this time I will show how I did a battery monitor that shows on the taskbar. The C source code uses the drawing functions DrawRect and Polygon to draw a battery symbol onto the taskbar. The drawing is updated every second by a background thread, that queries the battery status using GetSystemPowerStatusEx2. The thread then sends the battery flag reading using SendMessage with WM_USER.

Why did I write an app that already exists? If you look at your windows mobile taskbar, you will find, that the battery or the clock disappear from the taskbar. They will not be visible in all programs, depending on the space left on the taskbar. If you need to have a battery monitor that is shown all the time, this one may be for you. It will just show all the time on top of the taskbar.

The code

The drawings are done from arrays of POINT or RECT structures. So it is easy to change the drawings:

POINT pointsBatt[]={
	{22, 3},
	{4, 3},
	{4, 8},
	{2, 8},
	{2, 16},
	{4, 16},
	{4, 21},
	{22, 21},
	{22, 19},
	{6, 19},
	{6, 5},
	{22, 5}
};

//4 different bars inside batt
RECT rectBatt2[] = {
	{ 7, 7, 10, 17},
	{11, 7, 14, 17},
	{15, 7, 18, 17},
	{19, 7, 22, 17}
};
// a flash sign build as polygon
POINT pointsFlash[] = {
	{13, 2},
	{ 9,12},
	{14,10},
	{13,21},
	{17, 8},
	{12,10},
	{13, 2}
};

The drawings were first created on paper and then I entered the coordinates in these arrays (do you know an app, that will allow to draw on screen and then shows/saves an array of points? Let me know).

Depending on the return code send from the background thread, only parts of the drawing are done on the screen to reflect the actual status:

		case WM_UPDATESIGNAL: //background thread message, signal is in wParam
			iSignal = (int)wParam;
			DEBUGMSG(1, (L"Received WM_UPDATESIGNAL: %i\n", iSignal));
			//force a wm_paint message
			RECT rtUpdate;
			GetClientRect(g_hWnd, &rtUpdate);
			InvalidateRect(g_hWnd, &rtUpdate, true);
			break;
		case WM_PAINT:

			RECT rt;
			hdc = BeginPaint(hWnd, &ps);
				GetClientRect(hWnd, &rt); //draw text

				//the batt polygon
				hPen = CreatePen(PS_SOLID, 0, colorBatt);
				SelectObject(hdc, hBrBatt);
				SelectObject(hdc, hPen);
				x = sizeof(pointsBatt)/sizeof(POINT);
				Polygon(hdc, pointsBatt, x);

				switch(iSignal){
					case BATTERY_FLAG_HIGH:
						//draw all bars
						// the right bar
						FillRect(hdc, &rectBatt2[0], hbrGreen);
						// the mid bar
						FillRect(hdc, &rectBatt2[1], hbrGreen);
						// the left bars
						FillRect(hdc, &rectBatt2[2], hbrGreen);
						FillRect(hdc, &rectBatt2[3], hbrGreen);
						break;
					case BATTERY_FLAG_LOW:
						//draw two bar
						// the right bar
						FillRect(hdc, &rectBatt2[0], hbrYellow);
						// the mid yellow bar
						FillRect(hdc, &rectBatt2[1], hbrYellow);
						break;
					case BATTERY_FLAG_CRITICAL:
						//draw one bar
						// the right red bar
						FillRect(hdc, &rectBatt2[0], hbrRed);
						break;
					case BATTERY_FLAG_CHARGING:
						//draw flash sign
						//the flash symbol
						hPen = CreatePen(PS_SOLID, 0, colorYellow);
						SelectObject(hdc, hbrYellow);
						SelectObject(hdc, hPen);
						x = sizeof(pointsFlash)/sizeof(POINT);
						Polygon(hdc, pointsFlash, x);
						break;
					case BATTERY_FLAG_NO_BATTERY:
						//draw no bar
					case BATTERY_FLAG_UNKNOWN:
						//draw no bar
					default:
						//draw no bar
						//the batt polygon
						hPen = CreatePen(PS_SOLID, 0, colorUnknown);
						SelectObject(hdc, hbrUnknown);
						SelectObject(hdc, hPen);
						x = sizeof(pointsBatt)/sizeof(POINT);
						Polygon(hdc, pointsBatt, x);
						break;
				};
			EndPaint(hWnd, &ps);
			return 0;

I first searched a way to have a free floating window on top of all others before I realized, that I can use the window handle of the taskbar as the parent window handle of my window.

When you tap and release the stylus on the battery window, you will get a popup menu and there you can exit the application or open the options dialog. The current option dialog only controls the horizontal position of the battery window on the taskbar. Maybe you like to extend the options and enable the user to change colors etc.

The position itself is saved/loaded from the registry. If the rekistry key is missing, it will be created automatically:

void saveReg(DWORD dwVal){
	HKEY hKeyRes;
	DWORD dwDispo=0;
	LONG lRes=0;

	//ensure the regKey exists
	lRes = RegCreateKeyEx(HKEY_LOCAL_MACHINE, szRegKey, 0, NULL, REG_OPTION_NON_VOLATILE, 0, NULL, &hKeyRes, &dwDispo);
	//RegOpenKeyEx(HKEY_LOCAL_MACHINE, szRegKey, 0,0, &hKey);
	if( lRes == ERROR_SUCCESS ){
		//save the value
		DWORD dwType = REG_DWORD;
		lRes = RegSetValueEx(hKeyRes, L"position", 0, dwType, (byte*)&dwVal, sizeof(DWORD));
		if(lRes != ERROR_SUCCESS)
			DEBUGMSG(1, (L"RegSetValueEx failed: lRes=0x%0x\n", lRes));
		RegCloseKey(hKeyRes);
	}
	else{
		DEBUGMSG(1, (L"RegCreateKeyEx failed: lRes=0x%0x\n", lRes));
	}
}

DWORD readReg(){
	DWORD dwVal=80;
	HKEY hKey;
	DWORD dwDispo=0;
	LONG lRes=0;

	//ensure the regKey exists
	lRes = RegOpenKeyEx(HKEY_LOCAL_MACHINE, szRegKey, 0,0, &hKey);
	if( lRes == ERROR_SUCCESS ){
		//load the value
		DWORD dwType = REG_DWORD;
		DWORD dwSize = sizeof(DWORD);
		lRes = RegQueryValueEx (hKey, L"position", 0, &dwType, (byte*) &dwVal, &dwSize);
		if(lRes != ERROR_SUCCESS)
			DEBUGMSG(1, (L"RegGetValueEx failed: lRes=0x%0x\n", lRes));
		RegCloseKey(hKey);
		return dwVal;
	}
	else{
		DEBUGMSG(1, (L"RegOpenKeyEx failed: lRes=0x%0x\n", lRes));
		return 80;
	}
}

The code is written in Embedded Visual C 4, but you can use it in VS2005/2008 too.

If you only like to use the app, here is the ArmV4i executable (WM5/6 compatible)

Bookmark It

3.0        added code for read/write reg for CharTable and KeyTable
 REGEDIT4

 [HKEY_LOCAL_MACHINE\Software\Intermec\KeyToggleChar]
 "KeyTable"=hex:\
 30,31,32,33,34,35,36,37,38,39
 "CharTable"=hex:\
 C6,00,E6,00,D8,00,F8,00,C5,00,E5,00,C2,00,E2,00,C4,00,E4,00
 "LEDid"=dword:00000001
 "autoFallback"=dword:00000000
 "Timeout"=dword:00000003
 "StickyKey"=dword:00000074

 KeyTable     holds a list of 10 keys which will be 'remapped'
 CharTable    holds  list of UniChar codes to use as replacement
 LEDid        defines the ID of the LED to use for showing sticky state
 autoFallback defines, if the sticky state is reset after a mapped key is pressed.
              Use 0, if you dont want the sticky state to fallback after keypress
 TimeOut      defines a timout after which sticky state will be reset

To get the UTF-8 word entries for Unicode chars, see for example here: http://www.unicode.org/charts/charindex.html

The default mapping used here is defined as follows and maps the keys 0-9 to:

0x00c6, 0x00e6, 0x00d8, 0x00f8, 0x00C5, 0x00e5, 0x00c2, 0x00e2, 0x00c4, 0x00e4
 AE        ae     O/       o/      A°      a°      A^      a^      Ä       ä

The main problem I had was finding a function to get the window handle of the current input active window. Finally I found GetForegroundKeyboardTarget() (after some days of searching).

void sendKey(UINT vKey){
 TCHAR* lpText;
 lpText = new TCHAR[MAX_PATH];
 TCHAR strClass[MAX_PATH];
 int iRes;
 
 // hWndActive = GetFocusEx(); // a trick that does not deliver the focusses window
 hWndActive = GetForegroundKeyboardTarget();
 
 GetWindowText(hWndActive, lpText, MAX_PATH-1);
 iRes = GetClassName(hWndActive, strClass, MAX_PATH);
 DEBUGMSG(1, (L"### GetFocusEx() window, handle=0x%x, text='%s', class='%s'\n", hWndActive, lpText, strClass));
 
 if(hWndActive == NULL)
 return; //no window found
 
 WPARAM wParam = getCharForKey(vKey); // charTable[vKey-0x30];// 0x00c6; //AE
 LPARAM lParam = 1; //flags: repeat=1
 
/*
 wm_keydown, wParam=0, lParam=1
 wm_char, wParam=char code, lParam=1
 wm_keyup, wParam=0, lParam=c0000001;
*/
 
 LRESULT lRes=0;
 lRes = SendMessage(hWndActive, WM_KEYDOWN, 0, 1);
 DEBUGMSG(1, (L"SendMessage1 for 'WM_KEYDOWN' returned 0x%0x\n", lRes));
 lRes = SendMessage(hWndActive, WM_CHAR, wParam, 1);
 DEBUGMSG(1, (L"SendMessage2 for '0x%04x' returned 0x%0x\n", wParam, lRes));
 lRes = SendMessage(hWndActive, WM_KEYUP, 0, 0xc0000001);
 DEBUGMSG(1, (L"SendMessage3 for 'WM_KEYUP' returned 0x%0x\n", lRes));
 
 return;
}

I leaved my different finding in code (commented out), so you can see, what functions I have tried.

Bookmark It

I would like to bring back to mind the advantages of using a layout manager (LM) for your smart device .Net projects.

None of the layout Managers mentioned here are developed by me, I dont have the time to do so, especially, when there are great free solutions on the net.

On full .net framework you have the tablelayout panel where you can place your GUI elements and the elements will be arranged automatically, when the main form is resized.
OK, on smart devices, there will be less form resizing as mostly the forms are full screen by design (QVGA:240×320 or VGA:480×640) . But what happens to your form design, when the user rotates the screen from portrait to landscape orientation?

Sample Code:

		public DialogTest2 ()
		{
			int iWidth=240-26;
			int iMaxHeight=320-26-13;
			int iLeft=6;
			this.Size=new Size(240,320);
 
			TextBox textbox = new TextBox();
			textbox.Multiline=true;
			textbox.Size=new Size(iWidth,180);
			textbox.Top=6;
			textbox.Left=iLeft;
 
			Button btnExit=new Button();
			btnExit.Text="Back";
			btnExit.Click+=new EventHandler(btnExit_Click);
			btnExit.Top=iMaxHeight-26;
			btnExit.Left=iLeft;
			btnExit.Size=new Size(iWidth,26);
			btnExit.BackColor=Color.Gray;
 
			Button btnSwitch=new Button();
			btnSwitch.Text="Switch";
			btnSwitch.Click+=new EventHandler(btnSwitch_Click);
			btnSwitch.Top=iMaxHeight-26-26-13;
			btnSwitch.Left=iLeft;
			btnSwitch.Size=new Size(iWidth,26);
			btnSwitch.BackColor=Color.Gray;
 
			this.Controls.Add(textbox);
			this.Controls.Add(btnSwitch);
			this.Controls.Add(btnExit);
 
			this.BackColor=Color.Aqua;
			this.Text="DialogTest";
			this.AutoScroll=true;
		}

To adjust the positions and sizes of these 3 elements, you would need to change top, left and size property of every element in the forms resize event handler.

Positioning and sizing your GUI elements by hand

Now, you write code for the Form Resize event and set top, left and size of all your GUI elements by hand. That are three statements for every element and again for the two possible orientations. Having only four elements on your form, that will be 24 lines to code! Stupid coding, I thought and remembered the LMs available for JAVA. So I tried to find JAVA-like layout managers for DotNet. First, I did not find any, as the full framework already provides a LM and knowone seems to have a need for writing there own LM. But finally I found three LMs in the internet.

What is the purpose of a layout manager

A LM should arrange GUI elements inside a container (a form or dialog) according to some well known layout pattern. Instead of writing many code lines to position and resize your GUI elements, the LM will do this for you and place and size the elements according to the LMs rules. Of course you have to write additional statements to get an automatic layout, but you need only write once and it will work for portrait and landscape orientation.

Coding

All the LMs I found work similar. You have to add a container, possibly specify a layout, and then add your GUI elements to the LM.

The layout of your GUI elements

There are different layouts available:

• Flow Layout – positions GUI elements like written text, from left to right and top to bottom
• Table or Grid Layout – provides a table of cells in rows and columns. Each element is added to the next free cell
• Border Layout – the available space is devided into 5 zones, north, west, center, east and south. You can have one element (also containers with several elements) at each of these locations

The Layout Managers I found

1. Drop-dead easy LM (Link) (no further use here)
2. Grid Layout (Link) (no further use here, uses OnLayout)
3. Single Column FlowLayout (Link)(no further use here)
4. Layout Managers in C# (Link)(great compact framework compatible LM)
5. Simple Layout Managers (Link)(looks also impressive)
6. FlowLayoutPanel for CF (Link)(This was the first one I found in the net)
7. DotNetLayout (SourceForge Link)
8. GridLayout (Link)(easy to use GridLayout Manager, I only miss RowSpan feature)

Be aware of compact framework does not support OnLayout event. So some of the LMs listed above are not comaptible with CF.

Sample Codes

Example 1 using LayoutManagers

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90

using System;
using System.Drawing;
using System.Windows.Forms;
using Layout;
 
namespace PanelTest2
{
    public class DialogTest : Form
    {
        AreaPane panel;
        Button button;
        Button btn_Size;
        TextBox textbox;
        Label label;
        PictureBox picturebox;
        Panel panel1;
        public DialogTest ()
        {
            this.SuspendLayout();
            this.Text="PanelTest";
            panel = new ResizeablePane(this, ClientRectangle, new BorderLayout());
#region north
            Panel panelNorth=new Panel();
            ContainerBox boxNorth=new ContainerBox(panelNorth);
 
            label=new Label(); label.Name="label";
            label.Text="Label:";
            label.Top=26; label.Left=26;
            label.Size=new Size(100, 26);
            boxNorth.LayoutManager=new GridLayout(1,2);
            boxNorth.Add(label);//, BorderLayout.Direction.East);
 
            textbox=new TextBox(); textbox.Name="textbox";
            textbox.Left=150; textbox.Top=26;
            textbox.Size=new Size(100, 26);
            boxNorth.Add(textbox);//, BorderLayout.Direction.West);
#endregion
#region south
            Panel panelSouth=new Panel();
            ContainerBox boxSouth=new ContainerBox(panelSouth);
 
            button= new Button(); button.Name="button";
            button.Location=new Point(26,84);
            button.Size=new Size(52,26);
            button.Text="Exit";
            button.Click+=new System.EventHandler(this.button_clicked);
            btn_Size=new Button(); btn_Size.Name="btn_size";
            btn_Size.Location=new Point(108,84);
            btn_Size.Size=new Size(52,26);
            btn_Size.Text="Resize";
            btn_Size.Click+=new System.EventHandler(this.btn_Size_clicked);
 
            Button btnTest=new Button();
            btnTest.Text="Test";
            btnTest.Click += new EventHandler(btnTest_Click);
 
            boxSouth.LayoutManager=new GridLayout(1,3);
 
            boxSouth.Add(btnTest);
            boxSouth.Add(btn_Size);//, BorderLayout.Direction.East);
 
            boxSouth.Add(button);//, BorderLayout.Direction.East);
#endregion
#region center
            Panel panelCenter=new Panel();
            ContainerBox boxCenter=new ContainerBox(panelCenter);
            boxCenter.LayoutManager=new GridLayout(1,2);
 
            panel1=new Panel(); panel1.Name="panel1";
            panel1.BackColor=Color.Pink;
            panel1.Top=26;
            panel1.Left=180;
            panel1.Size=new Size(26,100);
 
            picturebox=new PictureBox(); picturebox.Name="picturebox";
            picturebox.BackColor=Color.Azure;
            picturebox.ForeColor=Color.Red;
            picturebox.Size=new Size(26,84);
            picturebox.Left=108;
            picturebox.Top=168;
 
            boxCenter.Add(panel1);
            boxCenter.Add(picturebox);
#endregion
            panel.Add(boxSouth, BorderLayout.Direction.South);
            panel.Add(boxNorth, BorderLayout.Direction.North);
            panel.Add(boxCenter, BorderLayout.Direction.Center);
 
            this.ResumeLayout(true);
        }

The code above will produce this automatic layout:

In the above code all GUI elements are created and sized manually. You can also design your Form with the Form designer in Visual Studio and then add the existing elements to the layout panels. But dont use foreach with the form elements, as the order is reversed or unwanted.

The code sample shows how to use a BorderLayout, panels to group elements and nesting layouts using a different LM like a grid layout.

Example 2 using GridLayout

...
using LayoutManagers; // <a class="linkification-ext" title="Linkification: http://dnetgridlayout.sourceforge.net/" href="http://dnetgridlayout.sourceforge.net/">http://dnetgridlayout.sourceforge.net/</a>
    public partial class Form1 : Form
    {
        //PanelWithLayout panelWithLayout1;
        TextBox txtBox1;
        TextBox txtBox2;
        Button btnTest;
        public Form1()
        {
            InitializeComponent();
            //Create a layout with 0 rows and 2 columns
            GridLayout myGridlayout = new GridLayout(0, 2, true);
 
            //the spacing between the elements
            myGridlayout.intHorizontalSpacing = 4;
            //the elements width should fill the column
            myGridlayout.bFillColumn = true;
            myGridlayout.dblRequiredHeight = 26;
 
            panelWithLayout1.AutoScroll = true;
            panelWithLayout1.setLayout(myGridlayout);
 
            txtBox1 = new TextBox();
            txtBox1.Width = panelWithLayout1.Width / 2;
            txtBox1.Height = 26;
 
            txtBox2 = new TextBox();
            txtBox2.Width = panelWithLayout1.Width / 2;
            txtBox2.Height = 26;
 
            //add a text box (in first row, as long as they fit)
            panelWithLayout1.addControl(txtBox1);
 
            //add a button
            btnTest = new Button();
            btnTest.Height = 26;
            btnTest.Width = panelWithLayout1.Width / 2;
            btnTest.Text = "Test";
            panelWithLayout1.addControl(btnTest);
 
            //add the button click event handler
            btnTest.Click += new EventHandler(btnTest_Click);
 
            //add the existing datagrid1 (see designer) spanning 2 columns
            dataGrid1.Width = panelWithLayout1.Width;
            dataGrid1.Height = <a class="linkification-ext" title="Linkification: http://panelWithLayout1.Height/3" href="http://panelWithLayout1.Height/3">panelWithLayout1.Height/3</a>;
            //add another textbox
            panelWithLayout1.addControl(txtBox2);
        }
...

Here are the screens of the gridLayout Test

As you can see, this also easy to use.

Example 3 also using GridLayout

This example uses the already available GUI elements of a form designed with VS2005 visual designer. So, this time you dont have to set sizes or positions, just add the existing GUI elements.

...
        public start_frm()
        {
            InitializeComponent();
            this.SuspendLayout();
            ResizeablePane LayoutPanel = new ResizeablePane(this, ClientRectangle, new GridLayout(0, 1, 4, 4));
            LayoutPanel.Add(btn_barcode);
            LayoutPanel.Add(btn_camera);
            LayoutPanel.Add(btnMSCamera);
            LayoutPanel.Add(btnSignature);
            LayoutPanel.Add(btnExit);
            LayoutPanel.Add(sStatusPos);
            LayoutPanel.Add(status);
            LayoutPanel.Add(lblVersion);
            this.ResumeLayout(true);
...

Here is the portrait screen view of the above code:

Conclusion

If you ever need to design GUIs for mobile devices with automatic portrait and landscape switching consider to use one of the Layout Managers available.

Bookmark It

Extend your keyboard to get function keys by pressing numbers.

With keytoggle you can define a ‘sticky’ key that will change the beahviour of the number keys. As long as the sticky key is ‘active’, number keys from 1 to 0 will produce the function keys F1 to F10. If sticky key is active, the left LED will light in green.

To start keytoggle just tap it or let it start by a link in StartUp. If keytoggle is loaded, you can see a small yellow arrow sign in the taskbar.

If you tap this symbol, you are asked, if you want to unload the app. If the registry does not have values defined for keytoggle, it will use default values. If you try to launch keytoggle a second time you will get a message box. Only one instance of keytoggle can run at a time.

Using the registry you can define the behaviour of the sticky key. In example, you can have the sticky key remain active until it is pressed again, let it ‘go off’ or ‘fallback’ after a period of time or let it fallback after a number key has been pressed. If the sticky key is pressed again, it will always fallback.

REGEDIT4
[HKEYLOCALMACHINE\SOFTWARE\Intermec\KeyToggle]
“Timeout”=dword:00000000
“autoFallback”=dword:00000001
“StickyKey”=dword:00000090

The values:

  StickyKey:
  default "144", decimal (VK_NUMLOCK)
  the VK_ value of the sticky key. Here you define, what you will have as the sticky key. VK_ values are defined as in winuser.h. In example the VK_ value for the NumLock key is 0x90 (144 decimal, see VK_NUMLOCK). The sticky key will be consumed by keytoggle and is not visible to the system any more. That means, that you should choose key value that you do not need in any app you use.

  Timeout:
  default "3", three seconds
  number of seconds the sticky key will remain active. If 0, the sticky key will not fallback automatically by time

  autoFallback:
  default "0"
  if 0 the sticky key will not fallback after pressing a number key. If 1 the sticky key will fallback after a number key is pressed

Launch arguments

  Keytoggle only supports one argument: "-writereg". This will create the registry keys used by keytoggle and fill them with the default values.

Usage Sample Scenario

You have a ITC Color device with a numeric keyboard but no function keys. You need the function keys to be able to use a web application within iBrowse. With keytoggle you can now for example use the Gold key as sticky key. You have to download the device reource kit to get a keyboard remapper tool from the ITC site and assign the Gold key to NUMLOCK.

With this remapping and keytoggle running you will to get a function keystroke simulated if you press the Gold key and then a number key.

Download installer: DOWNLOAD:KeyToggle Installer CAB - (Hits: 24, size: 5.19 kB)

Bookmark It

REGEDIT4
[HKEY_LOCAL_MACHINE\Software\RDP_autologin]
"FitToScreen"="1"
"FullScreen"="1"
"Status"="connecting..."
"Save Password"="1"
"Domain"=""
"Password"="Intermec+2004"
"Username"="rdesktop"
"Computer"="192.168.0.130"
"DesktopWidth"=640
"DesktopHeight"=480
"startOnExit"="\rdp_keepBusy.exe"
"execargs"="noRDPstart"
"UseMouseClick"="0"    //added with version 3 to switch between mouse and keyboard simulation

Some days ago I published my RDP_Autologin code: RDP_Autologin

As there were some screen metrics hardcoded and more and more devices come with a VGA screen the hardcoded QVGA values will not match. So I extended the first version and implemented some additional logic and settings.

First, the emulated screen tap has been adjusted to depend on the device screen width and height. For that I included the HIRES_AWARE resource to get the real screen size.

...
	DWORD dX = (0xFFFF / iScreenWidth) * (80); // changed from 13 to width=240, 1/3=80
	DWORD dY = (0xFFFF / iScreenHeight) * (iScreenHeight - 13);
...
BOOL getScreenSize(){
	int iScreenX = GetSystemMetrics(SM_CXSCREEN);
	int iScreenY = GetSystemMetrics(SM_CYSCREEN);
	DEBUGMSG(1, (L"\ngetScreenSize: x=%i, y=%i\n", iScreenX, iScreenY));
	if(iScreenX>0)
		iScreenWidth=iScreenX;
	if(iScreenY>0)
		iScreenHeight=iScreenY;
	if(iScreenX+iScreenY > 0){
		_itow(iScreenWidth, sScreenWidth, 10);
		_itow(iScreenHeight, sScreenHeight, 10);
		return TRUE;
	}
	else
		return FALSE;
}
...

The rdp file defaults for desktop width and height will also be calculated but you can set the default to use via the registry.

REGEDIT4

[HKEY_LOCAL_MACHINE\SOFTWARE\RDP_autologin]
"startOnExit"="\Windows\rdp_keepbusy.exe"
"DesktopHeight"=dword:000001E0
"DesktopWidth"=dword:00000280
"Computer"="192.168.0.2"
"FitToScreen"="0"
"FullScreen"="0"
"Username"="rdesktop"
"Password"="xxxx"
"Domain"=""
"Save Password"="1"
"Status"="connecting..."

As you can see, I added a line where you can specify an application that will be started at the end of the autoconnect process: startOnExit.

Downloads:

evc4 source code: DOWNLOAD:rdp autologin 2 evc4 source code - (Hits: 91, size: 113.37 KB)

Armv4i executable: DOWNLOAD:rdp autologin 2 Armv4i executable - (Hits: 98, size: 6.47 KB)

NEW: Version 3 adds new option to let the app use MouseClick or Keyboard Simulation

Exe File (ArmV4i) DOWNLOAD:RDP Autologin Version 3 - New version with Mouse or Keyboard simulation switch in registry: UseMouseClick=DWORD 0/1 (Hits: 8, size: 6.56 kB)

eVC4 source code DOWNLOAD:RDP AutoLogin version 3 source code - eVC4 source code for my RDP_AutoLogin version 3 (Hits: 10, size: 131.73 kB)

Bookmark It

This time I added support for Int*rm*c CN50, which does NOT support reading raw GPS data using a serial connection. Instead the CN50 streams the data to the ‘communication’ port.

BTW: did you know how easy it is to create skins for use with MyMobiler?

The stream reader part is outsourced into a thread, which is based on my background thread class bgThread2 (see http://www.hjgode.de/wp/2010/06/01/mobile-development-easy-to-use-background-thread-with-gui-update/). It was very easy to implement the stream reading within the thread, so the GUI is not blocked during read waits.

        #region TheTHREAD
        private void myThreadStart()
        {
            System.Diagnostics.Debug.WriteLine("Entering thread proc");
            int _i=0;
            try
            {
                do
                {
                    //The blocking function...
                    char[] cSep = new char[] { '\r', '\n' };
                    int iRes = 0;
                    do
                    {
                        string sRes = "";
                        iRes = GPS_Sample8.ReadFileClass.readFile(_sCOM, ref sRes);
                        string[] sAll = sRes.Split(cSep);
                        foreach (string s1 in sAll)
                        {
                            SendData(bgWnd.Hwnd, iRes, s1);
                        }
                        //Application.DoEvents();
                        Thread.Sleep(100);
                    } while (iRes == 0);

                    System.Diagnostics.Debug.WriteLine("Thread sleeps...");
                    _i++;
                    Thread.Sleep(100);
                } while (bRunThread);
            }
            catch (ThreadAbortException)
            {
                System.Diagnostics.Debug.WriteLine("Thread will abort");
                bRunThread = false;
            }
            catch (Exception ex)
            {
                System.Diagnostics.Debug.WriteLine("Exception in ThreadStart: " + ex.Message);
            }
            System.Diagnostics.Debug.WriteLine("ThreadProc ended");
        }
        #endregion

The main form code has some new functions to check for iCN50 and start/stop the stream reader background thread:

        #region CN50raw
        bgThread2 myStreamReaderThread;
        private void OpenStream()
        {
            //background thread already running?
            if (myStreamReaderThread == null)
            {
                string szPort="";
                szPort = GetGPSPort();
                if (szPort != "")
                {
                    //start a new thread
                    myStreamReaderThread = new bgThread2(szPort);
                    myStreamReaderThread.bgThread2Event += new bgThread2.bgThread2EventHandler(myStreamReaderThread_bgThread2Event);
                }
                else
                    AddRawText("No raw GPS port found");
            }
        }

        void myStreamReaderThread_bgThread2Event(object sender, bgThread2.BgThreadEventArgs bte)
        {
            AddRawText(bte.sString);
        }
        private void CloseStream()
        {
            if (myStreamReaderThread != null)
            {
                myStreamReaderThread.Dispose();
                Application.DoEvents();
                myStreamReaderThread = null;
            }
            Application.DoEvents();
            mnuRAWStart.Enabled = true;
            mnuRAWStop.Enabled = false;
        }
        #endregion

Downloads (VS2008 solution targeting WM6 SDK):

Bookmark It

        protected override void OnStart(string[] args)
        {
            AddLog("OnStart entered...");
            StartTFTPServer();
...
        }

        protected override void OnStop()
        {
            tftp.StopListener();
            if (TFTPServerThread != null)
                TFTPServerThread.Abort();
            TFTPServerThread = null;
            timerEvents.Stop();
            StopTFTPServer();
            GC.Collect();
 
        }

The StartTFTPServer code uses the same class file ServerSettings.cs as used in the desktop app project (tftpUtilSvcSettings), this is easier to manage:

            try
            {
                AddLog("CreateTFTPFromReg reading path");
                Path = svcSettings.sPath;
                LoggingLevel = svcSettings.LoggingLevel;
                DisplayLevel = svcSettings.SendEventLevel;
                FileAccess = svcSettings.FileAccess;
                AllowOptions = svcSettings.AllowOptions;
                RRQWRQStateCheck = svcSettings.RRQWRQStateCheck;
 
                AddLog("CreateTFTPFromReg reading IPstring");
                string IPstring = svcSettings.ServerIPAddr;
                if (IPstring.ToLower() == "any")
                    ServerIPAddr = System.Net.IPAddress.Any;
                else
                    System.Net.IPAddress.TryParse(IPstring, out ServerIPAddr);
 
                AddLog("CreateTFTPFromReg reading ServerUDPPort");
                ServerUDPPort = Convert.ToInt32(svcSettings.ServerUDPPort);
                iTimeout = Convert.ToInt32(svcSettings.Timeout);
                Resend = Convert.ToInt32(svcSettings.Resend);
 
                AddLog("CreateTFTPFromReg reading LoggingMethodInfo");
                LoggingMethodInfo[0] = svcSettings.LoggingMethod;
                LoggingMethodInfo[1] = svcSettings.LoggingOptions;
 
                //bool ShowAlert = false;
                AddLog("CreateTFTPFromReg reading BlockedIPs");
                BlockedIPs = svcSettings.BlockedIPs;
 
                AddLog("CreateTFTPFromReg starting new TFTPServer(...)");
                if (ServerIPAddr == null)
                    tftp = new TFTPServer(ServerUDPPort, Path, LoggingLevel, DisplayLevel, FileAccess, AllowOptions, RRQWRQStateCheck, Resend, iTimeout, LoggingMethodInfo, BlockedIPs);
                else
                    tftp = new TFTPServer(ServerUDPPort, Path, LoggingLevel, DisplayLevel, FileAccess, AllowOptions, RRQWRQStateCheck, Resend, iTimeout, LoggingMethodInfo, BlockedIPs, ServerIPAddr);
                returnval = true;
                AddLog("CreateTFTPFromReg everything seems to be OK");
            }
            catch (Exception ex)
            {
                AddLog("CreateTFTPFromReg EXCEPTION:" + ex.Message);
            }
            return returnval;

Changing service settings

As I would like to control the service later and change settings, I needed a way to ‘communicate’ with the server. I started with using the registry, but that did not work first. The service was unable to access [HKLM]Software\tftpUtilSvc. I researched and found the idea to use [HKU].Default\Software\tftpUtilSvc. Although the registry location location does not look nice (I mean service parameters have to reside below HKLM, but who cares), I tried the HKEY_USER.Default location and the service was able to load the values from the registry. Did not find any reference, why the service cannot access other root registry trees.

Here is my ugly “tftpUtilSvcSettings” application:

Although the tftpUtil code from sourceforge already contained code to write/read tftp settings to/from registry, it was not very clear and straight. I implemented a ServerSettings class and changed all direct references to the registry to this class. Inside the class, ServerSettings.cs, the values are saved/restored to the registry. As I did not like to change to much of the original tftpUtil class, there may be strange looking contructs and possible duplicate internal values.

During my debugging tests, the service did not start/stop correctly sometimes, due to errors in code. As the service remained in a disabled state, I had to rename all tftpUtilSvc service names to be able to test one more time without having to reboot my PC. If a service gets into this state, the service control manager is unable to remove the service as “InstallUtil.exe /u” was unable to uninstall the ‘broken’ service.

The service installer and controller

The tftp server service should later run on a Windows 2003 server and as I did not like to copy InstallUtil and possibly other files onto the server and then run InstallUtil.exe to install the service, I researched and found some cool CSharp code to install/uninstall/start/stop/restart a servce from code. This code has been integrated into the tftpUtilSvcSettings application (see ServiceUtils.cs) and so there is no real need for an installer. It is possible to just copy the files

nspring.dll
TFTPUtil.dll
tftpUtilSvc.exe
tftpUtilSvcSettings.exe

into one directory and then launch tftpUtilSvcSettings from there (assumed the DotNet runtimes are already installed). Then you can install and start the service directly from the tftpUtilSettings application.

A tftp test client

Included with the source code of tftpUtil of sourceforge is a project called “TFTPUtil Client GUI”. The name made me thing this is a tftp client I could use to test the server. BUT this is not a tftp client, it is just a dialog driven tftpUtil server. To test the tftp server I finally used also code found in internet: tftpClient at CodeProject.com. It is very simple code, no threading, no events, no progress indicator. I added a very simple interface and included the tftpClient class file to be able to do simple tests against the tftpUtil service. The tftpClient project is available as separate download for interested readers.

Downloads (Visual Studio 2008)

Bookmark It

The final solution uses a MessageWindow inside a control based class with a worker thread. As the control and the MessageWindow is part of the GUI thread, there is no need to use Invokes.

Inside the thread I use SendMessage to transfer background thread informations to the control, which then fires an event you can subscribe in the GUI thread.

The test app attached shows three threads running independently and all update the GUI frequently without blocking the GUI.

Here is my try to visualize my idea and solution

Normally you have to use such a construction:

// this variable will hold some text set by the worker thread
public string Message = "";
 
// Create a worker thread and then add items to the ListBox from the
// UI thread
public void DoThreading()
{
    // Create the worker thread and start it
    ThreadStart starter = new ThreadStart(this.UpdateListBox);
    Thread t = new Thread(starter);
    t.Start();
 
    // Loop 4 times, adding a message to the ListBox each time
    for(int i = 0; i < 4; i++);
    {
        this.listBox1.Items.Add("Message from UI thread");
        this.listBox1.Update();
        // Process any queued events on the UI thread
        Application.DoEvents();
        // Suspend processing for 1 second
        Thread.Sleep(1000);
    }
    this.listBox1.Items.Add("Last message from UI thread");
    this.listBox1.Update();
}
 
public void UpdateListBox()
{
    for(int j = 0; j < 5; j++)
    {
        // Set the message to be added to the ListBox from the worker
        // thread
        this.Message = "Worker thread loop count = " + j.ToString();
        // Invoke the WorkerUpdate method in the ListBox’s thread
        // context
        this.listBox1.Invoke(new EventHandler(WorkerUpdate));
        Thread.Sleep(700);
    }
}
// The delegate that’s called from the worker thread
// to update the ListBox
public void WorkerUpdate(object sender, EventArgs e)
{
    this.listBox1.Items.Add(this.Message);
    this.listBox1.Update();
}

As you can see from the code you need to inform the worker thread about the delegate to use for GUI updates. An GUI update from the worker thread is done with the help of UpdateListBox() and that invokes WorkerUpdate and that function finally updates the GUI. So four places are tight together with function names and definitions.

There are also other  ways to do it. For example with declaring a delegate first and then Invoke from background thread (see here):

delegate void setText(string sText);
...
private void ThreadProcSafe()
		{
			this.SetText("This text was set safely.");
		}
...
		private void SetText(string text)
		{
			// InvokeRequired required compares the thread ID of the
			// calling thread to the thread ID of the creating thread.
			// If these threads are different, it returns true.
			if (this.textBox1.InvokeRequired)
			{
				SetTextCallback d = new SetTextCallback(SetText);
				this.Invoke(d, new object[] { text });
			}
			else
			{
				this.textBox1.Text = text;
			}
		}

But this solution also needs your background thread to know the delegate name (SetText(string s)).

I also used this patterns  often, but finally I thought, I need an easy to use class that is independent from the main gui thread code and easy to re-use.

With my bgThread classes (yes there are three predefined ones), you can simply use this construction:

    public partial class ThreadTest2 : Form
    {
        private bgThread _bgThread;
...
        //start and link the thread
        private void btn_bgThreadTest_Click(object sender, EventArgs e)
        {
            _bgThread = new bgThread("192.168.128.5");
            _bgThread.bgThreadEvent += new bgThread.bgThreadEventHandler(_bgThread_bgThreadEvent);
        }
 
        void _bgThread_bgThreadEvent(object sender, bgThread.BgThreadEventArgs bte)
        {
            listBox1.Items.Insert(0, "bgThread: " + bte.iStatus.ToString();
        }
...
        //in your exit function you MUST dispose existing threads
            if (_bgThread != null)
                _bgThread.Dispose();
            Application.Exit();
...

As you can see, you have only a bgThread object and after creating a new instance just add the eventhandler. From the eventhandler you can directly update the GUI. You don’t have to define  a delegate.

All three bgThread classes work fine. They differ in the amount and type of data you have to transfer from the background thread to the GUI and so the BgThreadEventArgs differ in there definition.

The first class, bgThread1 class, only uses the IntPtr’s usable with SendMessage to transfer data. So you can transfer a maximum of two integers or so. That is enough for the most usage scenarios, where you only need to know status codes from the background thread. The attached bgThread1 class only impements one DWORD (or int).

The second class, bgThread class, uses a global lock and can transfer the data you desire. You have to redefine only the bgThreadEventArgs class to include the data types you would like to transfer. This class uses PostMessage and NOT SendMessage. It may happen, that the data and the message will get out of sync as PostMessage works asynchron.

The third class, bgThread2 class, uses WM_COPYDATA to transfer strutured data from the background thread. It is more complicated to use and extend than bgThread1 but uses SendMessage and no lock objects. Your data is delivered with SendMessage and so there will be no problem with data and message syncronization.

All classes define a class based on Component.

    public class bgThread : Component
    {
        public delegate void bgThreadEventHandler(object sender, BgThreadEventArgs bte);
        public event bgThreadEventHandler bgThreadEvent;
...
        internal BgThreadEventArgs _BGeventArgs;
        private Thread myThread;
...
        internal class bgThreadWndProc : MessageWindow
        {
            public bgThreadWndProc(bgThread Parent)
            {
                this._bgThread = Parent;
                hwndControl = Hwnd;
            }
            protected override void WndProc(ref Message m)
            {
                int iMsg = m.Msg;
                System.Diagnostics.Debug.WriteLine("WndProc called...");
                switch (iMsg)
                {
                    case msgID:
                        {
                            this._bgThread.NotifyData(m.WParam);
                            break;
                        }
                    default:
                        {
                            base.WndProc(ref m);
                            break;
                        }
                }
            }
        }//MsgWnd

Within the bgThread class there is a nested class based on MessageWindow.

The bgThread class constructor:

        public bgThread()
        {
            bgWnd = new bgThreadWndProc(this);
            myThread = new Thread(myThreadStart);
            bRunThread = true;
            myThread.Start();
        }

The thread proc itself. It may contain more or less blocking functions  calls.

       private void myThreadStart()
        {
            try
            {
                do
                {
                    //The blocking function...
                    //sample: call ping and return number of pings answered
                    int iReply = myPing.Ping(System.Net.IPAddress.Parse(_sIP));
                    //create a msg and send it to the messagewindow, the messagewindow will then inform the event subscribers
                    //we only need the number of ping replies (0 or 1) for the ping
                    Microsoft.WindowsCE.Forms.Message msg = Message.Create(bgWnd.Hwnd, msgID, new IntPtr(iReply), IntPtr.Zero);
                    MessageWindow.SendMessage(ref msg); //async Message send
                    Thread.Sleep(1000); //if you have fast 'blocking' functions you should sleep
                } while (bRunThread);
            }
            catch (ThreadAbortException)
            {
                System.Diagnostics.Debug.WriteLine("Thread will abort");
                bRunThread = false;
            }
            catch (Exception ex)
            {
                System.Diagnostics.Debug.WriteLine("Exception in ThreadStart: " + ex.Message);
            }
            System.Diagnostics.Debug.WriteLine("ThreadProc ended");
        }

The code that fires the event is not called from the thread. Instead it is called from the MessageWindow. The thread itself sends its ‘data’ via SendMessage to the bgThread ‘Component’ (which is part of the GUI!):

        private void NotifyData(IntPtr i1)
        {
            BgThreadEventArgs _bgThreadEventArgs;
            //is there any subscriber
            if (this.bgThreadEvent == null)
            {
                return;
            }
            try
            {
                int i = i1.ToInt32();
                _bgThreadEventArgs = new BgThreadEventArgs(i);
                this.bgThreadEvent(this, _bgThreadEventArgs);
            }
            catch (MissingMethodException)
            {
            }
        }

The above code snippets are taken from bgThread1 class, the one that only ‘transfers’ an integer to the GUI.

Possible extensions

• extend bgThread to Stop (pause) a bgThread
• extend bgThread and let it invoke a eventhandler that does the blocking stuff. Similar to the BackgroundWorker class available in internet. So you can define your background function inside the main code.

delegate void SetTextCallback(string text);

Bookmark It

The workflow theory is:

1. You scan a barcode or RFID TAG and the data is wedged into the SocketSend input textbox or you type some text into the textbox.
2. You connect your mobile device to the network, where you have a Windows PC running SocketWedge.
3. On the mobile within SocketSend you tap the transmit button and the data is send to SocketWedge.
4. SocketWedge receives the data and puts it in the keyboard message queue of a defined application.
5. The data is typed into the target application.

There are some code snippets you may find usefull for your apps:

On the Mobile (SocketSend)

• A background task that does ping a server and gives status back
• A ping coded for compact framework

On the PC (SocketWedge)

• Background threads that accept socket connections
• A nice keyboard input emulation using SendInput
• A hex encoder/decoder, yo you can use \xHH for hex input of special chars

Attached are the source codes for Visual Studio 2005 targeting Windows or Windows Mobile 6 SDK:

Executables:

Bookmark It

Create windows ce cab files which install a complete directory structure onto a device. I created
this small tool, as I needed an easy to use tool to create a pseudo installation cab.

The DLL will unzip a archiv file with all dirs onto a device. To build a new cab use the directory
structure you will find below cabdir.

Create your own install

First cleanup all files below cabdir\source. Then place all files into this directory as they should be
unpacked onto the device. For example for an application called barcodegame, I created the following
structure below source:

source
	+---Program Files
	¦   +---barcodegame
	¦       +---de
	¦       +---en
	+---Temp
	+---windows
		+---Start Menu
			+---Programs

Then let cabbuild.bat create a new archive with all files and then a cab with this packed file. For example
change to the cabdir (cd cabdir) and then launch

Buildcab.bat "BarcodeGame3"

The batch file will patch an inf file needed for cabwiz and zip all files below source. The batch will also
patch the file SetupDll_zip2.ini.tmpl which is used to specify the install location and the zip filename:

;SetupDll_zip2.ini
;directory values have to end with a \

[unzip]
basedir=\
reboot=0
zipfile=!APPNAME!.zip

The unzip directory could also be controlled by the registry but the ini file is looked up first.

If you like to control the unzip by the registry, you have to change the template file “cabtemplate.inf” and use these
settings:

"HKLM\Software\SetupDll_ZIP2"
	registry entries (all REG_SZ):
	unzipbasedir: "\"
	zipfilename: "install.zip"
	doreboot:"0"|"1"

Currently the batch and everything is adjusted to work with the ini template file “SetupDll_zip2.ini.tmpl”.

Dependencies

• The ini and inf template files are patched using the GNU tool sed.
• The zip file is created with 7za
• The cab is created with the MS SDK tool cabwiz.
• The setupdll is written with the help of iniutil.cpp/h and unzip.cpp/h

Coding techiques

• Windows native API.
• CreateWindow from DLL.
• Read ini files on Windows CE.
• Unzip file on Windwos Mobile.
• Update window from thread.

Advantages and Disadvantages

• No national windows support. No shellfolders
• Very easy to use

Download (source and ready to use): DOWNLOAD:SetupDll_Zip - (Hits: 146, size: 2.4 MB)

Bookmark It