Historical errata for OpenGL ES 2 for Android

Inside box “Rendering in a background thread”

“[…] we can also configure the surface view to render only on request by calling GLSurfaceView.setRenderMode(), with GLSurfaceView.RENDERMODE_CONTINUOUSLY as the argument.”

I think to render on request only the argument should read
“RENDERMODE_WHEN_DIRTY”.

Chapter “Installing the JDK and Android SDK”. You recommend using JDK 6. Is there any reason for not using JDK 7? JDK 6 reaches its end of life this month. See: www.oracle.com/technetwork/java/eol-135779.html.

page 98 (epub)
Under “How does a varying get blended at each fragment”, “…OpenGL will take the two values belong to a line…”
should be “…OpenGL will take the two values belong*ing* to a line…”

glVertexAttribPointer(int index, int size, int type, boolean normalized, int stride, int offset)

Wrong method signature, it should be:
glVertexAttribPointer(int index, int size, int type, boolean normalized, int stride, Buffer data)
Same error appear in the detail table right below, page 57

Second line from top:
A puck is built out of one cylinder top and one cylinder side…

Correct should be:
A puck is built out of one circle top and one cylinder side,,,

3rd paragraph from the bottom:

We first send the color on to the fragment shader,…

Correct should be:

We first send the color on to the vertex shader,…

First line:

First, we pass on the vertex position to the fragment shader and invert the position’s Z component; this is to give the fragment shader…

Correct should be:

First, we pass on the vertex position to the vertex shader and invert the position’s Z component; this is to give the vertex shader…

eclipse doesn’t like return 0 within the if block. it stops giving me an error if I place the return outside of the if block. I don’t know if its a logical error or a syntax error.

final int shaderObjectId = glCreateShader(type);
if (shaderObjectId == 0) {
if (LoggerConfig.ON) {
Log.w(TAG, “Could not create new shader.”);
}
return 0;
}

- You don’t speak about the default coordinate system (0,0,0) in center of the screen.

- You don’t speak about the default scale of coordinate (1,1,1) is the max by default (?)

I haven’t finish to read the book , but i don’t find these important information at the start.

In the validateProgram() method, call to glGetProgramInfoLog() should only be made iff InfoLogLength > 0 as illustrated below:

Log.v(TAG, “Results of validating program: ” + validateStatus[0] + “\
”);

final int[] logLen = new int[1];
GLES20.glGetProgramiv(programObjectId, GLES20.GL_INFO_LOG_LENGTH, logLen, 0);
if (logLen[0] > 0)
Log.v(TAG, “Log: ” + GLES20.glGetProgramInfoLog(programObjectId));
else
Log.v(TAG, “Log: ”);

In the last sentence of the paragraph:
“To
do this, we can take the difference between the current system time and
globalStartTime, and since System.nanoTime() returns the time in nanoseconds, we’ll
just need to divide the difference by 1 trillion to convert this into seconds”

But it should say 1 billion, not 1 trillion.

In the subchapter “Uploading and Compiling the Shader Source Code” the code-block “glCompileShader(shaderObjectId);” isn’t marked as the other ones. Can easily miss it, when only skimming for code.

The text says that we have time in nanoseconds, so we need to divide by TRILLION (10 ^ 12) to get seconds. That’s clearly a mistake, as we should divide by BILLION (10 ^ 9).

Later the code is shown dividing by BILLION, as expected, that confirms error earlier in text where it says TRILLION.

In context:

“To
do this, we can take the difference between the current system time and
globalStartTime, and since System.nanoTime()returns the time in nanoseconds, we’ll
just need to divide the difference by 1 trillion to convert this into seconds.”

And code:

floatcurrentTime=(System.nanoTime()-globalStartTime)/1000000000f;

Hope that helps :)

On page 104 the projection matrix is described.
Comparing with the implementation of the various elements on page 106 and 107 you will find that the element m[14] is wrong. It contains a term (2f*f*n), but should be (2*f*n).
Correct it will be:

m[14] = -((2*f*n) / (f-n));

Best regards,

Hi there!

If you take a look at the code for Heightmap.java, when you are trying to clamp the co-ordinates to ensure that you don’t access any pixels in the image out of bounds, the width - 1 and height - 1 variables should be reversed.

Specifically, the book has:
row = clamp(row, 0, width - 1);
col = clamp(col, 0, height - 1);

This should actually be:
row = clamp(row, 0, height - 1);
col = clamp(col, 0, width - 1);

Awesome book btw. Cheers.
- Ray.

The code where you unbind the texture after we are finished, the book currently has it as

glBindTexture(GL_TEXTURE_2D, 0);

Because we are generating a cube map, and because the beginning binding call was glBindTexture(GL_TEXTURE_CUBE_MAP, textureObjectIds[0]);, I believe the unbinding call should be:

glBindTexture(GL_TEXTURE_CUBE_MAP, 0);

The example still works, but I figure for the sake of consistency, this should be correct.

Thanks!
- Ray.

Possibly modify Figure 43 to use the xz axes.

The figure currently shows a unit circle in the xy plane, but in the description that follows (and in the code) the circle is orientated in the xz plane. This causes some confusion in the third paragraph, where the text says the unit circle y component maps to the y position.

final float yPosition =
(float)Color.red(pixels[(row*height) + col]) / (float)255;

it’s working with the same width and height, but it’s not working with the different length of width and height. I suggest that height should be replaced by width.

final float yPosition =
(float)Color.red(pixels[(row*width) + col]) / (float)255;

in page 112,the rotation matarix for round the x,the 7th element in figure is sin(a),but in example code, i am check the code from android.opengl.Matrix class,method rotateM implement,is use sin(a),not(sin), element 9 on book is sin(a),but code is -sin(a).

i am try write that code use figure from book,result is negative direction,i hope book can tell me why have this diffent and how to correct。

GLES20.GL_LINK_STATUS => GL_LINK_STATUS, since a static import is done of android.opengl.GLES20.*.
It is also used the correct way in the cal to glGetProgramiv() just above.

GLES20.GL_COMPILE_STATUS => GL_COMPILE_STATUS, since static import of android.opengl.GLES20.* is done.
Correctly used in the call to glGetShaderiv() at the end of page 55.

At he point where you get to compile, you compile and I didn’t get Figure 32, that was because you don’t say in the book that you have to switch to the old Matrix, on tableVertices, it was in the code where I realised that you have then used the Matrix with the old components X, Y, R, G ,B .
And that you have switched the POSITION_COMPONENT_COUNT back to 2.
So far your book is very understandable. Thanks.

Hello,

i can’t find the error.

The Table is black. No Textures