« first day (1888 days earlier)      last day (3122 days later) » 

Pip
1:11 AM
lol
 
 
1 hour later…
2:23 AM
woops
1
Q: Recover iOS App from itunes Connect

Jorge Bar MzaCan I recover the Xcode project of an app I uploaded into the AppStore? I lost my previous hard drive (where it was created and stored) and now I would like to keep improving my app but the only link I have to the project is iTunes Connect. Is there any way to do it?

 
if it's not backed up it's not important!
 
:D
 
I've backed up my project 3 times today because I don't trust Unity lol
 
:D
damn perforce unity plugin is buggy as fuck
the one from P4 for use with unity personal, not the one from unity you get with pro
 
:(
 
2:34 AM
0
Q: Why answers' authors do not up-vote the questions they answer

MAndI will admit: I am new here. Still, one thing that has intrigued me is that: why some one who bothers to give an answer to a given question does not up-vote that question? On the one hand, one might think that a question is worth a reply, but is not necessarily super useful. On the other hand, w...

Check out Josh's answer
 
voting works in mysterious ways~~~
 
Did you notice that Josh answered and did not upvote the question? :D
 
yes lol
 
user4704
3:03 AM
@Almo Quite.
 
3:20 AM
made my first multiple C# constructor today. so proud
 
 
4 hours later…
6:53 AM
0
Q: Merge [lag] and [latency]

congusbongusThe tags lag and latency seem to be about the same thing - network latency. Some of the questions even use the terms interchangeably, for example tagging using one but using the other in the question text. These tags should be synonyms for each other. However, I'm not sure what the parent tag sh...

 
7:26 AM
@Chris Constructor chaining?
 
just like, two different params you could pass through without requiring to null either
 
Code?
 
    public InputProfileKeybind(InputProfile newInputProfile, KeyCode newKeyCode) {
		currentInputProfile = newInputProfile;
		keyCode = newKeyCode;
		gamepadButtonType = InputControlType.None;
	}

	public InputProfileKeybind(InputProfile newInputProfile, InputControlType newGamepadButton) {
		currentInputProfile = newInputProfile;
		keyCode = KeyCode.None;
		gamepadButtonType = newGamepadButton;
	}
could never do something like that with UnityScript lol
 
Yeah for overloading operators
 
Neato :)
using System;
using System.Collections.Generic;


namespace Utilities
{

    public class Scanner
    {
        private Queue<String> m_buffer = new Queue<string>();
        private void FlushConsole()
        {
            if(m_buffer.Count == 0)
            {
                string[] str = Console.ReadLine().Split(' ');
                foreach(string s in str)
                {
                    m_buffer.Enqueue(s);
                }

            }

        }

        public void SkipNext()
I implemented a scanner in C# :)
 
7:41 AM
sweet
 
Do you think the cache feature is actually beneficial?
 
Probably
I would agree with your intuition that asking for "Is the next value an int" would usually be followed by "Okay give me that int then"
 
7:59 AM
Yeah and it parses the value regardless
 
 
2 hours later…
10:07 AM
Morning tips hat
 
hullo
 
10:28 AM
ö/
 
<o/
 
10:59 AM
Hello
 
I'm currently doing a basic little game engine and the sprite has some rather jagged movements when it tries scrolling, was wondering if this is a common problem that crops up
I'll probably do an actual question once I get back home, but I was wondering if that was an easy thing to fix maybe
 
I'm not sure what you mean from that description
 
Camera movement is not smooth ^
 
One possible interpretation, could also be that the sprite is not smooth while the camera is =p
 
11:07 AM
Yeah it's a bit hard to describe
 
Is the background scrolling smoothly but the sprite is not?
 
When the camera moves along with the sprite, it seems to go back and forth rather than a smooth motion
It moves fine when no scrolling is involved, though
 
Are you separating game ticks from frames? Sounds like a problem I've had with that once
What was it though, bleh
 
Hm, could be
IIRC the speed of animations is independant from frames
Maybe I should rewrite that
 
I think I was interpolating velocity in frames (i.e. renderPos = actualPos + velocity * tickAccumelator) and doing something wrong with velocity
I.e., moving 1 unit per tick while having velocity set to 2
The rendering would move 2 units and then on the next tick go back 1 tick for the actual movement
 
11:11 AM
(The engine I built it on currently is a mash up of three independant tutorials so it's kind of a mess)
Lots of cleaning up to do
 
Heh, that happens even when not following tutorials =p
Takes lots and lots of experience and planning to avoid that
 
The tile tutorial and the sprite tutorial were different, so I have to join up all that :p
Works okay so far fortunately
Though had some troubles programming the A* algorithm so I said fuck it and went with a greedy path algorithm
My sprites are in no hurry
 
They're just relaxing and having a drink or two with friends, no harm done right?
And yeah if you can't figure it out, greedy path works fine and it's pretty much a drop in replacement if you need more speed
(And if that's the bottleneck (unlikely) )
 
I'll probably rewrite it later yeah
Really it's not even that bad
I tried the whole concave obstacle, it didn't go all the way inside before turning back
You're alright, greedy algorithm
Though I guess that could be a problem for a really huge concave structure
 
Pip
morning (heading to school so cya all later) :P
 
11:26 AM
Hey! Bye! o/
 
I guess what I should do is instead of tying animations to the clock, I should just define a number of frames
 
@Slereah Depending on what you want to do, it might be just the wrong approach ... Typically, you want to define animation speed in frames per second (or alternatively, attach duration to each frame). For moving animations, movement speed is an issue so if you want to be able to vary it, the capability of (linearly!) scaling the animation speed up and down should also be available.
 
I guess I should try to see if it fixes the issue, first
Also is there a formula to compute to which direction a vector is closest for 8 directions
I managed to find it for 4, but 8 starts to get a bit tricky
That is, if a vector is closest to up, down, left, right, or the diagonals
 
11:58 AM
@Slereah Good question. Let me see if I can think up a solution ...
 
For 4 directions it's basically whether x and is superior or inferior to 0 and if x is inferior or superior to +- y
Basically which quadrant of a square it belongs to
But for 8 directions it's weird octants
 
General idea: Normalise the vector (make it length 1). Calculate the direction it's pointing at (atan(x, y)). The result is between 0 and 2*PI, with 0 being "to the positive x axis". Multiply by 4/PI, leading to a value between 0 and 8. Add 0.5 (the value is between 0.5 and 8.5 now). Round down to the nearest whole number. Bitwise AND with 7. Result is between 0 and 7.
I'm pretty sure you can create a method purely in the integer domain though, I just need to grab something to eat now. :)
 
no problem, thanks
I usually try to stick with integers because using floats for pixel coordinates feels a bit like a waste :p
 
12:22 PM
However if you're doing 8 directions you're gonna need some approximation of sqrt(2)
 
Maybe not exactly
 
Never exactly as sqrt(2) is irrational =p
 
the lines to consider aren't diagonals
 
Well it's like
For 4 directions, it's gonna be diagonals on a squre
[x]
Upper quadrant is up and so forth
But for 8 directions, diagonals are not involved
it's lines that are... 22.5° around the diagonals
$\pi/8$
I guess that if I use reals, that would be the easiest way to do it really
Compute the angle on a unit circle
Then it's just the angle / pi/8
well, 16
 
12:28 PM
Are the 8 directions equally distributed?
 
Yes
It's just a method to find what sprite to use
 
Then yes, barring rotating the world for no reason but to complicate all the things, they are the diagonals =p
 
I guess I either have to use floats or just do a fraction on the coordinates
 
user92578
Hi
 
Hi
from what I recall early games just used a look up table to not bother with angle problems :p
Who needs to compute things
 
user92578
12:37 PM
lol
 
user92578
yeah, just put some square roots into your calculator and copy them to a map
 
user92578
:D
 
user92578
instant +50 FPS
 
Computing trig functions was probably pretty tough to do before floating point units
I had to do an exponential heavy program once and basically half the time was spent computing them
Since exps aren't dealth with by most CPUs
 
user92578
I haven't done that much math stuff in programming
 
user92578
12:39 PM
I did write a two-point bezier value solver for easing
 
user92578
but that's pretty much it
 
Basically exp is done like
e^x = 2^k * e^r
 
user92578
Yeah doesn't look familiar even at a math level
 
And then e^r is somethin like (((a + r)*r + b) *r + c) ...
So it's pretty heavy for a single operation
 
Ok, the integer domain solution looks like this (warning: assumes you know basic complex numbers arithmetic to make sense): Interpret the (x, y) vector as a complex number v = x + iy. Determine in which quadrant it lies (simple check for the sign on x and y). Calculate v ^4 (which quadruples the angle the vector is at, but also makes it much longer, so watch out for overflows). Determine the result's quadrants in the same way for further refinement.
 
12:44 PM
I do yeah
Thanks
I'll give it a try tonight
Should also look into doing some splines to have less jagged path too
 
That, or using a simple lerp on the fly.
 
user92578
Now on to the search of a task that is productive but doesn't involve me working on trigger editing.
 
user92578
I don't know why, but that's just something I can't get myself to do
 
user92578
It's so messy and involves a lot of work
 
user92578
I'll watch a few Broforce dev logs for motivation
 
12:51 PM
the best of course would be to have a 1D game
Lineland simulator
No more path problems
 
@Slereah Swords and Soldiers is a 1D RTS
makes it very playable on console since selecting units with the controller isn't hard
 
heheh
then again, a lot of games are in effect 1D
old timey side scrollers if no jumping
 
... until you create a 1D racer where your acceleration/braking and turning strength are the input parameters (and your current speed plays a role too) and find out that you suddenly have to find a path through the multi-dimensional input controls space for the AI to play properly.
 
1:19 PM
0
Q: How to handle trivial questions?

Alexandre VaillancourtA user posted this question yesterday. Anyone with a little experience and not colour blind will spot the issue right away. It was so trivial that I started by giving the solution as a comment, but someone eventually suggested to post it as an answer. Often, a question can be answered simply...

 
need to tackle textures, triggers, and rope tomorrow
time for sleep now
 
night
 
@MartinSojka very nice integer domain solution
 
o hai mick
 
Nice thing being that I think C# already has some complex library by default
No need to do the classic school exercize of remaking a complex library :p
 
hey almo
 
Why are we doing math with complex numbers in games? (Excepting quaternions)
 
1:39 PM
Quaternions are homeomorphic to SO(4)!
 
lol i don't know why he's not just taking the angle, rounding it, and just computing a sincos
 
Since they're the even Clifford algebra of R^4
And the same is true of complex numbers and R^2!
 
@KevinvanderVelden Because complex number multiplication and 2D rotation are essentially the same thing.
 
i mean I'm all for overcomplicating problems for a 5% boost in efficiency
 
(Well, rotation and scale for "non-normalised" complex numbers .. but that's true for quaternions too)
 
1:41 PM
@MartinSojka okay, excepting 2d rotation/scale and quaternions
 
but only for fun, what if hes in a rush
 
@KevinvanderVelden Finding out which direction something is facing is a problem related to rotation.
 
if you want to get funny about it, quaternions are just complex-er numbers
 
@MickLH hence me immediately excepting them from the question
 
Not exactly
Complex number are complicated things
 
1:42 PM
well, in a funny sense, exactly
real, i, j, k
 
@Slereah Might one say, they're complex?
 
It's not just a matter of throwing in extra imainary numbers
Quite so
There are no such things as complex numbers with just two imaginary units, for instance
 
I consider the implications as part of the effect, so to each their own.
 
Now, dual quaternions? Those are just plain fun to work with. Too bad the hardware and library support for them is essentially non-existent at the moment.
 
lets write a library, I need something small but math heavy to get back into the swing of things
 
1:45 PM
One thing I was wondering about game development, unrelated to the thing I'm doing currently is
Are all 3D objects made with polygons in games?
 
that's like a semantics question
 
Technically some 3D shapes can be made directly
 
I've been thinking about splitting my engine out into a library of stuff everyone could use, and a front end that's just a thin implementation
 
And with some formulas can give pretty complex shapes
I'm wondering if any of them are used
Like for instance, a sphere
 
@Slereah technically? No. Practically: Purely with triangles
 
1:46 PM
@Slereah My engine supports a "depth override" in a shape definition
 
whatever you do, I think for the GPU to render it, it needs to be a list of verts
 
Spheres are pretty easy without resorting to polygons
 
Hardware acceleration (GPU) is made with triangles in mind
 
you might generate the verts from an equation
 
Doing multiple different techniques is bad for speed
 
1:47 PM
but the GPU can't do anything with just an equation
 
so you can totally use any procedural code you want, my engine renders the bounding box and evaluates your depth-override per pixel
 
(And higher N-gons are just triangles in disguise)
 
Hm, too bad
 
@Slereah Depending on which level we're talking about. For example, many particle systems are points (with additional data) for the engine and when sending them to the graphic card, but get turned into polygons (most often 2d "billboards" of two triangles) for rendering.
 
uh....
did anyone read that? lol
 
1:48 PM
@MickLH read what? =p
 
Was wondering because I came upon this while strolling the internet :
In mathematics, a super-ellipsoid or superellipsoid is a solid whose horizontal sections are super-ellipses (Lamé curves) with the same exponent r, and whose vertical sections through the center are super-ellipses with the same exponent t. Super-ellipsoids as computer graphics primitives were popularized by Alan H. Barr (who used the name "superquadrics" to refer to both superellipsoids and supertoroids). However, while some super-ellipsoids are superquadrics, neither family is contained in the other. Piet Hein's supereggs are special cases of super-ellipsoids. == Formulas == === Basic shape... ===
 
@MartinSojka so a geometry shader? =p
Still triangles
 
The superformula is a generalization of the superellipse and was first proposed by Johan Gielis in 2003. Gielis suggested that the formula can be used to describe many complex shapes and curves that are found in nature. The upcoming computer game No Man's Sky heavily employs the superformula for the procedural generation of its digital universe. In polar coordinates, with the radius and the angle, the superformula is: By choosing different values for the parameters , different shapes can be generated. The formula was obtained by generalizing the superellipse, named and popularized by Piet Hein...
Rather simple formulas to generate shapes
 
you can just formulate procedural shapes in my engine, no triangles required
no tricks needed
 
neat
 
1:49 PM
@KevinvanderVelden Exactly. Triangles at the end, but conceptually something else for the engine.
 
@MartinSojka aah enginewise, yeah true
 
You can provide, instead of triangles, a formula for Z in terms of X and Y
it renders directly on the pixel unit using the bounding box
you discard rays that miss
 
Pretty cool
 
it goes into the g-buffer, gets shaded like everything else
 
Rendering so many polygons to make a sphere when it's a pretty simple formula seemed a bit weird
 
1:50 PM
Hmm, I didn't think of that, deferred shading would make that easier
 
I thought so too...
 
I guess there's a tipping point in complexity of the shape where polygons are better than a formula, but still
 
Screen-space shaders can also create quite some neat effects without a single polygon involved.
 
@Slereah There's also the fact that @MickLH's method is evaluated per pixel every frame, while with triangles the triangles are generated and uploaded once and then just rendered
 
I have a tool also for converting formulas into Z = f(X, Y) form, it uses computer algebra for you since it's such simple manipulations
@KevinvanderVelden it ends up faster for basic stuff in most of my tests because it puts all the pressure on flops and none on bandwidth
 
1:53 PM
Yeah
 
similar deal with procedural materials, which you can use instead of textures, even complex formulas seem to be a better deal than even one memory lookup
 
Problem is of course that that is highly dependent on GPU hardware architecture
 
prolly cuz da bajillion flawps and the one true ram bottleneck
yeah definitely, I have been working for all users to experience vsync at maximum graphical fidelity
but GPUs being wildly different makes it extremely hard to predict when to change quality settings
 
just package a GPU with your game
 
I'm sorry I have to get a couple patents pending before I go on about this topic lol
This area is the only chance I have at making money on my engine
 
1:57 PM
heheh
 
Aah patents, the best way to make sure nobody uses your tech in the game industry
 
well if they do even better, because then you can sue!
 
I already have patent stuff infecting this engine
 
Doubt that he has the money to wage such a battle =p
 
user92578
Anyone seen my motivation? I can't seem to find it anywhere, I looked trough the drawers and everything.
 
1:58 PM
You can just turn it all off if you want, I have talked to the middleware companies
 
Depends on the damages I guess
I think if it's small enough they just throw some money at you to shut you up
 
Le sigh What I wouldn't give for the C linker over this C# shit
 
a friend of mine said software patents are like guns. Don't take one to the beach unless you intend to use it.
 
Someone sometime ago added a renamed dll of a project in the code
 
or something like that.
 
1:59 PM
It's like an advertisement for them, you get to edit your game with all the pretty algorithms on my license, but when you publish you either pay the fees or the algorithms turn off
 
Now sometimes C# tries to use both of them! Which is fine and just picks the first one it loads
Which is of course the wrong one
 
I fell for the C# trick myself
 
A proper fricking linker would give an error
 
Except it was "The program needs .net to run!"
Nooo
 

« first day (1888 days earlier)      last day (3122 days later) »