| --- Log opened Tue Oct 18 00:01:00 2011 |
| 00:01 | | Derakon [chriswei@Nightstar-f68d7eb4.ca.comcast.net] has quit [[NS] Quit: leaving] |
| 00:19 | | Stalker [Z@Nightstar-5aa18eaf.balk.dk] has quit [[NS] Quit: If the world didn't suck, we'd all fall off.] |
| 00:45 | | You're now known as TheWatcher[T-2] |
| 00:51 | < celticminstrel> | ...why isn't there a PI in <cmath>... |
| 00:54 | | You're now known as TheWatcher[zZzZ] |
| 00:54 | < celticminstrel> | Okay, so I can rotate this thing, but it's rotating too fast to be useful. |
| 00:54 | < gnolam> | There's M_PI. |
| 00:54 | < gnolam> | But it's not actually standard. |
| 00:54 | < gnolam> | (Even if it exists in... well, all the relevant compilers.) |
| 00:54 | < celticminstrel> | I need a good ratio for transforming mouse motion into rotation angle. |
| 00:54 | < gnolam> | "good ratio"? |
| 00:54 | < celticminstrel> | Does "all the relevant compilers" include MinGW? |
| 00:55 | < celticminstrel> | Well, currently my angle is being updated like this (maybe it's incredibly stupid, I dunno): |
| 00:55 | < celticminstrel> | theta += hypot(xdown - x,ydown - y) * (3.141592 / 180); |
| 00:55 | < celticminstrel> | In response to mouse motion (xdown and ydown are set when the mouse button is depressed). |
| 00:56 | < celticminstrel> | Which if I recall correctly takes the distance the mouse moved as degrees. |
| 00:56 | < celticminstrel> | A distance which is measured in pixels. |
| 00:58 | <~Vornicus> I don't see how hypot gets you an angle? |
| 00:58 | < gnolam> | atan2(). |
| 00:58 | < celticminstrel> | I was using the distance as an angle... |
| 00:59 | < celticminstrel> | Oh hey, atan2 works. |
| 00:59 | <~Vornicus> Why would the distance be an angle? |
| 00:59 | < gnolam> | Either track a single direction's offset from button down to button up, or create a virtual reference point (a certain distance away backwards from the current angle of the rotating object) upon button down. |
| 01:00 | < celticminstrel> | Not sure what that second one means. |
| 01:00 | < celticminstrel> | Also it seems that my rotation around y or x axis is broken. |
| 01:01 | < celticminstrel> | I did think it looked a bit off when I tested it statically... |
| 01:01 | < celticminstrel> | And with rotating around z it's still too fast even with atan2. :/ |
| 01:02 | <~Vornicus> Then divide it by like 100 |
| 01:03 | < celticminstrel> | Oh wait, maybe it's because I didn't remove multiplying by pi/180. |
| 01:03 | < celticminstrel> | Duh. |
| 01:05 | < celticminstrel> | Okay, only thing left to do is to make the direction of the mouse matter, but that can wait until later. |
| 01:05 | < celticminstrel> | Need to figure out why the other rotations failed. Maybe I got x/y and z mixed up. |
| 01:06 | < celticminstrel> | ...no, neither works. :( |
| 01:07 | < celticminstrel> | (For reference, the goal here is basically drawing a cube and transforming it in nearly every possibly way.) |
| 01:08 | | Attilla [Some.Dude@Nightstar-f29f718d.cable.virginmedia.com] has quit [Ping timeout: 121 seconds] |
| 01:11 | <~Vornicus> Oh I see what you really want |
| 01:13 | <~Vornicus> You want to use hypot(delta mouse) as your angle of rotation and vcross(delta mouse, +z) as your axis of rotation. |
| 01:13 | < celticminstrel> | ...vcross? |
| 01:14 | <~Vornicus> vector crtoss product |
| 01:14 | < celticminstrel> | That's not in the standard library is it? |
| 01:14 | <~Vornicus> In 3 dimensions, produces a vector perpendicular to the two input vectors. |
| 01:15 | < celticminstrel> | Currently I'm just picking one axis to rotate and rebuilding each time, though. |
| 01:15 | <~Vornicus> No, you'll have to build your own. But you need to anyway to get your rotation to work |
| 01:15 | < celticminstrel> | And z appears to work, but the other two look strange. |
| 01:15 | < celticminstrel> | As in, not like a cube at all. |
| 01:16 | < celticminstrel> | What does +z mean in your formula? |
| 01:16 | <~Vornicus> +z means "into the screen" |
| 01:16 | < celticminstrel> | Unit vector? |
| 01:16 | <~Vornicus> yeah, the unit vector into the screen |
| 01:17 | <~Vornicus> ANd then you normalize your vcross outputs and you'll have your axis of rotation according to the screen. |
| 01:22 | < celticminstrel> | The cross product seems to give (y, -x, 0). |
| 01:22 | < celticminstrel> | If I did this correctly. |
| 01:30 | <~Vornicus> Yes, that it will do. |
| 01:31 | <~Vornicus> Note: if you're rotating the camera instead of the cube (this is a common thing) you replace +z with the direction the camera is actually facing |
| 01:31 | < celticminstrel> | Is this why the rotate functions take a vector? I didn't really understand that part. |
| 01:32 | <~Vornicus> A rotate function that takes a vector and an angle will rotate the object by the angle, using the vector as the axis of rotation. |
| 01:58 | < celticminstrel> | Okay, fixed my x and y axis rotations. |
| 02:04 | | Reivles [orthianz@3CF3A5.E1CD01.C6689C.33956A] has quit [Ping timeout: 121 seconds] |
| 02:07 | < celticminstrel> | ...does projection just amount to zeroing one component? |
| 02:09 | < gnolam> | ... no |
| 02:10 | | Reiver [orthianz@3CF3A5.E1CD01.C6689C.33956A] has joined #code |
| 02:16 | < celticminstrel> | eg "project (x,y,z) onto the xy plane" gives (x,y,0). |
| 02:17 | < celticminstrel> | ? |
| 02:23 | | Kindamoody[zZz] is now known as Kindamoody |
| 02:26 | < McMartin> | That, yes. |
| 02:26 | < McMartin> | But only because everything interesting cancels out in that case. |
| 02:27 | < McMartin> | But one can project onto arbitrary planes, not just one that's flat and goes through the origin. |
| 02:30 | < celticminstrel> | ...I hadn't thought of that. <_< |
| 03:03 | | Derakon[AFK] is now known as Derakon |
| 03:04 | | * McMartin - having reimaged his work machine - heads home. |
| 03:04 | < McMartin> | celticminstrel: One of the more important planes that gets projected against is the surface of the screen. ;-) |
| 03:05 | < McMartin> | OpenGL's "Projection matrix" kinda represents that. |
| 03:07 | < celticminstrel> | Well, right now I'm just trying to compute a transformation matrix for an arbitrary rotation. It's not working. |
| 03:07 | < celticminstrel> | ...maybe I'm doing the whole thing backwards. |
| 03:08 | < celticminstrel> | Nope, nothing that simple. |
| 03:39 | < Derakon> | Dammit, Dungeons of Dredmor has me wanting to make my own for-pay roguelike. |
| 03:39 | < Derakon> | This is clearly a terrible idea. |
| 03:45 | | Reiver [orthianz@3CF3A5.E1CD01.C6689C.33956A] has quit [Ping timeout: 121 seconds] |
| 03:47 | | gnolam [lenin@Nightstar-202a5047.priv.bahnhof.se] has quit [[NS] Quit: Z?] |
| 03:49 | | Reiver [orthianz@3CF3A5.E1CD01.C6689C.33956A] has joined #code |
| 03:55 | | Vash [Vash@Nightstar-f03c5637.sd.cox.net] has joined #code |
| 03:56 | < McMartin> | celticminstrel: Remember that rotation matrices interact oddly when you compose them, since rotation is about the origin at the point in time of the rotation. |
| 03:57 | < McMartin> | So rotating something that's out there in space requires it to be translated to the origin first, then rotated, then put back. |
| 03:57 | < McMartin> | (Or, all objects start at the origin, are oriented, then placed) |
| 03:57 | < celticminstrel> | Well, I'm following the instructions in the textbook which do mention that... |
| 03:58 | < McMartin> | OK |
| 03:59 | < celticminstrel> | It's probably something being lost in translation to code. |
| 04:00 | < celticminstrel> | But basically I translate, rotate around x and y, rotate around z, reverse the rotations around x and y, and reverse the translation. In theory. |
| 04:01 | < McMartin> | Mrn. |
| 04:01 | <~Vornicus> Hooray, Forward Kinematics |
| 04:01 | < McMartin> | If it were me - and I weren't coding to an assignment spec - I'd actually skip most of those and just store the intended rotations elsewhere, recomputing the matrix from scratch. |
| 04:01 | < McMartin> | So that the initial setup is always "drawing at 0,0,0, unrotated" |
| 04:01 | < McMartin> | "Now rotate at z" |
| 04:01 | < McMartin> | "Now put on the x and y rotations" |
| 04:02 | < McMartin> | "Now translate it to where it should go" |
| 04:03 | < Derakon> | So basically you want to do T * R_xy * R_z * R_xy(-1) * T(-1)? |
| 04:03 | < Derakon> | Where (-1) means "inverse"? |
| 04:03 | < McMartin> | I think the goal - if I read the backscroll right - is "I want to put a rotation on this thing's local z-axis, in place." |
| 04:05 | < McMartin> | And given all the talk of inverses, it also sounds like the model itself is not intrinsically centered on (0,0,0). |
| 04:07 | < celticminstrel> | Derakon: That is what I'm doing. |
| 04:08 | < celticminstrel> | ...the textbook swaps the inverses with the non-inverses though. |
| 04:08 | < McMartin> | Mmm |
| 04:08 | < McMartin> | Yeah, that sounds very much like it assumes you're rotating something that's "already been placed" out there. |
| 04:08 | < McMartin> | Rotating about a point that isn't the origin, basically. |
| 04:09 | < celticminstrel> | Yeah. |
| 04:09 | < McMartin> | My solution for that for Sable was "never do this" >_> |
| 04:09 | < celticminstrel> | XD |
| 04:09 | < Derakon> | Heh. |
| 04:09 | < McMartin> | (Sable, for those who haven't seen it: https://hkn.eecs.berkeley.edu/~mcmartin/sable/ ) |
| 04:10 | < McMartin> | My old student project for Graphics class. |
| 04:10 | < celticminstrel> | Heh, his sample code uses new needlessly. |
| 04:10 | < celticminstrel> | Matrix* m = new Matrix(); /* do stuff with m */ delete m; |
| 04:11 | < Derakon> | And nothing derived from m is used? |
| 04:12 | < McMartin> | Even if it were, "new Matrix()" in the same scope fixes the type. |
| 04:12 | < McMartin> | That's not a factory method. |
| 04:12 | < McMartin> | (My guess: he's more used to Java than C++) |
| 04:12 | < celticminstrel> | Yeah, that seems like a good guess. |
| 04:13 | < Derakon> | Oh right, he could've just used a Matrix instead of a Matrix*. |
| 04:13 | < celticminstrel> | Yeah. |
| 04:13 | < Derakon> | Sorry, as I may have mentioned awhile back my C is really rusty. |
| 04:14 | < celticminstrel> | Matrix m; /* do stuff with m */ |
| 04:14 | < celticminstrel> | No need to delete. |
| 04:14 | < McMartin> | Also guaranteed leak-free, also ensures the destructor is run no matter how you leave the function. |
| 04:14 | < Derakon> | Yeah. |
| 04:14 | < McMartin> | Precise destructors - the thing I miss most in Java. |
| 04:15 | < McMartin> | Though really, C#'s IDisposable interface is better still. |
| 04:15 | < McMartin> | Since it can be tasked with arbitrary method pairing. |
| 04:15 | < McMartin> | C++'s "Boost" library includes some preprocessor jackassery to let you set a block of code to run on scope exit with the BOOST_SCOPE_EXIT macro, but. |
| 04:20 | < celticminstrel> | I think I did that sort of thing once by using boost's smart pointers with a custom something... deleter or whatever. |
| 04:20 | < McMartin> | Yup |
| 04:20 | < McMartin> | That's basically what BOOST_SCOPE_EXIT does. |
| 04:20 | < McMartin> | (For windows dev, CloseHandle is a classic) |
| 04:20 | < celticminstrel> | This was for SDL things. |
| 04:22 | < celticminstrel> | So the cube looks okay when the rotation angle is 0, though it's in the wrong place, but as soon as I change the angle it disappears entirely. |
| 04:22 | < celticminstrel> | Probably out of the viewport or something. |
| 04:22 | | Stalker [Z@Nightstar-3602cf5a.cust.comxnet.dk] has joined #code |
| 04:22 | < Derakon> | Yeah, shrink your translation down a bunch. |
| 04:22 | < celticminstrel> | Hm. |
| 04:24 | < celticminstrel> | The translation vector is also the axis of rotation, according to the textbook (unless I misread). |
| 04:25 | < Derakon> | Okay, sure, but shrink the magnitude of translation down so the object stays in view. |
| 04:26 | < celticminstrel> | Unit-vectorize it? |
| 04:26 | < Derakon> | Uh, sure. Or even less. |
| 04:26 | < Derakon> | Translation in a 4D matrix is the rightmost column, yes? So just put in, like, .1 for your translation |
| 04:28 | < celticminstrel> | The translation is being undone at the end though, I don't understand why this is even making a difference at all... |
| 04:29 | < Derakon> | Because you have your order of operations wrong or something. |
| 04:29 | < Derakon> | And the translation is the most likely transform to make the cube disappear. |
| 04:29 | < celticminstrel> | Huh, apparently it's translating only, no rotation. |
| 04:29 | < Derakon> | So you shrink down your transforms until you can see what's happening to the cube. |
| 04:31 | < Derakon> | Also remember that translating an object to <1, 0> and then rotating it 90? is equivalent to rotating 90? and then translation to <0, 1>. So "translating only" might actually be "rotating after translating". |
| 04:31 | < Derakon> | This is the kind of thing I internalized when I learned to use POV-Ray. |
| 04:34 | < celticminstrel> | I am currently trying to decode that last sentence. |
| 04:34 | <~Vornicus> WHenever sane, do your transformations in the order Scale Rotate Translate |
| 04:34 | < celticminstrel> | ...the textbook says I need to translate to origin to do the rotations, though. |
| 04:35 | < Derakon> | CM: okay, so, the center of rotation is always the origin. |
| 04:35 | < Derakon> | So if your cube is sitting at <1, 0>, imagine it as a cube that is tied by an invisible rope to the origin. |
| 04:35 | < Derakon> | Rotating holds the endpoint at the origin fixed and moves the cube. |
| 04:35 | < McMartin> | And then rotating is swinging it on that rope. |
| 04:35 | < celticminstrel> | Right. |
| 04:36 | | Reiver [orthianz@3CF3A5.E1CD01.C6689C.33956A] has quit [Ping timeout: 121 seconds] |
| 04:36 | < Derakon> | So unless your cube is already at the origin, rotating is actually rotating and translating. |
| 04:36 | < McMartin> | You can play awesome, sanity-blasting games with this to simulate sun/earth/moon systems, but, seriously, don't -_- |
| 04:36 | < Derakon> | Heh. |
| 04:36 | | Reiver [orthianz@3CF3A5.E1CD01.C6689C.33956A] has joined #code |
| 04:36 | < celticminstrel> | Are you saying that instead of moving it to the origin, rotating it, and moving it back, that I can rotate it first, then move it back to where it started? |
| 04:37 | < Derakon> | Uh, theoretically. |
| 04:37 | < Derakon> | But calculating the "move it back" would be hairy. |
| 04:37 | < celticminstrel> | That's what I thought, yeah. |
| 04:38 | <~Vornicus> Really it goes like this: 1. translate object so its center of rotation is at the origin. 2. scale. 3. rotate. 4. translate object to desired location |
| 04:39 | < Derakon> | (Also, scaling uses the same "tied to the origin" thing -- when you scale, the length of the rope changes too) |
| 04:40 | <~Vornicus> YOu can do all this with matrix multiplications but I don't remember which direction they go when you do that. |
| 04:40 | < celticminstrel> | Oh wait. The translation vector I'm using isn't what my textbook tells me to use. |
| 04:40 | < Derakon> | Whups! |
| 04:40 | < celticminstrel> | That makes more sense. |
| 04:40 | < Derakon> | http://en.wikipedia.org/wiki/Transformation_matrix has a bit on composing transformations. |
| 04:41 | < celticminstrel> | I was thinking it was a bit odd for the translation vector to be the same as the axis of rotation vector. <_< |
| 04:41 | < Derakon> | If you want to do transform A followed by transform B then you want BA. |
| 04:41 | < Derakon> | In other words, the transform closest to the object happens first. |
| 04:41 | < Derakon> | (BA * x, where x is the object; A happens first, then B) |
| 04:41 | < celticminstrel> | I... think that's what I'm doing. |
| 04:42 | < celticminstrel> | At least, the translate/scale/rotate commands all return "trans_matrix * *this". |
| 04:42 | < celticminstrel> | Then I think I'm calling them in the order I want them to happen. |
| 04:42 | | Reiver [orthianz@3CF3A5.E1CD01.C6689C.33956A] has quit [Ping timeout: 121 seconds] |
| 04:43 | | Reiver [orthianz@3CF3A5.E1CD01.C6689C.33956A] has joined #code |
| 04:49 | | Reiver [orthianz@3CF3A5.E1CD01.C6689C.33956A] has quit [Ping timeout: 121 seconds] |
| 04:50 | | Reiver [orthianz@3CF3A5.E1CD01.C6689C.33956A] has joined #code |
| 05:08 | | kwsn is now known as kw-sleep-n |
| 05:37 | < celticminstrel> | ...oh wow. I thought I was chaining the transformations, but instead I was just overwriting the previous one with each successive one. |
| 05:37 | < celticminstrel> | So only the last one took effect. |
| 05:37 | < Derakon> | Hee. |
| 05:38 | < celticminstrel> | Now I'm getting nans. |
| 05:38 | < celticminstrel> | So clearly I'm doing something very wrong. |
| 05:53 | < celticminstrel> | Oh apparently I have 0/0. |
| 05:53 | < celticminstrel> | On second thoughts, my input isn't what I was treating it as. |
| 05:59 | < Derakon> | Make certain you start with the identity matrix. :) |
| 05:59 | | Reiver [orthianz@3CF3A5.E1CD01.C6689C.33956A] has quit [Ping timeout: 121 seconds] |
| 06:02 | < celticminstrel> | Yeah, that's not the issue. It's calculating the cosine and sine of the x-axis component of the angle from the vector of the rotation axis that gives a nan. I'm not actually calling cos or sin; I'm using the formula from the text book which is basically the z component of the projected vector divided by its magnitude... both of which end up being zero. |
| 06:03 | < celticminstrel> | (And for sine, the y component instead of the z.) |
| 06:03 | < celticminstrel> | (The actual vector is (1,0,0).) |
| 06:08 | < celticminstrel> | (Because I haven't gotten to the point where it changes, yet.) |
| 06:08 | | Reiver [orthianz@3CF3A5.E1CD01.C6689C.33956A] has joined #code |
| 06:20 | | Derakon is now known as Derakon[AFK] |
| 06:31 | | celticminstrel [celticminst@Nightstar-5d22ab1d.cable.rogers.com] has quit [Connection closed] |
| 06:32 | | Reiver [orthianz@3CF3A5.E1CD01.C6689C.33956A] has quit [Ping timeout: 121 seconds] |
| 06:34 | | Reiver [orthianz@3CF3A5.E1CD01.C6689C.33956A] has joined #code |
| 07:49 | | Rhamphoryncus [rhamph@Nightstar-14eb6405.abhsia.telus.net] has quit [Client exited] |
| 07:55 | | Reiver [orthianz@3CF3A5.E1CD01.C6689C.33956A] has quit [Ping timeout: 121 seconds] |
| 07:56 | | Reiver [orthianz@3CF3A5.E1CD01.C6689C.33956A] has joined #code |
| 08:06 | | Reiver [orthianz@3CF3A5.E1CD01.C6689C.33956A] has quit [Ping timeout: 121 seconds] |
| 08:11 | | Reiver [orthianz@3CF3A5.E1CD01.C6689C.33956A] has joined #code |
| 08:15 | | Reiver [orthianz@3CF3A5.E1CD01.C6689C.33956A] has quit [Ping timeout: 121 seconds] |
| 08:15 | | Reiver [orthianz@3CF3A5.E1CD01.C6689C.33956A] has joined #code |
| 09:09 | | Kindamoody is now known as Kindamoody|out |
| 09:18 | | You're now known as TheWatcher |
| 09:32 | | Vash [Vash@Nightstar-f03c5637.sd.cox.net] has quit [[NS] Quit: I <3Lovecraft<3 Vorn!] |
| 10:14 | | Reiver [orthianz@3CF3A5.E1CD01.C6689C.33956A] has quit [Ping timeout: 121 seconds] |
| 10:15 | | Reiver [orthianz@3CF3A5.E1CD01.C6689C.33956A] has joined #code |
| 10:20 | | AnnoDomini [annodomini@Nightstar-2a577c2d.58.getinternet.no] has joined #code |
| 10:20 | | AnnoDomini [annodomini@Nightstar-2a577c2d.58.getinternet.no] has quit [[NS] Quit: leaving] |
| 10:21 | | AnnoDomini [annodomini@FFB3F3.4C5BE8.2014E2.DC0864] has joined #code |
| 10:43 | | Attilla [Some.Dude@Nightstar-f29f718d.cable.virginmedia.com] has joined #code |
| 11:27 | | Reiver [orthianz@3CF3A5.E1CD01.C6689C.33956A] has quit [Ping timeout: 121 seconds] |
| 11:27 | | Reiver [orthianz@3CF3A5.E1CD01.C6689C.33956A] has joined #code |
| 11:36 | | Reiver [orthianz@3CF3A5.E1CD01.C6689C.33956A] has quit [Ping timeout: 121 seconds] |
| 11:38 | | Reiver [orthianz@3CF3A5.E1CD01.C6689C.33956A] has joined #code |
| 11:41 | | gnolam [lenin@Nightstar-202a5047.priv.bahnhof.se] has joined #code |
| 12:15 | | AnnoDomini [annodomini@FFB3F3.4C5BE8.2014E2.DC0864] has quit [[NS] Quit: leaving] |
| 13:13 | | Reiver [orthianz@3CF3A5.E1CD01.C6689C.33956A] has quit [Ping timeout: 121 seconds] |
| 13:15 | | Reiver [orthianz@3CF3A5.E1CD01.C6689C.33956A] has joined #code |
| 13:31 | | Reiver [orthianz@3CF3A5.E1CD01.C6689C.33956A] has quit [Ping timeout: 121 seconds] |
| 13:35 | | Reiver [orthianz@3CF3A5.E1CD01.C6689C.33956A] has joined #code |
| 14:36 | | kw-sleep-n is now known as kwsn |
| 15:07 | | Reiver [orthianz@3CF3A5.E1CD01.C6689C.33956A] has quit [Ping timeout: 121 seconds] |
| 15:13 | | Reiver [orthianz@3CF3A5.E1CD01.C6689C.33956A] has joined #code |
| 16:12 | | Kindamoody|out is now known as Kindamoody |
| 16:58 | | Rhamphoryncus [rhamph@Nightstar-14eb6405.abhsia.telus.net] has joined #code |
| 17:20 | | Reiver [orthianz@3CF3A5.E1CD01.C6689C.33956A] has quit [Ping timeout: 121 seconds] |
| 17:25 | | Reiver [orthianz@3CF3A5.E1CD01.C6689C.33956A] has joined #code |
| 18:15 | | Kazriko [kazrikna@4CA975.462904.08E9A2.992A2D] has quit [Ping timeout: 121 seconds] |
| 18:44 | | Kazriko [kazrikna@4CA975.462904.08E9A2.992A2D] has joined #code |
| 18:47 | | Reiver [orthianz@3CF3A5.E1CD01.C6689C.33956A] has quit [Ping timeout: 121 seconds] |
| 18:52 | | Reiver [orthianz@3CF3A5.E1CD01.C6689C.33956A] has joined #code |
| 19:02 | | Rhamphoryncus [rhamph@Nightstar-14eb6405.abhsia.telus.net] has quit [Ping timeout: 121 seconds] |
| 19:02 | | Reiver [orthianz@3CF3A5.E1CD01.C6689C.33956A] has quit [Ping timeout: 121 seconds] |
| 19:08 | | Reiver [orthianz@3CF3A5.E1CD01.C6689C.33956A] has joined #code |
| 19:12 | | Reiver [orthianz@3CF3A5.E1CD01.C6689C.33956A] has quit [Ping timeout: 121 seconds] |
| 19:14 | | Rhamphoryncus [rhamph@Nightstar-14eb6405.abhsia.telus.net] has joined #code |
| 19:14 | | Kindamoody is now known as Kindamoody[zZz] |
| 19:16 | | Reiver [orthianz@3CF3A5.E1CD01.C6689C.33956A] has joined #code |
| 19:19 | | Reivles [orthianz@3CF3A5.E1CD01.C6689C.33956A] has joined #code |
| 19:20 | | Reiver [orthianz@3CF3A5.E1CD01.C6689C.33956A] has quit [Ping timeout: 121 seconds] |
| 19:27 | | Derakon [chriswei@Nightstar-f68d7eb4.ca.comcast.net] has joined #code |
| 19:28 | | Reivles [orthianz@3CF3A5.E1CD01.C6689C.33956A] has quit [Ping timeout: 121 seconds] |
| 19:28 | | * Derakon mutters at pattern matching. |
| 19:28 | < Derakon> | I want to grab everything up to the first instance of either "_t\d+" or "_p\d+". |
| 19:29 | < Derakon> | Something like "(.*)[_t|_p]" will bypass the first one (e.g. it will grab out "foo_t" from "foo_t5_p4". |
| 19:29 | < Derakon> | Er. It will grab "foo_t5". |
| 19:32 | | Reiver [orthianz@3CF3A5.E1CD01.C6689C.33956A] has joined #code |
| 19:36 | | Reiver [orthianz@3CF3A5.E1CD01.C6689C.33956A] has quit [Ping timeout: 121 seconds] |
| 19:39 | | Reiver [orthianz@3CF3A5.E1CD01.C6689C.33956A] has joined #code |
| 19:53 | < Simon_Shine> | Derakon, /(.*?)(?=_[tp]\d+)/ |
| 19:53 | < Simon_Shine> | Derakon, .*? is lazy so it tries to match as little as possible, and (?=...) is a look-ahead, so the regex pattern doesn't eat it. |
| 19:54 | < Simon_Shine> | Derakon, it's not a particularly efficient pattern. |
| 19:54 | < Derakon> | Nifty, thanks. |
| 19:55 | < Derakon> | It doesn't have to be hugely efficient; I'm only doing this once per file and there's never going to be more than maybe a few hundred files. |
| 19:55 | < Derakon> | What I was missing was the lookahead. |
| 19:57 | | Reiver [orthianz@3CF3A5.E1CD01.C6689C.33956A] has quit [Ping timeout: 121 seconds] |
| 19:57 | < Simon_Shine> | not really. /(.*?)[_tp]\d+/ would work so long as you only fetch the back-reference |
| 19:58 | < Simon_Shine> | I think it's the lazy quantifier that was missing. |
| 19:59 | < Derakon> | I knew about .*? but had trouble getting it to be the precise amount of lazy I needed. :) |
| 19:59 | < Simon_Shine> | ahh |
| 19:59 | < Simon_Shine> | it's "as greedy as it needs to be" |
| 19:59 | < Derakon> | Yeah. |
| 20:00 | < Simon_Shine> | if (something)*? *can* match, it will. if it can match in several ways, it will just return the first match. |
| 20:00 | < Derakon> | Right, what I was having trouble with was keeping it from finishing its match until it had grabbed the entire bit I wanted. |
| 20:01 | < Derakon> | E.g. with a string of abc_def_t01_p05 I kept getting "abc" as my match. |
| 20:03 | | Reiver [orthianz@3CF3A5.E1CD01.C6689C.33956A] has joined #code |
| 20:07 | | Reiver [orthianz@3CF3A5.E1CD01.C6689C.33956A] has quit [Ping timeout: 121 seconds] |
| 20:08 | | Reiver [orthianz@3CF3A5.E1CD01.C6689C.33956A] has joined #code |
| 20:53 | | * Derakon snerks at this comment on an article discussing the possibility that the universe has a significant nonzero angular momentum (i.e. it's spinning). |
| 20:54 | < Derakon> | The comment suggests that perhaps the astronomer missed a single proton somewhere rotating at immense speed... |
| 20:54 | < Derakon> | Or rather, counter-rotating. |
| 21:08 | | celticminstrel [celticminst@Nightstar-5d22ab1d.cable.rogers.com] has joined #code |
| 21:35 | | Reiver [orthianz@3CF3A5.E1CD01.C6689C.33956A] has quit [Ping timeout: 121 seconds] |
| 21:40 | | Reiver [orthianz@3CF3A5.E1CD01.C6689C.33956A] has joined #code |
| 21:44 | | Reiver [orthianz@3CF3A5.E1CD01.C6689C.33956A] has quit [Ping timeout: 121 seconds] |
| 21:46 | | Reiver [orthianz@3CF3A5.E1CD01.C6689C.33956A] has joined #code |
| 21:50 | | Reiver [orthianz@3CF3A5.E1CD01.C6689C.33956A] has quit [Ping timeout: 121 seconds] |
| 21:51 | | Reiver [orthianz@3CF3A5.E1CD01.C6689C.33956A] has joined #code |
| 21:55 | | McMartin [mcmartin@Nightstar-a47b9f7a.pltn13.sbcglobal.net] has quit [[NS] Quit: upgrade] |
| 21:58 | | McMartin [mcmartin@Nightstar-a47b9f7a.pltn13.sbcglobal.net] has joined #code |
| 22:02 | | Reiver [orthianz@3CF3A5.E1CD01.C6689C.33956A] has quit [Ping timeout: 121 seconds] |
| 22:03 | | Reiver [orthianz@3CF3A5.E1CD01.C6689C.33956A] has joined #code |
| 22:07 | | Reiver [orthianz@3CF3A5.E1CD01.C6689C.33956A] has quit [Ping timeout: 121 seconds] |
| 22:08 | | Reiver [orthianz@3CF3A5.E1CD01.C6689C.33956A] has joined #code |
| 22:42 | | Reiver [orthianz@3CF3A5.E1CD01.C6689C.33956A] has quit [Ping timeout: 121 seconds] |
| 22:42 | | Reiver [orthianz@3CF3A5.E1CD01.C6689C.33956A] has joined #code |
| 22:48 | < Rhamphoryncus> | Anybody know why a programming vector is called that? Is there some relation to the math vector that I'm not seeing? |
| 22:50 | < Derakon> | Probably because array was already taken and matrix implies at least two dimensoins. |
| 22:50 | < Derakon> | Er, dimensions. |
| 22:52 | < Rhamphoryncus> | But a vector is a specific structure that might be implemented using a tuple. How does that become more generic than a tuple? |
| 22:54 | < Rhamphoryncus> | http://stackoverflow.com/questions/581426/why-is-a-c-vector-called-a-vector |
| 22:54 | < Derakon> | Er, wait, are we talking about the C++ STL object or something else? |
| 22:55 | < Derakon> | Okay. |
| 22:55 | < Rhamphoryncus> | Says the math vector is fucking multidimensional |
| 22:55 | < Rhamphoryncus> | Like.. arbitrary number of dimensions |
| 22:55 | < Rhamphoryncus> | Thus an arbitrary amount of data |
| 22:55 | < Derakon> | Well, a mathematical vector is a single offset in any number of dimensions. |
| 22:56 | < Derakon> | A.k.a. [x, y, z] is a 3D vector. |
| 22:56 | < Derakon> | s/A.k.a./e.g./ |
| 22:56 | < Derakon> | Note also the 40-point answer on that page. |
| 22:57 | < Rhamphoryncus> | heh |
| 22:57 | < Rhamphoryncus> | Precisely backwards with array. Awesome. |
| 23:00 | < Rhamphoryncus> | So when it comes to designing my own language I should stay the fuck away from the term "vector" |
| 23:00 | < Derakon> | Hell, avoid keywords altogether~ |
| 23:00 | < Derakon> | Make every bit of syntax a number or symbol! |
| 23:02 | < Rhamphoryncus> | ? ftw? |
| 23:02 | < Namegduf> | There aren't enough numbers and/or symbols for C++ |
| 23:02 | < Namegduf> | Unless you get into Unicode, I guess |
| 23:02 | < Rhamphoryncus> | wait, I guess ? is preferred |
| 23:03 | < Rhamphoryncus> | just a little hard to type. ctrl-shift-u 3 b b |
| 23:03 | < Rhamphoryncus> | oops. ctrl-shift-u 3 b b space |
| 23:03 | < Derakon> | Which one is that? |
| 23:03 | < Derakon> | Oh, lambda. |
| 23:04 | < McMartin> | Derakon: Are you remaking APL here? |
| 23:04 | < Derakon> | What? No. That's Rhamph. |
| 23:04 | < McMartin> | 15:00 < Derakon> Make every bit of syntax a number or symbol! |
| 23:04 | < Derakon> | Assuming he listens to me, which he shouldn't. |
| 23:05 | < Rhamphoryncus> | I'm the nutjob making the language |
| 23:05 | < Rhamphoryncus> | Derakon is just giving "helpful" suggestions |
| 23:05 | < Derakon> | (As soon as I figure out that someone knows what they're doing, I switch from giving good advice to giving terrible advice) |
| 23:05 | < Rhamphoryncus> | heh |
| 23:05 | < McMartin> | Seriously, though, check out APL |
| 23:05 | < McMartin> | It is hilarious madness |
| 23:06 | < McMartin> | Especially since it also hails from the era when ASCII wasn't even all 7 bits all the time. |
| 23:06 | < Rhamphoryncus> | Ironically the original APL would have worked better since it would have had a special keyboard |
| 23:06 | < Rhamphoryncus> | vs that ctrl-shift-u 3 b b space madness |
| 23:06 | < Rhamphoryncus> | that wasn't ASCII ;) |
| 23:14 | | Reiver [orthianz@3CF3A5.E1CD01.C6689C.33956A] has quit [Ping timeout: 121 seconds] |
| 23:19 | < celticminstrel> | If it were my choice, I would've called std::vector "array" or something based on that word (maybe dynarray?). |
| 23:20 | | celticminstrel is now known as celmin|away |
| 23:21 | | Reiver [orthianz@3CF3A5.E1CD01.C6689C.33956A] has joined #code |
| 23:23 | < Rhamphoryncus> | You're in luck, new C++ will have another type with array in the name ;) |
| 23:24 | < Rhamphoryncus> | std::array |
| 23:24 | < Derakon> | Dynnari? |
| 23:24 | < Rhamphoryncus> | of course it's static size too |
| 23:41 | | Derakon [chriswei@Nightstar-f68d7eb4.ca.comcast.net] has quit [[NS] Quit: leaving] |
| 23:43 | < celmin|away> | Yeah, but std::array is just a wrapper for native arrays. |
| 23:43 | | celmin|away is now known as celticminstrel |
| --- Log closed Wed Oct 19 00:00:14 2011 |