Minetest logo

IRC log for #minetest, 2015-03-09

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

All times shown according to UTC.

Time Nick Message
00:04 Viper168 joined #minetest
00:53 Xenoth joined #minetest
00:53 Teckla left #minetest
00:53 QuailChaser joined #minetest
00:54 est31 Can somebody ban menche?
00:54 est31 or "QuailChaser"
00:55 Xenoth joined #minetest
00:55 alket hi est31
00:55 est31 Hi alket
01:00 Enke joined #minetest
01:39 luizrpgluiz joined #minetest
01:39 luizrpgluiz left #minetest
01:44 WarrTab Wasn't there some trick to force a lighting recalc for when lighting gets screwed up?
01:44 luizrpgluiz joined #minetest
01:44 luizrpgluiz hi all
01:44 WarrTab Apparently trres growing caused it for me...
01:45 est31 WarrTab are you using latest git?
01:45 est31 on the server
01:45 WarrTab Stable-0.4
01:45 est31 ugs dunno if paramats latest light fix is in there
01:46 VanessaE WarrTab: you can use worldedit to fix badly-lit areas if it's just a super-dark shadow that came out of nowhere
01:47 est31 if the worldedit on-board fix lights command doesnt work, do //replace air air
01:49 WarrTab Hm, ugh, all of these require that I install something that I'm not sure I want staying on the server...
01:49 WarrTab I hear worldedit is not for the faint of heart
01:50 luizrpgluiz left #minetest
01:51 WarrTab Crap, and I had JUST had this idea for floodlights
01:52 WarrTab Was trying to work out how I might code them to not grind a server to a halt if people use them everywhere
01:52 Zeno` joined #minetest
01:59 WarrTab So apparently you can do some pretty interesting stuff if you sneak and jump at the same time
02:00 Zeno` acrobatics!
02:01 WarrTab Like building a ladder out of blocks and climbing it really fast
02:01 WarrTab Or jumping up onto a 2-block-high wall
02:02 WarrTab Sneaking also seems to let you catch yourself while falling, stopping instantly with no fall damage
02:02 WarrTab You just need an edge of a node to catch onto
02:03 WarrTab Hey Zeno`
02:03 Zeno` hello, acrobat
02:03 WarrTab hm, sadly my android irc client has no tab completion, though I can imagine why they might have left it out
02:04 Zeno` I mean WarrTab
02:04 WarrTab Same diff
02:04 WarrTab Hey, since you reviewed pr 2462 already, feel like merging it?  :-)
02:05 stormchaser3000 left #minetest
02:05 WarrTab It feels awkward running clients with unofficial features long-term...
02:39 MinetestBot [git] Zeno- -> minetest/minetest: Stop formspecs closing with double-click in empty area 76177c8 http://git.io/ps7B (2015-03-09T12:35:37+10:00)
02:43 Zeno` 2462?
02:43 MinetestBot [git] Zeno- -> minetest/minetest: Stop formspecs closing with double-click in empty area e74b8da http://git.io/ps7x (2015-03-09T12:42:25+10:00)
02:51 paramat joined #minetest
02:52 Guest89181 left #minetest
02:57 whtemple1959 joined #minetest
03:01 WarrTab Zeno`:  https://github.com/minetest/minetest/pull/2462
03:01 whtemple1959 VanessaE: after single daddin my 5 year old son I am back to the challenge. I just saw an ostrich walk past sooo...Thank you very much for the advice on getting animals into our world
03:01 VanessaE no rpob
03:01 VanessaE no prob*
03:01 est31 VanessaE, which animal mod did you suggest?
03:02 VanessaE I didn't - I just explained how to open the setup dialog, in mobf :)
03:02 est31 ah
03:02 est31 ok :)
03:02 whtemple1959 est31: the same to you soooo...Thank you very much on the screwdriver usage to fix my slab problem. My stairs look beautiful now that the huge blocks are not hanging down in the foyer.
03:04 paramat2 joined #minetest
03:04 Zeno` Warr1024, I need approval from another core dev. I'll +1 it though so there only needs one other approval
03:04 WarrTab Thanks
03:04 whtemple1959 Well I still believe that you both have been most helpfull
03:04 Zeno` Warr1024, how does it look and perform on Android btw?
03:04 paramat2 FFS why am i paramat2?
03:04 VanessaE does my untested +1 count? :)
03:05 VanessaE paramat2: because you've been cloned.  :P
03:05 Zeno` untested +1's are a bit suspicious heh
03:05 WarrTab Zeno`: looks good, but performs rather meh, I usually leave it disabled... but then again, could be my old tablet.
03:05 paramat2 yeah i just registererd with nickserv
03:05 WarrTab I've never tried logging into a world w/ 64x txrs before on it, which is the setting I usually try.
03:06 WarrTab I guess I could test 32x...
03:06 paramat2 didn't get my password into the client in time, it said paramat in use
03:06 Zeno` Warr1024, I know that most of the rest of the code has a lack of comments but maybe your PR could do with a sprinkle of informative ones
03:06 Zeno` e.g. what u32 sx = (ctrx < 1) ? 0 : (ctrx - 1)  is doing
03:06 WarrTab I could do that, was just trying to be consistent.
03:07 WarrTab that's essentially a cheap box filter w/ non-wrapping bounds check
03:07 Zeno` Warr1024, the only thing consistent about the current codebase is that it's consistently lacking in comments where comments would be beneficial :)
03:08 Zeno` I like comments when they're helpful.  I don't understand some of my *own* code after several months of not looking at it and I'd rather not go through things line by line re-interpreting them heh
03:08 paramat2 apparently i am now identified as paramat, but appear here as second nick choice
03:08 WarrTab for each pixel in image, if fully transparent, average rgb values of neighbor pixels weighted by alpha, and set rgb of current pixel to that, if non-zero weight
03:08 WarrTab Ok, will add some comments...
03:09 WarrTab also not sure how much should go in minetest.conf.example as opposed to code.
03:09 Zeno` I don't mean useless comments. Judging from your code I think you know what comments are generally regarded as useful though ;)
03:09 WarrTab X = x + 1 // add x to 1
03:09 Zeno` yeah, like that :-o
03:09 paramat2 left #minetest
03:09 Zeno` lol
03:10 Zeno` :P
03:10 WarrTab "no, comments should explain WHY"... x = x + 1 // because I'm the programmer and you're a computer and you have to do what I say.
03:11 exio4 I've seen a lot of people write comments like those
03:12 exio4 I cried
03:14 WarrTab My favorite bad comments are the ones where a subsequent dev has gone in and added explanation as to why the code is bad, but didn't fix the code.
03:14 WarrTab Multi-line comments, // // // style, or /* * */ style?
03:15 est31 WarrTab, http://dev.minetest.net/Code_style_guidelines
03:16 WarrTab thanks, but it didn't answer that specific question...
03:16 whtemple1959 est31: if you feel like still playing fix it felix maybe you can help me with a couple of issues concerning the carpets?
03:17 WarrTab I guess it doesn't much matter?
03:17 est31 wow lol
03:17 exio4 WarrTab: at least those make sense
03:17 est31 being fix it felix  is fun when you are bored
03:18 est31 but yes I help you :)
03:18 exio4 WarrTab: I normally write comments on "hacky code" like "this code is supposed to do X, Y and Z, but using W instead of P and Q"
03:18 * est31 is partially bored, partially procrastinating network rewrite of technic
03:18 exio4 "because P and Q don't work with R"
03:19 WarrTab What I meant was if your comment is going to wrap a few lines, use // before each line, or one of the /* */ variants?
03:19 Zeno` I prefer (in minetest) // comments inside a code block that are one line
03:19 whtemple1959 ok then I hope you are bored...when I used the carpet my furniture floated. I attempted the screw driver fix but that had no effect. again I seem to be googling the wrong phrases because I find no answers to the problem...but I am sure the solution is as easy as the screwdriver one
03:20 Zeno` /* */ I use before the code block and also /* */ inside a code block if it looks nicer
03:20 Zeno` kind of subjective
03:20 Zeno` multiline comments I nearly always use /* */
03:20 Zeno` I think that's what most of the other devs do as well
03:21 Warr1024 cool
03:23 est31 whtemple1959 so you want to place the carpet at the bottom, and the furniture above?
03:24 est31 I guess that your issue is rooted in the fact that minetest is a voxel game.
03:25 whtemple1959 est31: well it seems that the carpet is at the bottom of the node..so when you place it and then try to set a cahir on the carpet the chair appears to float above the carpet
03:25 est31 whtemple1959, yes its because of minetest being a voxel game
03:26 est31 nothing to "fix here" I guess
03:26 est31 you can't have two nodes occupying the same position
03:26 paramat2 joined #minetest
03:27 whtemple1959 but I do not understand if the slabs can be rotated so they are on the top side of the node why can't the carpet be rotated in the same nammer thus allowing the chair to sit on the carpet?
03:27 paramat2 nick/channel temporarily unavailable, so i am now my own clone =/
03:28 WarrTab Not all nodes can be rotated
03:29 whtemple1959 well then I would have to ask why the designer chose to put the carpet on the bottom of the node rather than the top?
03:29 paramat2 i think you can create a node that appears below it's real position, Sokomine did with snow
03:30 paramat2 you'll need to create a floor node with carpet at it's top
03:31 whtemple1959 paramat2: oh you make me laugh your taxes I can do write codi I do not...
03:32 paramat2 WarrTab, did you get shadows when lots of saplings grew? i noticed that
03:32 whtemple1959 paramat2: but that does seque into my other question when I am looking up at the ceiling in the basement I see the carpet. would there be a way to put a slab at the bottm of a node and the carpet at the top?
03:33 WarrTab paramat2: yeah, exactly what happened to me, running stable-0.4 engine and _game.
03:33 paramat2 yeah, it would be a single node with top half carpet, bottom half floor
03:33 WarrTab Whtemple: no, you'd need 2 node spaces for that
03:33 whtemple1959 paramat2: now...is there simple instruction for an idiot to follow which would allow me to do just that?
03:33 Zeno` Warr1024, don't be afraid of functions if they're appropriate either
03:33 paramat2 nope
03:34 Zeno` the current code has a distinct lack of structure and functions
03:34 WarrTab I usually make my floors 2n thick to ensure I can make the ceiling and floor of different materials, if I'm going to want to do that.
03:34 paramat2 just make a node and split the texture top/bottom
03:34 paramat2 yeah good idea, i prefer 2 deep floors so i can have deep pile carpeting
03:35 WarrTab Zeno: I would, but I wanted to keep my code all in 1 place, I don't expect much of this to be reusable, and frankly my C++ probably isn't up to the task.
03:35 paramat2 2 floor + 3 air = 5 nodes per floor
03:36 WarrTab 4 air if you want the place to be cool in the summertime, and aren't worried about heating bills in the winter
03:37 Zeno` WarrTab, maybe it's not appropriate for that PR. I was being more general really :)
03:37 paramat2 carpet as the bottom of a room node is a bad idea
03:38 WarrTab well, there may be others someday, so thanks for the tip :-)
03:39 paramat2 i generate big buildings so minimise nodes per floor
03:41 WarrTab I usually use full-node flooring, like wool blocks
03:42 WarrTab Or inverted slabs to save material, but always something walkable.
03:42 WarrTab Now compiling to test
03:42 WarrTab All I changed was comments, but I still gotta be sure :-)
03:42 whtemple1959 paramat2: I like you sugestion to create a node with a slab at the bottom and carpet at the top. what would be even better if the slab could be edited to allow for different textures in the level below the carpet level. I am going to see if I can find out who made the carpet blocks and contact them. but thank you for the advice.
03:43 WarrTab the problem with your compound node idea is that each possinble combination would need to be a separate node, as I understsnd the engine
03:43 paramat2 yep
03:43 WarrTab So that's n**2 nodes for all materials
03:44 WarrTab If you had q00 different materials 9including color variations) it'd take 10,000 nodes to do the top/bottom combo
03:44 est31 yea the matrix is getting to much dimensions now
03:44 WarrTab Which is probably the main reason you don't already see this mod out there
03:46 Trixar_za joined #minetest
03:46 WarrTab https://github.com/Warr1024/minetest/commit/835987fc5110589762a9f2c8ae5a27fcf63ce32d
03:46 WarrTab Tbh I think I usually over-comment my code
03:47 WarrTab It actually serves a secondary function in indicating how confident I am that that code is correct.
03:47 WarrTab I.e. by investing an additional 80% in comments that'd have to be thrown away if the code is, I guess :-)
03:49 WarrTab I guess my comments didn't address the "why not just use a highres texturepack?" argument, but maybe enough is enough already...
03:50 whtemple1959 call me nieve but if plastic sheets, icu, and steel creates a stero or some plastic, speaker, and a digital clock creates an alarm clock then why couldn't 3 slab on th bottom and 3 wools across the dop make a floor node of stone and carpet?
03:51 paramat2 it could
03:51 whtemple1959 so to quote Cap't Picard....how do I make it so?
03:52 VanessaE you need to define a node with properties like that, and make a recipe that can craft those materials into it
03:53 WarrTab Meh, fixed some typos, added that explanation while I was at it.
03:54 whtemple1959 VanessA: I saw a tutorial on how to make a node...I am gone to torment myself before bed...but again thank you for the animals I saw some ducks and chicks... my sone will be so happy that Daddy got him pets...thank you so very much
03:55 WarrTab VanessaE: your issue re: formspec images scaling like crap, btw, might be a separate problem.
03:55 WarrTab I noticed that the android gui controls look shitty too, and they're very high-res images
03:56 WarrTab Seems like it shouldn't be too hard to enable blended downscaling, at least, for 2d gui elements
03:56 VanessaE WarrTab: it's everywhere in minetest formspecs
03:56 whtemple1959 Again to all of you thank you so much for the advice
03:56 VanessaE all image elements are scaled by only nearest neighbor - nothing better is used, let alone a better set of steps to get to the desired size
03:56 VanessaE (e.g. see my revised comment)
03:57 VanessaE that's a different issue from the thing you were solving - I only referenced your pull because it does a similar job for in-world images
03:57 VanessaE github just decided to make a connection between the two is all
03:57 WarrTab irrlicht really *should* offer a better scaling option; if it doesn't, then maybe we can do it in software and cahce the result; 2d manips are relatively cheap
03:58 VanessaE Zeno`: *poke*
03:58 VanessaE irrlicht apparently does offer better methods, but minetest doesn't use them
03:58 WarrTab Alternatively, it'd be nice to be able to just draw those using the existing 3d scaler; if my pr gets merged, we'll rarely have a need to turn off 3d filters :-)
03:59 WarrTab I wrote my first irrlich api calls in that pr, so maybe I should go diggjng in the formspec code to flex my newfound skill
04:00 paramat ah at last
04:00 VanessaE did you see the test image I posted?
04:01 WarrTab I swear, sometimes I don't know whether I enjoy digging nodes or digging lines of code in minetest more.
04:01 WarrTab Yeah, saw it
04:03 paramat left #minetest
04:03 WarrTab You mean the one in the issue report, right?
04:03 VanessaE https://github.com/minetest/minetest/issues/2419#issuecomment-77733241
04:03 VanessaE this
04:09 WarrTab Lanczos would be beyond the level of complexity I'd feel comfortable with in this context
04:09 WarrTab But just doing dimple pixel antialiasing shouldnt be too bad
04:09 WarrTab Code to scsle the image won't be hard
04:10 WarrTab Code to manage a cache of scaled images, so as not to rescale them on every frame, moreso.
04:10 VanessaE there's apparently an irrlicht function to do this, but I couldn't find too much about this whole thing - too much noise, not enough signal (and I don't understand C++ enough, let alone irrlicht)
04:11 WarrTab I didn't see anything that looked like what we want
04:12 WarrTab Irrlicht in general doesn't seem tomhave much support for the "scale up, but keep the pixellated aesthetic"
04:12 VanessaE that's why I suggested doing it in two steps.
04:12 WarrTab In my filter, I scaled up nearest and let the video caard scale down linear; that MIGHT work, but might need some tricky tuning
04:13 WarrTab I guess up-nearest-down-linear could be a good first implementation, until something more clever comes along
04:13 VanessaE yeah
04:13 VanessaE that'd look better than the current way, for sure
04:14 WarrTab It would be nice if I could just cache the up-nearest to a minimum size, like I do with my PR, then let irrlicht downsize it linearly as needed.
04:15 WarrTab Problem is I don't see a draw2dImage overload that lets me ask for a filter
04:15 VanessaE copyToScaled is what I ran across earlier
04:15 VanessaE useful?
04:15 WarrTab If I have to do the linear downscale to an jmsge instead of the screen, I need to manage thst memory use
04:15 VanessaE Scaling*
04:16 VanessaE good point
04:16 WarrTab copytoscaled I think is actually what I'm using, and assuming it's nearest, in my filter
04:16 VanessaE of course, if you have a limit set to say 64px, then there's probably not much memory usage there anyway
04:17 WarrTab well, it's not the static use so much as dynamic
04:17 WarrTab If i reuse that image on each draw, then it's fine
04:17 VanessaE right
04:17 WarrTab If each time I have to redo any of the software scaling, bye bye framerate
04:18 VanessaE right, didn't think of that.,
04:19 WarrTab Maybe a lot of users could handle the load, but I've got a couple machines myself that are getting their asses kicked just handling the mesh recalcs
04:19 WarrTab Plus, I code in C and C# (and others) but not C++
04:19 VanessaE yeah, better to cache the filtered/scaled results somewehre (disk) and reference those if possible.
04:19 VanessaE of course, there's no point in retaining the original imagery in memory anyway, is there?
04:20 WarrTab Ironically, I can handle malloc and free, but new and delete mystify me, and irrlicht's ref counting gives me the creeps
04:20 WarrTab Well, I'd need to keep the original if they're used for other purposes, including appearing at a different size on a different formspec...?
04:21 VanessaE shit.
04:21 WarrTab Btw, my "texture scaling filter" actually works on media, not textures...
04:21 WarrTab Meaning it should upsample any too-small formspec images, too
04:22 WarrTab Might be worth seeing if using it improves your situation already, if you can compile your own client from that commit
04:23 VanessaE I don't think it'll affect other things like a rendered cube in the formspec, though
04:23 VanessaE that's where the scaling nastiness really stands out
04:23 WarrTab yeah, I guess those render to an image, the stretch the image shittily, instead of rendering in 3d.
04:23 VanessaE yep
04:24 VanessaE render-to-texture is the method I think.
04:24 WarrTab Makes sense for performance reasons, and the ability to reuse that inventory "cube" image in other ways
04:24 WarrTab I wonder if render-to-texture isnthe way to get irrlicht to smoothly scale a 2d image...
04:25 WarrTab I thought inventory cubes were acrually isometric, though, not perspective
04:25 VanessaE so scaled results of that code would have to be cached on disk somehow, indexed/hashed by mod+formspec call or some such
04:25 VanessaE (is there a way to DO that? O_o)
04:25 WarrTab Isometric you can do with a 2d transformation matrix
04:25 VanessaE they are iso, yeah
04:26 VanessaE but idk exactly how they go from singular images to rendered cubes.  I've never looked at the code
04:26 WarrTab I'm just thinking in-memory cache for now.
04:26 VanessaE this is partly kahrl's domain also
04:26 VanessaE *pokes at him for signs of life*
04:26 WarrTab I'll have to dig in the code a bit more..
04:31 WarrTab it looks like what you want is copyToScaling to upscale to the smallest integer size that's at least as large as the target box, then copyToScalingBoxFilter to downscale to the exact tsrget size.
04:31 VanessaE that's more or less it yeah
04:31 WarrTab then probably memoizing those images in an in-memory cache keyed on (image name, target width, target height) would be ideal
04:32 VanessaE that would work, yeah
04:32 Pentium44 joined #minetest
04:32 VanessaE but there's one problem:  what do you do when the user resizes their window?
04:32 WarrTab cache misses
04:32 VanessaE oh ok
04:33 WarrTab If we're smart, we trap that event and flush the cache
04:33 VanessaE mmmm...
04:33 WarrTab If not, well, it'll cost a little wasted ram, but hopefully not much...
04:33 WarrTab And of course the cache could have some kind of purging mechanism for old entries...
04:34 LazyJ joined #minetest
04:35 VanessaE note that at 1600x1200, each final raw image will take a hair over 16k.
04:35 VanessaE (using Unified Inventory's inventory slots as a guide, they render to about 79x75 pixels at that window size)
04:35 WarrTab how are you calculating that?
04:36 VanessaE 79x75 pixel target size ^^^, so 75x75 pixels to make it square == 5625 pixels.  triple that since it would have to be treated as a raw image
04:36 VanessaE that gave me 16'875
04:36 WarrTab x4, not x3
04:36 VanessaE oh right
04:36 WarrTab Rgba, not rgb
04:36 VanessaE RGBA'
04:37 VanessaE so ~22.5 k
04:37 WarrTab A thousand of them would be 22mb
04:37 WarrTab not really problematic on most hardware, though...?
04:37 VanessaE that'd be about 120 MB of usage for Dreambuilder (there are about 5500 images there)
04:37 VanessaE nah, probably not
04:38 VanessaE especially if it's an optional filter
04:38 WarrTab all filters should always be optional :-)
04:39 VanessaE wait, do we count number of individual images, or the number of possible rendered items?  or just the number that would likely be loaded to the user's client at one time?
04:39 WarrTab I hate what plain mipmaps do to glass, and anything more than that makes either my fps or drawdist drop unacceptably on my tablet
04:39 VanessaE s/rendered items/rendered nodes and items/
04:40 WarrTab well, for this filter, it'd be unique on-screen images
04:40 VanessaE ok
04:40 WarrTab So different source images and different sizes
04:40 WarrTab But only the ones player has seen, if we're doing these on demand, which should be feasible
04:41 VanessaE ok so in Dreambuilder's case, that would be somewhere in the neighborhood of 2000 items or so (there are 1833 in the inventory at present), so maybe 60 MB at the worst likely case
04:41 VanessaE and that's if you don't expire old images
04:42 WarrTab and we can probab;y get away with a hard limit, throwing away oldest; I know I'm used to a couple seconds of think-time on opening a  new inventory...
04:42 WarrTab Yeah, if that's the case, we can probably just skip the limit for first implementation
04:43 VanessaE most of the inventory time is spent building meshes, so I doubt scaling a page of images will have more than a negligible impact.  a unified inventory image is only about 450 pixels' worth of target image space.
04:43 VanessaE er 450'000
04:44 VanessaE unified inventory PAGE is 450'000 pixels' worth.
04:44 VanessaE actually, smaller than that.
04:45 WarrTab well, I'm used to playing with only 16x textures, but my new filter expands those automatically, the way I've configured it, to 64x.
04:45 VanessaE 56x57 pixel rendered size for a creative inventory button on the right (at 1600x1200), so that's only about 255'360 pixels' worth.  (there are 80 items on a page)
04:45 WarrTab That's 16 times the memory, and it doesn't hurt.
04:46 WarrTab At 128x (64 times the memory) I feel it, though, and at 256x (256 times) it hurts alright
04:46 VanessaE that's fewer pixels than a single 512px image
04:46 VanessaE so the scaling delay would be next to nothing I think.
04:47 WarrTab But someone using a computer built within the last 5 years might shrug that off.
04:50 WarrTab In your test images, you upscaled to a power of 2, and used lanczos... can you retry using an integer non-power-of-2, like 5, and then linear downscale?
04:50 VanessaE sure
04:50 VanessaE sec.
04:51 WarrTab I think that's a lot more like what we can do reasonably cheaply in irrlicht, but SHOULD give you results about as good.
04:51 WarrTab A "box" resample filter treats pixels as if they're a grid of squares/rectangles of solid color.
04:52 WarrTab Lanczos is a much more complex model using points of light and contrast effects
04:52 VanessaE http://digitalaudioconcepts.com/vanessa/extra/gloopblocks_cement_pick.png
04:52 WarrTab Frankly, if I had my choice of scaler, I might try something like hqx or xBRZ
04:52 VanessaE wait.
04:53 VanessaE used the wrong downscaling algo
04:53 VanessaE there, now you can (re)load it.
04:53 WarrTab Looks good to me
04:53 VanessaE scaled up by 160px (so 10*) with NN, then down to 75px with linear.
04:54 WarrTab No shitty resize artifacts are evident
04:55 WarrTab Btw, if you want the pie-in-the-sky version of this, check out http://en.m.wikipedia.org/wiki/Image_scaling#Pixel_art_scaling_algorithms
04:55 VanessaE linear actually looks pretty reasonable.
04:56 WarrTab I made a texture pack by harvesting my 16px textures and running them through hq4x, and it was neat, even if the heuristic was a bit off in places
04:57 VanessaE some hdx-variant would look nice indeed.
04:57 VanessaE er hQx.
04:58 Erthome joined #minetest
04:59 WarrTab xBRZ looks absolutely awesome, but IIRC the original xBR was written in GLSL, and I haven't found any variant that offers a simple png->png cmdline version for me to play around with
04:59 VanessaE that method looks nice too
05:01 VanessaE too complex to use here though I guess
05:01 VanessaE http://sourceforge.net/projects/xbrz/
05:03 necromonger joined #minetest
05:03 necromonger left #minetest
05:06 OldCoder VanessaE, how does one make black dye?
05:06 WarrTab Crush coal?
05:07 VanessaE OldCoder: I don't remember in the default game, but in dreambuilder you can just craft coal
05:07 VanessaE actually same for the default game too
05:07 WarrTab it can be made the hard way too, I think, like mixing brown and blue
05:07 OldCoder I don't see that; in the games I have it looks like one needs a black flower
05:08 OldCoder But thank you both
05:08 OldCoder WarrTab mixing brown and blue using what mechanism?
05:08 WarrTab Craft them together
05:08 WarrTab should be shapeless
05:08 OldCoder Ah
05:09 OldCoder Brown and blue dyes?
05:09 WarrTab I think
05:09 OldCoder Thank you, worth looking into
05:27 LazyJ joined #minetest
05:38 OldCoder joined #minetest
05:39 barrydk joined #minetest
05:44 _schism_ joined #minetest
06:06 _schism_ joined #minetest
06:06 Viper168_ joined #minetest
06:34 Vazon joined #minetest
06:35 Zeno` joined #minetest
06:42 Zeno` do NOT panic
06:42 Zeno` I am back
06:47 Zeno` I repeat. Do NOT panic!
06:58 fuznuts joined #minetest
06:59 fuznuts I am in creative mode on a worldpainter map I made.  I used shapes to form the structure of buildings, but when I go to make a door, every time I take a block out of the side of the structure, the ones on top drop down to replace it..How do i get around this?
07:02 Miner_48er what are the top nodes?
07:46 fuznuts that is the thing....I am not sure if the higher numbers are the ones closest to the bottom, but that is what it seems like...
07:46 fuznuts which is kind of counterintuitive considering it means something completely opposite elsewhere in the process of starting a map.
07:48 gourytch joined #minetest
07:51 fuznuts the top nodes are sand though.  and i am thinking probably 150 down.
07:51 ThatGraemeGuy sand drops if there is nothing below it, what you are seeing is completely normal
07:54 srifqi joined #minetest
07:57 Haudegen joined #minetest
08:02 srifqi "how to place a node in singlenode-mapgen world?"
08:06 SylvieLorxu joined #minetest
08:09 aheinecke joined #minetest
08:10 BlandCereal joined #minetest
08:11 meldrian joined #minetest
08:11 meldrian Hellowww
08:11 BlandCereal Hi.
08:17 est31 joined #minetest
08:20 OldCoder joined #minetest
08:20 Zeno` est, you just got younger
08:20 Zeno` did you regenerate like The Doctor?
08:21 est I didn't see that series
08:21 est I just know some youtube videos with "I am the doctor"
08:21 Zeno` yeah, it's new. Not many people have
08:22 srifqi how to rename nickname?
08:22 Zeno` /nick <newname>?
08:23 rifqi ah, i see, thanks
08:24 srifqi Zeno`: i think i should learn MORE about irc
08:24 Zeno` srifqi, it is a dark and dangerous world full of voodoo
08:24 srifqi what?
08:24 srifqi XD
08:24 Zeno` lol :D
08:24 est bad part of irc is that every client and every network does their own stuff
08:25 srifqi still thinking about how to "squash" a pr
08:26 srifqi nvm
08:31 srifqi Zeno`: why you have a backslash at your nick?
08:31 est /msg NickServ info Zeno
08:32 est -NickServ- Last seen  : now
08:32 Zeno` backtick
08:32 Zeno` but... when I'm idle I am known as someone else..
08:32 Zeno` like now ;)
08:33 mazal joined #minetest
08:33 srifqi \er
08:33 \er yes?
08:33 srifqi why you change your nick to \er?
08:33 \er it's the nick I use when I am idle
08:33 \er i.e. away from keyboard :)
08:34 est slasher?
08:34 \er yep
08:34 srifqi ah, "slash"er
08:34 \er usually I am afk when I go smashing keyboards and stuff
08:34 \er so... slasher
08:34 est in fact you are a backslasher
08:34 \er lol
08:34 srifqi yay! finally know how to "squash" commits *offtopic
08:35 \er I can't register forwardslasher
08:35 srifqi backslasher
08:35 srifqi \er: backslaher
08:35 srifqi \er: *backslasher
08:35 srifqi argh...
08:35 \er squashing is a very important skill to learn
08:35 srifqi yes it is
08:35 srifqi and now i'm finally understand, and it works!
08:35 srifqi backslashv?
08:35 \v lol
08:36 z3n0 ok, I need to do some things
08:36 z3n0 bbiab
08:37 srifqi z3n0 = zeno XD
08:37 srifqi what does "bbiab" means?
08:37 est 1337$p33k
08:38 srifqi !calc 2^10
08:38 est 1024
08:38 srifqi est: why MinetestBot don't answer?
08:38 est srifqi, http://de.urbandictionary.com/define.php?term=bbiab
08:38 srifqi deutsch?
08:40 est ah one of those stupid url-locale-leaking services again
08:40 est ja deutsh
08:41 est deutsch*
08:41 srifqi kann du sprechen englisch? *am i right?
08:41 * z3n0 downloads arch 20150301
08:41 srifqi there is #minetest-de?
08:41 est yes
08:42 z3n0 getting sick and tired of fedora breaking my builds
08:42 est srifqi, kann is not right, if you are speaking to a person then its "kannst"
08:42 est and dont use infinitive "sprechen"
08:42 est or sorry its ok
08:42 est but words need to be arranged other way
08:43 srifqi okay, registered, /me opens inbox
08:43 est kannst du englisch sprechen?
08:43 est thats the right way
08:43 est but everybody would understand you :)
08:43 srifqi okay, i'm not german
08:43 z3n0 lol
08:43 srifqi just tried that
08:44 barrydk joined #minetest
08:44 * z3n0 will install a very minimal install of arch... command line only with build crap
08:44 z3n0 hopefully that solves my ongoing issues
08:44 est with?
08:45 srifqi !calc 2^12
08:45 z3n0 fedora keeps insisting on changing stuff when I update
08:45 srifqi !
08:45 srifqi !
08:45 srifqi :(
08:45 z3n0 I suppose I could make the files immutable
08:45 est thats what updating is for
08:45 z3n0 but then.. easier this way
08:45 z3n0 est, but it breaks my delicate android builds :)
08:45 srifqi !calc 2^2
08:46 z3n0 I will run headless virtualbox
08:46 z3n0 sounds fun anyway :)
08:46 est z3n0, then dont install your android toolchain into the system directories. thats no good practice anyway
08:46 z3n0 est, I'm not!
08:46 * est considers make install to be dangerous
08:46 z3n0 they are in opt/
08:47 srifqi !calc 2^10
08:47 z3n0 I'm not 100% stupid. Only 98%
08:47 srifqi 2%?
08:47 z3n0 yes, 2% insane
08:47 est /opt/local or opt bin?
08:47 est for me, everything is under ~
08:47 est ok some stuff is also under ~/../../opt
08:47 z3n0 est, that's my normal approach as well
08:47 est but thats also under ~ after all
08:47 est :p
08:48 z3n0 anyway my cpu can virtualise an init 3 environment that's probably just as fast as on metal
08:48 z3n0 so I think I'll do this as a learning experience anyway :)
08:49 z3n0 actually there is no reason it should be any slower
08:49 z3n0 nothing significant anywya
08:53 Trustable joined #minetest
08:57 srifqi z3n0: have you tried to squash 6 commit? how to do that?
09:01 srifqi nvm, works, even when there was a merge conflicts!
09:03 srifqi est: so, are you from germany?
09:04 Yepoleb joined #minetest
09:04 srifqi gotta go
09:04 srifqi !calc 2*10
09:06 prozacgod joined #minetest
09:06 book` joined #minetest
09:06 est31 joined #minetest
09:16 CW2 joined #minetest
09:21 FR^2 joined #minetest
09:39 Dartmouth joined #minetest
09:39 Dartmouth left #minetest
09:42 Dartmouth joined #minetest
09:57 ThatGraemeGuy joined #minetest
10:08 ImQ009 joined #minetest
10:35 SopaXorzTaker joined #minetest
10:49 twoelk joined #minetest
10:50 Tg1 joined #minetest
10:51 Haudegen joined #minetest
11:20 Calinou joined #minetest
11:53 Zeno` joined #minetest
11:56 Cerise joined #minetest
11:56 Cerise joined #minetest
11:58 srifqi joined #minetest
11:59 srifqi est
11:59 est srifqi, yes
12:00 srifqi is Zeno` still on?
12:00 est yup
12:00 est last action 7 minutes ago
12:00 est last "real" one 15 mins ago
12:01 srifqi what programming lang that you can?
12:02 Zeno` c, c++, asm, modula-2, pascal, cobol, python, lisp
12:02 Zeno` maybe basic
12:02 Zeno` hmm lua, c#
12:02 Zeno` sometimes fortran
12:02 srifqi Hi Zeno`
12:03 srifqi that's much
12:03 est most programming languages are the same
12:04 srifqi what ia the differences? the compiler?
12:04 srifqi is*
12:04 est compiler is different
12:04 est much dont even have one
12:04 est like asm
12:05 srifqi asm uses nasm?
12:05 est and the compiler for c# doesnt translate to "native" bytecode
12:05 est but for microsoft's recently open-sourced virtual machine
12:06 srifqi then how to build a exe file? from c or asm
12:07 est nasm is no compiler
12:07 est a compiler generates bytecode from cross-platform code
12:07 srifqi what it does?
12:07 est asm is not crossplatform
12:07 est except llvm asm ofc
12:08 srifqi what'a that?
12:08 srifqi s&
12:08 est llvm is the basis for clang
12:08 srifqi s*
12:08 est clang is a new and recently very successful compiler
12:08 srifqi woah, many things idk
12:08 Viper168_ joined #minetest
12:09 est direct rival to gcc
12:09 srifqi how about autohotkey?
12:09 est and clang
12:09 est sorry
12:09 est what?
12:09 est autohotkey?
12:10 srifqi you don't know it?
12:10 srifqi :’(
12:10 est whats that
12:10 Tg1 joined #minetest
12:11 est google sais its an automation tool for windows, do you mean that?
12:11 srifqi another programming language
12:11 srifqi maybe
12:12 srifqi so?
12:12 est I guess thats interpreted
12:12 Calinou it's more scripting than programming
12:12 est it doesnt seem to be a language that does real hard work
12:13 Calinou on GNU/Linux, you can use xdotool
12:13 est as in requiring alot of cpu
12:13 Calinou which can click, press keys, move mouse and type for you
12:13 ekem its a scripting language
12:13 ekem for yes, windows
12:13 ekem desktop managers and things will use autohotkey
12:14 srifqi ekem: how to convert it into exe?
12:14 ekem you probably wont find a method to accomplish that really, but you can package it in an exe to install it
12:14 srifqi ???
12:15 ekem the scripts you use for autohotkey are interpreted by autohotkey
12:15 ekem so you would use an exe to deploy autohotkey.exe for a user or check if its installed and then place the .ahk files
12:15 est if you package it into an exe, you need to add some parts of autohotkey too in order to make it work
12:16 Dartmouth joined #minetest
12:16 ekem ah i get a better idea of what you want to do
12:16 ekem you just compile the scripts
12:16 ekem install ahk and you will see
12:17 srifqi ekem: just curious. only that :-)
12:17 ekem ah ok, like est says its just packages the scripts with the interpretor
12:19 srifqi so, what programming language(s) are you doing now?
12:19 srifqi working on*
12:19 ekem who is that directed towards
12:20 Drangue joined #minetest
12:20 srifqi all who cares about my chat
12:21 srifqi nobody?
12:21 ekem wat are you learning srifqi
12:22 srifqi for now?
12:22 ekem forever
12:22 srifqi ummm...
12:23 srifqi many things. as i'm still 9th grade (in Indonesia)
12:24 srifqi !calc 2*2
12:24 ekem oh right on
12:24 ekem !help
12:24 MinetestBot https://github.com/sfan5/minetestbot-modules/blob/master/COMMANDS.md
12:24 ekem reveals: !c
12:25 srifqi ah, forget that.
12:25 srifqi thanks
12:25 ekem !c dy/dx pi^x
12:25 MinetestBot SyntaxError: invalid syntax (<string>, line 1)
12:25 ekem it derped
12:25 srifqi !c 2^10==1024
12:25 MinetestBot False
12:25 srifqi !c 2^10
12:25 MinetestBot 8
12:25 AnotherBrick xor :D
12:26 srifqi ah ^ means xor hehehe
12:26 ekem !g differentiate pi^x
12:26 MinetestBot ekem: http://www.derivative-calculator.net/
12:26 ekem eh close enough
12:26 ekem !g 2*2
12:26 MinetestBot ekem: http://www.twoplustwo.com/
12:26 srifqi !c 2*2
12:26 MinetestBot 4
12:27 srifqi !dev formspec
12:27 MinetestBot "Formspec defines a menu." - http://dev.minetest.net/formspec
12:27 srifqi why we are testing ALL command? XD
12:27 srifqi pr #1990
12:27 srifqi eh?
12:28 srifqi why ShadowBot?
12:28 srifqi ShadowBot
12:28 srifqi ekem?
12:30 srifqi est, have you github account?
12:31 est yes
12:31 est est31
12:35 srifqi why 31???
12:35 est est was taken
12:35 est why srifqi?
12:36 srifqi nvm
12:36 est note, that I have this nick (est) only since today
12:36 est before, my nick was est31
12:36 srifqi oh, okay
12:36 est I'll perhaps keep est31
12:36 srifqi kickall mod :-)
12:37 srifqi found in your repo
12:37 est yup :D
12:37 srifqi another just forks, hehehe
12:38 est I have a password manager
12:38 est https://github.com/est31/minetest-pwm
12:38 srifqi ah, missed that
12:38 srifqi sorry
12:40 Guest4022 joined #minetest
12:42 srifqi !btc eur
12:42 MinetestBot 1 BTC = 257.1600 €
12:43 srifqi woah, 257€
12:44 srifqi !hash md5 srifqi
12:44 MinetestBot ee07df7744b6f53656875b75fdf893b8
12:46 est what a pity that there are no known printable strings that trigger a collision
12:46 est I mean, there is this: http://www.mscs.dal.ca/~selinger/md5collision/
12:46 est but thats binary
12:49 turtleman_ joined #minetest
12:50 srifqi that's why passwords are strings not binary :D
12:51 est collision isnt bad
12:51 est for passwords
12:51 est only second preimage attack
12:51 srifqi wat?
12:51 est but md5 is very strong
12:52 srifqi than sha1?
12:52 est collision: you can chose A and B, and need to get H(A)==H(B)
12:52 srifqi yes, knew it
12:52 est second preimage: you have H(B), but don't know B, and now you search A to get H(A) == H(B)
12:53 srifqi !hash sha256 srifqi
12:53 MinetestBot 3491e6ed5188d7d30389d3d9180a14e1d4f94d483d8a3866e64732b603d80470
12:53 srifqi !hash md2 srifqi
12:53 MinetestBot srifqi: Unknown hash functions, supported: sha256, sha224, sha1, sha512, md5, sha384
12:53 srifqi :'(
12:54 SylvieLorxu joined #minetest
12:55 srifqi AnotherBrick?
12:56 Calinou !hash sha512 srifqi
12:56 MinetestBot 4b6c200028fcb419de4e3d1b3e81a28ed8472061cbc866fe0a8d30fdbc6ac068f0c245ac4eab7fc57ae54c2d8c248d04cab13cd3f0eda1c838456251c2c72c35
12:56 Calinou a very secure srifqi
12:57 srifqi nice!
12:57 srifqi sha16777216? hehehe
12:57 srifqi !hash sha16777216 srifqi
12:57 MinetestBot srifqi: Unknown hash functions, supported: sha256, sha224, sha1, sha512, md5, sha384
12:58 srifqi lol
12:59 srifqi hello MinetestBot
12:59 MinetestBot Hello srifqi
13:00 srifqi *feels lonely
13:00 srifqi XD
13:02 srifqi is this channel for chit-chat?
13:02 Dartmouth yup
13:02 Zeno` chitty chitty bang bang, chitty chitty bang bang, I love you
13:02 srifqi ah,
13:02 Zeno` ooo ahh
13:03 srifqi wat?
13:03 srifqi what is that sentence?
13:03 srifqi lyric?
13:03 Zeno` https://www.youtube.com/watch?v=ZTTzcXSLjhI
13:03 srifqi song?
13:04 Zeno` it's at the top of the charts at the moment
13:07 srifqi who is the singer's name
13:08 srifqi Zeno`
13:09 Zeno` hmm
13:09 Zeno` not sure
13:09 Zeno` Dick Van Dyke?
13:11 srifqi where does this songs come from
13:11 srifqi Zeno`
13:11 Zeno` http://en.wikipedia.org/wiki/Chitty_Chitty_Bang_Bang
13:11 Calinou updated Ubuntu build
13:12 exio4 hi cute people
13:12 exio4 how is your sweet life goin'
13:13 srifqi exio4: fine, thanks. how about you?
13:13 srifqi Zeno`: okay, understand
13:13 exio4 same, I would say, thank you! <3
13:14 srifqi Calinou: how large does ubuntu OS take?
13:14 srifqi :-)
13:14 srifqi 3GB?
13:14 Zeno` 3gb?!
13:14 Zeno` even fedora won't use that much
13:15 srifqi too much?
13:15 exio4 take what
13:15 exio4 like ram usage, hdd?
13:15 srifqi space in harddrive
13:15 exio4 I don't really care about _that_ that much
13:15 srifqi 756MB?
13:16 srifqi how large?
13:16 exio4 we need ZFS
13:16 srifqi i want to install it, but i must allocate space required
13:16 srifqi so?
13:17 exio4 just give it 15 or 30 gb
13:18 srifqi that's what i want. thanks
13:19 exio4 it'll give you room for having also files there and relatively don't worry about it
13:20 srifqi so linux based files are small(er than windows hehehe)
13:20 exio4 est: why wouldn't nasm be a compiler
13:21 exio4 after all, a compiler is a tool that compiles between two programming languages
13:24 est exio4, its an assembler
13:24 srifqi joined #minetest
13:24 exio4 still a compiler
13:24 srifqi ?????
13:24 srifqi evening
13:25 exio4 hi long time no see! ;P
13:25 srifqi XD
13:25 exio4 est: would you consider a program that compiles Python code to C a what?
13:25 srifqi just a few minutes
13:25 exio4 as an example
13:25 exio4 srifqi: it felt like ten years!
13:25 srifqi lol
13:26 exio4 an assembler is an specific compiler
13:27 srifqi any pear (php) that can parse JSON?
13:28 srifqi !g srifqi
13:28 MinetestBot srifqi: https://forum.minetest.net/viewtopic.php?f=9&amp;t=10940
13:29 est exio4 ok convinced
13:31 srifqi !tw srifqi
13:31 MinetestBot srifqi: Give me a link, a @username, or a tweet id
13:31 srifqi !tw @srifqi
13:31 MinetestBot Sorry, couldn't get a tweet from https://twitter.com/srifqi?1425907903.9976962 (@srifqi)
13:33 srifqi maybe MinetestBot should add !fb command
13:33 jordan4ibanez joined #minetest
13:35 srifqi jordan4ibanez
13:35 jordan4ibanez Hello
13:35 srifqi hello
13:40 Ritchie joined #minetest
13:45 jordan4ibanez I am just so confused as to how to go about adding ncurses to my build options in ncurses
13:50 srifqi jordan4ibanez, what ya doing?
13:51 jordan4ibanez Trying to learn how to use ncurses so we can have terminal input on the server
13:51 srifqi woah
13:51 srifqi good luck
13:52 est good luck with devs convincing of yet another dependency
13:52 jordan4ibanez It would have to run on a seperate thread as to not stop the main loop
13:52 jordan4ibanez Oh I don't care if they want it, I don't even know if I can do it
13:59 FreeFull joined #minetest
14:10 srifqi joined #minetest
14:14 srifqi how to use ShadowBot?
14:16 est ~source
14:16 ShadowBot My source is at https://github.com/ProgVal/Limnoria -- Custom plugins are at https://github.com/ShadowNinja/Limnoria-plugins
14:17 est ~rainbow srifqi
14:17 ShadowBot est: srifqi
14:22 jordan4ibanez Man, I always thought c++ was very complicated, it's much easier when you break it down, it almost feels like python
14:27 marktraceur You're looking at a piece of code without pointers, I bet.
14:28 Zeno` pointers are not complicated
14:29 exio4 what's complicated is OOP
14:29 Zeno` especially if you've programmed in assembly
14:29 exio4 what is*
14:29 exio4 also, mutable data
14:29 AnotherBrick also lambdas
14:29 est mutable data??
14:30 exio4 lambdas what
14:30 est lambdas: here functions
14:30 exio4 est: wut
14:30 AnotherBrick lambda function
14:30 exio4 yes, what with lambdas
14:30 exio4 anonymous functions are nice
14:30 Zeno` if you know lisp or any other functional language lambdas are not complicated either
14:30 exio4 high-order functions are the king of code reuse
14:30 est jup did use functional features of lua today
14:30 est code not released yet
14:31 est but its really just cool
14:31 Zeno` c++ is a crappy language though
14:31 Zeno` C is much nicer
14:31 AnotherBrick if you use lisp - everything is lambda, in C++ it's a perfect way of overcomplicating code
14:31 exio4 C++ is crappy because it does everything in its own weird way
14:31 exio4 AnotherBrick: if you use lisp, you said anything
14:32 exio4 also, "lisp" alone doesn't mean anything
14:32 est and c++ just wants to do everything
14:32 exio4 it just means a big family of languages
14:32 exio4 where you've got things like common lisp, which are imperative
14:32 Zeno` it's kind of why I oppose c++11
14:32 exio4 and scheme on the other side
14:32 Zeno` (one of the reasons anyway)
14:32 Tux[Qyou] joined #minetest
14:33 Zeno` minetest is a game engine... not some kind of generic reusable library
14:33 exio4 I think C++11 is a better language than previous C++s
14:33 exio4 Zeno`: ?
14:33 exio4 Zeno`: so what?
14:33 Zeno` exio4, so I am sick of the constantly moving target that C++ is
14:33 Zeno` not because I don't understand it, but because it's constantly moving
14:34 est I'm mostly indifferent in this discussion
14:34 Zeno` If I was 100% in charge of minetest I'd allow classes and single inheritance... not much else
14:34 est I dont like that code can only be compiled when you have arch linux or debian experimental
14:35 Zeno` maybe c11 auto is nice
14:35 est auto is nice yes
14:35 est but a bit dangerous
14:35 est operators are nice too
14:35 Zeno` as long as they're not overused
14:36 exio4 Zeno`: I think inheritance is worse than removing control structures and only allowing goto
14:36 est ofc dont abuse anything. even gotos have some meaning somewhere if you dont want to open a loop for nothing
14:36 Zeno` exio4, subclassing just once is not all that bad
14:36 exio4 subtyping is different
14:37 Zeno` gotos in C and C++ are not long jumps so they're not so bad if used appropriately
14:37 exio4 I don't mind subtyping even though it's typing rules are complex
14:37 est ok thats too much c++ for me, I have never properly learnt it :)
14:37 Zeno` exceptions... nearly always used incorrectly
14:37 est better an exception than an assert
14:38 exio4 btw Zeno`, about lambdas, https://github.com/minetest/minetest/blob/master/src/network/networkpacket.cpp this could be using a few lambas abstracting the code reuse
14:38 exio4 an exception should mean something REALLY BAD happened
14:38 Zeno` est, assert is nothing like an exception :)
14:38 exio4 like, textures corrupted, database thread crashed
14:38 est Zeno`, yes I just wanted to point to your assert fixes
14:38 Zeno` exio4, unfortunately most (a lot?) of people do not realise that
14:38 behalebabo joined #minetest
14:38 Zeno` est, you're a tease
14:38 exio4 you would throw an exception and "wish luck to anyone suffering that"
14:39 exio4 that is why I hate python too
14:39 exio4 the language is built around "welp, anything should throw exceptions if it didn't work like it should"
14:39 exio4 and I really think nullable types shouldn't be the default :D
14:39 est thats why lua is so great
14:40 est i really like the "everything is nil"
14:40 Zeno` I'm pretty sure there is an article somewhere where Stroustrup talks about where he wished he never added them to the language
14:40 exio4 exceptions or NULL?
14:40 exio4 oh
14:40 exio4 exceptions
14:40 Zeno` exceptions
14:41 exio4 anyway, I think exceptions are cool "when allowed", because they allow you to say "something really really bad happened"
14:41 jordan4ibanez I just don't understand it
14:41 jordan4ibanez http://paste.ofcode.org/sCerdTzzrg2TdhkHU9waEM
14:41 exio4 right now the code I am working on uses exceptions when invariants that should be mantained by the database are violated
14:41 jordan4ibanez When I do clear or erase at that point it won't print
14:41 exio4 aka something is really fkd :P
14:43 MinetestForFun joined #minetest
14:47 Zeno` exio4, if people used them for just that there would not be a problem ;)
14:47 exio4 heh
14:47 exio4 Zeno`: have you played with Haskell?
14:48 exio4 (yes, I probably asked this more than once)
14:48 Zeno` here: https://github.com/minetest/minetest/commit/5b8855e83c0d1cc7aef21492e7fe862b7d06917e       <-- before that exceptions in minetest were even used in place of conditionals!
14:49 Zeno` (and there are still some instances where they are used instead of conditionals)
14:49 exio4 working with exceptions is relatively painful in Haskell
14:49 Zeno` exceptions should always result in "cleaning up and aborting the program"
14:49 exio4 unless you use it _that_ way in a extremly controlled manner
14:49 Zeno` unfortunately most people don't realise that
14:50 Zeno` I think proper exception handling is called... fail semantics?
14:50 exio4 Zeno`: I am a fan on specific datatypes like Maybe for this
14:50 exio4 data Maybe a = Nothing | Just a
14:50 Zeno` that commit I pasted above increased performance by about 200%
14:50 SylvieLorxu Zeno`: For what it's worth, there are languages where catching exceptions is considered "good practice" as opposed of using if/else, like Python
14:51 SylvieLorxu But I think exceptions are also a lot cheaper in Python
14:51 SylvieLorxu In most languages, they're expensive
14:51 exio4 they aren't
14:51 exio4 they're cheap if you don't throw them
14:52 CraigyDavi joined #minetest
14:52 exio4 still, most things in Python can throw exceptions because the language is based around idiot things like broken scoping
14:53 Zeno` http://en.wikipedia.org/wiki/Coding_by_exception  ... http://en.wikipedia.org/wiki/Anti-pattern  ... http://en.wikipedia.org/wiki/Exception_handling#Termination_semantics
14:53 exio4 sowwy I just don't like Pyton
14:54 Zeno` termination semantics is the term I was looking for
14:54 exio4 Python *
14:54 luizrpgluiz joined #minetest
14:54 exio4 I don't really mind a performance hit if it makes the code easier to work with
14:54 exio4 exceptions are just "I have no idea where this is gonna end doing anything"
14:54 luizrpgluiz hi all
14:54 jordan4ibanez I used to think that way
14:54 jordan4ibanez Now I live in a VAN DOWN BY THE RIVER
14:55 Zeno` a performance hit is ok if used properly. Using exceptions in place of conditionals is not
14:55 Zeno` also if the loop has to execute in < 60ms of course a 200% performance hit (or more!) is not ok at all
14:55 est ______init_____
14:56 Zeno` I think it was more than 200% in fact. I wish I'd kept my profiles. It was ridiculous the amount of stack unwinding being done
14:57 Zeno` I also don't think   if (getNode() == OK) {} is much harder to understand than try { getNode(); } catch sfjskjfs {}
14:58 Zeno` if people have problems with if and if else then... *shrug* :D
14:58 luizrpgluiz but performance depends largely on computer
14:59 Zeno` luizrpgluiz, not in the case of exceptions
14:59 Zeno` if an exception has to be caught then the stack must be unwound for every instance where that exception is caught
14:59 Zeno` well, it actually has to unwind to find where it's caught as well
15:00 Elinvention joined #minetest
15:00 Zeno` in the case of "ok the program is going to terminate after the exception is thrown" it's not an issue
15:01 Zeno` which is why I don't really have a problem with termination semantics for exceptions (but even then they're not really necessary)
15:02 est they are important for example when you have a client server model where the clients are people that don't know the server is offline now
15:02 luizrpgluiz in my case, my home computer is an Intel Celeron D 512 mb ram and S3 ProSavage3
15:02 Zeno` it's the "resumable semantics" that seem to be so prevalent in C++ that are a real performance killer (and stupid at the same time)
15:02 Ataron joined #minetest
15:03 Zeno` est, there is no other way to tell the client they're offline?
15:03 Zeno` err, that the server is offline
15:03 * Zeno` looks under the hood of exceptions. Hmm, what's this I see. A solution!
15:03 Zeno` heheh
15:04 Zeno` anyway, I am actually asleep so I dunno why I'm typing
15:04 est lol
15:04 exio4 Zeno`: sleep then <3
15:04 Zeno` I *am* asleep
15:05 Zeno` I've been asleep for an hour at least
15:05 est sleep well
15:05 Zeno` I'm the only one here who types in their sleep?
15:06 AnotherBrick You're not really typing. This whole chat is your nightmare
15:06 Zeno` Surely I'm not the only one
15:06 GaboXandre joined #minetest
15:07 Zeno` Unless I am the Highlander (There can be only one!). I suspect I'm not though
15:07 Zeno` actually, I'm gonna watch that movie right now
15:07 Zeno` O/
15:07 AnotherBrick When I type being asleep comes out something likeeeeeeeeeeeeeeeeeeeeeeeeeeeeee <- that
15:07 Zeno` :)
15:12 Tuxedo[Qyou] joined #minetest
15:16 luizrpgluiz left #minetest
15:24 luizrpgluiz joined #minetest
15:25 Tux[Qyou] joined #minetest
15:36 turtleman_ joined #minetest
15:37 SopaXorzTaker joined #minetest
15:38 luizrpgluiz left #minetest
15:48 Warr1024 joined #minetest
15:51 roniz joined #minetest
15:52 alket joined #minetest
15:52 SopaXorzTaker joined #minetest
15:55 Dartmouth joined #minetest
15:56 Tuxedo[Qyou] joined #minetest
15:57 phantombeta joined #minetest
16:06 Tux[Qyou] joined #minetest
16:24 fusion44 joined #minetest
16:32 ekem <luizrpgluiz> in my case, my home computer is an Intel Celeron D 512 mb ram and S3 ProSavage3 lol wat
16:33 ekem im checking my computing privilege so hard right nwo
16:34 ImQ009 joined #minetest
16:47 jordan4ibanez joined #minetest
17:00 Calinou joined #minetest
17:00 Dartmouth_ joined #minetest
17:18 ElectronLibre joined #minetest
17:22 Krock joined #minetest
17:24 Krock Howdy, cowboys?
17:31 AnotherBrick how badly size of world database file is affecting server performance? i understand that it depends on tons of factors, like, HDD you use, RAM you have, etc..
17:32 Krock an niternet conection would be useful aswell
17:32 Krock *internet
17:32 VanessaE AnotherBrick: generally the size of a map/database won't affect performance at all
17:32 VanessaE the engine does not load the whole map into memory, so RAM is only a function of things like the mods you run, and whether or not the engine has bugs that result in increased RAM usage
17:33 Krock VanessaE, well, a 1 MB map might be faster than a 1 GB map (but only when looking at reading speed)
17:33 VanessaE disk storage is purely a function of what you store there, including the aforementioned map.
17:33 VanessaE Krock: well yeah but I doubt you'll see a 1MB map in practice :P
17:33 Krock VanessaE, yes and the 1 MB-map server will be much slower because the players keep genreating more and more areas
17:33 VanessaE (smallest maps tend to be a hundred or so megs)
17:34 raymondillo joined #minetest
17:34 Krock the database size depends on the amount of mods
17:34 VanessaE Krock: not necessarily in either case :)
17:34 VanessaE ShadowNinja's creative server doesn't have many mods, and it uses flat mapgen
17:34 Krock s/mods/different placed nodes/
17:35 AnotherBrick thanks a lot Vanessa and Krock!
17:35 VanessaE yeah, that much is true
17:35 Krock np, I love to confuse people
17:35 VanessaE AnotherBrick: the biggest concern you should have is CPU usage
17:35 Krock I hope I succeed
17:35 VanessaE and that comes from complex mods e.g. mesecons, technic, pipewroks
17:35 VanessaE works*
17:36 ShadowNinja Size depends on: backend, space explored, regularity (compressability) of map, maybe more.
17:42 PilzAdam joined #minetest
17:43 est31 joined #minetest
17:45 raymondillo left #minetest
17:58 Haudegen joined #minetest
18:01 Player_2 joined #minetest
18:10 compunerd joined #minetest
18:16 turtleman_ joined #minetest
18:20 Miner_48er joined #minetest
18:20 Viper168 joined #minetest
18:26 nrzkt joined #minetest
18:30 Menche joined #minetest
18:39 Haudegen joined #minetest
18:39 ImQ009 joined #minetest
18:41 Carnaxus joined #minetest
18:43 roniz joined #minetest
18:43 Carnaxus 'Allo everyone.  Just wondering if a mobile version of Minetest is planned?
18:43 sfan5 we do have an android version
18:44 sfan5 it's a bit buggy though
18:44 jordan4ibanez a bit
18:44 sfan5 http://minetest.net/download#android
18:44 Carnaxus I'll be gentle.
18:47 Carnaxus You're probably all going to laugh at me, but I've never installed a non-PlayStore app before.  Are there any extra steps I need to take beyond unzipping it?
18:47 Dumbeldor joined #minetest
18:49 Carnaxus ?
18:51 fling joined #minetest
18:53 Calinou Carnaxus, unzip using a file manager, then run the .apk
18:53 Calinou you may need to enable Unknown sources in the Security settings of your device
18:54 Calinou Minetest is not (yet?) available in F-Droid either
18:54 Calinou but still, check F-Droid out, because it's a very useful store
18:54 Calinou https://f-droid.org
18:55 Carnaxus Thanks!
18:55 ElectronLibre joined #minetest
18:56 Krock sfan5, do you need a pull to add my script I sent you?
18:57 Krock well, I couldn't test it so it might be better when you look at it first
19:05 Ritchie joined #minetest
19:08 jojoa1997 joined #minetest
19:13 turtleman_ joined #minetest
19:15 Ritchie joined #minetest
19:15 Calinou http://www.pnotepad.org/
19:15 Calinou I did not know about this
19:15 Calinou apparently Windows-only
19:15 jojoa1997 Windows Master Race
19:16 Calinou http://wiki.torque3d.org/coder:compiling-in-linux ⇒ looks like Torque 3D is getting official GNU/Linux support
19:17 sfan5 "in linux"
19:17 sfan5 Krock: which script
19:18 Vazon_ joined #minetest
19:18 Krock sfan5, http://irc.minetest.ru/minetest/2015-03-07#i_4176428
19:18 Calinou sfan5, “As of version 3.7, Torque 3D supports being run under Linux Ubuntu” :P
19:18 sfan5 did you look at my away message
19:19 sfan5 Krock: it said you should use /msg when you want to be sure i get something
19:19 Krock sfan5, no, I don't /whois everyone.
19:20 sfan5 actually
19:20 sfan5 i don't expect you to
19:20 sfan5 but this give thing is stupid
19:20 Krock :3
19:20 Krock However, what do you think about it?
19:21 sfan5 Krock: come to ##minetestbot for a sec please
19:22 Ritchie joined #minetest
19:29 Haudegen joined #minetest
19:33 MinetestBot joined #minetest
19:33 MinetestBot joined #minetest
19:39 roniz joined #minetest
19:46 roniz_ joined #minetest
20:10 roniz joined #minetest
20:12 roniz_ joined #minetest
20:16 roniz__ joined #minetest
20:23 roniz_ joined #minetest
20:31 proller joined #minetest
20:32 _schism_ joined #minetest
20:41 Robert_Zenz joined #minetest
20:46 alket joined #minetest
20:51 meldrian joined #minetest
20:52 alket joined #minetest
20:57 roboman2444 joined #minetest
21:03 Tux[Qyou] joined #minetest
21:04 BluebirdGreycoat joined #minetest
21:26 ImQ009 joined #minetest
21:26 ElectronLibre left #minetest
21:34 Jordach joined #minetest
21:35 Haudegen joined #minetest
21:49 Warr1024 joined #minetest
21:53 jordan4ibanez joined #minetest
21:57 turtleman_ joined #minetest
22:01 Jeija joined #minetest
22:04 preludelinux joined #minetest
22:13 cheapie joined #minetest
22:17 kilbith joined #minetest
22:18 kilbith left #minetest
22:19 Jeija left #minetest
22:31 Hirato joined #minetest
22:32 cheapie joined #minetest
22:35 Taoki joined #minetest
22:40 MinetestForFun joined #minetest
22:42 srifqi joined #minetest
22:43 twoelk|2 joined #minetest
23:00 jordan4ibanez joined #minetest
23:07 jojoa1997 joined #minetest
23:07 alket joined #minetest
23:08 srifqi joined #minetest
23:14 meldrian joined #minetest
23:16 Miner_48er joined #minetest
23:18 luizrpgluiz joined #minetest
23:19 luizrpgluiz hi
23:25 Wayward_One joined #minetest
23:38 Xenoth joined #minetest
23:47 luizrpgluiz left #minetest
23:50 jalcine no

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