Sunday, May 31, 2009

Programmatically change Web project settings from dynamic port to static

If you have a solution consisting of many Web applications that currently use dynamic ports for debugging and now you want to modify them to be static, it will be tedious and extremely time-consuming to go in each project and change it. I developed an application to resolve this. For the future sake, I create a custom project template so that I don't have to go in the project properties and change it whenever a new Web application is created. Indeed, with my own project template, I can customize other things like libraries, references and macros I always use in a project.

Currently the application supports VS.2008. It takes the solution file (.sln) as an input to locate the project locations, and then goes in each indiviual project file (e.g., .csproj and .vbproj) to alter the setting. It only targets Web application including WCF. but not Website because the Web project properties for a Website is stored in .sln file.

The idea behind this application is very simple.

Define the settings

First, I create a setting file called vs.xml.   Currently the port is default to 2332.   The text in light green background is similar to the text you may find in your Web project file.

<?xml version="1.0" encoding="utf-8" ?>
<VisualStudio target="Visual Studio 2008 ">
  <SolutionFile version="10.00" desc="Microsoft Visual Studio Solution File, Format Version">
    <Languages>
      <Langauge enabled="true">
        <Type>CSharp</Type>
        <GUID>{FAE04EC0-301F-11d3-BF4B-00C04F79EFBC}</GUID>
        <extension>.csproj</extension>
      </Langauge>
      <Lanaguage enabled="true">
        <Type>VisualBasic</Type>
        <GUID>{F184B08F-C81C-45f6-A57F-5ABD9991F28F}</GUID>
        <extension>.vbproj</extension>
      </Lanaguage>
    </Languages>
  </SolutionFile>
  <ProjectFile desc="Microsoft Visual Studio Project File">
    <Web> 

<!-- Begin Settings --> <ProjectExtensions> <VisualStudio> <FlavorProperties GUID="{349c5851-65df-11da-9384-00065b846f21}"> <!-- Define Web Project Settings --> <WebProjectProperties> <UseIIS>False</UseIIS> <AutoAssignPort>False</AutoAssignPort> <DevelopmentServerPort>2332</DevelopmentServerPort> <DevelopmentServerVPath>/</DevelopmentServerVPath> <IISUrl> </IISUrl> <NTLMAuthentication>False</NTLMAuthentication> <UseCustomServer>False</UseCustomServer> <CustomServerUrl> </CustomServerUrl> <SaveServerSettingsInUserFile>False</SaveServerSettingsInUserFile> </WebProjectProperties> <!-- End of Web Project Settings --> </FlavorProperties> </VisualStudio> </ProjectExtensions> <!-- End of Settings -->

</Web> </ProjectFile> </VisualStudio>

The above vs.xml can be altered to support VS.2005. But I am going to describe the details here.

Define data structure to contain the settings

Create a data structure to contain all the Web properties; they are:

KeyValue
UseIISFalse
AutoAssignPortFalse
DevelopmentServerPort2332
DevelopmentServerVPath/
IISUrl
NTLMAuthenticationFalse
UseCustomServerFalse
CustomServerUrl
SaveServerSettingsInUserFile

I use a Dictionary to capture this.

    private Dictionary<String, String> props;
When the application reads vs.xml, it will automatically load all the settings into this props Dictionary.

Function to alter the settings

The function SaveTo as shown below will update the settings into the project file based on the settings in props.   Note that my current version is slightly different now.

    /// <summary>
    /// Saves the settings to a specific file.  
    /// The project file must have ProjectExtensions/VisualStudio/FlavorProperties/WebProjectProperties
    /// nodes in order to be qualified for process.
    /// If the required elements under WebProjectProeperties node are missing from the specific document, 
    /// it will insert them.  
    /// The original elments in the document under that node remain intact if they don't match.
    /// </summary>
    /// <param name="fileName">the exact full filename of the project file for save</param>
    public void SaveTo(String fileName) {
      String namespaceUri = null, prefix = null;
      bool hasNamespace = this.HasNamespace(fileName, out namespaceUri, out prefix);

      XmlDocument xmlDoc = new XmlDocument();
      xmlDoc.Load(fileName);
 
      XPathNavigator navigator = xmlDoc.CreateNavigator();
      XmlNamespaceManager namespaceManager = new XmlNamespaceManager(navigator.NameTable);
      if (hasNamespace) {
        SetupNamespaceEnv(prefix);
        namespaceManager.AddNamespace(prefix, namespaceUri);
      }

      XPathNavigator result = null;

      if (navigator.HasChildren) 
        result = navigator.SelectSingleNode(GetTargetNodeXPathExpression(), namespaceManager);
      else 
        throw new XmlException("Invalid file format, or it is not a Web project file.");

      if (result == null)
        throw new XmlException("It is not a Web project file.");
      

// up to here, "result" should be at node "WebProjectProperties" Dictionary dup = this.CloneProperties(); if (result.MoveToFirstChild()) { do { result.SetValue(dup[result.Name]); dup.Remove(result.Name); } while (result.MoveToNext()); this.changed = false; // all properties updated. if (dup.Count > 0) { // Missing properties we need to save, do INSERT to file now...; result.MoveToParent(); // position back to "WebProjectProperties" node this.AppendChild(ref result, dup); } } else { // no child found at "WebProjectProperties" node this.AppendChild(ref result, dup); // add a child nod with all the needy settings } xmlDoc.Save(fileName); // Now is time to save to disk.

if (hasNamespace) { this.ReverseNamespaceEnv(); } }

The main purpose of this function does is to update the project settings:
  • First, it locates the <WebProjectProperties> node to begin the process.
      <ProjectExtensions>
        <VisualStudio>
          <FlavorProperties GUID="{349c5851-65df-11da-9384-00065b846f21}">            
            <WebProjectProperties>
            ...
            </WebProjectProperties>
          </FlavorProperties>
        </VisualStudio>
      </ProjectExtensions>
    
  • When the node is found, the function will temper the settings accordingly. All settings are stored inside the <WebProjectProperties> node. However,
    • If there is empty inside the <WebProjectProperties> node, all settings will be appended into it.
    • According to props, any missing setting(s) will be added accordingly.
    • Any extra settings cannot be found in props will remain intact.

Describe basic project file structure in Solution (.sln)

If you use a text editor to open .sln file and view it, you will find that each project may contain the following structure.

 Project("language_GUID") = "project_name", "project_location_or_file", "project_GUID"
   ProjectSection(...) = ...
   ...
   EndProjectSection
 EndProrject
A regular Windows or Web project will not contain ProjectSection. But if a Web project is type of Website, ProjectSection is used for storing WebsiteProperties because it doesn't own a project file. A WCF client may use WebsiteProperties for its ProjectDependencies. In sum, most projects only contain two lines: The first Project line and the EndProrject line. For example,
Project("{FAE04EC0-301F-11D3-BF4B-00C04F79EFBC}") = "MyTest", "Test\MyTest.csproj", "{7C0275FE-A91D-42E9-9EF1-5CCBC9D54C3C}"
EndProrject

Put everything together

All project locations registered with the solution are relative to the solution. Thus the location Test\MyTest.csproj in the above example is a relative path to its solution. As soon as the project location is found, convert it to the absolute path and then feed it to the function SaveTo, SaveTo will update the settings automatically.

In order to accomplish this, I create a Collection to contain all the project data found in the solution file.

    private Collection<ProjectIdentifier> projectCollection;
Each project data is defined in the following class:
  public class ProjectIdentifier {
    private String langGuid = null;
    private String name = null;
    private String location = null;
    private String guid = null;

    public ProjectIdentifier(String langGuid, String name, String location, String guid) {
      this.langGuid = langGuid;
      this.name = name;
      this.location = location;
      this.guid = guid;
    } 
    ...
  }

When everything is ready, loop through the projectCollection to update the project files:

  foreach (ProjectIdentifier project in c) {
    ...
    if (System.IO.File.Exists(project.Location)) {
      try {            
        ...
        if (project.IsWebProject()) {
         ...
            this.Backup(project.Location, this.txtTag.Text.Trim());

          propsBuilder.SaveTo(project.Location);
        ...
      }
      catch (Exception ex) {
        String err = String.Format("Error occurs: {0}{1}{2}",
                                   ex.Message, Environment.NewLine, ex.StackTrace);
        ...
      }
    }
    else { // file not found or is not a file.
      ...
    }
  } // foreach

Request my application

Before execution, my application will first back up the project file by default while it traverses the project folder. You can override this setting if you don't like it. The backup files can be programmatically removed in a later time when you feel comfortable with the changes. I don't know how useful for you but it certainly helps me to resolve a lot of tedious tasks.   Please feel free to leave me a message via the comment box for requesting this application if you think that it will be helpful to your issue.   Don't forget to leave your email for response. All application requests will NOT be left on comment.

Update August 6, 2009
On or off there are people requesting it. Here are the links where I put for storage so I am no longer to manage and distribute them myself. Please feel free to download it yourself. The code has been changed since this post.
Download: [ Source ] [ Binary ]

Note that: if you request it, then you use it as your own risk. I am not responsible for any damages that may occur.

Update July 15, 2011
Sorry, this application is no longer to download. It is hard for me to ensure that the files are still alive.

Additional notes for exporting project template


Template Directory Setting

The default location on XP for VS.2008 for template storage is at
      %USERPROFILE%\My Documents\Visual Studio 2008\Templates

You can change this setting at the VS.2008 menu bar: Tools --> Options.
Projects and Solutions

Export Template and Location

Create a blank project, edit it, import and reference your vital libraries.  When you're done, go File menu to choose Export Template.... Follow the wizard to complete the rest.

A project template is always exported to the default output folder where you cannot change:
      %USERPROFILE%\My Documents\Visual Studio 2008\My Exported Templates

The location you set at Projects and Solutions is primarily for the IDE to locate and load your custom templates.

If you accept the default setting Automatically import the template into Visual Studio when you export the template, IDE will automatically save a copy under that template folder where Projects and Solutions specifies.

Additional Template Folders

Project Templates Location
For the first time project template creation, three (3) additional folders will be automatically created for your template organization:
  • Visual Basic
  • Visual C#
  • Visual Web Developer
And your template file will stay with them at the same directory level. In my example, it is MyC#WebApplication.zip.

I won't bother to put the template into one of those folders. You can drag the zip file into them if you want to organize it. As long as the template is under the template folder, the IDE is smart enough to loop through its sub-folders and find your template.

Edit Template Project Description

If you want to display a descriptive name on the IDE, go to the designated project templates folder and then do the editing. In my case, it is locate at H:\dev\Visual Studio 2008\Templates\ProjectTemplates.
  • Extract MyTemplate.vstemplate from the zip file and then open it for edit.
  • Locate <Name> and give descriptive name, e.g., My ASP.NET Web Application in C#
    <VSTemplate Version="2.0.0" xmlns="http://schemas.microsoft.com/developer/vstemplate/2005" Type="Project">
      <TemplateData>
        <Name>My ASP.NET Web Application in C#</Name>
        <Description>A C# Web Application with a specific port (.NET Framework 3.5)</Description>
        <ProjectType>CSharp</ProjectType>
        ...
    
  • Save it and then zip it back to the original zip file.
Now try to add a new project, you will see the new template shown under My Templates. Add New Template

Wednesday, May 27, 2009

Problem of copying big file within a VM from the host

Have you ever experienced the following error?

Cannot copy <file_name>: The specified network name is no longer available.

Regardless of which Microsoft Windows, this problem occurs whenever you use Bridged mode on VMware network. Everything works fine with Host-only or NAT.

On VMware Server 1, I can disable the automatic bridging and then reconfigure other virtual adapter (not vmnet1 or vmnet8) as Bridged network by binding it to a physical adapter. This problem will go away. But this technique doesn't work for VMware Server 2. Thus, the problem still remains unresolved. What I can do is to copy files within the host to the VM, instead of inside the VM and do copying.

I don't believe that this post is correct. VMnet1 is Host-only. Indeed, I am not sure how we can change the nature of vmnet1 and vmnet8 to Bridged. Since the author of that post was using Host-only on vmnet1, certainly copying big file problem would not be an issue.

The other thing I would like to mention is that any newly added virtual adapter is always Host-only if the VMware Server is running on Windows. I am not sure if it holds true for other platforms.

Monday, May 25, 2009

Bridged vs Host-only vs NAT

I have been using VMware since 2002. I mostly use it for my development testing. There was no free version from VMware at that time. Luckily I got exposed to both WorkStation and GSX at work. Now I can have my free virtual server installed at home. All my VMware servers are running on Microsoft platform. I have both VMware Server 1 and 2 running on different machines. I like VMware Server 1 more than Server 2 because of its simple interface (which is VMware Server Console), a bit light weight to me and no restart needed when a configuration is changed.

Most my friends are confused by the 3 different modes network setting: Bridged, Host-only and NAT. Like most people, they use the default setting, Bridged. In the following, I am going to briefly distinguish the difference among these 3 network modes. I would also point out some issues existing in each mode, from a user point of view. Those issues applies to the free versions both VMware Server 1 and 2 running on Microsoft Windows platforms. For detail of how these 3 modes work or are configured, please consult VMware documentation ( Server 1 [Install, Manual] | Server 2 User Guide ).

Each VMware network mode is associated with a named virtual adapter. You can add and remove any virtual adapter you want. On Windows, the newly added virtual adapter is only bound to Host-only network. It may not be the case if you are running the server on other platforms but I don't know. In addition, you can configure a custom adapter to assoicate with each physical network adapter to fit your need.

These three basic VMware network are listed as follows.

Mode Virtual Adapter
Bridged : vmnet0 this is the default.
Host-only : vmnet1
NAT : vmnet8

Note that the issues listed in the following table are only applied to Microsoft Windows. I am not sure if they hold true for other platforms.

Bridged (vmnet0)

Communication and Visibility: The virtual machine (VM) acts as if a regular PC ran on the same physical network. You can go Internet and do whatever you want unless your software is hardware dependent. All VMs on vmnet0 are visible to others on the same physical network and vice versa.

Performance: In my experience, the VMware Bridged network is kind of slow. On Microsoft Windows, it is much slower on a workgroup network, compared to a domain.

DHCP or Additional Service: No DHCP service is provided. If you don't have a router or DHCP server running on your network, this mode may not suit for you. In this case, you should consider either Host-only or NAT network depending on if you want Internet access.

Issues:

  • Big File Copy: Regardless of VMware Server 1 or 2, the file copying from the host inside a VM will fail when the file size is large enough. While the file is being copied from the network, the following error may occur: "Cannot copy <file_name>: The specified network name is no longer available." (Also see this)
  • Leaking IPs to physical network: The VM's IP is assigned by the physical network (e.g. your router or DHCP server). Thus, this problem doesn't apply here.

Host-only (vmnet1)

Communication and Visibility:The network communication of this mode is limited between the host and the VMs on the same vmnet1 adapter. Those VMs cannot make a connection beyond the host; thus there is no internet access capability. They remain hidden behind the host. You cannot configure it to expose any VM on vmnet1 to the physical network either.

Performance: The performance of Host-only (or vmnet1) is very fast. I usually use this setting for software testing especially when I don't need an Internet access. For example, I could run a database server inside a VM for a simple test.

DHCP or Additional Service: This mode provides DHCP sevice. If you don't have a router or a DHCP sever running on your network, this one comes in handy. This service can be disabled.

Issues:

  • Big File Copy: Unlike Bridged, I don't experience any file copy problem regardless of size (also see this).
  • Leaking IPs to physical network: VWware DHCP may assign IPs to the PCs on the physical network and cause problems. I encounter this problem when I have my VM Server running first and then my other physical PC joins the physical network later.

NAT (vmnet8)

Communication and Visibility: All VMs using the adapter vmnet8 form a private network with the host. They all have internet access capability, but they are not visible to others beyond the host as if they are sitting behind their own firewall with the host. You can expose the VM for a particular access by port forwarding via NAT configuration provided by the VMware network utility (in the desktop Start menu: VMware -> VMware Server -> Manage Virtual Networks).

Because of its internet capability, I use it often for Web application testing and development too.

Performance: To me, it is faster than Bridged and could be as fast as Host-only.

DHCP or Additional Service: This mode provides DHCP sevice. If you don't have a router or a DHCP sever running on your network, this one comes in handy. This mode also provides NAT service for further network configuration. Both services can be disabled.

Issues:

  • Big File Copy: Unlike Bridged, I don't experience any file copy problem regardless of size (also see this).
  • Leaking IPs to physical network: VWware DHCP may assign IPs to the PCs on the physical network and cause problems. I encounter this problem when I have my VM Server running first and then my other physical PC joins the physical network later.

Disabling VMware DHCP and NAT on Windows

If you are only using Bridged network, I would recommend to have both VMware DHCP and NAT services disabled. They are not used by Bridged.
On Windows, you can disable it via your Services console (services.msc) from Administrative Tools. These two services on Windows are called VMnetDHCP (for VMware DHCP Service) and VMWare NAT Nervice respectively.
  • Locate the service you want to disable.
  • Right click to select Properties.
  • From the Startup type dropdown box, select Disabled.
  • Click OK button.

Or run the following commands at the command prompt.
   sc config "VMnetDHCP" start= disabled 
   sc config "VMware NAT Service" start= disabled
Note that
  • start= must be in one single word.
  • There must be a space between the equal sign (=) and the word disabled.

Sunday, May 24, 2009

Apache ignores DocumentRoot in VirtualHost

Lasts night I got stuck in configuring a virtual host on Apache 2.2. Regardless of which port I set (including the port 80), the Apache server ignored my DocumentRoot setting for the virtual host. Until I read this post, it reminded me to check all wildcard settings. The problem I had was slightly different from the one in the post.

In that post, the problem was from the wildcard setting on the ServerName. For me, I was missing a port number next to the wildcard when I was setting it at the VirtualHost directive for localhost. Everything works fine after the port number had been added.

Please don't make the same mistake as I did. Even you are using port 80, you should always put the port number next to it. Otherwise, your virtual DocumentRoot will NOT be recognized.

The followings are my virtual host settings in httpd-vhost.conf.

Incorrect Correct
NameVirtualHost *:9090

# Missing port number to localhost or the default site
<VirtualHost *>
  DocumentRoot "C:\servers\Apache Software Foundation\Apache2.2\htdocs"
  ServerName localhost
</VirtualHost>

<Directory "h:/vhosts">
  Order Deny,Allow
  Allow from all
  Options Indexes FollowSymLinks
</Directory>

<VirtualHost *:9090>
   # Note: no wildcard supported in ServerName
   ServerName examples.local
   ErrorLog "logs/examples.local-error.log"
   CustomLog "logs/examples.local-access.log" common
   DocumentRoot "h:/vhosts/examples"
</virtualHost>
NameVirtualHost *:9090

# Port number is given to the localhost or the default site
<VirtualHost *:9090>
  DocumentRoot "C:\servers\Apache Software Foundation\Apache2.2\htdocs"
  ServerName localhost
</VirtualHost>

<Directory "h:/vhosts">
  Order Deny,Allow
  Allow from all
  Options Indexes FollowSymLinks
</Directory>

<VirtualHost *:9090>
  # Note: no wildcard supported in ServerName
  ServerName examples.local
  ErrorLog "logs/examples.local-error.log"
  CustomLog "logs/examples.local-access.log" common
  DocumentRoot "h:/vhosts/examples"
</virtualHost>

Tuesday, May 19, 2009

SOAP and REST-based WCF basic settings in Web.config

This only serves my own notes because I always forget it when I switch one type of Web service to another.
SOAPREST
<system.serviceModel>
   <services>
     <service
            behaviorConfiguration="ServiceBehavior">
       <endpoint
                address=""
                binding="wsHttpBinding"
                contract=...>
       ...
       </endpoint>
       ...
     </service>
   </services>
...
</system.serviceModel>
<system.serviceModel>
   <services>
     <service
            behaviorConfiguration="ServiceBehavior">
       <endpoint
              address=""
              binding="webHttpBinding"
              behaviorConfiguration="RESTBehavior"
              contract=...>
       ...
       </endpoint>
       ...
     </service>
  </services>
  ...
  <behaviors>
    ...
     <endpointBehaviors>
         <behavior name="RESTBehavior">
             <webHttp/>
             ...
       </behavior>
     </endpointBehaviors>
  </behaviors>
  ...
</system.serviceModel>

Friday, May 15, 2009

Blogger cannot display the image

Please wait, loaing rat_cheese.gif...
Using the images stored with blogger could be tricky. There is no doubt that the image can be displayed correctly if you use exactly the code provided by blogger.
<a onblur="try {parent.deselectBloggerImageGracefully();} catch(e) {}"
     href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEh0dTd1RVdaar9KTLCks9J022j3MxwYgjx7fIjXCTaj1COwHcxCAEge4DKoFhoFNuFs9mGxkc4yXFby9JiZbzIuwBFSe9Qb9vKkgqZxrUR-gZdcFhksqi4IlVW8UZylPZAcge3Dw2UlaTM/s1600-h/rat_cheese.gif">
     <img style="display:block; margin:0px auto 10px; text-align:center;cursor:pointer; cursor:hand;width: 230px; height: 320px;"
       src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEh0dTd1RVdaar9KTLCks9J022j3MxwYgjx7fIjXCTaj1COwHcxCAEge4DKoFhoFNuFs9mGxkc4yXFby9JiZbzIuwBFSe9Qb9vKkgqZxrUR-gZdcFhksqi4IlVW8UZylPZAcge3Dw2UlaTM/s320/rat_cheese.gif"
       border="0" alt=""
       id="BLOGGER_PHOTO_ID_5336176003028964418" />
As you may notice in the above code, the references specified in href and src are no much difference except for one directory hierarchy:
  • href uses the image in s1600-h:     the image looks original or bigger.
  • src uses the image in s320:     the image is shrunk or smaller.
With the image specified in s320 directory, the image is displayed without a problem:
<img style="display:block; margin:0px auto 10px; text-align:center;"
     src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEh0dTd1RVdaar9KTLCks9J022j3MxwYgjx7fIjXCTaj1COwHcxCAEge4DKoFhoFNuFs9mGxkc4yXFby9JiZbzIuwBFSe9Qb9vKkgqZxrUR-gZdcFhksqi4IlVW8UZylPZAcge3Dw2UlaTM/s320/rat_cheese.gif"
     border="0" alt="Using s320"
     id="BLOGGER_PHOTO_ID_5336176003028964418" />
With the image specified in s1600-h directory, the image cannot be displayed:
<img style="display:block; margin:0px auto 10px; text-align:center;"
     src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEh0dTd1RVdaar9KTLCks9J022j3MxwYgjx7fIjXCTaj1COwHcxCAEge4DKoFhoFNuFs9mGxkc4yXFby9JiZbzIuwBFSe9Qb9vKkgqZxrUR-gZdcFhksqi4IlVW8UZylPZAcge3Dw2UlaTM/s1600-h/rat_cheese.gif"
     border="0" alt="Using s1600-h"
     id="BLOGGER_PHOTO_ID_5336176003028964418" />
In order to reference the bigger image, you have to use the undocumented directory s1600 for image reference, not s1600-h:
<img style="display:block; margin:0px auto 10px; text-align:center;"
     src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEh0dTd1RVdaar9KTLCks9J022j3MxwYgjx7fIjXCTaj1COwHcxCAEge4DKoFhoFNuFs9mGxkc4yXFby9JiZbzIuwBFSe9Qb9vKkgqZxrUR-gZdcFhksqi4IlVW8UZylPZAcge3Dw2UlaTM/s1600/rat_cheese.gif"
     border="0" alt="Using s1600"
     id="BLOGGER_PHOTO_ID_5336176003028964418" />

Sunday, May 3, 2009

LINQ and ConnectionString

There are some issues you should know when you are working with Database Markup Language(DBML) or generating LINQ to SQL classes in VS.2008 IDE:

Why is the dropdown box of "Choose your context object" always empty or why can't I find my context object while I am configuring LinqDataSource?

Why do I have this error "No parameterless constructor defined for this object"?

Why does the actual connection string suddenly appear inside the parameterless contructor of my DataContext class?

Why can't I add a database table to DBML / LINQ designer?



  1. Why is the dropdown box of "Choose your context object" always empty or why can't I find my context object while I am configuring LinqDataSource?

    Please be sure the DLL containing the context object existed and you have a reference to it. If the LINQ class is newly created, please compile the project first before you try to configure LinqDataSource.

  2. Why do I have this error "No parameterless constructor defined for this object"?

    If you remove the ConnectionString manually from the Connection element of the DBML source file first (use external text editor to remove it) and then set the Application Settings to False (also see How to Turn off Application Settings for LINQ), your parameterless constructor of the DataContext class (e.g., Book.designer.cs) will immediately be removed by the IDE. Then the above error will occur when you recompile and run the application.

    Normally enabling Application Settings on LINQ will have the parameterless constructor generated for you automatically, as long as its ConnectionString property is not empty. Indeed, you won't be able to remove the ConnectionString property inside the IDE.  But you remove it anyway outside it. When this happens, the original parameterless contructor will lose forever regardless of Application Settings. Thereafer, adding back the ConnectionString property to the file will not resolve it. And the other problem dicussed in the next question may occur.

    To resolve this, you need to manually add the original parameterless constructor back.

  3. Why does the actual connection string (like below) suddenly appear inside the parameterless contructor of my DataContext class? For example,
        public BookDataContext() :
            base("Data Source=.\\SQLEXPRESS;AttachDbFilename=|DataDirectory|\\Books.mdf;Integrated Security=True;User Instance=True", mappingSource)
        {
            OnCreated();
        }

    Situation 1:    If you set the Application Settings to False while the ConnectionString property exists in the DBML source file, your original parameterless constructor of that DataContext class will be altered by hard-coding with the actual connection string that can be found in your .config file (e.g., web.config).

    Situation 2:    The application had Application Settings set to False and the correct default parameterless constructor some time ago. Later, you make a change by adding a new table to the LINQ. As soon as you drag the new table to the DBML or LINQ designer, your original, correct parameterless contructor will be altered by hard-coding with the actual connection string.

    Currently, I don't know any other easy way to deal with it. If you know, please jot me a note. What I do is

    1. First, ensure Application Settings for LINQ is set to False.
    2. Then, manually alter or copy the parameterless contructor back to the original
    3. .

  4. Why can't I add a database table to DBML / LINQ designer?

    This happens when

    • your Application Settings for LINQ is set to False,
    • ConnectionString element doesn't exist in the DBML source file, and
    • the parameterless constructor in LINQ class is missing.

    The IDE needs to know where to read the connection string in order to access the database. Please ensure you have the parameterless contructor defined correctly while remaining Application Settings for LINQ to False and the absence of ConnectionString in the DBML source file.

Sample of Parameterless Constructor in a ContextClass

public BookDataContext() :     base(global::System.Configuration.ConfigurationManager.ConnectionStrings["BookConnectionString"].ConnectionString, mappingSource) {     OnCreated(); } where BookDataContext is my DataContext class, BookConnectionString is the name of the connection string found in .config file. You should change them to fit yours accordingly.

How to Turn off Application Settings for LINQ
The default setting of LINQ to SQL class is always True. You can do the following to disable it.

On VS IDE
  • Go and double-click to open the DBML file from Solution Explorer.
  • Then open the file Properties window if it is not open. Be sure that the focus is on the DBML designer tab; otherwise, you won't see or cannot continue with the next step.
  • Expand Connection and then set Application Settings to False.

Saturday, April 25, 2009

MagicJack: pros and cons

MagicJack (MJ) is a USB Internet phone device using the processor chip from Tiger Jet acting as an adapter that allows you to plug in a regular phone and then use, or directly use the provided softphone via your computer with a speaker and a microphone ready. The device is Windows compliance so there is no driver needed. Any new driver update will be automatically detected and distributed via Windows update.

I purchased it because I was suddenly required 100% travel due to the department transfer and I began disliking my new task. I was thinking to leave the company by starting my job search journey. Since I had the company cell phone, I could simply forward all my personal calls to my cell because my home phone was a basic land line without any feature. Thus, I would like to have an economic solution. MagicJack is the solution I chose. If you have stable high-speed Internet connection, the phone quality will be as good as a regular land line. MJ works like other VoIPs like Vonage but with much lower cost. You can use it wherever you are including internationally, and unlimited calls are free to made within/to the US and Canada.

I am a MagicJack user and use it every day now. I would like to share some thought and experience here. You may find them useful to decide if MagicJack is right for you.

Pros Cons

Driver Needed No, MagicJack is compatible on Windows and Windows compliance. Thus, there is no driver needed. Any new driver update will be automatically detected and distributed via Windows update.

Easy Installation Unlike its advertisement on TV (i.e., no installation), MagicJack requires software download and installation. You must sign-up an account as soon as you have softphone installed and then activated it before you can make a call. This is based on my experience. I have 3 laptops and 1 desktop at home; they all required manually software download and installation.

Cost and Features Low cost. The first year subscription including the device itself is $39.95 + shipping and handling ($9.95 for USA user). Up to this moment I am writing (April 25, 2009), each subsequent subscription per year is only $19.95 and $59.95 for 5 years. Unlike a landline or other VoIPs, the price here is final without additional tax or subcharge. The cost includes all the needy features: unlimited local and long distance call, 411 directory assistance, voice mail, call waiting, 3-way calling, call forwarding, 911 service, and Do-Not-Disturb.

Total Call Minutes No limit on total call minutes. Unlimited calls can be made within the United State and Canada. But the duration of each call is limited to 2 hours.

Quality of the Call and Connectivity The service provided by MagicJack is VoIP relying on Internet connection. Thus, your Internet connection stability is the key to your call quality and connectivity. I have a stable, high-speed Internet; I found the call quality is as good as a landland. I usually talk to my family about 2 hours during weekends (but there is time limit per call; 2 hours max). Within these 2 hours, my call is never interrupted nor dropped. I am happy with the call quality and I have no problem on the call connectivity.

Phone Number Like other VoIP phone service providers (e.g., Vonage), MagicJack provides you a phone number. You can choose the number that you want including the area code.

Windows Only MagicJack cannot run on any other platforms but windows only.

Computer-Dependent MagicJack goes offline when your computer is off. In addition, you have to log in onto the system and let the software (i.e., softphone) load before use. MagicJack provides you a call log (via softphone) but the missing call will not be captured when the computer is off unless the person leaves you a message. Any left message will be directly delivered to your email box via a WAV file.

Software-Dependent You cannot make a call if your softphone is not ready. You have to wait until your softphone is completely loaded. Seeing "Ready to Call" on the softphone doesn't mean you can make a call. You may still have to wait a few seconds. Although you are able to use a regular phone to make calls, the softphone is the actual interface sitting between your phone and the device (i.e., MagicJack). Thus, if the softphone is not running, your phone will not work, and you will never know if there is a missing call unless the person leaves you a message. Only when the softphone is up and running, all calls timestamp will be kept in your call log.

Any Hardware Conflict Yes, at least I found that MagicJack is conflict with my USB TV tuner. As soon as I plug the TV tuner in, the MagicJack softphone won't work without unplugging and re-plugging. The same problem occurs when you start or reboot the system with the TV tuner and MagicJack attached. If you make a call without hearing the phone dialing after you have keyed in all the numbers, it indicates either your softphone is not ready or you have hardware conflict issue.

Consume More Drive Labels MagicJack is a USB device but it consumes 2 disk/drive labels. If you have a lot of external USB drives (e.g., USB thumb drives), some of them may not be recognized because MagicJack takes their original drive labels. To resolve this, you need to specifically change the drive labels either on MagicJack itself or the USB drives via Computer Management.

Configurability No, there is not much you can configure MagicJack except for the call forwarding via their Web-based interface. You cannot configure the greeting message if you don't like it (sorry, I cannot find a way to do it). And you cannot change the time zone either. Thus, the time shown on the call log doesn't reflect your actual time. I have no idea which time zone is.

Customer Service Unlikely you will speak to a live person over the phone but live-chat with their technical support 24/7 or via email query. Frankly, I don't find them much helpful on the issue I had. MagicJack will provide you a contact number when you have a billing issue but you need to chat with the technical support first. The attitude of the customer representative in the billing department is horrible. The one I talked to was impatient and unpleasant. The more I spoke to her, the more I got frustratetd. See my another comment on MagicJack.

Time Limit Per Call Each call is limited to 2 hours. Otherwise, the call is automatically terminated and then what you can hear is the dial tone. You can continue with your conversation by placing the call again.

Other Issues

  • Dial Tone Problem: Having a dial tone doesn't mean you can make a call immediately! Sometimes you may need to unplug the MagicJack and plug it back in order to resolve the issues. Without a dial tone when you try to make a call happens daily. From my experience, it could occur once a day at least. Hanging up and then pressing the button again sometimes may resolve the problem. Otherwise, replugging Magicjack should work.
  • Late Response: Don't immediately make a call when the softphone is first loaded and indicates "Ready to Call". Give more seconds before use; otherwise, most likely you need to hang up and call again. There is always a delay on the softphone.
  • Popup Softphone: Owing to its software dependent, every time a call runs in or you try to make a call, the softphone will pop up even if you have minimized it to the system tray. It is very irritating. In addition, after having hung up the call, 2 or 3 seconds later, the softphone will pop up in front of every application. I find it very disturbing.
  • Cannot Do System Standy or Hibernate: You cannot put your system to sleep by doing either standy or hibernate when you have MagicJack attached.
  • MagicJack Won't Auto-Start after Logoff: If you log off your account and then log in later, you need to manually start MagicJack.


Updated on October 17, 2009
An additional note: since MagicJack is via VoIP protocol to send voice, your traditional fax machine may not work with VoIP line. I have recently run into this problem. For the detail of why faxing and modems don't work well over VoIP, see here.

Thursday, April 23, 2009

Limitation highlight of SQL server editions

The following applies to SQL 2005 Editions.
Feature Express Developer or Enterprise Workgroup or Standard
Is free? Yes No No
Redistributable? Yes No No
Work with
Team Foundation Server?
No No - Developer
Yes - Enterprise
Yes
Concurrent Connection 5 No Limit -
subject to
@@MAX_Connections
No Limit -
subject to
@@MAX_Connections
Database size 4 GB No Limit No Limit
CPU supported 1 Maximum
OS supported
2 - Workgroup
4 - Standard
RAM Supported 1 GB Maximum OS supported 3GB - Workgroup
Maximum
OS supported
for Standard
Others features comparison list can be found here.

Tuesday, April 21, 2009

Batch or VBScript cannot run as scheduled task on Windows 2008/Vista, why?

Have you ever experienced the following problem?

The script file runs fine as a scheduled task under Windows 2003/XP but it fails on Windows 2008/Vista. If you run the script manually on the command prompt on Windows 2008/Vista, it will work perfectly without a problem. It always fails when it runs as a scheduled task regardless of logon. Task Start Failed always occurs with the default error code 2147750687.

The problem spent me almost 2 months to figure out while I was busy with something else. It happened almost 8 months ago from now when everything was being moved from Windows 2003 to Windows 2008.

The followings were the scheduled task settings on Windows 2008/Vista:

  • Run under the local Administrator account
  • Run whether user is logged on or not
  • Run with highest privileges - actually this one shouldn't be needed since it is run under the local Administrator.

Task Scheduler 2.0 took place on Windows 2008 as well as Windows Vista. There are a lot of differences from its predecessor on Windows 2003 or Windows XP. Here is one of them. When you run your script as a scheduled task, the scheduler (2.0) will create an account impersonating as the user you configure/specify for the task. The account is always resided in or runs from %systemroot%\system32\. Therefore, if your script internally uses relative path, it is going to fail 100% because of path/file not found or permission denied. For that reason, there are two options I would suggest:

  • Modify all the relative paths to be fully qualified paths or absolute paths; e.g., c:\tasks\scheduled\output.txt
  • Modify the script to take the directory as an argument when you run it; e.g., c:\tasks\scheduled\taskA.vbs c:\tasks\myappPath\.

In my case, as soon as I've fixed the relative path problems, all scripts with Task Scheduler 2.0 run like a charm.

In sum, checking if the problem is related to a relative path on your mystery task may save you from a million troubles.

Monday, April 13, 2009

How to attach MDF without LDF into SQL 2005 - Part 2

In my part 1, I mentioned to use VS.2008 to generate the data/schema script first and then import the data back to the database. Here is another way to attach MDF back to the SQL 2005 sever without LDF. In this approach, you don't need to install Visual Studio but you may need another tool for help depending on your result by running the following SQL statement.

In the following illustration, I will use AdventureWorks.mdf as an example.

Execute the following statement to attach your MDF on SQL Management Studio:

USE master
GO
CREATE DATABASE AdventureWorks
ON PRIMARY (FILENAME = N'C:\Program Files\Microsoft SQL Server\MSSQL.1\MSSQL\Data\AdventureWorks.mdf')
FOR ATTACH_REBUILD_LOG
GO

If everything works fine, you may get the similar message below and have your database attached after execution:

File activation failure. The physical file name "C:\Program Files\Microsoft SQL Server\MSSQL.1\MSSQL\Data\AdventureWorks_log.ldf" may be incorrect.
New log file 'C:\Program Files\Microsoft SQL Server\MSSQL.1\MSSQL\Data\AdventureWorks_log.LDF' was created.
Otherwise, you may have the following similar error:
File activation failure. The physical file name "C:\Users\<SomeUserName>\Documents\MyCode\Test\App_Data\AdventureWorks_log.ldf" may be incorrect.
The log cannot be rebuilt when the primary file is read-only.
Msg 1813, Level 16, State 2, Line 3
Could not open new database 'AdventureWorks'. CREATE DATABASE is aborted.

Depending on how or where you get your MDF file (because your MDF may not be detached properly), your LDF rebuilt may fail and so may your database attach command. In this case, you need to download and use the free tool SQL Server Express Utility (SSEUtil) to interact with the server and create the LDF file for you. This tool also works well with the full version of SQL Server 2005. There are two (2) big steps you need to perform in order to make your MDF finally available on SQL Management Studio.

Step 1: Use SSEUtil to attach MDF into the database server

First, you can copy the MDF to the location where you usually store for your SQL server if you like, e.g., C:\Program Files\Microsoft SQL Server\MSSQL.1\MSSQL\Data\, or keep wherever it is now.

Then at the command prompt, run sseutil's attach database command. Before doing it, please be sure to add your path to sseutil.exe. Otherwise, you need to use its absolute path name,., e.g., c:\util\sseutil.exe.

sseutil -s .\SQLExpress -a "C:\Program Files\Microsoft SQL Server\MSSQL.1\MSSQL\Data\AdventureWorks.mdfAdverureWorks
where
  • -s to specify the instance of SQL server.
  • -a to attach the database by specifying its MDF file path name.
  • The last parameter is optional. It is the name of database you want to call (e.g., AdvantureWorks).
There are several options that you can use with sseutil, like -user and -pwd to specify the usr name and its password for connection. Please see the readme file of SSEUtil for details.

or

You can use sseutil console to interact with the server.

  • To start a sseutil console, at the command prompt, type
    sseutil -c
  • You will see
    Console mode. Type 'help' for more information.
    1>
  • To attach a MDF, type
    !attach "C:\Program Files\Microsoft SQL Server\MSSQL.1\MSSQL\Data\AdventureWorks.mdf" AdventureWorks
  • Type quit when you're done. This step is necessary to enable you to add the database into the sysdabases in the next step by removing the lock on file.


Step 2: Add MDF into sysdatabases
Now you have MDF attached to the server and have LDF recreated. But the database will still remain hidden from SQL Server Management Studio. In order to make it visible in SQL Server Management Studio, you need to go through the database attach wizard to add the database back to sysdatabases; or in the management studio, run either one of the following SQL statements:
  • USE master
    GO
    EXEC SP_ATTACH_DB
    @filename1=N'C:\Program Files\Microsoft SQL Server\MSSQL.1\MSSQL\Data\AdventureWorks.mdf',
    @dbname=N'AdventureWorks

    or

  • USE master
    GO
    CREATE DATABASE AdventureWorks
    ON PRIMARY (FILENAME = N'C:\Program Files\Microsoft SQL Server\MSSQL.1\MSSQL\Data\AdventureWorks.mdf')
    FOR ATTACH
    GO
I hope that you find this guide useful.

Sunday, April 12, 2009

How to attach MDF without LDF into SQL 2005

SQL 2005 does not support DBCC REBUILD_LOG. Thus, you cannot use it to rebuild the LDF. Here is how I do to attach a database only using MDF without LDF present.

If you are only interested in attaching MDF to the SQL 2005 server without data extraction and replication, please see my part 2, which works without Visual Studio installed.

I use VS.2008 to do the trick. VS.2005 doesn't have this ability. This method probably may work in Visual Web Developer 2008 Express Edition.

Open VS.2008 and do the followings:
  1. New a Web site or use the existing one.
  2. Add the file to the project's App_Data folder.
  3. Right click this folder in the Solution Explorer and select Add Existing Item...
  4. Locate the MDF file you want to use.
  5. Then click Add.
  6. Double-click on the MDF file that you just added; it will open Server Explorer for you.
  7. Locate and right-click the MDF file data connection on Server Explorer.
  8. Select Publish to provider...
  9. Go through the wizard; when the wizard asks you where to publish, I select Script to file. It will generate a SQL file that contains all the database schema and its data.
On the SQL 2005 Server (mine is Express version),
  1. Create a database with the same name as your MDF.
  2. Run the SQL file you just created by Server Explorer in VS.2008.
Your database should be revived now!

How to print PDF background as you see on file

Have you ever experienced the following problem when you print a PDF file? The background images appear like regular text. It is not what you see on file when you view it.

You may have a PDF look like this when you view it:

PDF as View

Whenever you try to print, you will get this:

PDF Print without the Setting of Print as image

There is nothing wrong with the PDF. To fix this, use Print as image at your printer option. You may find this option at your Advanced button or tab. Then you should print something as below:

Print as Image

Note that the printer used here is a black and white laser printer, not color print.

Friday, April 10, 2009

Favorite torrent sites

Good torrent sites Bad torrent sites
isohunt.com

torrentportal.com

fulldls.com

If you go there, be careful of running out of CPU usage and you have to kill your browser in order to stop it

How to enable script debugging on Internet Explorer

If you are debugging your code with Internet Explorer, be sure to enable your setting of script debugging.

On Internet Explorer, at the menu bar,

  • Choose Tools -> Internet Options
  • Click Advanced
  • Ensure to uncheck Disable script debugging (Internet Explorer) in the Browsing category.

How to enable script debugging on Internet Explorer

Windows Firewall Settings and IIS for ASP.NET

If you are using the built-in internal development/test server in Visual Web Developer Express or VS.NET for development, testing and debugging, there is nothing you should worry about the Windows Firewall settings. However, if you would like to test your Web application over the network on a local IIS Web server, you need to have the following setting changed on Windows Firewall setting:

On Windows XP,

  • Go Control Panel -> Windows Firewall
  • Select Advanced tab
  • On the Network Connection Services selection list, select one of the connection item, e.g., Local Area Connection, Wireless Network Connection, ... and etc.
  • Then click Settings...
  • Ensure that Web Server (HTTP) is checked.

On Windows Vista,

  • Go Windows Control Panel
  • Select the Exception tab
  • Ensure that World Wide Web Services (HTTP) is checked.

Sunday, April 5, 2009

Can you be that creative?

I wish I could make one of these for fun. When I saw them, I was stunned.

Quoted from the email text I received:

These egg shells were cut with a high intensity precision Laser Beam. This gives a very good idea of what can be achieved with a Laser Beam. This gives you an idea what laser surgery performed on one's eye is all about. Is it any wonder how one's vision can be improved in just a few moments? Science is sometimes wonderful, and it's still on the frontier of gaining new knowledge. Inc redible what can be done with an eggshell and a laser beam.

Saturday, April 4, 2009

Can Your Brain Switch?

turn

Please look at it for at least 30 seconds.

Someone has sent this picture to me recently and asked me what I saw.

Well, my colleague who is left-handed saw both ways after a few seconds. I have another colleague who primarily uses her right hand also saw her turning both ways. Mine is even more interesting. At first, I only saw her turning clock-wise. But after a while, I saw her turning the other way around and then in the half way, she was bouncing back. I saw her turning back and forth a few times. Then I saw her turning completely anti-clockwise. I cannot believe my eyes. It is neat. Now I even make her turn whichever way I want or make her bounce back left and right repeatedly. How about you? Can you make her turn? Which direction do you see?

The following is the explanation of seeing her turning different way quoted from the original text. I don't know how true it is. You're the judge then. To me, I am not sure if there is a trick on the picture. I believe everyone can see her turn both ways. Please give it some time to find your way to see her turn the other way

If you see this lady turning in clockwise you are using your right brain.

If you see it the other way, you are using left brain.

Some people do see both ways, but most people see it only one way.


If you try to see it the other way and if you do see, your IQ is above 160 which is almost a genius.
Then see if you can make her go one way and then the other by shifting the brain's current?


BOTH DIRECTIONS CAN BE SEEN
This was proved at Yale University, over a 5 year study on the human brain and it's functions. Only 14% of the US population can see her move both ways.

LEFT BRAIN FUNCTIONS
uses logic
detail oriented
facts rule
words and language
present and past
math and science
can comprehend
knowing
acknowledges
order/pattern perception
knows object name
reality based
forms strategies
practical
safe
RIGHT BRAIN FUNCTIONS
uses feeling
"big picture" oriented
imagination rules
symbols and images
present and future
philosophy & religion
can "get it" (i.e. meaning)
believes
appreciates
spatial perception
knows object function
fantasy based
presents possibilities
impetuous
risk taking

Friday, April 3, 2009

Cheated by MagicJack

I ordered MagicJack (MJ) on Feb 22, 2009 Sunday when I saw their ad on TV. I am sure that I remember what the ad said because the ad was repeated several times. I wasn't making my decision right away to order it. I made it almost last minute.

What the TV ad said is: $39.95 for the device and one year free subscription. The ad sounds more attractive to say that making an order via the phone number advertised on the TV now would waive shipping and handling. The product provided money back guarantee for 30 days free trial. It also emphasized that within the trial period, the customer could send it back at their expense without paying the shipping cost. Thus, it was absolutely no risk.

Well, I was cheated. Besides $39.95, there is always $6.95 shipping and handling fee. Their representative claimed, the ad maybe said that I could get my money back including shipping cost. The S&H is not free. Okay, I listened incorrectly. The following is the page when I first activated MJ and logged in. It clearly said that S&H is free. Even now, the page is still there. But they said they don't know what I am talking about and claimed the page could be wrong. Gosh, unbelievable!

Their automatic ordering system is horrible and not easy to use. They cannot get my name right and they have my incorrect email address. There is no confirmation number when I placed an order so I wouldn't know if the order is done successfully. When I first received MJ, I called the customer service for installation. It is not as smooth as it is claimed on TV. At the same time, I inquired my account information and asked when I would receive my billing statement; the customer service replied that there is nothing to worry. When I activated MJ, I would sign up an account at the same time. The billing statement will be sent to my account later. Up to this moment, I haven't received my statement yet but my credit card was charged already on the 30th day from the date I placed an order.

I called their billing department for questions. Their real person attitude is horrible. She helped nothing but referred me back to the customer service (via live chat). My account was billed under different name and different email address. It was why I have never got my bill. Lucky them, they have my credit card right!

Chatting to their customer service twice about account data correction, they said they had informed the accounting department to have my file updated and may try re-send me the statement. An email for correction will reach me within 24-48 hours. 4 days already and I hear nothing from them. Their issue service is very bad!

Luckily their phone service is not bad - it is the only comfort.