EVPODC Getting Started Part 2 (Configuration)

When you first enter the application, you see the Queries menu (Figure 1). However it’s too early for making queries because we don’t have data yet.

Figure 1

For the working system we need to create users, folders and routes. Then we’ll be able to create documents.

To create users we click Administration, then Users. We see that there are three existing users. They are builtin users. Don’t delete them. Unfortunately, the system allows you to delete anyone. Although the application behavior is unpredictable if you do so.

Say this is a development company and we have a team containing 2 developers, a requirement writer, a tester and a team leader. Please, note that the only field required is Login. You also need to fill in the email address if you want to use notifications. I am going to use them for this sample. So I created 5 new users as you can see in the DropDown list. (figure 2)

Figure 2

Now we need to add the users to different security roles to restrict access because we don’t want all to see everything. We go to Administration / Roles. There are two builtin roles: Administrators and Users. All the new created users are included in the group Users by default. We can leave them as they are and add new roles. I want to create the following roles:

• Requirement Department: Bill Johnson (requirement writer)
• Quality Assurance: John Mayer (tester)
• Developers: Kevin Black (developer), Stephen Smith (developer)
• Managers: Dexter Brown (team leader)

See figure 3

Figure 3

You might be asking: where are the permissions? The answer is that it’s too early for permissions. You confugure them for each folder and document type and we haven’t created neither. That’s what we are going to do now.

Go to Administration / Document Folders

This interface is easy to understand. When you input Folder Name and click Create New Folder, a subfolder is added in the folder that is selected on the left. For our scenario I create a complex folder structure (see figure 4) for the purposes of our sample.

Figure 4

That’s how it looks like in the app (figure 5)

Figure 5

Now you see where permissions are. These are default permissions for Development Team folder. I assume I need to give you a clue about security settings in the app because it’s flexible but not clear at first sight. The security system is similar to the NTFS folder security and other known applications. Each security object, a folder in this case, has security settings for many security roles. In figure 5 we have settings for Administrators and Users because they are builtin roles. Each security role defined in the folder has a set of settings you can see under Users. I clicked href Users. That’s why we are viewing the settings for this role. As you can see Users have all permissions in new folders by default. Note that if a role is not in the role list, it doesn’t have access to the folders at all. So our new roles don’t have any permissions in the folder we look at but don’t forget that our users are members of the Users role. That’s why they can do eveyithing at the moment.

 

Let’s start from configuring the root folder Documents. For that you need to click the link at the very bottom of the page Root folder permissions…

In the root folder our team is going to store annual leave requests and equipment requests. These document types are useful for everybody. Therefore the rule is that everybody can create a document in this folder. People can also read documents of their collegues in the folder because we need them to plan vacations together. We don’t want to give access to equipment requests, but we’ll talk about that later because I am going to configure that in the document type permissions specifically for equipment requests. So the settings for the role Users in the Documents folder will be as in the figure 6.

Figure 6

According to the settings:

members of Users can:

• create new documents (Create checkbox)
• read other users’ documents (Read checkbox) including their own (Read own checkbox)
• update their own documents (Update own)
• see the folder on the list of folders (Browse). However, it is not applicable to the root folder because it’s never on the list of folders. You are always within in.

cannot:

• udate others’ document (Update)
• delete documents (Delete) including their own (Delete own).

The reason why we don’t allow the users to delete their own documents is because we want to guaranty that no annual leave request was deleted by a malicious employee.

It’s time to configure a role that is responsible for these documents and has more permissions. They are Managers. We add the role to the list and set the settings (see figure 7).

Figure 7

These settings say that Managers can do everything in the folder. In fact they update when they accept request or reject it. They might delete document if it’s no longer needed. You got the idea of the folder security configuration.

Next thing we have to do is to create the document types equipment requests and annual leave requests. We need to go to Folders/DocumentTypes and click Create document type. The first tab of the document type form for equipment requests you can see in Figure 8.

Figure 8

Note Prefix and Next number fields. Each document has a unique number. When you create the first document of this type, its number is going to be EQ00001. After that the next document of this type is going to be EQ00002 etc. 

Next we fill in the route tab. 

 

 

 

EVPODC getting started part 1 (installation)

Here I am trying to give you essential information how the open source workflow system EVPO DC works and what you need to do to explore its capabilities. I am sure that is what people do when they assess a new application and decide whether it suits them or not.

After installing the app by using the windows setup, you have a shortcut in you Start Menu. It is a URL called “EVPO Document Circulation”. When you run it for the first time, you see the configuration wizard (figure 1).

Figure 1

The system needs administrative access to a mysql database server to deploy its database. Evpodc like any workflow application needs a database for working. The wizard doesn’t create or modify anything at this step. It just checks the access to the database server.

At step 2 you will be asked the name of the database containing the data for the system. If the database doesn’t exist or “Create new database” is checked, a new database will be created.

Step 3. Two accounts will be created in MySql server for accessing the database. It doesn’t really matter what passwords you input, because you don’t have to remember them. The system will store the passwords in its configuration files and will be using them later.

Step 4. There is a lot of text. The idea is that the app must have permissions to access that directory. Usually the ASP.NET account set by default cannot modify files in directories. This is not the case if you install Evpodc by using setup.exe. The server is run in Utildev cassini. It uses the NT AUTHORITY\SYSTEM account that have all the necessary permissions. So you don’t have to worry about it. Just click next. The directory will be created for you.

Step 5. The system asks you to reveal all your secrets . Actually it needs to know this information because the app notifies people by email when they need to do something. This feature is optional. (see Figure 2)

Don’t forget to check Enable SSL if you use a gmail account.

Step 6. In the next screen the wizard asks you to click the button resetting the application and applying the settings you have just entered. It also says the account name and the password of the administrator. You will need them for your first login. Click reset and that’s it. Setup is finished.

Customize property editor in the VS designer

Sometimes, the VS property grid in the XAML designer (the code name is Cider) is not enough for custom WPF controls. Once, I needed to update a property of my control automatically depending on a value of another property. First, I searched for an example of such an action in WPF controls of the framework. I found nothing similar. The property grid in VS is plain and one property never depends on another. However the property grid in Expression Blend is much more interesting. Therefore, there should be a way to extend the designer. The way starts from the CategoryAttribute and EditorAttribute atributes. Let’s see an example of a custom WPF control that extends Label. The extended control renders Content and adds prefix before the content and suffix after the content. So our control has two additional properties: Prefix and Suffix. The control is for names. Therefore, Prefix can only be “Mr” or “King”. When the user selects Prefix, Suffix gets an initial value depending on the selected Prefix. Then the user can edit Suffix if she is not happy with the value set automatically. Say the prefix “Mr” correspons to the suffix “Junior” and the prefix “King” corresponds to the suffix “III”. So if the user inputs “Arthur” to Content and selects the prefix “King”, the suffix will be set to “III” automatically. See the picture of the control at runtime below:

It looks identically in the designer. The following picture is the part of the property grid containing Content, Prefix and Suffix:

Note, Suffix can be edited after selecting Prefix because the king is not necessary III. He can be I, II or even IV. But if you set the prefix to Mr, the edited value of suffix will be overriden with Junior.

Let’s see the class declaration of the custom control (Control1).

 [Editor(typeof(ExtCategoryEditor), typeof(ExtCategoryEditor))]
    public class Control1 : Label
    {

The Editor attribute says that we provide a editor that can be embedded into the VS property grid for the control. In fact the part of the property grid under “Ext” is a grid that ExtCategoryEditor provides. Our editor is only for the Ext category. In Control1 class we also need to say which properties correspond to “Ext ” category, so that they may get into our editor. We simply mark properties with CategoryAttribute for this:

        [Category("Ext")]
        public string Prefix
        {
            get { return (string)GetValue(PrefixProperty); }
            set { SetValue(PrefixProperty, value); }
        }

You probably think that ExtCategoryEditor is the class that contains the Grid control we see in the property grid. That’s not right. ExtCategoryEditor is a class derived from CategoryEditor that has a set of methods. One of them gives DataTemplate containing the Grid control. Look at the ExtCategoryEditor class:

    internal class ExtCategoryEditor : CategoryEditor
    {
        public override bool ConsumesProperty(PropertyEntry property)
        {
            string[] properties = new string[]
            {
                "Prefix",
                "Suffix",
                "Self"
            };

            if (properties.Contains(property.PropertyName))
            {
                return true;
            }
            else
            {
                return false;
            }
        }

        public override DataTemplate EditorTemplate
        {
            get
            {
                return (DataTemplate)EditorTempaleForExt.Current["ExtDataTemplate"];
            }
        }

        public override object GetImage(System.Windows.Size desiredSize)
        {
            return null;
        }

        public override string TargetCategory
        {
            get { return "Ext"; }
        }
    }

EditorTemplate is the method giving Grid, which is in EditorTemplate. The Editor template is declared in ResourceDictionary EditorTemplateForExt. Its XAML declaration is below:

    <DataTemplate x:Key="ExtDataTemplate">
        <Grid>
            <Grid.ColumnDefinitions>
                <ColumnDefinition/>
                <ColumnDefinition/>
            </Grid.ColumnDefinitions>
            <Grid.RowDefinitions>
                <RowDefinition/>
                <RowDefinition/>
            </Grid.RowDefinitions>
            <TextBlock Grid.Row="0" Grid.Column="0" Text="Prefix"/>
            <ComboBox Grid.Row="0" Grid.Column="1"
                      DisplayMemberPath="Text"
                      SelectedValuePath="Text"
                      SelectedValue="{Binding Prefix, ElementName=uiPropertyCourier}"
                      SelectionChanged="ComboBox_SelectionChanged">
                <ComboBox.Items>
                    <TextBlock Text="King"/>
                    <TextBlock Text="Mr"/>
                </ComboBox.Items>
            </ComboBox>

            <TextBlock Grid.Row="1" Grid.Column="0" Text="Suffix"/>
            <TextBox Name="uiSuffix" Grid.Row="1" Grid.Column="1" Text="{Binding Suffix, ElementName=uiPropertyCourier}" />

            <local:PropertyCourier  x:Name="uiPropertyCourier"
                Visibility="Collapsed"
                Prefix="{Binding [Prefix].PropertyValue.Value}"
                Suffix="{Binding [Suffix].PropertyValue.Value}">

            </local:PropertyCourier>
        </Grid>

    </DataTemplate>

You can see that “Binding [<Property name>].PropertyValue.Value” in Microsoft’s examples. It means that DataContext of your DataTemplate and the elements inside it is an instance of some internal Microsoft class. The class has an Indexer that receives string as a parameter. It actually gets the property name. I won’t write the names of those Microsoft classes because they declared with internal keyword and are not supposed to be used by developers outside Microsoft. However you can see them in the debugger. Just add watch to the DataContext property.

This is ComboBox_SelectionChanged method:

        private void ComboBox_SelectionChanged(object sender, SelectionChangedEventArgs e)
        {
            ComboBox comboBox = (ComboBox)sender;
            PropertyCourier propertyCourier =
                (PropertyCourier)comboBox.FindName("uiPropertyCourier");
            string value = (string)comboBox.SelectedValue;
            switch (value)
            {
                case "King":
                    propertyCourier.Suffix = "III";
                    break;
                case "Mr":
                    propertyCourier.Suffix = "Junior";
                    break;
            }
        }

PropertyCourier is an empty FrameworkElement with two dependency properties: Prefix and Suffix. It is just an intermediate object that pass over values from our controls to the instances of the internal microsoft classes. The point is that you can not update properties in Control1 directly because of two reasons. Firstly, It is not convenient to get access to the Control1 object from the method. Secondly, the values won’t be serialized. Values should pass through internal Cider classes to be serialized. I use PropertyCourier as a convenient way to update values through Cider classes.

For more details download the code.

How to load referenced assemblies from any place you want

This is HowTo for those who need it. The app.config file below defines rules for the main assembly. It sets where the referenced assemblies could be found at runtime. The main assembly references other two assemblies and gets one of them from network (\\someserver\shared\Summator.dll) and another one from the local machine in a subdirectory.

Application config file is below:

<configuration>
  <runtime>
    <assemblyBinding xmlns="urn:schemas-microsoft-com:asm.v1">
      <dependentAssembly>
        <assemblyIdentity
            name="Summator"
            publicKeyToken="03840bb05afb3f17"
            culture="neutral" />
        <codeBase
            version="1.0.0.0"
            href="file:///\\someserver\shared\Summator.dll"/>
      </dependentAssembly>
      <probing privatePath="SUB1\SUB2\SUB3\">
      </probing>
    </assemblyBinding>
  </runtime>
</configuration>

If you reference an assembly from your local machine, you are not limited by subdirectories of the main application because you can use the following method to mount your drives to one of the subdirectories:

http://support.microsoft.com/kb/307889

PS: Configuration files can be conveniently edited by the mscorcfg.msc tool (Control Panel – Administrative Tools – Microsoft .NET Framework 2.0 Configuration, MyComputer, Application, Right click, Add…).

The puzzle about a car and two goats

A collegue of mine gave me the following puzzle to solve:

Suppose you’re on a game show, and you’re given the choice of three doors. Behind one door is a car, behind the others, goats. You pick a door, say number 1, and the host, who knows what’s behind the doors, opens another door, say number 3, which has a goat. He says to you, “Do you want to pick door number 2?” Is it to your advantage to switch your choice of doors?

In the evening I wrote a simple console application that proves the solution practically running the game 10000 times. You can compile it in command line “csc.exe GoatSelector.cs” and execute GoatSelector.exe.

The application shows you validation of the .NET random engine first. There are percentages of 0, 1, 2 numbers got from the Random class executed for 10000 times. Then it shows percentage of success when you “pick door number 2” (switching your choice) and percentage of success when you stay with your first selected door. Download the code

Effect of XML:Space=”preserve”

I was working with XAML created by Expression Blend. At runtime I noticed that a TextBlock moved on the right and one line down from the position it was in the designer. It took me some time to find the cause of the discrepancy. Below there is XAML of two TextBlock elements:

<TextBlock Height="17" VerticalAlignment="Top" xml:space="preserve"><Run xml:space="preserve">first line</Run></TextBlock>
<TextBlock Margin="0,22,0,0" Height="23" VerticalAlignment="Top">
    Second line
</TextBlock>

In the designer it looks like this:

When the application is running, the picture is different:

As you see, the second line moved on the right and one line down.

The only unusual part of XAML is the xml:space=”preserve” attribute, which is set at the first TextBlock and the Run elment inside it. This attribute says that all the indentation and spaces in XAML code should be saved at the time of rendering. But according to MSDN the attribute scope is only the content of the element where the attribute is defined. It’s defined at the first TextBlock and the contained Run element. If you remove the attribute from those elements everything renders correctly. But why does the attribute from the first TextBlock influences the second TextBlock. In the debugger I stopped the application and saved XAML of the “second line” TextBlock and it really had xml:space=”preserve” attribute. I assume it somehow came from the “first line” TextBlock.

It is apparently a bug in WPF. The bug only apears If you have more than one nested elements with the xml:space =”preserve” attribute defined at both. The next sibling element after those will have this attribute set to “preserve” implicitly.

The solution is simple: you need to set xml:space=”default” explicitly at the next element. So our XAML will be:

<TextBlock Height="17" VerticalAlignment="Top" xml:space="preserve"><Run xml:space="preserve">first line</Run></TextBlock>
<TextBlock Margin="0,22,0,0" Height="23" VerticalAlignment="Top" xml:space="default">
    Second line
</TextBlock>

Download the code for this post

regular expressions for XML tags

Regular expressions are powerful and can substitute XML libraries for simple tasks. Say, you need to select all elements with a specific name from an XML file. Below is a sample doing that. The program reads XMLFile11, selects three different elements and prints them and their content to the console.

XMLFile1:

<?xml version="1.0" encoding="utf-8" ?>
<Main>
  <Item Name="item1"/>

  <Item Name ="item2">
    <Components>
      <Component/>
      <Component/>
      <Component Name="component10">
        <SubComponent/>
      </Component>
    </Components>

  </Item>
  <Item>
    <Item></Item>
    <Item></Item>
  </Item>
</Main>

C# code:

    class Program
    {
        public Program(string Xml)
        {
            _xml = Xml;
        }

        string _xml;

        void PrintTags(string tagName)
        {
            string expression =
                @"(<{0}\/>)|" + // gets <tagName/>
                @"(<{0}\s[^>]*?\/>)|" + //<tagName[space]BlaBla.../>
                @"(<{0}>[\s\S]*?<\/{0}\s*>)|" + //<tagName>BlaBla...</tagName>
                @"(<{0}\s[\s\S]*?>[\s\S]*?<\/{0}\s*>)"; //<tagName[space]BlaBla...>BlaBla...</tagName>

            Regex regex = new Regex(String.Format(expression, tagName));
            Match match = regex.Match(_xml);
            do
            {
                Console.WriteLine("tag: {0}", tagName);
                Console.WriteLine(match.Value);
                match = match.NextMatch();
            } while (match.Success);
        }

        void Run()
        {
            PrintTags("Item");
            PrintTags("Component");
            PrintTags("Components");
        }

        static void Main(string[] args)
        {
            Program program = new Program(File.ReadAllText("XMLFile1.xml"));
            program.Run();
            Console.Read();
        }
    }

I had to define expresions for four cases (see comments in the PrintTags method).
The following is the output:

Output:

tag: Item
<Item Name="item1"/>
tag: Item
<Item Name ="item2">
    <Components>
      <Component/>
      <Component/>
      <Component Name="component10">
        <SubComponent/>
      </Component>
    </Components>

  </Item>
tag: Component
<Component/>
tag: Component
<Component/>
tag: Component
<Component Name="component10">
        <SubComponent/>
      </Component>
tag: Components
<Components>
      <Component/>
      <Component/>
      <Component Name="component10">
        <SubComponent/>
      </Component>
    </Components>

However this code won’t work if XML file contains nested elements with identical names. See example below.

XML:

<Item>
    <Item></Item>
    <Item></Item>
</Item>

If you call the PrintTags method with “Item”, you will get <Item><Item></Item>. It happens because the regular expression doesn’t count opened and closed tags.

Code for this post