPS with fancy leaves and around 24 FPS with simple leaves
00:04 kahrl at dreambuilder seed 0 mgv6 spawn facing north with viewing_range_nodes_min = _max = 150
00:05 Sokomine wild bee hives love to eat fps
00:06 kahrl I redid the one that I first measured without the bee hive
00:06 VanessaE heh
00:06 kahrl didn't observably change FPS
00:06 VanessaE she's just being a wiseass :)
00:07 kahrl :)
00:09 RealBadAngel kahrl, so gain is similiar as for me
00:10 kahrl I have to say though, simple leaves are quite ugly
00:10 kahrl https://imgrush.com/6jvuaVsfWxda.png
00:11 VanessaE opaque leaves are worse :)
00:11 VanessaE note to self:  submit a replacement mesh for that beehive.
00:11 Warr1024 opaque leaves could pretty easily be made much less ugly.
00:12 Warr1024 Someone just needs to tweak the texture and replace the black gaps with something less ugly, then tune the alpha down to like 1%.
00:12 RealBadAngel kahrl, check leaves on moretrees
00:12 RealBadAngel they do look way better in most cases
00:12 VanessaE Warr1024: I did that for a couple of textures RBA used in one of his screenshots.
00:12 Warr1024 using a small but non-zero alpha will prevent image optimizers from discarding the RGB data, but will cause them to be invisible when normal transparency thresholds the alpha channel.
00:13 VanessaE Warr1024: the problem is image optimizers usually delete the color of a transparent pixel
00:13 kahrl VanessaE: I think opaque is better than simple
00:13 VanessaE right, what you said.
00:13 asl97 opaque leaves are nice
00:13 VanessaE kahrl: well, there's no accounting for taste :D
00:13 Warr1024 VanessaE: make them only semi-transparent, e.g. opacity 1% or something.
00:13 RealBadAngel each one here has different opinion which one is better
00:13 Warr1024 then you could probably merge something like that upstream to mt_game.
00:14 RealBadAngel luckily theres a dropdown for it :)
00:14 Warr1024 removing the blackness in the leaves texture would make the game look better for some users, and no different for others, so it sounds like a win...
00:14 RealBadAngel btw, i just turned leaves to meshnodes
00:14 Warr1024 all we would need would be to agree on WHAT goes in there :-)
00:14 RealBadAngel theyre way too dense
00:15 VanessaE RealBadAngel: using that .obj I sent?
00:15 RealBadAngel https://imgrush.com/4MK7MCRGusQH
00:15 RealBadAngel VanessaE, yup
00:15 VanessaE wow
00:15 VanessaE may as well be opaque at that point
00:16 Warr1024 I like the look of that, though I doubt I'd be able to spare the framerate, myself.
00:16 Warr1024 I actually DO use fast leaves, though I hate the black gaps.
00:18 Warr1024 RealBadAngel: if the leaf texture were a bit less dense, then what you have there would make for a pretty neat volumetric effect.
00:19 kahrl ok yeah, moretrees are fine with simple leaves, due to the number of air nodes between leaves
00:20 RealBadAngel VanessaE, are beaches on by default?
00:20 RealBadAngel *beeches
00:20 VanessaE yes
00:20 VanessaE OHY
00:20 VanessaE no
00:20 VanessaE they're not
00:20 paramat lol
00:21 VanessaE I was gonna say, beaches are paramat's department :)
00:22 VanessaE beeches are disabled by default, don't remember why I did that.  probably something to do with being unable to disable default trees, at least at one time I think it was
00:22 VanessaE don't remember now.
00:24 est31 about weblate: I think we should configure the weblate bot to automatically push and pull changes
00:25 est31 this needs celeron55's level of access to it.
00:26 est31 https://docs.weblate.org/en/latest/​admin/continuous.html#github-setup
00:28 paramat seeing kahrl's screenshot, simple leaves are very ugly for default trees, this cannot be the default. now i'm not even keen on having the option, sorry
00:28 paramat left #minetest-dev
00:30 RealBadAngel https://imgrush.com/ILx-ZVKn3b3Y.png
00:30 RealBadAngel ^^default trees
00:30 RealBadAngel with simple + moretrees oak texture
00:31 VanessaE RealBadAngel: is that with the cube mesh in place?
00:31 RealBadAngel no, its glasslike
00:31 VanessaE oh ok
00:32 RealBadAngel default leaves texture has too many holes in it to look good
00:32 VanessaE right.  so it seems like a small matter to compensate for the lost density in "simple" mode by using a "thicker" texture then.
00:32 kahrl it's not bad from above since there is a greenish background, but from ground level there's just the sky as background
00:33 VanessaE RealBadAngel: any way you could make the client load a different leaves texture (if one exists) when in simple mode?
00:33 RealBadAngel ofc i can
00:33 VanessaE server-supplied texture of course
00:34 RealBadAngel no problemo, such code is here for ages
00:34 RealBadAngel special tiles
00:34 VanessaE probably need not be complicated - default_leaves.png, default_leaves_simple.png  or something
00:34 RealBadAngel i can switch to them when simple selected
00:34 VanessaE right
00:35 Warr1024 doesn't the current behavior just append ^[noalpha?
00:35 VanessaE and fall back to the regular tile if special_tiles isn't configured for that node
00:35 VanessaE Warr1024: close to it yeah
00:35 Warr1024 btw, does anyone actually use that texture filtering stuff I added once upon a time?
00:35 RealBadAngel VanessaE, can you make a more dense texture that would be similiar to original one?
00:36 VanessaE RealBadAngel: not really, no.  I suck at pixel art
00:38 asl97 Warr1024: if you mean rotating and layering, i sort of did
00:41 asl97 nevermind about the render 6 faces when any of the face is expose idea,  it make the edge of the tree too dense and doesn't effect the center enough
00:42 OldCoder joined #minetest-dev
00:58 RealBadAngel https://imgrush.com/RvCoJY2fft-k.png
00:58 RealBadAngel how about that?
00:59 VanessaE nice.
00:59 VanessaE that's about the same density as "fancy" mode I'd say
01:00 RealBadAngel https://imgrush.com/VBqw3VcYxVQX.png
01:01 RealBadAngel thats the texture
01:01 VanessaE looks like an entirely different texture than before.
01:03 RealBadAngel its made by copy + paste with an offset
01:10 RealBadAngel est31, kahrl, have you tried the texture?
01:10 est31 RealBadAngel, which?
01:11 est31 the ice?
01:11 est31 or the shader fix
01:11 kahrl sorry, bit busy, but looks good on the screenshot
01:12 RealBadAngel https://imgrush.com/VBqw3VcYxVQX.png this, as default_leaves.png and simple leaves
01:13 RealBadAngel or maybe this one: https://imgrush.com/sA_zAuziASES
01:14 RealBadAngel its more juicy
01:19 VanessaE too regular.
01:19 VanessaE I mean, too much of a pattern
01:32 RealBadAngel ok, let somebody else make the texture for the game
01:32 RealBadAngel ive coded the part with special tiles
01:33 RealBadAngel if theyre defined, switching to simple will cause using them instead od regular tiles
01:34 RealBadAngel if not, the very same texture will be used for both modes
01:38 RealBadAngel meanwhile, im off to bed
01:38 RealBadAngel cya
03:51 sloantothebone joined #minetest-dev
04:24 Sokomine joined #minetest-dev
04:54 DFeniks joined #minetest-dev
04:54 hmmmm https://github.com/kwolekr/minetest/commit​/47c7116b3677874e964357bfc8125202e8b0ce65
04:55 hmmmm sorry it took so long.
04:56 hmmmm also created PR 2893 for this
05:01 est31 this is the point where git tries to be smarter than it should be https://github.com/minetest/minetest/pull/2893/​files#diff-f20c82d93887e9a0f4d4eeb66cc7eaa7L133
05:02 est31 I first thought WTF where is the getMap call
05:03 hmmmm use the split view
05:03 hmmmm a change this big is too inconvenient to look at with the unified diff
05:08 est31 wow you dont need this typedef struct MinimapModeDef {} MinimapModeDef; stuff anymore very convenient
05:08 hmmmm not with C++
05:12 est31 whats behind this DSTACK(__FUNCTION_NAME); stuff
05:12 est31 and why did you remove it
05:13 hmmmm it's old and not needed anymore
05:13 hmmmm it's for DebugStacker
05:13 hmmmm DebugStacker is in all honesty quite useless... seeing as how it's applied to functions in a random manner
05:13 est31 to deliver stack traces even without debug symbols?
05:13 hmmmm yes
05:14 hmmmm the thing it is includes only location, no parameters or anything
05:14 hmmmm and if something calls that function without a DSTACK() then you only get that one location
05:14 est31 is it activated right now? is it the debugstacker which delivers these broken "stcks" if you compile in release mode?
05:14 hmmmm this is not terribly useful since we have minidumps now
05:14 hmmmm yeah, I think they do persist in release mode
05:15 est31 ok then
05:17 est31 enqueueBlock isn't used now anymore
05:18 est31 why did you remove it here https://github.com/minetest/minetes​t/pull/2893/files?diff=split#diff-f​20c82d93887e9a0f4d4eeb66cc7eaa7R277
05:18 hmmmm oh my
05:18 hmmmm that's a bug
05:19 hmmmm well, I was going to remove enqueueBlock and have pushBlockUpdate call deferUpdate on its own, but I figured that's probably not as symmetric of a design since popBlockUpdate() doesn't wait on updates
05:19 hmmmm so I changed it all back and evidently forgot something
05:20 est31 ok
05:21 hmmmm interesting how it still worked even with such a critical bug
05:21 hmmmm I didn't notice it at all tbh
05:22 est31 thats what reviews are for
05:23 hmmmm no I mean the minimap still functioned even though it wasn't receiving updates
05:23 hmmmm that must be because of the flood of setPos() calls it receives
05:23 est31 it gets its updates from other points I think
05:23 hmmmm hmm
05:23 hmmmm well I guess it's not too bad
05:24 hmmmm it only signals the event if a whole integer position changed (since pos/old_pos are v3s16)
05:24 est31 perhaps you got your updates thanks to this https://github.com/kwolekr/minetes​t/blob/master/src/minimap.cpp#L325
05:25 est31 but I guess the bug is present when you dont move, dont change the minimap mode, only turn around, and see the world appearing behind you, but the minimap not updating
05:25 hmmmm still if it were up to me I'd force an update once only ~3 nodes of distance?
05:25 hmmmm how about you?  is setPos() too chatty with the update thread?
05:26 est31 depends on when setPos gets called
05:26 hmmmm I think every single frame
05:26 est31 then it should be limited I think
05:26 hmmmm indeed
05:26 hmmmm it's limited by pos != old_pos
05:27 hmmmm so movement will set it off, but not standing around
05:27 est31 question is however whether movement will appear blocky because of that
05:27 hmmmm and it's implicitly updated once every whole node
05:27 hmmmm due to integer rounding
05:29 est31 do you know how fast you are with fast mode?
05:30 hmmmm sanic fast
05:30 hmmmm i have no idea
05:30 est31 just wondering whether minimap can keep up with that tempo
05:30 hmmmm i'm gonna just try it
05:31 hmmmm also removing that silly dependency on Player
05:31 VanessaE it keeps up at 20m/sec standard run speed at least.
05:31 VanessaE (barely)
05:31 est31 20 nodes per sec
05:31 VanessaE *nod*
05:31 est31 so minimap rendering can take up to 50 ms until its lagging behind
05:31 hmmmm heh
05:32 hmmmm 20 nodes per second?!?
05:32 VanessaE hmmmm: yep.
05:32 hmmmm that's a bit fast
05:32 hmmmm are you suure
05:32 VanessaE measured it once a couple years ago
05:34 VanessaE 197 nodes in 10 seconds.
05:34 VanessaE so still what it was.
05:34 hmmmm don't you accelerate as you're running
05:35 est31 what do you mean with "The minimap's internal buffers are statically sized"
05:35 hmmmm a car on a drag strip running a quarter mile in 15 seconds doesn't mean the car goes at 60 mph
05:35 VanessaE that was straight speed, measured by holding down "E" first, then jamming down on "W" as soon as I was ready.  Then letting off only "W" when the clock hit 10 seconds.
05:35 hmmmm they're 512 * 512 regardless of what size may be needed
05:35 VanessaE (but keeping E held down for a moment)
05:36 hmmmm I noticed something as I was dicking around one day
05:36 est31 ah that, I remember
05:36 hmmmm your player's speed doesn't make a difference in how far you jump
05:36 hmmmm is this... the way it's supposed to work?
05:36 est31 well, not a biggie, but would be better to be fixed.
05:36 est31 (unless I miss sth)
05:37 VanessaE however, accelerating from walk to run takes like half a second to get to full run speed, and a bit shy of 2 seconds to stop.
05:38 VanessaE speed makes no difference in jumping?  well height no, but I'm pretty sure it matters with horizontal distance.
05:39 VanessaE a bit less than 4 nodes horizontal distance with a walking jump, about 30 or so nodes with a running jump.
05:39 hmmmm i tested it
05:39 hmmmm i built a ramp and ran off a cliff
05:39 hmmmm noticed that i only need to go 3 nodes back to get the full distance
05:39 VanessaE yeah
05:39 hmmmm whereas if I built up speed from ~100 nodes back there was literally no difference in the distance jumped
05:39 hmmmm that's.... not how real life works
05:39 VanessaE oh yeah, there's no inertia.
05:40 hmmmm if this were changed, how would it affect gameplay?
05:40 VanessaE but then again, there's all kinds of things you can do in Minetest that doesn't work in real life, like falling 300 meters and landing perfectly safely in say 30 meters deep of water :)
05:40 VanessaE I dunno, I think it would irritate some people.
05:41 VanessaE it'll affect parkour maps/games for sure though
05:41 VanessaE (those seem strangely popular)
05:42 VanessaE for my particular uses of fast, a bit longer time to accelerate would be beneficial actually
05:43 est31 hmmmm, here we copy the whole QueuedMinimapUpdate, no? https://github.com/minetest/minetest/pull/2893/​files#diff-f20c82d93887e9a0f4d4eeb66cc7eaa7R74
05:43 Hunterz joined #minetest-dev
05:43 est31 before, we only passed a pointer
05:44 est31 ah but data is only a pointer
05:44 est31 then its good
05:44 hmmmm right
05:44 hmmmm allocating a chunk of memory each update just to save 4 bytes :).. wtf
05:44 hmmmm to save copying 4 bytes*
05:45 hmmmm and in the process use 8 more
05:45 est31 ?
05:46 est31 well, I'm too lazy to calculate myself, I believe you
05:46 est31 even if we did save, new is expensive as you said
05:46 hmmmm it depends on a lot of things
05:47 hmmmm new could be very fast if the implementation is using a low-fragmentation heap along with thread local arenas to eliminate the need for locking
05:47 hmmmm which you're probably not going to find in windows
05:48 hmmmm low fragmentation heap is a feature that you need to explicitly activate
05:48 est31 ok
05:49 est31 your PR is WIP? or you want to make another one with more improvements?
05:50 est31 because except the missing update it looks good
05:50 est31 havent tested it yet though
05:51 hmmmm it's not WIP anymore
05:51 hmmmm I made my last set of improvements minor enough to stay in the same commit
05:52 celeron55 est31: if you need something to be configured in github, /msg me with details
05:58 est31 regarding the newline change, its no issue you added newlines to the getName() method, I just wanted to point out that I copied the approach from unittest https://github.com/minetest/minetest/blob/​master/src/unittest/test_collision.cpp#L27
05:59 est31 want*
05:59 est31 (its more correct style anyway)
06:00 hmmmm hrmm
06:00 hmmmm why did I write it like that
06:00 hmmmm which do you like better?
06:02 est31 I like it most the way I've commited it, as it was. Perhaps can we add an exception to the rules for single line methods that fit inside the line limit?
06:02 hmmmm yeah, same
06:03 hmmmm except the single line method needs to be a simple statement
06:03 hmmmm like returning a value
06:04 est31 or a simple setter method
06:04 hmmmm the exception is just an exception
06:04 hmmmm you could still follow the basic rules
06:04 hmmmm so either-or works
06:04 est31 ok
06:05 hmmmm like how we allow if (foobar)\n\tblah(); as well as if (foobar) {\n\tblah();\n}
06:05 est31 yea
06:05 est31 do we allow if (foobar) blah(); ?
06:07 hmmmm no
06:07 est31 ok
06:08 est31 yea the missing update is noticeable
06:08 hmmmm https://www.kernel.org/doc​/Documentation/CodingStyle
06:08 hmmmm Don't put multiple statements on a single line unless you have something to hide:
06:08 hmmmm est31:  I updated it
06:08 hmmmm are you using the latest version of that PR?
06:11 est31 now it isnt present anymore
06:14 est31 glasslike leaves is deadly for small trees
06:15 nrzkt joined #minetest-dev
06:15 hmmmm content_mapblock.cpp has some atrocious code
06:15 hmmmm have you looked at makeCuboid?
06:16 jin_xi joined #minetest-dev
06:20 est31 ?
06:22 hmmmm MMMMMMM
06:23 hmmmm https://github.com/minetest/minetest/commit/a9​3838707a9aab104c20b6a971d8c622184a508f?diff=un​ified#diff-b5b39c3a29359759ac6f0293dc37a453R96
06:23 hmmmm do you notice something wrong here?
06:25 est31 sizeof(irr_loglev_conv) gives you the size of the array in bytes
06:25 hmmmm yup
06:25 est31 every entry in the array takes one byte.
06:26 est31 doesnt it?
06:26 hmmmm does it really
06:26 hmmmm no
06:26 hmmmm enum entries are ints
06:26 hmmmm ints are 4 bytes
06:26 est31 whats sizeof(LogMessageLevel)
06:26 hmmmm but even so, the only time you can assume sizeof(T) == 1 is when the type is char or unsigned char, or u8
06:26 hmmmm or c8
06:26 est31 whats c8
06:26 est31 only know s8
06:27 hmmmm fixed-width char
06:27 hmmmm that's an irrlicht thing
06:27 est31 ok
06:27 hmmmm i think it's only used for video::SColor
06:28 est31 sizeof(irr_loglev_conv) / sizeof(LogMessageLevel) would work though
06:28 est31 ?
06:28 hmmmm even better.  you could use ARRLEN()
06:28 est31 a smart compiler would optimize the division away
06:29 est31 didnt knew of ARRLEN
06:29 est31 is that std C?
06:29 hmmmm no
06:29 hmmmm every time you start a new project, that's pretty much the first macro you write
06:29 hmmmm #define ARRAYLEN(x) (sizeof(x) / sizeof((x)[0]))
06:30 est31 thats nice
06:31 est31 assumes though that the length is at least 1
06:31 hmmmm i think that is a fair assumption
06:32 est31 yea, one has to bear it in mind though when changing code
06:32 hmmmm can you name a circumstance where length is not at least 1?
06:34 est31 you could define an array of configurations, and write later on a loop over the array
06:34 est31 then you notice you dont need the loop
06:34 est31 and instead of commenting out the loop, you comment out the elements in the array
06:34 est31 but thats already bad style
06:35 est31 and style guidelines are made to protect you from right these bugs
06:35 hmmmm so you're concerned about an out-of-bounds array access on a compiler preprocessor directive?
06:36 hmmmm sizeof() never gets executed in code
06:36 hmmmm it's always replaced by a constant value by the compiler
06:36 est31 makes sense
06:36 est31 ok then
06:37 est31 ARRAYLEN is free to use
06:37 est31 good to know
06:42 est31 next pr to look at : #2891
06:42 ShadowBot https://github.com/minetest/minetest/issues/2891 -- Use UTF-8 instead of narrow by est31
06:42 est31 its essentially a big huge s/narrow/utf8/g
06:44 RealBadAngel joined #minetest-dev
06:46 hmmmm iconv() is quite slow for what it does
06:46 hmmmm are you really sure you want to add all that for no tangible benefit whatsoever?
06:46 est31 that has actual benefit
06:47 hmmmm ?
06:47 est31 I think e.g. #2883 can be fixed with that
06:47 ShadowBot https://github.com/minetest/minetest/issues/2883 -- Support Unicode in tooltip names
06:49 VanessaE by the way, offtopic, whose bright idea was it to bind en/disable camera updates to the 'n' key?
06:50 est31 if we use "narrow", we have a **very** restricted character set
06:51 hmmmm maybe... maybe only replace narrow_to_wide with utf8_to_wide in non-constant expressions
06:51 est31 windows people will get stuff converted into this ANSI thingy
06:53 est31 makes sense
06:53 est31 I've wanted to remove that function entirely at some point
06:53 est31 but if there is an actual speed benefit
06:53 est31 I'll do a benchmark
06:55 Hunterz utf8 will be fine when will be there font with utf8 symbols for example
06:55 est31 ?
06:56 est31 fonts are another matter
06:56 Hunterz for example star as prefix for admin
06:56 Hunterz in the chat
06:56 est31 erm, sorry but that doesnt have to do anything with utf8
06:56 Hunterz ok
06:58 asl97 VanessaE: the default doesn't seem to have change but for some reason, the mapping is wrong
06:58 asl97 https://github.com/minetest/minetest/b​lob/master/src/defaultsettings.cpp#L57
07:02 VanessaE celeron55: you about?
07:02 VanessaE asl97: odd.  well mine's a Release build
07:03 VanessaE so "none" is probably being interpreted as a string, hence the 'n;
07:03 VanessaE 'n' *
07:04 asl97 make sense, you can make the bug report
07:04 est31 hmmmm, for me utf8 is about as double as slow as narrow
07:05 est31 but note, 10 000 calls to utf8_to_wide can be handled in 24 ms.
07:06 est31 unless my compiler tried to optimize stuff it shouldnt
07:07 Darcidride joined #minetest-dev
07:07 est31 https://github.com/est31/minetest/commit/​5e6f40f9f4f7e8c0086eaea3c7e64b67f66642de
07:07 VanessaE offtopic:  can someone please fix this G*d damned forum blacklist bug (moderators/admins should be exempt from the blacklist)?
07:11 Darcidride joined #minetest-dev
07:11 RealBadAngel hmmm, what if we make different model of coding stuff for minetest? we will gather  some ideas and you will code that for us :P
07:13 est31 That was a cleanup
07:13 est31 spares new and removes own minimapupdatequeue
07:14 est31 (not needed)
07:17 est31 I'll merge #2893, it has two +1 unless RealBadAngel or sb else wants to check for issues with it.
07:17 ShadowBot https://github.com/minetest/minetest/issues/2893 -- Clean-up Minimap code by kwolekr
07:17 hmmmm wait
07:17 hmmmm who is the second +1?
07:17 hmmmm and it's not me, btw
07:17 est31 nrzkt
07:17 nrzkt est31
07:17 est31 heh
07:17 hmmmm oh... on the PR i guess?
07:17 est31 no, nrzkt i am first!
07:17 hmmmm i don't read github as much as irc
07:18 nrzkt :p
07:18 est31 :)
07:18 est31 you are the third +1 :p
07:18 est31 whatever
07:18 hmmmm alright, pushed
07:18 est31 k
07:19 hmmmm what do you guys think about 2886
07:20 RealBadAngel hmmmm, you have asked me about size, 512. told you that you should change it to 256
07:20 hmmmm huh??
07:21 est31 hmmmm, I dont know that uch about that part of the code, and it seems complicated
07:21 hmmmm est31:  if it's only that much of a difference then maybe I calculated it wrong or something
07:21 RealBadAngel shall i find the logs?
07:21 est31 hmmmm, perhaps its slow on freebsd?
07:21 hmmmm err, @the utf8 performance
07:21 hmmmm yeah it is
07:21 est31 can you test it?
07:21 hmmmm I did test it
07:21 hmmmm this is with GNU iconv
07:21 est31 whats the numbers
07:22 hmmmm like... 8-11 microseconds per conversion
07:22 hmmmm not horrible
07:22 hmmmm just something to not put in a tight loop
07:22 RealBadAngel also, https://github.com/minetest/minetest/pull/2893/​files#diff-f20c82d93887e9a0f4d4eeb66cc7eaa7R182
07:22 hmmmm it's definitely something that's not worth it for converting a string
07:22 RealBadAngel why the loops are on the same level?
07:22 est31 why not
07:23 est31 "dimensions"
07:23 hmmmm RealBadAngel, whenever there's iteration over dimensions or coordinates you can do that to reduce the amount of indentations
07:24 RealBadAngel about the names, mmblock ? is_radar ? excuse me, is what? where?
07:24 RealBadAngel another loop: https://github.com/minetest/minetest/pull/2893/​files#diff-f20c82d93887e9a0f4d4eeb66cc7eaa7R399
07:25 est31 hmmmm, what about if I make some of the constants translateable? Of course not those that are sent by the server.
07:26 Darcidride joined #minetest-dev
07:26 hmmmm yeah, I guess os
07:26 hmmmm so
07:27 hmmmm just don't make strings that are obviously never going to be utf-8 encoded ever pass through the utf-8 to wide version
07:30 VanessaE I just pulled and built (including hmmmm's just-pushed commit), I don't see anything obviously wrong with the minimap.
07:30 RealBadAngel hmmmm, and thx you havent asked me about merging this. you have said that i will be able to review it. when i was reading the commit you have just merged it.
07:31 hmmmm I had +2
07:31 hmmmm besides, what's wrong with it?
07:31 hmmmm are you mad?
07:31 VanessaE well there is *one* possible glitch - as you ascend/descend, the depth of the map doesn't change smoothly.  seems to update a mapblock at a time
07:32 RealBadAngel next time i will never ask for your vote and what do you think about the code
07:32 RealBadAngel you have even changed code formatted up to your suggestions
07:32 hmmmm VanessaE:  I didn't change anything related to that
07:32 hmmmm RealBadAngel:  Like what
07:32 VanessaE hmmmm: ok.
07:32 hmmmm RealBadAngel:  I didn't expect you to write the code looking like it does now initially
07:33 hmmmm that's just some cleaning and perfecting things
07:33 hmmmm like what I did with Hud
07:34 hmmmm RealBadAngel:  By the way what was that quote you were going to show me before??
07:34 RealBadAngel nvm, i wont waste my time, will make a pr to fix it
07:35 RealBadAngel also i need to check if functionality hasnt changed
07:35 hmmmm RealBadAngel:  why are you taking my code review so personally
07:35 hmmmm besides, you submitted it after only changing like 1/4th of what I asked for in the first place
07:36 proller joined #minetest-dev
07:36 RealBadAngel you havent asked me how it works, your questions were like "what radar is"
07:36 VanessaE RealBadAngel: it hasn't changed.  seems exactly the same to me.  but check that someone didn't break something regarding height/depth scanning.  Also, right-shift + F9 does not toggle the shape.
07:36 hmmmm RealBadAngel:  my point is that "radar" is a vague name for a boolean option
07:36 hmmmm something like "is_radar" means a lot more
07:37 RealBadAngel such massive changed pushed in a hurry, have i heard some1 complainin a lot bout it?
07:37 RealBadAngel minimap is not a radar, minimap can be in radar mode
07:38 est31 hmmmm, isn't it tempting to completely remove narrow_to_wide at some point?
07:39 RealBadAngel so rather "radar_enabled"
07:39 hmmmm est31:  not really
07:39 hmmmm est31:  then again I don't bother much with character set conversions
07:39 hmmmm maybe it's more of a pressing issue for you than it is me
07:40 est31 how much slower is iconv for you than narrow_to_wide?
07:41 hmmmm i never compared the two
07:41 hmmmm i'll do that
07:41 RealBadAngel est31, simple leaves are faster circa 13% for me, 16% for kahrl
07:41 hmmmm hold on
07:42 RealBadAngel also ive coded already support for simple to use alternative texture
07:42 est31 ?
07:42 RealBadAngel https://imgrush.com/sA_zAuziASES.png
07:42 RealBadAngel this is an example texture with simple mode
07:43 Amaz joined #minetest-dev
07:43 est31 well, its ugly
07:43 est31 but if people do report speedups
07:44 est31 and if some even like it
07:44 RealBadAngel http://pastie.org/10279179
07:44 RealBadAngel using alternative is as easy as that ^
07:44 est31 ?
07:44 est31 what are special_tiles?
07:44 RealBadAngel if no alternative given, original is used for all modes
07:44 RealBadAngel special tiles are the 2nd set of tiles each node has
07:45 RealBadAngel theyre exactly the same as tiles, used when you need more textures or whatever
07:50 RealBadAngel est31, anyway, makin a good texture for simple leaves is up for mtgame team, im not any good with 16px
07:51 est31 RealBadAngel, still, they are ugly
07:51 RealBadAngel some like apples some peas
07:51 est31 I +1 if you dont change the default
07:51 RealBadAngel for paramat best lookin are opaque
07:51 est31 leave default to fancy
07:52 RealBadAngel i will update the pr in a few minutes, need to edit config example and that lua stuff
07:52 RealBadAngel btw, the dropdown stuff is copied from other dds
07:53 est31 yea seen it
07:53 RealBadAngel so you may want to clean the rest after im done
07:53 est31 yea
08:01 Yepoleb_ joined #minetest-dev
08:02 est31 so you +1 #2891 hmmmm ?
08:02 ShadowBot https://github.com/minetest/minetest/issues/2891 -- Use UTF-8 instead of narrow by est31
08:03 est31 I will make the maximum viewing range thing translateable
08:03 hmmmm yea
08:03 hmmmm +1
08:03 est31 and perhaps I'll make an extra PROJECT_NAME_WIDE
08:04 est31 automatically converted out of  PROJECT_NAME
08:04 RealBadAngel est31, https://github.com/minetest/minetest/p​ull/2892#discussion-diff-34115348R377 i will leave that as is
08:05 RealBadAngel one should change them all in chain
08:05 RealBadAngel not only my dd
08:06 est31 the break isn't that hard I think.
08:06 est31 but its mostly a nitpick
08:06 est31 ill do a cleanup there later
08:06 rubenwardy joined #minetest-dev
08:10 RealBadAngel ok
08:10 est31 hmmmm, which file is in your opinion a good place to put a static variable
08:11 est31 for the project name
08:11 est31 just in wchar
08:11 hmmmm iunno
08:11 hmmmm anything works reall
08:11 hmmmm really*
08:11 est31 I want to automatically generate it from the normal char'd project name
08:12 est31 or I don't do it, its not worth those microseconds
08:12 est31 Tesseract can do these things better
08:13 cib0 joined #minetest-dev
08:15 RealBadAngel est31, pushed the changes
08:15 RealBadAngel will wait just for travis and merge it
08:16 est31 man this gettext is stupid
08:16 est31 you have no way to format your string
08:19 est31 I'll push without gettext
08:19 est31 pushed RealBadAngel
08:27 est31 ok about translation PRs
08:28 est31 I think we should decline such PRs and point to transifex
08:28 VanessaE I'd say run them through google translate, and if nothing really nasty shows up, just merge them.
08:29 est31 we have weblate now
08:29 VanessaE yes
08:29 VanessaE never used it myself, though.
08:29 est31 I myself am not sure how to handle it
08:29 est31 I think there is util/updatepo.sh
08:29 VanessaE but you're gonna be hard-pressed to get someone who didn't use it in the course of a translations PR to then use it to verify their existing PRs or whatever
08:29 est31 has to be run regularly
08:30 est31 right too
08:33 est31 I guess people should just cherry pick the translation commits from weblate
08:35 est31 nrzkt, can you look at #2843
08:35 ShadowBot https://github.com/minetest/minetest/issues/2843 -- Update Japanese Translation by Rui914
08:36 est31 after that I'll run updatepo.sh to show the stuff in weblate
08:39 nrzkt I'm not japanese :p
08:40 Calinou joined #minetest-dev
08:42 RealBadAngel hi Calinou, could you make texture for simple leaves maybe?
08:44 Calinou I'm not good at making pretty leaves, look at how they look in Carbone
08:44 Calinou feel free to take the texture from it :P
08:44 Calinou https://forum.minetest.net/v​iewtopic.php?f=4&t=9428
08:54 julienrat joined #minetest-dev
08:54 julienrat left #minetest-dev
09:03 yang2003 joined #minetest-dev
09:06 RealBadAngel https://github.com/minetest/minetest_game/pull/555
09:06 RealBadAngel its tweaked ice texture
09:06 VanessaE +1
09:06 yang2003 Good
09:08 AnotherBrick joined #minetest-dev
09:08 est31 RealBadAngel, can you drag& drop the images?
09:08 RealBadAngel just added screenshots to compare
09:08 est31 links can rot
09:08 RealBadAngel ok
09:09 RealBadAngel est31, do you expect the pr to hang here for ages?
09:09 est31 no, but every PR gets archived.
09:09 est31 I know in this case its perhaps not needed
09:10 est31 but always better that way
09:14 RealBadAngel done
09:15 RealBadAngel pushing now #2892
09:15 ShadowBot https://github.com/minetest/minetest/issues/2892 -- Add new leaves style - simple (glasslike drawtype) by RealBadAngel
09:36 cib0 joined #minetest-dev
09:38 chchjesus joined #minetest-dev
09:45 asl97 ice with border (glasslike_frame): https://cloud.githubusercontent.com/assets/4752645​/8567543/bd53c9f8-2598-11e5-8574-dbc990e24ebe.png
09:46 RealBadAngel do you know of course that glasslike framed is way slower?
09:47 RealBadAngel this way you disable grouping nodes in fastfaces rows also
09:47 RealBadAngel not quite good idea for a node being massively used in the world
09:51 asl97 i think i lost 50% fps lol, but it does look better though
09:51 RealBadAngel it does, but price is too high
09:52 asl97 hmmm.... why is it so slow though, there should be a cache or something
09:52 RealBadAngel because of autoconnecting
09:53 RealBadAngel and more vertices and textures used
09:53 RealBadAngel you will be able to add such border when connected textures will go in
09:55 RealBadAngel http://www.minecraftforum.net/forums/m​apping-and-modding/resource-packs/reso​urce-pack-discussion/1256343-a-texture​-artists-guide-to-mcpatchers-features
09:56 RealBadAngel i want to code such thing one day
10:04 sloantothebone joined #minetest-dev
10:24 twoelk joined #minetest-dev
11:12 sloantothebone joined #minetest-dev
11:30 proller joined #minetest-dev
11:34 sloantothebone joined #minetest-dev
11:35 sloantothebone joined #minetest-dev
11:37 sloantothebone joined #minetest-dev
11:38 sloantothebone joined #minetest-dev
11:40 sloantothebone joined #minetest-dev
11:47 sloantothebone joined #minetest-dev
11:49 sloantothebone joined #minetest-dev
11:50 sloantothebone joined #minetest-dev
11:51 werwerwer joined #minetest-dev
11:51 sloantothebone joined #minetest-dev
11:53 sloantothebone joined #minetest-dev
11:54 sloantothebone joined #minetest-dev
11:55 sloantothebone joined #minetest-dev
11:57 sloantothebone joined #minetest-dev
11:58 sloantothebone joined #minetest-dev
11:59 sloantothebone joined #minetest-dev
12:01 sloantothebone joined #minetest-dev
12:03 Krock joined #minetest-dev
12:12 julienrat joined #minetest-dev
12:12 julienrat left #minetest-dev
12:17 sloantothebone joined #minetest-dev
12:19 sloantothebone joined #minetest-dev
12:49 julienrat joined #minetest-dev
12:50 julienrat left #minetest-dev
12:51 proller joined #minetest-dev
13:13 H-H-H joined #minetest-dev
13:36 deltib_ joined #minetest-dev
13:39 twoelk|2 joined #minetest-dev
13:46 RealBadAngel why the fuck github wants to merge all changed trees into minetest_game?
13:46 RealBadAngel i cannot make separate pulls at all
13:49 Amaz RBA, I doubt that this is the case, but are you making the simple_leaves branch from the new_ice branch? Or has the ice commit somehow got onto your master?
13:50 RealBadAngel its in separate branch
13:50 Amaz Well, then, it's obviously just git being it's usual nightmareish self :)
13:52 VanessaE joined #minetest-dev
13:53 RealBadAngel ok, made it finally
13:54 RealBadAngel VanessaE, can you review default:leaves simple texture?
13:55 VanessaE the texture looks fine, no time to do a proper review though
13:55 VanessaE bbl
14:28 sloantothebone joined #minetest-dev
14:30 sloantothebone joined #minetest-dev
14:31 sloantothebone joined #minetest-dev
14:32 sloantothebone joined #minetest-dev
14:34 sloantothebone joined #minetest-dev
14:34 deltib_ joined #minetest-dev
14:35 sloantothebone joined #minetest-dev
15:04 CraigyDavi joined #minetest-dev
15:05 hmmmm joined #minetest-dev
15:35 proller joined #minetest-dev
15:48 hmmmm hey wait a minute
15:48 hmmmm you can't just remove a global setting like that
15:49 hmmmm now peoples' configs are going to be broken
15:56 hmmmm wow, simple leaves are ugly
16:14 kahrl joined #minetest-dev
16:23 proller joined #minetest-dev
16:39 jin_xi joined #minetest-dev
16:48 kaeza joined #minetest-dev
16:53 Robert_Zenz joined #minetest-dev
16:56 MinetestForFun joined #minetest-dev
17:18 paramat joined #minetest-dev
17:21 paramat hi sfan5, can we merge game#551 on your single +1 since it is a mesh thing checked by RBA and not a new feature?
17:21 ShadowBot https://github.com/minetes​t/minetest_game/issues/551 -- Convert stairs in meshnode by kilbith2
17:23 paramat also game#555 it's not great but will do for now, current one is bad and needs to go
17:23 ShadowBot https://github.com/minetes​t/minetest_game/issues/555 -- Tweaked ice texture - seamless one. by RealBadAngel
17:25 rubenwardy joined #minetest-dev
17:31 sfan5 paramat: merging 551 w/o two approvals is fine with me
17:33 paramat okay thanks, i might merge this later
17:36 paramat wow, i'm bisecting the mystery memory leak, present in 0.4.9 and i'm still searching backwards
17:41 H-H-H joined #minetest-dev
17:57 RealBadAngel joined #minetest-dev
18:04 paramat now pushing game 551
18:10 paramat complete
18:13 sloantothebone joined #minetest-dev
18:14 sloantothebone joined #minetest-dev
18:15 sloantothebone joined #minetest-dev
18:16 sloantothebone joined #minetest-dev
18:17 sloantothebone joined #minetest-dev
18:18 sloantothebone joined #minetest-dev
18:20 sloantothebone joined #minetest-dev
18:21 sloantothebone joined #minetest-dev
18:22 sloantothebone joined #minetest-dev
18:23 sloantothebone joined #minetest-dev
18:24 sloantothebone joined #minetest-dev
18:25 sloantothebone joined #minetest-dev
18:26 sloantothebone joined #minetest-dev
18:28 sloantothebone joined #minetest-dev
18:29 hmmmm sloantothebone, please fix your connection
18:29 sloantothebone !!!
18:29 sloantothebone I just got banned in a channel for my connection! Am I banned forever??
18:30 sfan5 no
18:30 sfan5 you need to ask the channel admins to remove the ban
18:30 sloantothebone Ok look if u have a problem with my connection kick me
18:31 sloantothebone There must be some kind of bot that kicks people with bad connection
18:31 sloantothebone I'm sxorry, really dont want to spam you guys
18:31 hmmmm it seems to have subsided
18:31 Amaz Maybe turn irc off till it is fixed?
18:35 sloantothebone How do I make it turn off automatically if bad connection
18:35 kaeza sfan5, do you have a copy of libgcc_s_sjlj-1.dll you could host on your server?
18:36 sfan5 mine is probably not compatible with yours
18:36 sfan5 you should have that file come with gcc
18:36 kaeza apparently my MinGW doesn't come with it
18:36 sfan5 did you do find . -name '*.dll'?
18:38 kaeza `find /c/MinGW -name 'libgcc_s_sjlj*.dll'` turns up nothing
18:39 kaeza everything MinGW-related is under that dir
18:39 sfan5 using cygwin?
18:39 kaeza no, plain MinGW
18:39 sfan5 no
18:39 Calinou sloantothebone, which client do you use? perhaps disable automatic reconnection or increase its delay.
18:40 sfan5 >/c/MinGW
18:40 sfan5 thats not an usual linux path
18:40 kaeza sfan5, using mgw natively under win32, not cross-compilation
18:41 sfan5 kaeza: anyway, download this https://sfan5.pf-control.de/minet​est-builds/latest.php?b=minetest and copy it from the bin/ dir
18:42 kaeza sfan5, it's not specifically for me; I'm almost finished doing a modified buildbot so MT can be compiled natively under Windows with MinGW with a simple command
18:42 sloantothebone Calinou I increased the delay x3
18:42 kaeza so it would benefit to have a static URL from which to fetch a working libgcc dll
18:42 sfan5 i see
18:49 sfan5 kaeza: win32 https://sfan5.pf-control.de​/dll32/libgcc_s_sjlj-1.dll and https://sfan5.pf-control.de/dll32/libstdc++-6.dll
18:49 sfan5 kaeza: win64 https://sfan5.pf-control.d​e/dll64/libgcc_s_seh-1.dll and https://sfan5.pf-control.de/dll64/libstdc++-6.dll
18:50 kaeza sfan5, thank you. testing
18:55 kaeza surprise, surprise, doesn't work :(
18:56 paramat left #minetest-dev
18:59 kaeza disabling leveldb removes the dependency though, so I guess building leveldb from sources should fix it
19:00 kaeza or I could just leave leveldb disabled
19:00 CraigyDavi joined #minetest-dev
19:16 err404 joined #minetest-dev
19:31 kaeza unrelated: https://github.com/minetest/minetest/pull/2895
19:44 paramat joined #minetest-dev
19:53 paramat i'd like to push this later #2894 hmmmm you're probably busy but just letting you know. i want to keep this simple, no flags =)
19:53 ShadowBot https://github.com/minetest/minetest/issues/2894 -- Biome API decorations: 'spawnby' searches a 2 node high volume by paramat
19:55 RealBadAngel paramat, shall i fix the accacia leaves texture for ya?
19:55 hmmmm paramat, looks good but what about updating the documentation?
19:55 hmmmm and are you sure that solves kilbith's problem?
20:03 proller joined #minetest-dev
20:04 sloantothebone I cannot connect to a minetest server with my public ip, I think its because of my ISP what do I do?
20:07 Robert_Zenz joined #minetest-dev
20:10 sfan5 sloantothebone: you need to forward ports, also this channel is strictly for minetest development, ask such questions in #minetest
20:10 selat joined #minetest-dev
20:20 paramat thanks, will update docs. and am consulting kilbith
20:20 paramat RBA i wanna do it, thanks anyway
20:29 paramat yes i'm fairly sure this solves kilbith's problem while also working for the other usage relative to the ground beneath the decoration
20:48 paramat updated
20:50 dv- joined #minetest-dev
20:50 dv- ENABLE_FREETYPE=FALSE is broken?
20:53 paramat will push when checks are done
20:54 CraigyDavi joined #minetest-dev
21:01 RealBadAngel paramat, added screenshot and jungle leaves simple
21:02 * paramat looks
21:11 paramat they look okay in the screenshot, i'll test in-game too
21:23 troller joined #minetest-dev
21:42 paramat now pushing 2894
21:47 paramat complete
21:48 paramat left #minetest-dev
22:30 est31 joined #minetest-dev
22:36 troller joined #minetest-dev
22:42 est31 now weblate people have edited the japanese translation
22:42 est31 it conflicts with #2843
22:42 ShadowBot https://github.com/minetest/minetest/issues/2843 -- Update Japanese Translation by Rui914
22:42 est31 VanessaE, seems there is a reason to decline translation PRs
23:20 kaeza we have working weblate again?
23:20 kaeza translation PRs were temporary because weblate was broken
23:23 kaeza est31, I didn't use NULL in GetTempPath because MSDN doesn't even specify if it is a valid value
23:24 est31 kaeza, it should work, because if the length is 0, a dereference of the pointer should be illegal
23:32 kaeza est31, pull updated
23:34 est31 looks good, nerging
23:34 est31 merging*
23:52 RealBadAngel joined #minetest-dev
23:54 AnotherBrick joined #minetest-dev

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