bug in 2d visualization

[CarloLucibello]

I obtain a weird result when trying to visualize multiple 2d points, as in

    w, r = glscreen()
    s = Vec2f0(100)
    N = 10
    points = Point2f0[rand(Point2f0, -10f0:eps(Float32):10f0) for x=1:N]
    # points = Point2f0[Point2f0(x,0f0) for x=-10f0:0.1f0:10f0]
    v = visualize(points, scale=s)
    view(v)
    r()

The resulting render is a single point in the bottom left corner of the window:

The same happens when running the testsuite, except for two cases,

push!(TEST_DATA2D, visualize(particle_robj[:positions], scale=s, color=particle_color, style=Cint(OUTLINED), shape=Cint(ROUNDED_RECTANGLE)))

and

curve_data(i) = Point2f0[Point2f0(sin(x/i)*250, x) for x=1:1024]
push!(TEST_DATA2D, visualize(const_lift(curve_data, bounce(20f0:0.1f0:1024f0)), :lines))

which are displayed correctly

Cheers,
Carlo

[SimonDanisch]

Well this is sort of to expected, since the units are in pixels.
Have you tried rand(Point2f0, -500f0:eps(Float32):500f0) ?
It should be okay in the testsuite, though. Can you attach a screenshot?

[CarloLucibello]

thank you for helping out

Have you tried rand(Point2f0, -500f0:eps(Float32):500f0) ?

the same happens

Here is tests' screenshot, see those empty spaces

How do you translate the camera btw? I can only rotate and zoom

[SimonDanisch]

Oh wait, I think I've seen that bug.
Can you try to replace the call to glscreen with this:

w,r = glscreen(debugging=true)

Also, this might be fixed in the next2 branch.
If you wish, I can look what packages you will need to checkout.
It should be somethin like:

Pkg.checkout("GLVisualize", "next2")
Pkg.checkout("GLAbstraction", "sd/next²")
Pkg.checkout("GeometryTypes", "sd/typealias")
Pkg.checkout("FixedSizeArrays")
Pkg.checkout("GLWindow")

[SimonDanisch]

The middle mouse button gives you translate... I'll need to rework the camera, since a lot of people don't actually use a mouse with a middle button...

[CarloLucibello]

I tried the debugging option but it didń't print any additional information

on the branch next2 my script yields this

Now I have one single triangle and a change in the camera mode (from orthogonal to perspective I think)

[SimonDanisch]

Can you try runtest.jl?

[CarloLucibello]

[SimonDanisch]

Great, this looks like it should be fixed :)
Try

using GLVisualize, GeometryTypes
w, r = glscreen()
s = Vec2f0(100)
N = 10
points = Point2f0[Point2f0(x,0f0) for x=linspace(0, N*s[1], N)]
v = visualize(points, scale=s)
#v = visualize((Circle(Point2f0(0), scale[1]), points))
view(v, method=:orthographic_pixel) #it's actually a bug that you need to define the method.
r()

[CarloLucibello]

nice, it works!

[SimonDanisch]

Awesome, thanks for trying this out! I'll leave this open until next2 is merged!

[ashleylid]

Hi - is this still open? I get an error when running the test:

ERROR: LoadError: ArgumentError: StructsOfArrays not found in path

I am rendering in a loop and would like to clear the screen re-rendering in the same window, is there an easy way to do that? Because with every render the loop slows.

[SimonDanisch]

So are you on the next branch?
I've updated the README in that branch. It now contains all the packages you need.
Note, that I just pushed some changes to the ArraysOfStructs fork, which are not compatible yet -.-
You need to be on this commit:
JuliaArrays/StructsOfArrays.jl#3
SimonDanisch/StructsOfArrays.jl@333e5f5

[ashleylid]

I need the sd/screenbuffer so have just been merging with master. When I am on either branch of GLVisualize:

Pkg.test("GLVisualize")
INFO: Computing test dependencies for GLVisualize...
INFO: No packages to install, update or remove
INFO: Testing GLVisualize
surface
Barplot
3d dot particles
some more funcitonality from Meshes
particles
sierpinski
vectorfielddata
volume_data
ERROR: LoadError: LoadError: syntax: "<<" is not a unary operator

I was just running the test file out of interest, to make sure everything was working. But my origional quest was just to find a way to clear the window before rerunning my loop. I found this: empty!(w.renderlist) but it doesnt seem to do anything.. ie the render slows with every iteration. Which hurts after 1000 renders - just crashes.

I dont want to bog you down with stuff thats not that important, although it would be nice to fix what the problem is - what branch should I be on?

[ashleylid]

Which branch should I be on? I am a little further along with the test something about << not being a unary operator - but I put this on another issue..