Here is the situation: You innocently start selecting the tail end of a word and suddenly and automatically the editor selects the whole thing!!!  This drives me nuts.  Maybe it was a long word, and I only wanted the last part.  Similarly, you start selecting a paragraph and Word selects the whole [insert strong expletive] thing.  Is this annoying you? Take a deep cleansing breath and relax -- here is how you permanently disable this nasty, nasty behavior and get back into the driver's seat.

Outlook 2010

File->Options ->Mail ->Compose Messages section-> Editor Options…->Advanced

Word 2010

File->Options->Advanced->Editing Options – will show the same dialog, although the settings for Outlook (which uses Word for editing) are kept separate – you need to make the change in both Outlook and Word if you want consistent behavior.

Tips:

Smart cursoring is tells the editor to keep the insertion point in the same relative place as you scroll using the scrollbar – a useful feature IMHO.

Triple clicking a paragraph selects the entire paragraph – again a useful feature

malware

Unfortunately some of the older posts's pictures are gone as I upgraded to a newer version of WordPress - I will fix that later.

Once everything is configured and you actually connect to the POP3 box, Windows/Outlook gives you an error message telling you:

"The server you are connected to is using a security certificate that could not be verified. A certificate chain processed, but terminated in a root certificate which is not trusted by the trust provider. Do you want to continue using this server?"

So what is really happening? SSL certificates are weird things – the way I understand them is that when you buy a certificate from a Certification Authority like Verisign, the certificate that you get points to a root certificate from Verisign. That Verisign root certificate (and other root certificates from other certification authorities is installed by default by Microsoft. When your computer makes a connection via SSL to the server, it can and does check the server's certificate. Following the chain of certificates to the root, it finds the Verisign Root certificate. Since this certificate matches the root certificate already installed, everything is good.

Why are we getting that self-signed certificate error, then? What happened is that it is possible for people to generate a certificate that is self-signed, meaning it does not link to a trusted Certification Authority. In this case, the root certificate does not match anything on the system, and a warning message is displayed. Hostpapa must have just used self-signed certificates, making things harder for the users.

I have found two different methods to get the certificate, either using OpenSSL or using Firefox 3. I'm sure there are others, but both of these methods work.

Method#1 to get the Certificate: Using OpenSSL

Download OpenSSL for Windows and install it:

Open a command line prompt (start->run->cmd)

At the command prompt (c:\>) type:

cd \program files\gnuwin32\bin

Then type in the following commands:

openssl s_client -connect hp12.hostpapa.com:995 > file.txt

[replace hp12.hostpapa.com with the server name showing in cPanel]

Ctrl+C (to exit the program)

openssl x509 -in file.txt -outform DER -out hostpapa.der

start hostpapa.der

This starts the certificate import wizard.

Click Install Certificate

Click Next

Click Next

Method #2 for getting the SSL Certificate, using only Firefox

You will need the Firefox 3 web browser installed for this to work. Note the number of dialogs and clicks required to get you where you want. You will feel like a criminal doing this but it is perfectly legitimate and safe.

1. Open your Firefox browser, in url box type about:config and press enter.
2. In next screen you will get a warning message click on "I will be careful, I promise!" button.
3. In next screen search for "network.security.ports.banned.override".
4. If this string is not there write click anywhere in the firefox text window and click on new and then select string.
5. You will get a box says " Enter the preference name" in its text box give the name "network.security.ports.banned.override" and press ok.
6. In next screen you have to give the string value. This string value is port number of your port.
7. If you want to allow all port then please give 0-65535. It will allow all the port. If you want to give some specific port then you can give those specific port using comma separated value example 101, 102, 103 etc. Choose only 995 for HostPapa, because POP3 SSL runs on this port
8. Start Firefox and enter the following URL: https://hp12.hostpapa.com:995

Click "Or you can add an exception"

Click Add an exception

Click Get Certificate

Click View, select the Details tab

Click Export

Choose x.509 Certificate, click save, and close all the Firefox dialogs

Double Click the hp12.hostpapa.der certificate, and choose to install the certificate. Method #1 shows the dialogs on Windows Vista that show up when importing the certificate.

Note that you are importing the certificate only for the current user. If you want to import for all users on the machine, you must use the MMC

Start->Run->MMC.exe

File->Add /Remove SnapIn -> Locate Certmgr.msc in c:\windows\system32

Indicate Local machine

Open ConsoleRoot->Certificates(LocalComputer)->Trusted Root Certification Authorities->Certificates

Right-click on Certificates->All Tasks->Import

Follow the wizard to import the saved certificate

The UI was designed in Expression Blend, although I admit that I often went back to editing the XML manually. Maybe this is how you are supposed to use Blend, or I just don't know yet how to do this in Blend. One thing to note is that even though the application is localized, the UI still shows up both in Blend and in the Visual Studio 2008 editor. For the curious, I am using Approach #1 from my CodeProject Article on localization. (You will need to download and install the ResXFileCodeGeneratorEx addin to compile the sample.)

I must admit that I borrowed some ideas from Josh Smith's article on Model View Controller Unit Testing WPF. Let us focus on the commands for now:

public static class Commands
{
    public static readonly RoutedUICommand Button1Command;
    public static readonly RoutedUICommand Button2Command;
 
    static Commands()
    {
        Button1Command = new RoutedUICommand(Strings.Button1, "Button1Command",
            typeof (Commands));
        Button2Command = new RoutedUICommand(Strings.Button2, "Button2Command",
            typeof (Commands));
    }
}

So this is just a static class containing static commands that are initialized as soon as the class is referenced for the first time. Note that the names come from a string resource, and this helps in the localization. Now how do we use the commands?

This code is part of the MainWindow.xaml, right under the Window tag. What it does is tell WPF that we wish to intercept the commands we list under the <Window.CommandBindings> tag. Now the command itself comes from the static class I just showed. Note the use of the 'DemoApplication' namespace identifier; for this to work we have to define/import the namespace which is done as an attribute of the Window tag like so:

xmlns:DemoApplication="clr-namespace:DemoApplication;assembly="

The command also has two attributes, the CanExecute and Executed, which point to code-behind found in the MainWindow.cs. The CanExecute determines if the command is available, and the Executed delegate is fired when the command is invoked. Here is what the implementation looks like:

private void Button1_CanExecute(object sender, CanExecuteRoutedEventArgs e)
{
    e.CanExecute = true;
}
 
private void Button1_Executed(object sender, ExecutedRoutedEventArgs e)
{
    MessageBox.Show(Strings.TextIs + mViewModel.EditBoxContents, Strings.ButtonPressed);
}
 
private void Button2_CanExecute(object sender, CanExecuteRoutedEventArgs e)
{
    e.CanExecute = true;
}
 
private void Button2_Executed(object sender, ExecutedRoutedEventArgs e)
{
    mViewModel.EditBoxContents = Strings.ResetText;
}

Now we have the commands defined and created and hooked up the delegates that actually handle the commands (and if the command is available), it is time to hook up the command to both to the buttons and to the menus.

 
<Menu DockPanel.Dock="Top">
  <MenuItem Header="{x:Static Properties:Strings.TopLevelMenu}" >
    <MenuItem Header="{x:Static Properties:Strings.Button1}"
              Command="{x:Static DemoApplication:Commands.Button1Command}" >
      <MenuItem.Icon>
        <Image Source="Icons/control_play.png"/>
      </MenuItem.Icon>
    </MenuItem>
    <MenuItem Header="{x:Static Properties:Strings.Button2}"
              Command="{x:Static DemoApplication:Commands.Button2Command}" >
      <MenuItem.Icon>
        <Image Source="Icons/control_stop.png"/>
      </MenuItem.Icon>
    </MenuItem>
  </MenuItem>
</Menu>
 
<Button Content="{x:Static Properties:Strings.Button1}"
        Command="{x:Static DemoApplication:Commands.Button1Command}"
        Grid.Row="1"
        VerticalAlignment="Top" />
<Button Content="{x:Static Properties:Strings.Button2}"
        Command="{x:Static DemoApplication:Commands.Button2Command}"
        Grid.Column="1"
        Grid.Row="1"
        VerticalAlignment="Top" />
 

The localized app in French looks as follows:


The English version is here:


Enough for now - good night!

 
pre
{
	border: 1px solid #ccc; padding: 0em 0.5em;
	background-color: #FBEDBB;
	padding: 5pt;
	font: 9pt "Courier New", Courier, mono;
	white-space: pre;
	overflow:auto;
}
 

The following is an example of what C# code looks like:

 
public partial class Window1
{
    private readonly ApplicationViewModel mViewModel;
 
    public Window1()
    {
        InitializeComponent();
        mViewModel = new ApplicationViewModel();
        DataContext = mViewModel;
    }
 
    private void Button1_CanExecute(object sender, CanExecuteRoutedEventArgs e)
    {
        e.CanExecute = true;
    }
 
    private void Button1_Executed(object sender, ExecutedRoutedEventArgs e)
    {
        MessageBox.Show("Text is: " + mViewModel.EditBoxContents, "Button Pressed");
    }
 
    private void Button2_CanExecute(object sender, CanExecuteRoutedEventArgs e)
    {
        e.CanExecute = true;
    }
 
    private void Button2_Executed(object sender, ExecutedRoutedEventArgs e)
    {
        mViewModel.EditBoxContents = "Reset";
    }
}
 

This is what XAML code looks like:

 
<Application x:Class="DemoApplication.App"
    xmlns="http://schemas.microsoft.com/winfx/2006/xaml/presentation"
    xmlns:x="http://schemas.microsoft.com/winfx/2006/xaml"
    StartupUri="Window1.xaml">
    <Application.Resources>
    </Application.Resources>
</Application>
 

Normally I use Word 2007 to write the blog entries, it is very convenient. However, you can't use it to post code, as it does not make the round trip (loading, editing, saving) without royally screwing up.

All this leads me to think about Unit Testing and WPF, and the question: Is unit testing worthwhile for UI? And at what point in time (or complexity) does it make sense to use the MVC or Model View-View-Model-Model pattern?

What do you think? – leave a comment.

Creating the Patch

1. Do a SVN Update to get the latest bits from the trunk.
2. Go to the top-level folder in your project tree in Windows Explorer, right click and choose Create patch.

   

3. The Create Patch dialog comes up – up, listing (a) items you have made changes are checked, (b) items you have on your local system, but that are not in the repository are unchecked.

   
   

4. Any files with the checkmark will go into the patch (actually only the diffs go into the patch file). Click OK, and you are prompted to save.
   
5. Congratulations, you have created your patch!

    

Applying the patch

1. Go to the machine you wish to apply the patch to, and do a SVN update. Also do a SVN Revert, to minimize chances of conflicts later on. Now your machine's sandbox should match the repository.
2. Copy the patch file from removable media or from the network to your local machine.
3. Choose SVN Apply patch.
   
4. Locate and open your .patch or .diff file:
   
5. Tortoise will launch its merge program (you can configure what program to launch under tortoise options, unified diff viewer), and the file patch window. Right click anywhere in the "File Patches" window, but not on any specific file, and choose patch all.
   
6. As soon as you choose this command, Tortoise will silently patch your files, and you're done. To verify its work, you can attempt to do a commit and see what the changes are.

    

Happy patching!

I use the following image as the desktop background when I log in. I like the blue sky tones suggesting freedom, but when you open the window, there is darkness. Sometimes things are deceptive…

If you would like to have this picture as your background on Windows XP when you log in [it only changes the background while the login dialog is present!], please follow the following steps: [Note: I'm pretty sure this only works if you get a login dialog, not the login screen where you click icons to log in]. If you wish to modify such an icon-based (Fast User Switching) login screen, it is possible to use another tool such as Logon Studio. If you get a logon dialog, the following instructions apply:

1. Download the Magritte.bmp file from here. [You may need to right click on the link and save, see step 2]
2. Save the file in your Windows directory (usually c:\windows), you must have administrative privileges to do this.
3. Download installwallpaper.reg to your desktop (save it there) and double click it. The system will ask if you want to merge the contents to the registry. Click 'yes' to accept. [Verify the contents with a text editor, if you are paranoid, before clicking 'yes']. This will add the image to your background of the login screen, and change the background colour from blue to black. Delete the .reg file from your desktop.
4. To see the image, log off.

Then the city moved to Telephone registration.  You got a family number, and child number, course number and you did everything by phone, provided you could get past the busy tone.

Then came the internet, and signing up for courses became much easier.  Go to your computer at 7:00am, pick the courses and voila. 

But that's not fair - people without internet access can't sign up as easily and are treated as second class citizens.  Thats why the city of Toronto came up with a Maximum number of Sessions for their website.  A means to throttle the users into more manageable chunks while being more fair.  You go to the site http://efun.toronto.ca/torontofun and, if you are exceeding the number, you get a polite message about how the number of sessions was exceeded, and the reason thereof.

On the other hand, we have a busy life, and need to get our kids to school, finishing up with the registration.  So I quickly (<5min) wrote the script below using ruby  ruby182-15.exe (object oriented scripting language) and Watir watir-1.4.1.exe (a framework for testing web applications using ruby). Once those two things are installed, load the following into your favourite editor...

require 'watir'
test_site = 'http://efun.toronto.ca/torontofun'
ie = Watir::IE.new
ie.goto(test_site)
while ie.contains_text("Maximum Sessions")
    ie.refresh()
end

Even if you're not a programmer, this is pretty clean and easy to understand. Saved it as efun.rb, and double clicked.  Internet explorer went to the site, and did the equivalent of hitting the refresh button at the rate of 3 to 5 refresh cycles per second, flooding the server with requests.  In less than 5 minutes, we had the normal start screen showing. The script had done its job.  The little hacker in me rejoiced and I'm actually proud of this little script.

Now is this ethical and fair?  I'm not sure. We could have been clicking refresh 2x per second manually to try to get in; I just automated the whole thing.  In any case I posted this if you were curious...

Therefore, as God's chosen people, holy and dearly loved, clothe yourselves with compassion, kindness, humility, gentleness and patience. Col 3,12