How to get started with developing Internet Explorer extensions?

Does anyone here have experience with/in developing IE extensions that can share their knowledge? This would include code samples, or links to good ones, or documentation on the process, or anything.

I really want to do this, but I'm hitting a giant wall with lousy documentation, lousy code/example code/lack thereof. Any help/resources you could offer would be greatly appreciated.

Specifically, I would like to start with how to get access to/manipulate the DOM from within a IE extension.

EDIT, even more details:

Ideally, I would like to plant a toolbar button that, when clicked, popped a menu up that contains links to external sites. I would also like to access the DOM and plant JavaScript on the page depending on some conditions.

What is the best way to persist information in an IE extension? In Firefox/Chrome/Most modern browsers, you use window.localStorage , but obviously with IE8/IE7, that's not an option. Maybe a SQLite DB or such? It is okay to assume that .NET 4.0 will be installed on a user's computer?

I don't want to use Spice IE as I want to build one that is compatible with IE9 as well. I've added the C++ tag to this question as well, because if it's better to build one in C++, I can do that.


Man... this has been a lot of work! I was so curious about how to do this, that I did it myself.

First of all... credit is not all mine. This is a compilation of what I found, on these sites:

  • CodeProject article, how to make a BHO;
  • 15seconds, but it was not 15 seconds, it took about 7 hours;
  • Microsoft tutorial, helped me adding the command button.
  • And this social.msdn topic, that helped me figure out that the assembly must be in the GAC.
  • This recent MSDN blog post contains a fully-working example
  • many other sites, in the discovery process...
  • And of course, I wanted my answer to have the features you asked:

  • DOM traversal to find something;
  • a button that shows a window (in my case to setup)
  • persist the configuration (I will use regitry for that)
  • and finally execute javascript.
  • I will describe it step by step, how I managed to do it working with Internet Explorer 8 , in Windows 7 x64 ... note that I could not test in other configurations. Hope you understand =)

    Creating a Working Internet Explorer 8 Addon

    I am using Visual Studio 2010 , C# 4 , .Net Framework 4 , so some of these steps might be slightly different for you.

    Created a class library. I called mine InternetExplorerExtension.

    Add these references to the project:

  • Interop.SHDocVw
  • Microsoft.mshtml
  • Note: These references may be in different places in each computer.

    this is what my references section in csproj contains:

    <Reference Include="Interop.SHDocVw, Version=1.1.0.0, Culture=neutral, PublicKeyToken=90ba9c70f846762e, processorArchitecture=MSIL">
      <SpecificVersion>False</SpecificVersion>
      <EmbedInteropTypes>True</EmbedInteropTypes>
      <HintPath>C:Program Files (x86)Microsoft Visual Studio 9.0Common7IDEPrivateAssembliesInterop.SHDocVw.dll</HintPath>
    </Reference>
    <Reference Include="Microsoft.CSharp" />
    <Reference Include="Microsoft.mshtml, Version=7.0.3300.0, Culture=neutral, PublicKeyToken=b03f5f7f11d50a3a">
      <EmbedInteropTypes>True</EmbedInteropTypes>
    </Reference>
    <Reference Include="System" />
    <Reference Include="System.Data" />
    <Reference Include="System.Drawing" />
    <Reference Include="System.Windows.Forms" />
    <Reference Include="System.Xml" />
    

    Create the following files:

    IEAddon.cs

    using System;
    using System.Collections.Generic;
    using System.Runtime.InteropServices;
    using System.Windows.Forms;
    using Microsoft.Win32;
    using mshtml;
    using SHDocVw;
    
    namespace InternetExplorerExtension
    {
        [ComVisible(true)]
        [ClassInterface(ClassInterfaceType.None)]
        [Guid("D40C654D-7C51-4EB3-95B2-1E23905C2A2D")]
        [ProgId("MyBHO.WordHighlighter")]
        public class WordHighlighterBHO : IObjectWithSite, IOleCommandTarget
        {
            const string DefaultTextToHighlight = "browser";
    
            IWebBrowser2 browser;
            private object site;
    
            #region Highlight Text
            void OnDocumentComplete(object pDisp, ref object URL)
            {
                try
                {
                    // @Eric Stob: Thanks for this hint!
                    // This will prevent this method being executed more than once.
                    if (pDisp != this.site)
                        return;
    
                    var document2 = browser.Document as IHTMLDocument2;
                    var document3 = browser.Document as IHTMLDocument3;
    
                    var window = document2.parentWindow;
                    window.execScript(@"function FncAddedByAddon() { alert('Message added by addon.'); }");
    
                    Queue<IHTMLDOMNode> queue = new Queue<IHTMLDOMNode>();
                    foreach (IHTMLDOMNode eachChild in document3.childNodes)
                        queue.Enqueue(eachChild);
    
                    while (queue.Count > 0)
                    {
                        // replacing desired text with a highlighted version of it
                        var domNode = queue.Dequeue();
    
                        var textNode = domNode as IHTMLDOMTextNode;
                        if (textNode != null)
                        {
                            if (textNode.data.Contains(TextToHighlight))
                            {
                                var newText = textNode.data.Replace(TextToHighlight, "<span style='background-color: yellow; cursor: hand;' onclick='javascript:FncAddedByAddon()' title='Click to open script based alert window.'>" + TextToHighlight + "</span>");
                                var newNode = document2.createElement("span");
                                newNode.innerHTML = newText;
                                domNode.replaceNode((IHTMLDOMNode)newNode);
                            }
                        }
                        else
                        {
                            // adding children to collection
                            var x = (IHTMLDOMChildrenCollection)(domNode.childNodes);
                            foreach (IHTMLDOMNode eachChild in x)
                            {
                                if (eachChild is mshtml.IHTMLScriptElement)
                                    continue;
                                if (eachChild is mshtml.IHTMLStyleElement)
                                    continue;
    
                                queue.Enqueue(eachChild);
                            }
                        }
                    }
                }
                catch (Exception ex)
                {
                    MessageBox.Show(ex.Message);
                }
            }
            #endregion
            #region Load and Save Data
            static string TextToHighlight = DefaultTextToHighlight;
            public static string RegData = "SoftwareMyIEExtension";
    
            [DllImport("ieframe.dll")]
            public static extern int IEGetWriteableHKCU(ref IntPtr phKey);
    
            private static void SaveOptions()
            {
                // In IE 7,8,9,(desktop)10 tabs run in Protected Mode
                // which prohibits writes to HKLM, HKCU.
                // Must ask IE for "Writable" registry section pointer
                // which will be something like HKU/S-1-7***/Software/AppDataLow/
                // In "metro" IE 10 mode, tabs run in "Enhanced Protected Mode"
                // where BHOs are not allowed to run, except in edge cases.
                // see http://blogs.msdn.com/b/ieinternals/archive/2012/03/23/understanding-ie10-enhanced-protected-mode-network-security-addons-cookies-metro-desktop.aspx
                IntPtr phKey = new IntPtr();
                var answer = IEGetWriteableHKCU(ref phKey);
                RegistryKey writeable_registry = RegistryKey.FromHandle(
                    new Microsoft.Win32.SafeHandles.SafeRegistryHandle(phKey, true)
                );
                RegistryKey registryKey = writeable_registry.OpenSubKey(RegData, true);
    
                if (registryKey == null)
                    registryKey = writeable_registry.CreateSubKey(RegData);
                registryKey.SetValue("Data", TextToHighlight);
    
                writeable_registry.Close();
            }
            private static void LoadOptions()
            {
                // In IE 7,8,9,(desktop)10 tabs run in Protected Mode
                // which prohibits writes to HKLM, HKCU.
                // Must ask IE for "Writable" registry section pointer
                // which will be something like HKU/S-1-7***/Software/AppDataLow/
                // In "metro" IE 10 mode, tabs run in "Enhanced Protected Mode"
                // where BHOs are not allowed to run, except in edge cases.
                // see http://blogs.msdn.com/b/ieinternals/archive/2012/03/23/understanding-ie10-enhanced-protected-mode-network-security-addons-cookies-metro-desktop.aspx
                IntPtr phKey = new IntPtr();
                var answer = IEGetWriteableHKCU(ref phKey);
                RegistryKey writeable_registry = RegistryKey.FromHandle(
                    new Microsoft.Win32.SafeHandles.SafeRegistryHandle(phKey, true)
                );
                RegistryKey registryKey = writeable_registry.OpenSubKey(RegData, true);
    
                if (registryKey == null)
                    registryKey = writeable_registry.CreateSubKey(RegData);
                registryKey.SetValue("Data", TextToHighlight);
    
                if (registryKey == null)
                {
                    TextToHighlight = DefaultTextToHighlight;
                }
                else
                {
                    TextToHighlight = (string)registryKey.GetValue("Data");
                }
                writeable_registry.Close();
            }
            #endregion
    
            [Guid("6D5140C1-7436-11CE-8034-00AA006009FA")]
            [InterfaceType(1)]
            public interface IServiceProvider
            {
                int QueryService(ref Guid guidService, ref Guid riid, out IntPtr ppvObject);
            }
    
            #region Implementation of IObjectWithSite
            int IObjectWithSite.SetSite(object site)
            {
                this.site = site;
    
                if (site != null)
                {
                    LoadOptions();
    
                    var serviceProv = (IServiceProvider)this.site;
                    var guidIWebBrowserApp = Marshal.GenerateGuidForType(typeof(IWebBrowserApp)); // new Guid("0002DF05-0000-0000-C000-000000000046");
                    var guidIWebBrowser2 = Marshal.GenerateGuidForType(typeof(IWebBrowser2)); // new Guid("D30C1661-CDAF-11D0-8A3E-00C04FC9E26E");
                    IntPtr intPtr;
                    serviceProv.QueryService(ref guidIWebBrowserApp, ref guidIWebBrowser2, out intPtr);
    
                    browser = (IWebBrowser2)Marshal.GetObjectForIUnknown(intPtr);
    
                    ((DWebBrowserEvents2_Event)browser).DocumentComplete +=
                        new DWebBrowserEvents2_DocumentCompleteEventHandler(this.OnDocumentComplete);
                }
                else
                {
                    ((DWebBrowserEvents2_Event)browser).DocumentComplete -=
                        new DWebBrowserEvents2_DocumentCompleteEventHandler(this.OnDocumentComplete);
                    browser = null;
                }
                return 0;
            }
            int IObjectWithSite.GetSite(ref Guid guid, out IntPtr ppvSite)
            {
                IntPtr punk = Marshal.GetIUnknownForObject(browser);
                int hr = Marshal.QueryInterface(punk, ref guid, out ppvSite);
                Marshal.Release(punk);
                return hr;
            }
            #endregion
            #region Implementation of IOleCommandTarget
            int IOleCommandTarget.QueryStatus(IntPtr pguidCmdGroup, uint cCmds, ref OLECMD prgCmds, IntPtr pCmdText)
            {
                return 0;
            }
            int IOleCommandTarget.Exec(IntPtr pguidCmdGroup, uint nCmdID, uint nCmdexecopt, IntPtr pvaIn, IntPtr pvaOut)
            {
                try
                {
                    // Accessing the document from the command-bar.
                    var document = browser.Document as IHTMLDocument2;
                    var window = document.parentWindow;
                    var result = window.execScript(@"alert('You will now be allowed to configure the text to highlight...');");
    
                    var form = new HighlighterOptionsForm();
                    form.InputText = TextToHighlight;
                    if (form.ShowDialog() != DialogResult.Cancel)
                    {
                        TextToHighlight = form.InputText;
                        SaveOptions();
                    }
                }
                catch (Exception ex)
                {
                    MessageBox.Show(ex.Message);
                }
    
                return 0;
            }
            #endregion
    
            #region Registering with regasm
            public static string RegBHO = "SoftwareMicrosoftWindowsCurrentVersionExplorerBrowser Helper Objects";
            public static string RegCmd = "SoftwareMicrosoftInternet ExplorerExtensions";
    
            [ComRegisterFunction]
            public static void RegisterBHO(Type type)
            {
                string guid = type.GUID.ToString("B");
    
                // BHO
                {
                    RegistryKey registryKey = Registry.LocalMachine.OpenSubKey(RegBHO, true);
                    if (registryKey == null)
                        registryKey = Registry.LocalMachine.CreateSubKey(RegBHO);
                    RegistryKey key = registryKey.OpenSubKey(guid);
                    if (key == null)
                        key = registryKey.CreateSubKey(guid);
                    key.SetValue("Alright", 1);
                    registryKey.Close();
                    key.Close();
                }
    
                // Command
                {
                    RegistryKey registryKey = Registry.LocalMachine.OpenSubKey(RegCmd, true);
                    if (registryKey == null)
                        registryKey = Registry.LocalMachine.CreateSubKey(RegCmd);
                    RegistryKey key = registryKey.OpenSubKey(guid);
                    if (key == null)
                        key = registryKey.CreateSubKey(guid);
                    key.SetValue("ButtonText", "Highlighter options");
                    key.SetValue("CLSID", "{1FBA04EE-3024-11d2-8F1F-0000F87ABD16}");
                    key.SetValue("ClsidExtension", guid);
                    key.SetValue("Icon", "");
                    key.SetValue("HotIcon", "");
                    key.SetValue("Default Visible", "Yes");
                    key.SetValue("MenuText", "&Highlighter options");
                    key.SetValue("ToolTip", "Highlighter options");
                    //key.SetValue("KeyPath", "no");
                    registryKey.Close();
                    key.Close();
                }
            }
    
            [ComUnregisterFunction]
            public static void UnregisterBHO(Type type)
            {
                string guid = type.GUID.ToString("B");
                // BHO
                {
                    RegistryKey registryKey = Registry.LocalMachine.OpenSubKey(RegBHO, true);
                    if (registryKey != null)
                        registryKey.DeleteSubKey(guid, false);
                }
                // Command
                {
                    RegistryKey registryKey = Registry.LocalMachine.OpenSubKey(RegCmd, true);
                    if (registryKey != null)
                        registryKey.DeleteSubKey(guid, false);
                }
            }
            #endregion
        }
    }
    

    Interop.cs

    using System;
    using System.Runtime.InteropServices;
    namespace InternetExplorerExtension
    {
        [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, [MarshalAs(UnmanagedType.IUnknown)]out IntPtr ppvSite);
        }
    
    
        [StructLayout(LayoutKind.Sequential, CharSet = CharSet.Unicode)]
        public struct OLECMDTEXT
        {
            public uint cmdtextf;
            public uint cwActual;
            public uint cwBuf;
            [MarshalAs(UnmanagedType.ByValTStr, SizeConst = 100)]
            public char rgwz;
        }
    
        [StructLayout(LayoutKind.Sequential)]
        public struct OLECMD
        {
            public uint cmdID;
            public uint cmdf;
        }
    
        [ComImport(), ComVisible(true),
        Guid("B722BCCB-4E68-101B-A2BC-00AA00404770"),
        InterfaceTypeAttribute(ComInterfaceType.InterfaceIsIUnknown)]
        public interface IOleCommandTarget
        {
    
            [return: MarshalAs(UnmanagedType.I4)]
            [PreserveSig]
            int QueryStatus(
                [In] IntPtr pguidCmdGroup,
                [In, MarshalAs(UnmanagedType.U4)] uint cCmds,
                [In, Out, MarshalAs(UnmanagedType.Struct)] ref OLECMD prgCmds,
                //This parameter must be IntPtr, as it can be null
                [In, Out] IntPtr pCmdText);
    
            [return: MarshalAs(UnmanagedType.I4)]
            [PreserveSig]
            int Exec(
                //[In] ref Guid pguidCmdGroup,
                //have to be IntPtr, since null values are unacceptable
                //and null is used as default group!
                [In] IntPtr pguidCmdGroup,
                [In, MarshalAs(UnmanagedType.U4)] uint nCmdID,
                [In, MarshalAs(UnmanagedType.U4)] uint nCmdexecopt,
                [In] IntPtr pvaIn,
                [In, Out] IntPtr pvaOut);
        }
    }
    

    and finally a form, that we will use to configure the options. In this form place a TextBox and an Ok Button . Set the DialogResult of the button to Ok . Place this code in the form code:

    using System.Windows.Forms;
    namespace InternetExplorerExtension
    {
        public partial class HighlighterOptionsForm : Form
        {
            public HighlighterOptionsForm()
            {
                InitializeComponent();
            }
    
            public string InputText
            {
                get { return this.textBox1.Text; }
                set { this.textBox1.Text = value; }
            }
        }
    }
    

    In the project properties, do the following:

  • Sign the assembly with a strong-key;
  • In the Debug tab, set Start External Program to C:Program Files (x86)Internet Exploreriexplore.exe
  • In the Debug tab, set Command Line Arguments to http://msdn.microsoft.com/en-us/library/ms976373.aspx#bho_getintouch
  • In the Build Events tab, set Post-build events command line to:

    "C:Program Files (x86)Microsoft SDKsWindowsv7.0ABinNETFX 4.0 Toolsx64gacutil.exe" /f /i "$(TargetDir)$(TargetFileName)"
    
    "C:WindowsMicrosoft.NETFrameworkv4.0.30319RegAsm.exe" /unregister "$(TargetDir)$(TargetFileName)"
    
    "C:WindowsMicrosoft.NETFrameworkv4.0.30319RegAsm.exe" "$(TargetDir)$(TargetFileName)"
  • Attention: as my computer is x64, there is a specific x64 inside the path of gacutil executable on my machine that may be different on yours.

    64bit IE Needs 64bit-compiled and 64bit-registered BHO. Use 64bit RegAsm.exe (usually lives in C:WindowsMicrosoft.NETFramework64v4.0.30319RegAsm.exe)

    How this addon works

    It traverses all DOM tree, replacing the text, configured using the button, by itself with a yellow background. If you click on the yellowed texts, it calls a javascript function that was inserted on the page dynamically. The default word is 'browser', so that it matches a lot of them! EDIT: after changing the string to be highlighted, you must click the URL box and press Enter... F5 will not work, I think that it is because F5 is considered as 'navigation', and it would require to listen to navigate event (maybe). I'll try to fix that later.

    Now, it is time to go. I am very tired. Feel free to ask questions... may be I will not be abled to answer since I am going on a trip... in 3 days I'm back, but I'll try to come here in the meantime.


    Another cool approach would be to check out:

    http://www.crossrider.org

    It's a framework based on JS with jquery which lets you develop browsers extensions for IE, FF and Chrome using a single common JS code. Basically the framework does all the nasty work and you're left with writing your applications code.


    The state for IE extensions is actually pretty sad. You have the old model of IE5 Browser Helper Object (yeah, those infamous BHOs that everyone liked to block back in the day), toolbars and the new accelerators for IE. Even then, compatibility will break sometimes. I used to maintain an extension for IE6 that broke with IE7, so there are some things that have changed. For the most part, as far as I know (I haven't touch BHOs in years) you still need to code them using Active Template Libraries (kind of like an STL for Microsoft's COM) and well as such is only for C++. You could do COM Interop with C# and get away with doing it in C# but its probably going to be too hard for what it is worth. Anyway, if you are interested in coding your own extension for IE (which is plausible if you want to have your extensions available in all major browsers) here are the official Microsoft Resources.

    http://msdn.microsoft.com/en-us/library/aa753587(v=vs.85).aspx

    And for the accelerators that are new in IE8 you could check this one.

    http://msdn.microsoft.com/en-us/library/cc289775(v=vs.85).aspx

    I agree the documentation is terrible, and the APIs are quite outdated. Still I hope this helps.

    EDIT: I guess I can throw one last source of information here. I was looking through my notes of back when I was working on BHOs. And this is the article that got me started with them. It is kind of old, but has a good explanation of the ATL interfaces that you will be using when working with IE BHOs (IObjectWithSite for example). I think it is pretty well explained and helped me a lot back then. http://msdn.microsoft.com/en-us/library/bb250436.aspx I also checked the example that GregC posted. It does work with at least IE8, and it is compatible with VS 2010, so if you want to do C# you can get started there and take a look at Jon Skeet's Book. (C# in Depth 2nd edition) Chapter 13 has a good deal of information about the new features in C# 4 that you can use to make the interaction with COM nicer. (I would still recommend you doing your addin in C++)

    链接地址: http://www.djcxy.com/p/40226.html

    上一篇: 语言如何扩展自己?

    下一篇: 如何开始开发Internet Explorer扩展?