GetOrientation problems - Android Software Development

It has been a while since I wrote a program for the last time, and when I got my Desire I decided it would be a good moment to start again.
After a few very simple test programs I decided to go a little further and try to access the sensors.
With a little research I found out I had to use public static float[] getOrientation (float[] R, float[] values) . This is what I found in the documentation:
float[] android.hardware.SensorManager.getOrientation(float[] R, float[] values)
public static float[] getOrientation (float[] R, float[] values)
Since: API Level 3
Computes the device's orientation based on the rotation matrix.
When it returns, the array values is filled with the result:
values[0]: azimuth, rotation around the Z axis.
values[1]: pitch, rotation around the X axis.
values[2]: roll, rotation around the Y axis.
All three angles above are in radians and positive in the counter-clockwise direction.
Parameters
R rotation matrix see getRotationMatrix(float[], float[], float[], float[]).
values an array of 3 floats to hold the result.
Returns
The array values passed as argument.
Click to expand...
Click to collapse
I didn't understand the use of float[] R part, so I decided to just give null as argument
Code:
float[] values = new float[3];
SensorManager.getOrientation(null, values);
TextView xText = (TextView) findViewById(R.id.TextView02);
TextView yText = (TextView) findViewById(R.id.TextView03);
TextView zText = (TextView) findViewById(R.id.TextView04);
xText.setText(String.valueOf(values[0]));
yText.setText(String.valueOf(values[1]));
zText.setText(String.valueOf(values[2]));
But my program fails when it comes to the line where I call getOrientation, so my guess is I shouldn't give null as argument. How should I use getOrientation()?

Related

RIL

XDA Developers,
Years ago I downloaded the RIL .dll and API from this site. Then it was only half complete and I had to patch it my self.
I am sure development has been completed since then.
Can anybody tell me if RIL is still maintained by this site? In which case, where can I get it?
Specifically I need signal quality on XDA II upwards. If anybody can help me?
Kind Regards,
Ben
After posting, my text RIL was shown with a link to most of the information.
There is one think I can't understand. The stucture returned as as follows:
int nMinSignalStrength; // @field TBD
int nMaxSignalStrength; // @field TBD
int nLowSignalStrength; // @field TBD
int nHighSignalStrength; // @field TBD
Which have the values (On XDA IIs) of -113, -51, -110, -60
Would any member know what the meaning of these values is?
I have then tried to convert the quality to a percentage. But the percentage always reads way to high, or zero.
My guess is that these are DB and therefore logarithmic. Would any member know how to convert to a linar scale?
My guess is something like: log10(n / 3) where n is any of the above or retuned figure.
Any ideas would be very useful!
Regards, Ben.
Hi,
Code:
HRESULT RIL_GetSignalQuality(
HRIL hRil // @parm handle to RIL instance returned by <f RIL_Initialize>
);
returns the following structure:
Code:
typedef struct rilsignalquality_tag {
DWORD cbSize; // @field structure size in bytes
DWORD dwParams; // @field indicates valid parameters
int nSignalStrength; // @field TBD
int nMinSignalStrength; // @field TBD
int nMaxSignalStrength; // @field TBD
DWORD dwBitErrorRate; // @field bit error rate in 1/100 of a percent
int nLowSignalStrength; // @field TBD
int nHighSignalStrength; // @field TBD
} RILSIGNALQUALITY, *LPRILSIGNALQUALITY;
Why don't you just use nSignalStrength? Sounds pretty simple and linear to me? Nothing to calculate...
In trying to make my posting simple. I think I forgot to clarify my problem
First is the problem that the nSignalStrength falls between two values.
But there are two fields it can fall between:
nMinSignalStrength <= nSignalStrength <= nMaxSignalStrength
nLowSignalStrength <= nSignalStrength <= nHighSignalStrength
Which one should be used? Why are there two?
Secondly, I want to show a percentage result between one of the above. But these figures are, I belive, Decibels (BD). Each 3 DB = a doubling of the value. So 1 = 10%, 4 = 20%, 7 = 40% etc...
Therefore a liniar percent placement of nSignalStrengh tells me nothing. Most values are close to Max, and then suddenly zero.
My math is a little rusty I was hoping somebody may have a nice function for returning the linear range from the logarithmic rage….
Thanks again to any members who can offer some help
Ben
Hmm... but what if min and max are just values currently encountered in your local cell? When you move to another cell you may receive different min/max values. Or maybe these are the values of the farest and nearest cell? I don't know either, but your explanation of min/max sounds worse to me than does mine...
This could be correct. The values I have are on my XDA IIs using O2 are:
Min -113
Low -110
High -60
Max -51
Therefore:
Min < Low < nSignalStrength < High < Max
So I am using Low = 0% and High = 100%. But this returnes figures of above 50% when signal is quite low.
I think the linear conversion is exp(value / 3)
Therefore percent is:
percent = (exp(nSignalStrength / 3) - exp(Low / 3)) / (exp(high / 3) - exp(Low / 3)) * 100;
Which seems to give better figures. But I am not sure how accurate it is..
Any experts on signal quality out there, I'd love to hear from them!
Ben
If anybody is following this thread, this *seems* to return a good percentage for signal quality. I am not sure of the quality or accuracy. But it works
static double dValue, dMax;
dValue = (int)data->nSignalStrength; // (int) to convert twos complement signed integer correctly.
dMax = (int)data->nHighSignalStrength; // (int) to convert twos complement signed integer correctly.
dValue -= (int)data->nLowSignalStrength;
dMax -= (int)data->nLowSignalStrength;
dValue = pow(dValue / -3.0, 2);
dMax = pow(dMax / -3.0, 2);
dValue /= dMax;
dValue *= 100;
if (dValue > 100) dValue = 100; // never
if (dValue < 0) dValue = 0; // never
Regards,
Ben
Hi there, i've been following this thread with interest I have a very limited knowlege of C++ but not even on the PPC.
Would you mind attaching or even PMing your cpp so i could possibly learn more ?
I find i learn more by examples, and am quite interested in making an application that can disable the radio then re-enable it after a set ammount of time (so i can swap between sims, i have a dual sim adapter)
Best thing to do is follow the sample code supplied by these nice gues from xda-developers, by clicking on this RIL link.
BUT replace the ril.h in the .zip archive with the ril.h you will find from the link.
(nb: if the author of this .zip archive is reading, it's way out of date to the ril.h on this site. ps, hows the ril development going?)
If you get the sample code working, this will give you some idea of what is possible. Your options may not be possible. But look at the ril.h and the functions listed. If one of them does what you want, give it a go.
Regards,
Ben

Simulate Button Click

Hi,
I am trying to simulate a PictureBox.Click event. I have searched these forums and the ones on MSDN with many different combinations of search terms.
However I cannot find anything!
Basically what I am trying to achieve is to fire the picturebox's click event from within my code. What I have read so far seems to indicate that I need to make a call to SendMessage (COM interop?) to actually make windows perform the click.
This is for the compact framework version 1.0.
Any help you can give would be great because this is all very new to me, I'm a web application developer by trade so i'm a fish out of water on this one!
OK, the other method that I am investigating is the use of the mouse_event as demonstrated in this article by Daniel Moth.
However I am struggling to find the namespace Win32Api anywhere in the framework so I'm struggling with that also.
*Update*
I have been able to simulate a click using mouse_event in a call to the coredll using the following code:
Code:
[DllImport("coredll")]
static extern bool SetCursorPos(int X, int Y);
[DllImport("coredll")]
static extern void mouse_event(uint dwFlags, uint dx, uint dy, uint dwData, int dwExtraInfo);
[Flags]
public enum MouseEventFlags
{
LEFTDOWN = 0x00000002,
LEFTUP = 0x00000004,
MIDDLEDOWN = 0x00000020,
MIDDLEUP = 0x00000040,
MOVE = 0x00000001,
ABSOLUTE = 0x00008000,
RIGHTDOWN = 0x00000008,
RIGHTUP = 0x00000010
}
bool tempVal = SetCursorPos(x, y);
mouse_event((uint)MouseEventFlags.LEFTDOWN, 0, 0, 0, 0);
mouse_event((uint)MouseEventFlags.LEFTUP, 0, 0, 0, 0);
however I have one outstanding problem (that I know of!).
I am struggling to capture the corrext X and Y co-ordinates of the control. I have tried many different methods and they all return values of 0 for both axis.
How do you guys do it?
Many Thanks
I haven't answered till now since I don't know .NET
But since you found your way to using native APIs I think I can help you.
This is how I would do it in C/C++:
Code:
RECT wndRect; //this is a structure that contains window top, left, bottom and right coordinates.
GetWindowRect(FindWind(L"[I]window class[/I]", L"[I]window name[/I]"), &wndRect);
x = wndRect.left + 1; //add 1 to window position to make sure the click is inside
y = wndRect.top + 1;
GetWindowRect returns the window position in screen coordinates for top left and bottom right corners.
For FindWindow to work you need to know the class and name of the window you want to find. You can find them using a utility called SPY++ which comes with any Microsoft C++ compiler.
The class is window type (so it will probably be something like 'PictureBox' or 'Image') and window name is most likely blank.
Hope this helps.
Thanks fo your help
I don't know if your code would have done the trick or not but I managed to work my way through the different class definitions to find what i needed.
Code:
int x = selectedButton.PointToScreen(selectedButton.Bounds.Location).X + 2;
int y = selectedButton.PointToScreen(selectedButton.Bounds.Location).Y + 2;
This still doesn't work but I really don't have time to work out why, all I know is that the call to SetCursorPos() returns false. I know the mouse_event code works because if I position the cursor over the Start button it opens the menu.
Time is of an essence so for now I'm just going to have to drop this and pray that I have time to finish this when towards the end of my project.
/me thinks writing my first Mobile 5.0 application in 4 days (I'm a web application developer by trade) was bad planning on the management's fault anyway
Thanks for your input though
use this method signature:
[DllImport("coredll")]
public static extern void mouse_event(int dwFlags, int dx, int dy, int dwData, int dwExtraInfo);
int instead of uint or long.
For me it worked.

[Q] Converting an editable text box to be used as a integer

I have a edit able text box that is next to a button lets say onclick of the button it will take the text from the editable text box and converted to an integer so it can be subracted by.. oh lets say 32 so my first method was this.
number = EText.getText();
textOutcome.setText(number - 32);
this gave me an error of
The operator - is undefined for the argument type(s) Editable, int
my second attempt was
textOutcome.setText((Integer.parseInt(EText.getText().toString()) - 32));
This gave me no errors and it ran fine untill i clicked the button which causes the command above to run. and it force closed my application.
Does any one have any ideas i can do, or any sample code?
Thanks,
Blue.
Bumping the topic.
Please can anyone help me with this..!
Have you tried stepping through the code with the debugger? Otherwise, the code you posted looks fine. There is probably something else at play here.
Basicall it force closes whea subtraction is made..
Sent from my Samsung Epic 4g.
Ahh the debugger found something early in the code thanks for the idea !
You have to do any mathematical arithmetic with the integer type, not with string, So when you extract the number from the EditText you have to convert the string into a integer, or whatever type of number it may be.
WHen you have the integer then you can apply the arithmetic.
when this is done you then have to convert the total into a string type again.
Hope this helps.
So basically im going to write a little of code off the top of my head which obviously can have errors. Since im not testing it.
EditText txbox = (EditText)findViewById(R.id.EditText01);
String txdata = txbox.getText().toString();
int txval = Integer.ParseInt(txdata);
//you should make sure that the value is an integer, if not i think //Integer.ParseInt() will just cut off the rest if its irrational.
//Now you can do arithmetic//
int b = 2;
int t;
t = txval - b;
//The total value is in variable t.
So lets convert this into a string and put it back into an object like edittext or textview
txbox.setText(String.valueOf(t));
//Thats all, hope that helps.

Set Device Time from Server

Hi,
Does anybody know how to pull the time from a server and use that time to update the devices clock?
I want to lock the Clock down on the device so the best way to ensure the time is correct is to sync it with a server.
I'm using VB.net.
Give this a spin.
It sets the get the UTC time, from http://www.timeapi.org/ , but have a look at the site to see what else it can do.
Your device should deal with your timezone offset from UTC automatically.
I used .NET Reflector to translate into VB from C#, hence the odd definition of some variables.
Code:
Public Structure SYSTEMTIME
Public wYear As UInt16
Public wMonth As UInt16
Public wDayOfWeek As UInt16
Public wDay As UInt16
Public wHour As UInt16
Public wMinute As UInt16
Public wSecond As UInt16
Public wMilliseconds As UInt16
End Structure
Declare Function GetSystemTime Lib "CoreDll.dll" _
(ByRef lpSystemTime As SYSTEMTIME) As UInt32
Declare Function SetSystemTime Lib "CoreDll.dll" _
(ByRef lpSystemTime As SYSTEMTIME) As UInt32
Private Shared Sub Main(ByVal args As String())
Dim Buffer As Byte() = New Byte(&H19 - 1) {}
Dim DateItems As String() = New String(8 - 1) {}
Dim Separators As Char() = New Char() { "-"c, "-"c, "T"c, ":"c, ":"c, "+"c, ":"c }
Dim Now As New SYSTEMTIME
Program.GetSystemTime((Now))
Dim ResponseStream As Stream = WebRequest.Create("http://www.timeapi.org/utc/now").GetResponse.GetResponseStream
ResponseStream.Read(Buffer, 0, &H19)
ResponseStream.Close
DateItems = Encoding.UTF8.GetString(Buffer, 0, &H19).Split(Separators)
Now.wYear = Convert.ToUInt16(DateItems(0))
Now.wMonth = Convert.ToUInt16(DateItems(1))
Now.wDay = Convert.ToUInt16(DateItems(2))
Now.wHour = Convert.ToUInt16(DateItems(3))
Now.wMinute = Convert.ToUInt16(DateItems(4))
Now.wSecond = Convert.ToUInt16(DateItems(5))
Now.wMilliseconds = 0
Program.SetSystemTime((Now))
End Sub
The original C# code is here. You will have to convert the C# 'using' statements to VB 'Imports'
Code:
using System;
using System.Net;
using System.Text;
using System.IO;
using System.Runtime.InteropServices;
namespace GetTime
{
class Program
{
[DllImport("coredll.dll")]
private extern static void GetSystemTime(ref SYSTEMTIME lpSystemTime);
[DllImport("coredll.dll")]
private extern static uint SetSystemTime(ref SYSTEMTIME lpSystemTime);
private struct SYSTEMTIME
{
public ushort wYear;
public ushort wMonth;
public ushort wDayOfWeek;
public ushort wDay;
public ushort wHour;
public ushort wMinute;
public ushort wSecond;
public ushort wMilliseconds;
}
static void Main(string[] args)
{
byte[] Buffer = new byte[25];
string DateString;
string[] DateItems = new string[8];
char[] Separators = new char[7] { '-', '-', 'T', ':', ':', '+', ':' };
SYSTEMTIME Now = new SYSTEMTIME();
GetSystemTime(ref Now);
Stream ResponseStream = WebRequest.Create("http://www.timeapi.org/utc/now").GetResponse().GetResponseStream();
ResponseStream.Read(Buffer, 0, 25);
ResponseStream.Close();
DateString = Encoding.UTF8.GetString(Buffer, 0, 25);
DateItems = DateString.Split(Separators);
Now.wYear = Convert.ToUInt16(DateItems[0]);
Now.wMonth = Convert.ToUInt16(DateItems[1]);
Now.wDay = Convert.ToUInt16(DateItems[2]);
Now.wHour = Convert.ToUInt16(DateItems[3]);
Now.wMinute = Convert.ToUInt16(DateItems[4]);
Now.wSecond = Convert.ToUInt16(DateItems[5]);
Now.wMilliseconds = 0;
SetSystemTime(ref Now);
}
}
}
Here's how it works: The call to the URL returns the date/time as a char buffer.
"2011-12-02T14:56:38+00:00" as an example.
After converting this to a string we convert it into an array of strings in DateItems[] as
2011
12
02
14
56
38
00
00
using Separators[] to split the fields apart.
These are then converted to ushort values in the time structure, before setting the time on the device with it.
Nice code stephj .
Cheers!!!
Thanks for that stephj, will give it a go.
i've tried this and i'm getting a :
The remote server returned an error: (407) Proxy Authentication Required.
any ideas?
I tried the above program on an emulated device connected to the net via its host PC's connection and it worked a treat.
To get through a proxy, you will need to add the following code.
Code:
Program.GetSystemTime((Now))
**************INSERT THIS ****************
Dim Proxy as new WebProxy("http://ProxyServer:80/",true)
proxy.Credentials = CredentialCache.DefaultCredentials
WebRequest.DefaultWebProxy = proxy
**************INSERT ENDS ****************
Dim ResponseStream As Stream = WebRequest.Create("http://www.timeapi.org/utc/now").GetResponse.GetResponseStream
Where the first parameter of the WebProxy constructor is the name of your proxy server, and the port it uses. Here at work, we can get away with just using our proxy server's internal network DNS name of "Internet:8080", to access the outside web through a Microsoft ISA Proxy/Firewall. Note in this case the proxy uses a different port number of 8080 as opposed to the default http port of 80. If yours is not 80 you will have to provide it. You may have to do some groundwork to track down the name of your proxy server, and the port it uses. I would start by looking at the proxy setup in Settings->Connections->'Set up my proxy server' for starters.
CredentialCache.DefaultCredentials, picks up your default login/credentials and passes them to the proxy server. If it works properly, you should go through it, seamlessly.
Good luck!
Hi steph,
i tweaked this slightly and set up a webreference in my program, i then pull the time down from that in the format of "2011-12-12T09:43:14", it then splits it.
The problem i am having is that sometimes it seems to add an hour to the time.
Any ideas?
What time zone is your device operating under?
Start->Settings->[System]->Clock & Alarms->[Time]
What's the zone, and is it Home or Visiting? - I'll see if I can duplicate it.
GMT London and is set as Home.
As a test I set the time on the device to: 01/06/1999 and time 01:00:00. When i run the program the first time i press my 'Update Time' button and it returns 12/12/11 and 12:53:04, so 1 hour ahead. If i press the button again I get the correct time.
I notice on your original post that the time returned had +00:00 on the end but you dont seem to use it when you break the string down, is this not required?
Thanks for the help, much appreciated
At least I can duplicate it, so that is a start. It only seems to throw the hour forward if the date has changed, which explains why the second call to it corrects the time.
If you just knock the time back a couple of hours, a single call of the program sets the correct time.
Interesting. I'll try and get to the bottom of it, might take a day or so, on and off.
Edit: Got it!
The 1st of June 1999 is in BST one hour in front of GMT/UTC. The first call changes the date and time using BST, leaving the time pushed forward one hour. The date is then set to today's date which is in GMT/UTC. The second call sets the time against the GMT Timezone which corrects the time.
Set the date time to 04:00 yesterday and it works fine.
I'll see if we can fix it a bit better. More later.........
Here's the cure: It's C#, I'll leave you to convert it to VB.
The secret is to set the time twice, but stall the thread for ten seconds inbetween. The device will sort itself out in the gap, as various time and date housekeeping tasks are triggered. The second SetTime(), will carry out the final fix of the time before the program finally ends.
This should also work in BST when changing from a GMT/UTC date, and should also work with all global time zones.
Add this:
Code:
using System.Threading;
SetSystemTime(ref Now);
Thread.Sleep(10000);
SetSystemTime(ref Now);
that's great steph, thanks again.
There is a minor problem with the previous code:- If the minute happens to roll over in the 10 seconds while the process is asleep, then the change of minute will be lost.
Here's a better solution. Create a DateTime object in which to store the returned date and time from the WebRequest().
Use it to set the system date/time, and then advance it forward ten seconds.
Set the date/time with it again, after the ten second sleep() has elapsed.
Seems to work O.K. Post bug reports to this thread, if you find it doesn't.
Code:
using System;
using System.Net;
using System.Text;
using System.Threading;
using System.IO;
using System.Runtime.InteropServices;
namespace GetTime
{
class Program
{
[DllImport("coredll.dll")]
private extern static uint SetSystemTime(ref SYSTEMTIME lpSystemTime);
private struct SYSTEMTIME
{
public ushort wYear;
public ushort wMonth;
public ushort wDayOfWeek;
public ushort wDay;
public ushort wHour;
public ushort wMinute;
public ushort wSecond;
public ushort wMilliseconds;
}
static void Main(string[] args)
{
byte[] Buffer = new byte[25];
string DateString;
string[] DateItems = new string[8];
char[] Separators = new char[7] { '-', '-', 'T', ':', ':', '+', ':' };
SYSTEMTIME Now = new SYSTEMTIME();
Stream ResponseStream = WebRequest.Create("http://www.timeapi.org/utc/now").GetResponse().GetResponseStream();
ResponseStream.Read(Buffer, 0, 25);
ResponseStream.Close();
DateString = Encoding.UTF8.GetString(Buffer, 0, 25);
DateItems = DateString.Split(Separators);
DateTime SaveTime = new DateTime(Convert.ToInt32(DateItems[0]),Convert.ToInt32(DateItems[1]),Convert.ToInt32(DateItems[2]),Convert.ToInt32(DateItems[3]),Convert.ToInt32(DateItems[4]),Convert.ToInt32(DateItems[5]));
Now.wYear = Convert.ToUInt16(SaveTime.Year);
Now.wMonth = Convert.ToUInt16(SaveTime.Month);
Now.wDay = Convert.ToUInt16(SaveTime.Day);
Now.wHour = Convert.ToUInt16(SaveTime.Hour);
Now.wMinute = Convert.ToUInt16(SaveTime.Minute);
Now.wSecond = Convert.ToUInt16(SaveTime.Second);
Now.wMilliseconds = 0;
SetSystemTime(ref Now);
SaveTime.AddSeconds(10);
Thread.Sleep(10000);
Now.wYear = Convert.ToUInt16(SaveTime.Year);
Now.wMonth = Convert.ToUInt16(SaveTime.Month);
Now.wDay = Convert.ToUInt16(SaveTime.Day);
Now.wHour = Convert.ToUInt16(SaveTime.Hour);
Now.wMinute = Convert.ToUInt16(SaveTime.Minute);
Now.wSecond = Convert.ToUInt16(SaveTime.Second);
SetSystemTime(ref Now);
}
}
}
C# .NET CF 2.0 WinMo 5.0 onwards executable is included in the zip file.
It is a console application, so don't expect anything much to happen until the 'hourglass' disappears.
If the first SetSystemTime() ends up changing the date across a Daylight Saving Boundary date, then, during the sleep() period, the device's housekeeping date/time/alarm services will probably retrigger any outstanding task and event reminders for 'today'.
P.S. GetSystemTime() can be dropped it is not required.
works like a charm steph, nice one. thanks.
@stephj, Thanks for the updated code .
Cheers!!!
Seem to be having a problem with this since UK has put the clocks forward an hour, every time I update using this code my device is adding an hour to the time.
The device is an M3 Mobile.
What are the settings on the Start-> Settings->Clock and Alarms->Time[Tab]
It should still have the Home radio button active and the time zone set to GMT London, Dublin. The mobile device should take care of daylight saving itself.
Is your M3 running WinMo 6.1 or 6.5? But I can't see it making much difference.
Works OK on a stock 6.1 Kaiser (Vodafone v1615)
Update: Also seems to work OK on the WM 6.5.3 Professional emulator.
wubbledoos said:
Hi,
Does anybody know how to pull the time from a server and use that time to update the devices clock?
I want to lock the Clock down on the device so the best way to ensure the time is correct is to sync it with a server.
I'm using VB.net.
Click to expand...
Click to collapse
Well on my Kaisers (running either wm 6.1 or 6.5) I use SKTsync to sync my device clock to one of the NIST or SNTP servers. It's freeware and it does it all for you, all you do is run the app. It's soooo simple. I also run an app called CT Scheduler lite. I use it to automatically run SKTsync every night.

Easy question about buffers for someone who actually understands them...

I'm a very accomplished coder when using interpreted languages, but this low-level memory management stuff is proving hard for me to get my head around.
I'm trying to work through the OpenGL ES sample code on the official developer site and I cannot understand why they create a ByteBuffer for the vertex list and then immediately turn around and create a FloatBuffer from the ByteBuffer. Why not just allocate and reference the FloatBuffer directly? Aren't there now two sets of data that basically refer to the same set of numbers? Here's the snippet to which I am directly referring (perhaps the reason lies elsewhere in the code, though?):
public class Triangle {
private FloatBuffer vertexBuffer;
// number of coordinates per vertex in this array
static final int COORDS_PER_VERTEX = 3;
static float triangleCoords[] = { // in counterclockwise order:
0.0f, 0.622008459f, 0.0f, // top
-0.5f, -0.311004243f, 0.0f, // bottom left
0.5f, -0.311004243f, 0.0f // bottom right
};
....
public Triangle() {
// initialize vertex byte buffer for shape coordinates
ByteBuffer bb = ByteBuffer.allocateDirect(
// (number of coordinate values * 4 bytes per float)
triangleCoords.length * 4);
// use the device hardware's native byte order
bb.order(ByteOrder.nativeOrder());
// create a floating point buffer from the ByteBuffer
vertexBuffer = bb.asFloatBuffer();
// add the coordinates to the FloatBuffer
vertexBuffer.put(triangleCoords);
// set the buffer to read the first coordinate
vertexBuffer.position(0);
}
}​
It is basicallly due to the way Java uses data types. In the language of gods (C) we can just cast types and all is good.
What is happening there is a workaround basically.. machines have a word order (not sure if you have ever heard of the term back-words or byte ordering before) the ByteBuffer has a method which shuffles data to put it into the CPU byte order (ByteOrder.nativeOrder()) which can then be sent to the GPU.
Imagine the value 0xFFEE on an Intel CPU that would be stored as the 2 bytes 0xEE 0xFF or "backwards" if you will, it is called a Little endian CPU. An ARM CPU may store it is 0xFF 0xEE in memory as it is Big Endian
Since Java has it's own convention that it always uses, the ByteBuffer is there to provide a method to re-order it to whatever the CPU needs.
Hope my explanation is not more confusing! I am half-asleep and was just heading to bed :-S

Categories

Resources