Introduction

A browser extension software lives inside the browser and adds new features making the browser suitable for specific personal or business needs. The most common business needs that are easily accomplished with browser plug-ins are: modify default search, add side frames, inject new content into existing webpage, communicate with remote servers or Web Services, search highlight, show ads, and any other task that an installed software program can do. In this article, I will explain how to inject new static and dynamic content into an existing webpage.

Background

Browser extensions allow developers to provide easy access to their browser enhancements by adding elements (like an Explorer Bar) to the default user interface. Introduced in Microsoft Internet Explorer 4.0, this feature enables developers to create Explorer Bars and add entries into the standard context menus. Beginning with Internet Explorer 5, this feature allows developers to add entries into the Tools menu and buttons to the toolbar.

A Browser Helper Object (BHO) runs within Internet Explorer and offers additional services, often without any obvious user interface. For example, a BHO might highlight terms of interest to the user, such as addresses.

For more background information, see: http://msdn.microsoft.com/en-us/library/ms976373.aspx.

The Business Use Case

You want to develop an Internet Explorer extension which shows some fixed layer containing promotional offers and discounts for the goods you are selling. To achieve that, you need to develop a custom IE plug-in using BHO that will track user browsing activities and display your discounts based on some content keywords filtering.

So far so good. You know what should be done. Now let's discover how it is done.

Step by Step Walk-Through

I will explain here a step by step guide on how to accomplish this task as it is better to learn how to do it instead of just take a ready solution which you don't know how it works, and the moment you touch something to make it suit your needs, it breaks down and you have no idea why. I will use .NET C# as a programming framework and language.

1. Open MS Visual Studio 2010 and create a new project "Visual C#" -> "Windows" -> "Class Library" named "IEPlugin"

A new project is opened for you containing a single class Class1. Rename the class to BHO. You can use F2 on the CS file in the Solution Explorer.

From this high-level overview of Browser Helper Objects, one concept emerges clearly: a BHO is a dynamic-link library (DLL) capable of attaching itself to any new instance of Internet Explorer and, under certain circumstances, also Windows Explorer. Such a module can get in touch with the browser through the container's site. In order to do that, we must implement the IObjectWithSite interface.

2. Implement the IObjectWithSite interface

Now you have to define the IObjectWithSite interface which the BHO class will implement. The IObjectWithSiteinterface provides simple objects with a lightweight siting mechanism (lighter than IOleObject). Often, an object must communicate directly with a container site that is managing the object. Outside ofIOleObject::SetClientSite, there is no generic means through which an object becomes aware of its site. TheIObjectWithSite interface provides a siting mechanism. This interface should only be used when IOleObject is not already in use. By using IObjectWithSite, a container can pass the IUnknown pointer of its site to the object through SetSite. Callers can also get the latest site passed to SetSite by using GetSite.

Add references and imports to the following components:

 Collapse

using SHDocVw;
using mshtml;
using System.IO;
using Microsoft.Win32;
using System.Runtime.InteropServices; 

Define the IOleObjectWithSite interface just above your BHO class declaration like this:

 Collapse

[
    ComVisible(true),
    InterfaceType(ComInterfaceType.InterfaceIsIUnknown),
    Guid("FC4801A3-2BA9-11CF-A229-00AA003D7352")
]
public interface IObjectWithSite
{
    [PreserveSig]
    int SetSite([MarshalAs(UnmanagedType.IUnknown)]object site);
    [PreserveSig]
    int GetSite(ref Guid guid, out IntPtr ppvSite);
}

Make sure you don't change the GUID!

3. Make the BHO class implement the IOleObjectWithSite interface

The SetSite() method is where the BHO is initialized and where you would perform all the tasks that happen only once. When you navigate to a URL with Internet Explorer, you should wait for a couple of events to make sure the required document has been completely downloaded and then initialized. Only at this point can you safely access its content through the exposed object model, if any. This means you need to acquire a couple of pointers. The first one is the pointer to IWebBrowser2, the interface that renders the WebBrowser object. The second pointer relates to events. This module must register as an event listener with the browser in order to receive the notification of downloads and document-specific events.

 Collapse

[
        ComVisible(true),
        Guid("2159CB25-EF9A-54C1-B43C-E30D1A4A8277"),
        ClassInterface(ClassInterfaceType.None)
]

public class BHO : IObjectWithSite
{
  private WebBrowser webBrowser;

  public int SetSite(object site)
  {
    if (site != null)
    {
        webBrowser = (WebBrowser)site;
        webBrowser.DocumentComplete += 
          new DWebBrowserEvents2_DocumentCompleteEventHandler(
          this.OnDocumentComplete);
    }
    else
    {
        webBrowser.DocumentComplete -= 
          new DWebBrowserEvents2_DocumentCompleteEventHandler(
          this.OnDocumentComplete);
        webBrowser = null;
    }

    return 0;

  }

  public int GetSite(ref Guid guid, out IntPtr ppvSite)
  {
    IntPtr punk = Marshal.GetIUnknownForObject(webBrowser);
    int hr = Marshal.QueryInterface(punk, ref guid, out ppvSite);
    Marshal.Release(punk);
    return hr;
  }

  public void OnDocumentComplete(object pDisp, ref object URL)
  {
    HTMLDocument document = (HTMLDocument)webBrowser.Document;
  }

}

The BHO site is the COM interface used to establish a communication. Define a GUID attribute for your BHO as it will be used later on during registration. We have implemented the SetSite and GetSite methods and added an empty event handler for the document loaded event. Thus, we will receive an event every time the document is loaded into the browser.

4. Implement the OnDocumentComplete method to inject the JavaScript code and the div element

 Collapse

public void OnDocumentComplete(object pDisp, ref object URL)
{
    HTMLDocument document = (HTMLDocument)webBrowser.Document;

    IHTMLElement head = (IHTMLElement)((IHTMLElementCollection)
                           document.all.tags("head")).item(null, 0);
    IHTMLScriptElement scriptObject = 
      (IHTMLScriptElement)document.createElement("script");
    scriptObject.type = @"text/javascript";
    scriptObject.text = "\nfunction hidediv(){document.getElementById" + 
                        "('myOwnUniqueId12345').style.visibility = 'hidden';}\n\n";
    ((HTMLHeadElement)head).appendChild((IHTMLDOMNode)scriptObject);

    string div = "<div id=\"myOwnUniqueId12345\" style=\"position:" + 
                 "fixed;bottom:0px;right:0px;z-index:9999;width=300px;" + 
                 "height=150px;\"> <div style=\"position:relative;" + 
                 "float:right;font-size:9px;\"><a " + 
                 "href=\"javascript:hidediv();\">close</a></div>" +
        "My content goes here ...</div>";

    document.body.insertAdjacentHTML("afterBegin", div);
}

First we inject the JavaScript that we will use to close the div popup. Then we inject the HTML for the div element in the body with style that is displayed in the bottom right corner of the browser. And that is it. You may want to filter our URLs and content for which you display the div. You can go like this:

 Collapse

if (URL.ToString().Contains("www.google.com")) 
{
    // Show div only in here ... } 

5. Register your BHO to be loaded by Internet Explorer

 Collapse

public const string BHO_REGISTRY_KEY_NAME = 
   "Software\\Microsoft\\Windows\\" +     "CurrentVersion\\Explorer\\Browser Helper Objects";  [ComRegisterFunction] public static void RegisterBHO(Type type) {     RegistryKey registryKey =        Registry.LocalMachine.OpenSubKey(BHO_REGISTRY_KEY_NAME, true);      if (registryKey == null)         registryKey = Registry.LocalMachine.CreateSubKey(                                 BHO_REGISTRY_KEY_NAME);      string guid = type.GUID.ToString("B");     RegistryKey ourKey = registryKey.OpenSubKey(guid);      if (ourKey == null)     {         ourKey = registryKey.CreateSubKey(guid);     }      ourKey.SetValue("NoExplorer", 1, RegistryValueKind.DWord);      registryKey.Close();     ourKey.Close(); }  [ComUnregisterFunction] public static void UnregisterBHO(Type type) {     RegistryKey registryKey =        Registry.LocalMachine.OpenSubKey(BHO_REGISTRY_KEY_NAME, true);     string guid = type.GUID.ToString("B");      if (registryKey != null)         registryKey.DeleteSubKey(guid, false); }

The Register method simply tells IE which is the GUID of your extension so that it could be loaded. The "No Explorer" value simply says that we don't want to be loaded by Windows Explorer.

Now all you have to do is create a simple installation project that will install your IE plug-in or install it manually.

That is all about it! Quick and simple.

首先简单说一下这个 /3GB 开关 (知道的可以不看)

默认情况下，Windows 可以对总计 4 千兆字节 (GB) 的虚拟地址空间进行寻址。默认情况下，此地址空间中的 2 GB 为内核（操作系统）保留，另外 2 GB 是为用户模式程序保留的。当你将 /3GB 开关放入操作系统的 Boot.ini 文件中时，你就重新分配了虚拟地址空间，给用户模式程序提供 3 GB 的空间，同时将内核限制为 1 GB。

设置办法：C:\boot.ini 文件做如下修改：

[boot loader]
timeout=30
default=multi(0)disk(0)rdisk(0)partition(2)\WINNT
[operating systems]
multi(0)disk(0)rdisk(0)partition(2)\WINNT="????" /3GB

如下操作系统可以支持 /3GB 开关

Windows XP Professional
Windows Server 2003
Windows Server 2003, Enterprise Edition
Windows Server 2003, Datacenter Edition
Windows 2000 Advanced Server
Windows 2000 Datacenter Server
Windows NT Server 4.0, Enterprise Edition

Windows VISTA , Windows 7 和 Windows server 2008 也支持这个开关

详见

http://www.microsoft.com/whdc/system/platform/server/PAE/PAEmem.mspx

设置完后重新启动系统，这时应用程序就可以申请超过3G的内存了。一切好像到这里就该结束了，然而并非如此。

当我运行我在.net framework 下做的测试程序时，我发现3GB开关打开后，这个测试程序依然无法申请超过2G的内存，内存申请到1.5G以上时就无法再分配内存了。

查找资料后我发现，操作系统在支持/3GB参数后应用程序也要做相应的修改，告诉操作系统可以按照/3GB方式运行才行，我想这很可能是出于对应用程序兼容性方面的考虑。

为了告知操作系统这个应用程序可以支持/3GB方式，我们需要往exe 文件头中添加一个 IMAGE_FILE_LARGE_ADDRESS_AWARE 标志。添加的方式很简单：

在你的系统的 Program Files\Microsoft Visual Studio 8\VC\bin 目录下找到 editbin 这个可执行文件，在命令行下执行：

editbin /LARGEADDRESSAWARE yourapplication.exe 这里的yourapplication.exe 需要输入的是你的.net 应用程序的路径名加文件名。

执行了这条语句后，我再次运行测试程序，内存可以申请到 2.5G了，比2GB模式下多了1G内存可以使用。

最后多罗嗦一句，为什么.net 应用程序在 2GB 情况下只能申请最多 1.4-1.6 GB 的内存？

这是因为.net 的垃圾回收器在工作时需要拷贝 live objects，.net framework 需要为它保留一定空间的内存来完成这些拷贝工作。这也是为什么Microsoft 建议asp.net 应用程序内存分配的上限最好设置为800M的原因.

.net 应用程序如何优化性能，参见下面链接，大家有兴趣可以去看看，这里不再多说了，离主题有点远了。

Improving .NET Application Performance and Scalability