Develop Reference

Running FxCop analyzers from command line

I'm trying to use the new FxCop analyzers, but they're only available as NuGet packages or as VSIX extensions. I'd like to be able to run them directly, either from inside a C# program or from the command line. Anyone have any advice? Even general info on where you can find the executables for NuGet or VSIX would help.
(I know about fxcopcmd.exe, but that's the legacy version, and it works only on built .exes or .dlls. If at all possible, I need something that works before building.)

Answering my own question in case anyone else has to deal with this. I found a solution, but fair warning, it's not pretty.
I took an example C# solution from Github, loaded it up in Visual Studio, and used NuGet to install the FxCop analyzers. This installed the analyzers, and changed the solution's .csproj files to reference them. In my case, I found a copy of the analyzers in C:\users\myname.nuget\packages.
I compared the modified .csproj files to the originals, to see what changes had been made during installation. I recommend anyone following along make this comparison themselves, but in my case, the changes were:
Five Import elements at the top referencing various .props files.
An empty NuGetPackageImportStamp element.
Two new ItemGroups near the bottom, the first containing a single element named "None", the second containing various Analyzer elements referencing .dlls.
A new Target to ensure that the .props files actually existed.
I wrote a C# program that took an arbitrary solution, found all the .csproj files inside, and manually added those new XML elements to them. I skipped the one-element ItemGroup and the Target without any problems.
Ideally you would then (from inside the same C# program) call msbuild on the .sln file, save every output line matching the regex "): warning CA\d\d\d\d: " (i.e. the warnings that FxCop generated), and restore the original .csproj files. I did that all manually. Here's the code for the XML manipulation, though:
static void addAnalyzersToCsProj(string file)
{
string[] packages = new string[]
{
#"C:\users\myname\.nuget\packages\microsoft.codeanalysis.fxcopanalyzers\3.0.0\build\Microsoft.CodeAnalysis.FxCopAnalyzers.props",
#"C:\users\myname\.nuget\packages\microsoft.codeanalysis.versioncheckanalyzer\3.0.0\build\Microsoft.CodeAnalysis.VersionCheckAnalyzer.props",
#"C:\users\myname\.nuget\packages\microsoft.codequality.analyzers\3.0.0\build\Microsoft.CodeQuality.Analyzers.props",
#"C:\users\myname\.nuget\packages\microsoft.netcore.analyzers\3.0.0\build\Microsoft.NetCore.Analyzers.props",
#"C:\users\myname\.nuget\packages\microsoft.netframework.analyzers\3.0.0\build\Microsoft.NetFramework.Analyzers.props",
};
var root = XElement.Load(file);
var ns = "";
for (var i = 0; i < 5; i++)
{
XElement packageImport = new XElement(ns+"Import");
packageImport.SetAttributeValue("Project", packages[i]);
string condition = "Exists('" + packages[i] + "')";
packageImport.SetAttributeValue("Condition", condition);
root.AddFirst(packageImport);
}
var propertyGroup = root.Descendants(ns + "PropertyGroup").First();
var stamp = new XElement(ns+"NuGetPackageImportStamp", "");
propertyGroup.Elements().Last().AddAfterSelf(stamp);
var newGroup = new XElement(ns+"ItemGroup");
// do we need to include the "None Include="packages.config"" thing?
string[] libraries = new string[]
{
#"C:\users\myname\.nuget\packages\microsoft.codeanalysis.versioncheckanalyzer\3.0.0\analyzers\dotnet\cs\Microsoft.CodeAnalysis.VersionCheckAnalyzer.resources.dll",
#"C:\users\myname\.nuget\packages\microsoft.codeanalysis.versioncheckanalyzer\3.0.0\analyzers\dotnet\Microsoft.CodeAnalysis.VersionCheckAnalyzer.dll",
#"C:\users\myname\.nuget\packages\microsoft.codequality.analyzers\3.0.0\analyzers\dotnet\cs\Humanizer.dll",
#"C:\users\myname\.nuget\packages\microsoft.codequality.analyzers\3.0.0\analyzers\dotnet\cs\Microsoft.CodeQuality.Analyzers.dll",
#"C:\users\myname\.nuget\packages\microsoft.codequality.analyzers\3.0.0\analyzers\dotnet\cs\Microsoft.CodeQuality.CSharp.Analyzers.dll",
#"C:\users\myname\.nuget\packages\microsoft.netcore.analyzers\3.0.0\analyzers\dotnet\cs\Microsoft.NetCore.Analyzers.dll",
#"C:\users\myname\.nuget\packages\microsoft.netcore.analyzers\3.0.0\analyzers\dotnet\cs\Microsoft.NetCore.CSharp.Analyzers.dll",
#"C:\users\myname\.nuget\packages\microsoft.netframework.analyzers\3.0.0\analyzers\dotnet\cs\Microsoft.NetFramework.Analyzers.dll",
#"C:\users\myname\.nuget\packages\microsoft.netframework.analyzers\3.0.0\analyzers\dotnet\cs\Microsoft.NetFramework.CSharp.Analyzers.dll",
};
foreach (string lib in libraries)
{
XElement analyzer = new XElement(ns+"Analyzer");
analyzer.SetAttributeValue("Include", lib);
newGroup.AddFirst(analyzer);
}
Console.WriteLine(root.Elements().Last().ToString());
root.Elements().Last().AddAfterSelf(newGroup);
root.Save(file, SaveOptions.None);
// and do we need to include the error checking target?
}
As far as I can tell, it works, though I have no idea what would happen if you tried to do it on a solution that already has the analyzers installed normally.
Running the FxCop analyzers through msbuild seems inefficient, but I haven't found a better way to do it. They look like they're built to only work within a compiler. I hope I'm wrong, and I would still appreciate any advice on how to run the analyzers automatically without having to build the whole project.

"Injecting" MySQL Data into application [duplicate]

Is it possible to embed a pre-existing DLL into a compiled C# executable (so that you only have one file to distribute)? If it is possible, how would one go about doing it?
Normally, I'm cool with just leaving the DLLs outside and having the setup program handle everything, but there have been a couple of people at work who have asked me this and I honestly don't know.

I highly recommend to use Costura.Fody - by far the best and easiest way to embed resources in your assembly. It's available as NuGet package.
Install-Package Costura.Fody
After adding it to the project, it will automatically embed all references that are copied to the output directory into your main assembly. You might want to clean the embedded files by adding a target to your project:
Install-CleanReferencesTarget
You'll also be able to specify whether to include the pdb's, exclude certain assemblies, or extracting the assemblies on the fly. As far as I know, also unmanaged assemblies are supported.
Update
Currently, some people are trying to add support for DNX.
Update 2
For the lastest Fody version, you will need to have MSBuild 16 (so Visual Studio 2019). Fody version 4.2.1 will do MSBuild 15. (reference: Fody is only supported on MSBuild 16 and above. Current version: 15)

Just right-click your project in Visual Studio, choose Project Properties -> Resources -> Add Resource -> Add Existing File…
And include the code below to your App.xaml.cs or equivalent.
public App()
{
AppDomain.CurrentDomain.AssemblyResolve +=new ResolveEventHandler(CurrentDomain_AssemblyResolve);
}
System.Reflection.Assembly CurrentDomain_AssemblyResolve(object sender, ResolveEventArgs args)
{
string dllName = args.Name.Contains(',') ? args.Name.Substring(0, args.Name.IndexOf(',')) : args.Name.Replace(".dll","");
dllName = dllName.Replace(".", "_");
if (dllName.EndsWith("_resources")) return null;
System.Resources.ResourceManager rm = new System.Resources.ResourceManager(GetType().Namespace + ".Properties.Resources", System.Reflection.Assembly.GetExecutingAssembly());
byte[] bytes = (byte[])rm.GetObject(dllName);
return System.Reflection.Assembly.Load(bytes);
}
Here's my original blog post:
http://codeblog.larsholm.net/2011/06/embed-dlls-easily-in-a-net-assembly/

If they're actually managed assemblies, you can use ILMerge. For native DLLs, you'll have a bit more work to do.
See also: How can a C++ windows dll be merged into a C# application exe?

Yes, it is possible to merge .NET executables with libraries. There are multiple tools available to get the job done:
ILMerge is a utility that can be used to merge multiple .NET assemblies into a single assembly.
Mono mkbundle, packages an exe and all assemblies with libmono into a single binary package.
IL-Repack is a FLOSS alterantive to ILMerge, with some additional features.
In addition this can be combined with the Mono Linker, which does remove unused code and therefor makes the resulting assembly smaller.
Another possibility is to use .NETZ, which does not only allow compressing of an assembly, but also can pack the dlls straight into the exe. The difference to the above mentioned solutions is that .NETZ does not merge them, they stay separate assemblies but are packed into one package.
.NETZ is a open source tool that compresses and packs the Microsoft .NET Framework executable (EXE, DLL) files in order to make them smaller.

ILMerge can combine assemblies to one single assembly provided the assembly has only managed code. You can use the commandline app, or add reference to the exe and programmatically merge. For a GUI version there is Eazfuscator, and also .Netz both of which are free. Paid apps include BoxedApp and SmartAssembly.
If you have to merge assemblies with unmanaged code, I would suggest SmartAssembly. I never had hiccups with SmartAssembly but with all others. Here, it can embed the required dependencies as resources to your main exe.
You can do all this manually not needing to worry if assembly is managed or in mixed mode by embedding dll to your resources and then relying on AppDomain's Assembly ResolveHandler. This is a one stop solution by adopting the worst case, ie assemblies with unmanaged code.
static void Main()
{
AppDomain.CurrentDomain.AssemblyResolve += (sender, args) =>
{
string assemblyName = new AssemblyName(args.Name).Name;
if (assemblyName.EndsWith(".resources"))
return null;
string dllName = assemblyName + ".dll";
string dllFullPath = Path.Combine(GetMyApplicationSpecificPath(), dllName);
using (Stream s = Assembly.GetEntryAssembly().GetManifestResourceStream(typeof(Program).Namespace + ".Resources." + dllName))
{
byte[] data = new byte[stream.Length];
s.Read(data, 0, data.Length);
//or just byte[] data = new BinaryReader(s).ReadBytes((int)s.Length);
File.WriteAllBytes(dllFullPath, data);
}
return Assembly.LoadFrom(dllFullPath);
};
}
The key here is to write the bytes to a file and load from its location. To avoid chicken and egg problem, you have to ensure you declare the handler before accessing assembly and that you do not access the assembly members (or instantiate anything that has to deal with the assembly) inside the loading (assembly resolving) part. Also take care to ensure GetMyApplicationSpecificPath() is not any temp directory since temp files could be attempted to get erased by other programs or by yourself (not that it will get deleted while your program is accessing the dll, but at least its a nuisance. AppData is good location). Also note that you have to write the bytes each time, you cant load from location just 'cos the dll already resides there.
For managed dlls, you need not write bytes, but directly load from the location of the dll, or just read the bytes and load the assembly from memory. Like this or so:
using (Stream s = Assembly.GetEntryAssembly().GetManifestResourceStream(typeof(Program).Namespace + ".Resources." + dllName))
{
byte[] data = new byte[stream.Length];
s.Read(data, 0, data.Length);
return Assembly.Load(data);
}
//or just
return Assembly.LoadFrom(dllFullPath); //if location is known.
If the assembly is fully unmanaged, you can see this link or this as to how to load such dlls.

.NET Core 3.0 natively supports compiling to a single .exe
The feature is enabled by the usage of the following property in your project file (.csproj):
<PropertyGroup>
<PublishSingleFile>true</PublishSingleFile>
</PropertyGroup>
This is done without any external tool.
See my answer for this question for further details.

The excerpt by Jeffrey Richter is very good. In short, add the libraries as embedded resources and add a callback before anything else. Here is a version of the code (found in the comments of his page) that I put at the start of Main method for a console app (just make sure that any calls that use the libraries are in a different method to Main).
AppDomain.CurrentDomain.AssemblyResolve += (sender, bargs) =>
{
String dllName = new AssemblyName(bargs.Name).Name + ".dll";
var assem = Assembly.GetExecutingAssembly();
String resourceName = assem.GetManifestResourceNames().FirstOrDefault(rn => rn.EndsWith(dllName));
if (resourceName == null) return null; // Not found, maybe another handler will find it
using (var stream = assem.GetManifestResourceStream(resourceName))
{
Byte[] assemblyData = new Byte[stream.Length];
stream.Read(assemblyData, 0, assemblyData.Length);
return Assembly.Load(assemblyData);
}
};

To expand on #Bobby's asnwer above. You can edit your .csproj to use IL-Repack to automatically package all files into a single assembly when you build.
Install the nuget ILRepack.MSBuild.Task package with Install-Package ILRepack.MSBuild.Task
Edit the AfterBuild section of your .csproj
Here is a simple sample that merges ExampleAssemblyToMerge.dll into your project output.
<!-- ILRepack -->
<Target Name="AfterBuild" Condition="'$(Configuration)' == 'Release'">
<ItemGroup>
<InputAssemblies Include="$(OutputPath)\$(AssemblyName).exe" />
<InputAssemblies Include="$(OutputPath)\ExampleAssemblyToMerge.dll" />
</ItemGroup>
<ILRepack
Parallel="true"
Internalize="true"
InputAssemblies="#(InputAssemblies)"
TargetKind="Exe"
OutputFile="$(OutputPath)\$(AssemblyName).exe"
/>
</Target>

The following method DO NOT use external tools and AUTOMATICALLY include all needed DLL (no manual action required, everything done at compilation)
I read a lot of answer here saying to use ILMerge, ILRepack or Jeffrey Ritcher method but none of that worked with WPF applications nor was easy to use.
When you have a lot of DLL it can be hard to manually include the one you need in your exe. The best method i found was explained by Wegged here on StackOverflow
Copy pasted his answer here for clarity (all credit to Wegged)
1) Add this to your .csproj file:
<Target Name="AfterResolveReferences">
<ItemGroup>
<EmbeddedResource Include="#(ReferenceCopyLocalPaths)" Condition="'%(ReferenceCopyLocalPaths.Extension)' == '.dll'">
<LogicalName>%(ReferenceCopyLocalPaths.DestinationSubDirectory)%(ReferenceCopyLocalPaths.Filename)%(ReferenceCopyLocalPaths.Extension)</LogicalName>
</EmbeddedResource>
</ItemGroup>
</Target>
2) Make your Main Program.cs look like this:
[STAThreadAttribute]
public static void Main()
{
AppDomain.CurrentDomain.AssemblyResolve += OnResolveAssembly;
App.Main();
}
3) Add the OnResolveAssembly method:
private static Assembly OnResolveAssembly(object sender, ResolveEventArgs args)
{
Assembly executingAssembly = Assembly.GetExecutingAssembly();
AssemblyName assemblyName = new AssemblyName(args.Name);
var path = assemblyName.Name + ".dll";
if (assemblyName.CultureInfo.Equals(CultureInfo.InvariantCulture) == false) path = String.Format(#"{0}\{1}", assemblyName.CultureInfo, path);
using (Stream stream = executingAssembly.GetManifestResourceStream(path))
{
if (stream == null) return null;
var assemblyRawBytes = new byte[stream.Length];
stream.Read(assemblyRawBytes, 0, assemblyRawBytes.Length);
return Assembly.Load(assemblyRawBytes);
}
}

You could add the DLLs as embedded resources, and then have your program unpack them into the application directory on startup (after checking to see if they're there already).
Setup files are so easy to make, though, that I don't think this would be worth it.
EDIT: This technique would be easy with .NET assemblies. With non-.NET DLLs it would be a lot more work (you'd have to figure out where to unpack the files and register them and so on).

Another product that can handle this elegantly is SmartAssembly, at SmartAssembly.com. This product will, in addition to merging all dependencies into a single DLL, (optionally) obfuscate your code, remove extra meta-data to reduce the resulting file size, and can also actually optimize the IL to increase runtime performance.
There is also some kind of global exception handling/reporting feature it adds to your software (if desired) that could be useful. I believe it also has a command-line API so you can make it part of your build process.

Neither the ILMerge approach nor Lars Holm Jensen's handling the AssemblyResolve event will work for a plugin host. Say executable H loads assembly P dynamically and accesses it via interface IP defined in an separate assembly. To embed IP into H one shall need a little modification to Lars's code:
Dictionary<string, Assembly> loaded = new Dictionary<string,Assembly>();
AppDomain.CurrentDomain.AssemblyResolve += (sender, args) =>
{ Assembly resAssembly;
string dllName = args.Name.Contains(",") ? args.Name.Substring(0, args.Name.IndexOf(',')) : args.Name.Replace(".dll","");
dllName = dllName.Replace(".", "_");
if ( !loaded.ContainsKey( dllName ) )
{ if (dllName.EndsWith("_resources")) return null;
System.Resources.ResourceManager rm = new System.Resources.ResourceManager(GetType().Namespace + ".Properties.Resources", System.Reflection.Assembly.GetExecutingAssembly());
byte[] bytes = (byte[])rm.GetObject(dllName);
resAssembly = System.Reflection.Assembly.Load(bytes);
loaded.Add(dllName, resAssembly);
}
else
{ resAssembly = loaded[dllName]; }
return resAssembly;
};
The trick to handle repeated attempts to resolve the same assembly and return the existing one instead of creating a new instance.
EDIT:
Lest it spoil .NET's serialization, make sure to return null for all assemblies not embedded in yours, thereby defaulting to the standard behaviour. You can get a list of these libraries by:
static HashSet<string> IncludedAssemblies = new HashSet<string>();
string[] resources = System.Reflection.Assembly.GetExecutingAssembly().GetManifestResourceNames();
for(int i = 0; i < resources.Length; i++)
{ IncludedAssemblies.Add(resources[i]); }
and just return null if the passed assembly does not belong to IncludedAssemblies .

It may sound simplistic, but WinRar gives the option to compress a bunch of files to a self-extracting executable.
It has lots of configurable options: final icon, extract files to given path, file to execute after extraction, custom logo/texts for popup shown during extraction, no popup window at all, license agreement text, etc.
May be useful in some cases.

I use the csc.exe compiler called from a .vbs script.
In your xyz.cs script, add the following lines after the directives (my example is for the Renci SSH):
using System;
using Renci;//FOR THE SSH
using System.Net;//FOR THE ADDRESS TRANSLATION
using System.Reflection;//FOR THE Assembly
//+ref>"C:\Program Files (x86)\Microsoft\ILMerge\Renci.SshNet.dll"
//+res>"C:\Program Files (x86)\Microsoft\ILMerge\Renci.SshNet.dll"
//+ico>"C:\Program Files (x86)\Microsoft CAPICOM 2.1.0.2 SDK\Samples\c_sharp\xmldsig\resources\Traffic.ico"
The ref, res and ico tags will be picked up by the .vbs script below to form the csc command.
Then add the assembly resolver caller in the Main:
public static void Main(string[] args)
{
AppDomain.CurrentDomain.AssemblyResolve += new ResolveEventHandler(CurrentDomain_AssemblyResolve);
.
...and add the resolver itself somewhere in the class:
static Assembly CurrentDomain_AssemblyResolve(object sender, ResolveEventArgs args)
{
String resourceName = new AssemblyName(args.Name).Name + ".dll";
using (var stream = Assembly.GetExecutingAssembly().GetManifestResourceStream(resourceName))
{
Byte[] assemblyData = new Byte[stream.Length];
stream.Read(assemblyData, 0, assemblyData.Length);
return Assembly.Load(assemblyData);
}
}
I name the vbs script to match the .cs filename (e.g. ssh.vbs looks for ssh.cs); this makes running the script numerous times a lot easier, but if you aren't an idiot like me then a generic script could pick up the target .cs file from a drag-and-drop:
Dim name_,oShell,fso
Set oShell = CreateObject("Shell.Application")
Set fso = CreateObject("Scripting.fileSystemObject")
'TAKE THE VBS SCRIPT NAME AS THE TARGET FILE NAME
'################################################
name_ = Split(wscript.ScriptName, ".")(0)
'GET THE EXTERNAL DLL's AND ICON NAMES FROM THE .CS FILE
'#######################################################
Const OPEN_FILE_FOR_READING = 1
Set objInputFile = fso.OpenTextFile(name_ & ".cs", 1)
'READ EVERYTHING INTO AN ARRAY
'#############################
inputData = Split(objInputFile.ReadAll, vbNewline)
For each strData In inputData
if left(strData,7)="//+ref>" then
csc_references = csc_references & " /reference:" & trim(replace(strData,"//+ref>","")) & " "
end if
if left(strData,7)="//+res>" then
csc_resources = csc_resources & " /resource:" & trim(replace(strData,"//+res>","")) & " "
end if
if left(strData,7)="//+ico>" then
csc_icon = " /win32icon:" & trim(replace(strData,"//+ico>","")) & " "
end if
Next
objInputFile.Close
'COMPILE THE FILE
'################
oShell.ShellExecute "c:\windows\microsoft.net\framework\v3.5\csc.exe", "/warn:1 /target:exe " & csc_references & csc_resources & csc_icon & " " & name_ & ".cs", "", "runas", 2
WScript.Quit(0)

If you are using .NET Core 3.0
You can do this with the dotnet publish command with PublishSingleFile property:
dotnet publish -r win-x64 -c Release /p:PublishSingleFile=true
The only downside is you end up with a single EXE file with a huge size.

It's possible but not all that easy, to create a hybrid native/managed assembly in C#. Were you using C++ instead it'd be a lot easier, as the Visual C++ compiler can create hybrid assemblies as easily as anything else.
Unless you have a strict requirement to produce a hybrid assembly, I'd agree with MusiGenesis that this isn't really worth the trouble to do with C#. If you need to do it, perhaps look at moving to C++/CLI instead.

Generally you would need some form of post build tool to perform an assembly merge like you are describing. There is a free tool called Eazfuscator (eazfuscator.blogspot.com/) which is designed for bytecode mangling that also handles assembly merging. You can add this into a post build command line with Visual Studio to merge your assemblies, but your mileage will vary due to issues that will arise in any non trival assembly merging scenarios.
You could also check to see if the build make untility NANT has the ability to merge assemblies after building, but I am not familiar enough with NANT myself to say whether the functionality is built in or not.
There are also many many Visual Studio plugins that will perform assembly merging as part of building the application.
Alternatively if you don't need this to be done automatically, there are a number of tools like ILMerge that will merge .net assemblies into a single file.
The biggest issue I've had with merging assemblies is if they use any similar namespaces. Or worse, reference different versions of the same dll (my problems were generally with the NUnit dll files).

Try this:
https://github.com/ytk2128/dll-merger
here you can merge all 32 bit dlls/exe - even its not ".net" dlls - so for me better then ilmerge for example ...

Embed .net dll in c# .exe [duplicate]

Is it possible to embed a pre-existing DLL into a compiled C# executable (so that you only have one file to distribute)? If it is possible, how would one go about doing it?
Normally, I'm cool with just leaving the DLLs outside and having the setup program handle everything, but there have been a couple of people at work who have asked me this and I honestly don't know.

I highly recommend to use Costura.Fody - by far the best and easiest way to embed resources in your assembly. It's available as NuGet package.
Install-Package Costura.Fody
After adding it to the project, it will automatically embed all references that are copied to the output directory into your main assembly. You might want to clean the embedded files by adding a target to your project:
Install-CleanReferencesTarget
You'll also be able to specify whether to include the pdb's, exclude certain assemblies, or extracting the assemblies on the fly. As far as I know, also unmanaged assemblies are supported.
Update
Currently, some people are trying to add support for DNX.
Update 2
For the lastest Fody version, you will need to have MSBuild 16 (so Visual Studio 2019). Fody version 4.2.1 will do MSBuild 15. (reference: Fody is only supported on MSBuild 16 and above. Current version: 15)

Just right-click your project in Visual Studio, choose Project Properties -> Resources -> Add Resource -> Add Existing File…
And include the code below to your App.xaml.cs or equivalent.
public App()
{
AppDomain.CurrentDomain.AssemblyResolve +=new ResolveEventHandler(CurrentDomain_AssemblyResolve);
}
System.Reflection.Assembly CurrentDomain_AssemblyResolve(object sender, ResolveEventArgs args)
{
string dllName = args.Name.Contains(',') ? args.Name.Substring(0, args.Name.IndexOf(',')) : args.Name.Replace(".dll","");
dllName = dllName.Replace(".", "_");
if (dllName.EndsWith("_resources")) return null;
System.Resources.ResourceManager rm = new System.Resources.ResourceManager(GetType().Namespace + ".Properties.Resources", System.Reflection.Assembly.GetExecutingAssembly());
byte[] bytes = (byte[])rm.GetObject(dllName);
return System.Reflection.Assembly.Load(bytes);
}
Here's my original blog post:
http://codeblog.larsholm.net/2011/06/embed-dlls-easily-in-a-net-assembly/

If they're actually managed assemblies, you can use ILMerge. For native DLLs, you'll have a bit more work to do.
See also: How can a C++ windows dll be merged into a C# application exe?

Yes, it is possible to merge .NET executables with libraries. There are multiple tools available to get the job done:
ILMerge is a utility that can be used to merge multiple .NET assemblies into a single assembly.
Mono mkbundle, packages an exe and all assemblies with libmono into a single binary package.
IL-Repack is a FLOSS alterantive to ILMerge, with some additional features.
In addition this can be combined with the Mono Linker, which does remove unused code and therefor makes the resulting assembly smaller.
Another possibility is to use .NETZ, which does not only allow compressing of an assembly, but also can pack the dlls straight into the exe. The difference to the above mentioned solutions is that .NETZ does not merge them, they stay separate assemblies but are packed into one package.
.NETZ is a open source tool that compresses and packs the Microsoft .NET Framework executable (EXE, DLL) files in order to make them smaller.

ILMerge can combine assemblies to one single assembly provided the assembly has only managed code. You can use the commandline app, or add reference to the exe and programmatically merge. For a GUI version there is Eazfuscator, and also .Netz both of which are free. Paid apps include BoxedApp and SmartAssembly.
If you have to merge assemblies with unmanaged code, I would suggest SmartAssembly. I never had hiccups with SmartAssembly but with all others. Here, it can embed the required dependencies as resources to your main exe.
You can do all this manually not needing to worry if assembly is managed or in mixed mode by embedding dll to your resources and then relying on AppDomain's Assembly ResolveHandler. This is a one stop solution by adopting the worst case, ie assemblies with unmanaged code.
static void Main()
{
AppDomain.CurrentDomain.AssemblyResolve += (sender, args) =>
{
string assemblyName = new AssemblyName(args.Name).Name;
if (assemblyName.EndsWith(".resources"))
return null;
string dllName = assemblyName + ".dll";
string dllFullPath = Path.Combine(GetMyApplicationSpecificPath(), dllName);
using (Stream s = Assembly.GetEntryAssembly().GetManifestResourceStream(typeof(Program).Namespace + ".Resources." + dllName))
{
byte[] data = new byte[stream.Length];
s.Read(data, 0, data.Length);
//or just byte[] data = new BinaryReader(s).ReadBytes((int)s.Length);
File.WriteAllBytes(dllFullPath, data);
}
return Assembly.LoadFrom(dllFullPath);
};
}
The key here is to write the bytes to a file and load from its location. To avoid chicken and egg problem, you have to ensure you declare the handler before accessing assembly and that you do not access the assembly members (or instantiate anything that has to deal with the assembly) inside the loading (assembly resolving) part. Also take care to ensure GetMyApplicationSpecificPath() is not any temp directory since temp files could be attempted to get erased by other programs or by yourself (not that it will get deleted while your program is accessing the dll, but at least its a nuisance. AppData is good location). Also note that you have to write the bytes each time, you cant load from location just 'cos the dll already resides there.
For managed dlls, you need not write bytes, but directly load from the location of the dll, or just read the bytes and load the assembly from memory. Like this or so:
using (Stream s = Assembly.GetEntryAssembly().GetManifestResourceStream(typeof(Program).Namespace + ".Resources." + dllName))
{
byte[] data = new byte[stream.Length];
s.Read(data, 0, data.Length);
return Assembly.Load(data);
}
//or just
return Assembly.LoadFrom(dllFullPath); //if location is known.
If the assembly is fully unmanaged, you can see this link or this as to how to load such dlls.

.NET Core 3.0 natively supports compiling to a single .exe
The feature is enabled by the usage of the following property in your project file (.csproj):
<PropertyGroup>
<PublishSingleFile>true</PublishSingleFile>
</PropertyGroup>
This is done without any external tool.
See my answer for this question for further details.

The excerpt by Jeffrey Richter is very good. In short, add the libraries as embedded resources and add a callback before anything else. Here is a version of the code (found in the comments of his page) that I put at the start of Main method for a console app (just make sure that any calls that use the libraries are in a different method to Main).
AppDomain.CurrentDomain.AssemblyResolve += (sender, bargs) =>
{
String dllName = new AssemblyName(bargs.Name).Name + ".dll";
var assem = Assembly.GetExecutingAssembly();
String resourceName = assem.GetManifestResourceNames().FirstOrDefault(rn => rn.EndsWith(dllName));
if (resourceName == null) return null; // Not found, maybe another handler will find it
using (var stream = assem.GetManifestResourceStream(resourceName))
{
Byte[] assemblyData = new Byte[stream.Length];
stream.Read(assemblyData, 0, assemblyData.Length);
return Assembly.Load(assemblyData);
}
};

To expand on #Bobby's asnwer above. You can edit your .csproj to use IL-Repack to automatically package all files into a single assembly when you build.
Install the nuget ILRepack.MSBuild.Task package with Install-Package ILRepack.MSBuild.Task
Edit the AfterBuild section of your .csproj
Here is a simple sample that merges ExampleAssemblyToMerge.dll into your project output.
<!-- ILRepack -->
<Target Name="AfterBuild" Condition="'$(Configuration)' == 'Release'">
<ItemGroup>
<InputAssemblies Include="$(OutputPath)\$(AssemblyName).exe" />
<InputAssemblies Include="$(OutputPath)\ExampleAssemblyToMerge.dll" />
</ItemGroup>
<ILRepack
Parallel="true"
Internalize="true"
InputAssemblies="#(InputAssemblies)"
TargetKind="Exe"
OutputFile="$(OutputPath)\$(AssemblyName).exe"
/>
</Target>

The following method DO NOT use external tools and AUTOMATICALLY include all needed DLL (no manual action required, everything done at compilation)
I read a lot of answer here saying to use ILMerge, ILRepack or Jeffrey Ritcher method but none of that worked with WPF applications nor was easy to use.
When you have a lot of DLL it can be hard to manually include the one you need in your exe. The best method i found was explained by Wegged here on StackOverflow
Copy pasted his answer here for clarity (all credit to Wegged)
1) Add this to your .csproj file:
<Target Name="AfterResolveReferences">
<ItemGroup>
<EmbeddedResource Include="#(ReferenceCopyLocalPaths)" Condition="'%(ReferenceCopyLocalPaths.Extension)' == '.dll'">
<LogicalName>%(ReferenceCopyLocalPaths.DestinationSubDirectory)%(ReferenceCopyLocalPaths.Filename)%(ReferenceCopyLocalPaths.Extension)</LogicalName>
</EmbeddedResource>
</ItemGroup>
</Target>
2) Make your Main Program.cs look like this:
[STAThreadAttribute]
public static void Main()
{
AppDomain.CurrentDomain.AssemblyResolve += OnResolveAssembly;
App.Main();
}
3) Add the OnResolveAssembly method:
private static Assembly OnResolveAssembly(object sender, ResolveEventArgs args)
{
Assembly executingAssembly = Assembly.GetExecutingAssembly();
AssemblyName assemblyName = new AssemblyName(args.Name);
var path = assemblyName.Name + ".dll";
if (assemblyName.CultureInfo.Equals(CultureInfo.InvariantCulture) == false) path = String.Format(#"{0}\{1}", assemblyName.CultureInfo, path);
using (Stream stream = executingAssembly.GetManifestResourceStream(path))
{
if (stream == null) return null;
var assemblyRawBytes = new byte[stream.Length];
stream.Read(assemblyRawBytes, 0, assemblyRawBytes.Length);
return Assembly.Load(assemblyRawBytes);
}
}

You could add the DLLs as embedded resources, and then have your program unpack them into the application directory on startup (after checking to see if they're there already).
Setup files are so easy to make, though, that I don't think this would be worth it.
EDIT: This technique would be easy with .NET assemblies. With non-.NET DLLs it would be a lot more work (you'd have to figure out where to unpack the files and register them and so on).

Another product that can handle this elegantly is SmartAssembly, at SmartAssembly.com. This product will, in addition to merging all dependencies into a single DLL, (optionally) obfuscate your code, remove extra meta-data to reduce the resulting file size, and can also actually optimize the IL to increase runtime performance.
There is also some kind of global exception handling/reporting feature it adds to your software (if desired) that could be useful. I believe it also has a command-line API so you can make it part of your build process.

Neither the ILMerge approach nor Lars Holm Jensen's handling the AssemblyResolve event will work for a plugin host. Say executable H loads assembly P dynamically and accesses it via interface IP defined in an separate assembly. To embed IP into H one shall need a little modification to Lars's code:
Dictionary<string, Assembly> loaded = new Dictionary<string,Assembly>();
AppDomain.CurrentDomain.AssemblyResolve += (sender, args) =>
{ Assembly resAssembly;
string dllName = args.Name.Contains(",") ? args.Name.Substring(0, args.Name.IndexOf(',')) : args.Name.Replace(".dll","");
dllName = dllName.Replace(".", "_");
if ( !loaded.ContainsKey( dllName ) )
{ if (dllName.EndsWith("_resources")) return null;
System.Resources.ResourceManager rm = new System.Resources.ResourceManager(GetType().Namespace + ".Properties.Resources", System.Reflection.Assembly.GetExecutingAssembly());
byte[] bytes = (byte[])rm.GetObject(dllName);
resAssembly = System.Reflection.Assembly.Load(bytes);
loaded.Add(dllName, resAssembly);
}
else
{ resAssembly = loaded[dllName]; }
return resAssembly;
};
The trick to handle repeated attempts to resolve the same assembly and return the existing one instead of creating a new instance.
EDIT:
Lest it spoil .NET's serialization, make sure to return null for all assemblies not embedded in yours, thereby defaulting to the standard behaviour. You can get a list of these libraries by:
static HashSet<string> IncludedAssemblies = new HashSet<string>();
string[] resources = System.Reflection.Assembly.GetExecutingAssembly().GetManifestResourceNames();
for(int i = 0; i < resources.Length; i++)
{ IncludedAssemblies.Add(resources[i]); }
and just return null if the passed assembly does not belong to IncludedAssemblies .

It may sound simplistic, but WinRar gives the option to compress a bunch of files to a self-extracting executable.
It has lots of configurable options: final icon, extract files to given path, file to execute after extraction, custom logo/texts for popup shown during extraction, no popup window at all, license agreement text, etc.
May be useful in some cases.

I use the csc.exe compiler called from a .vbs script.
In your xyz.cs script, add the following lines after the directives (my example is for the Renci SSH):
using System;
using Renci;//FOR THE SSH
using System.Net;//FOR THE ADDRESS TRANSLATION
using System.Reflection;//FOR THE Assembly
//+ref>"C:\Program Files (x86)\Microsoft\ILMerge\Renci.SshNet.dll"
//+res>"C:\Program Files (x86)\Microsoft\ILMerge\Renci.SshNet.dll"
//+ico>"C:\Program Files (x86)\Microsoft CAPICOM 2.1.0.2 SDK\Samples\c_sharp\xmldsig\resources\Traffic.ico"
The ref, res and ico tags will be picked up by the .vbs script below to form the csc command.
Then add the assembly resolver caller in the Main:
public static void Main(string[] args)
{
AppDomain.CurrentDomain.AssemblyResolve += new ResolveEventHandler(CurrentDomain_AssemblyResolve);
.
...and add the resolver itself somewhere in the class:
static Assembly CurrentDomain_AssemblyResolve(object sender, ResolveEventArgs args)
{
String resourceName = new AssemblyName(args.Name).Name + ".dll";
using (var stream = Assembly.GetExecutingAssembly().GetManifestResourceStream(resourceName))
{
Byte[] assemblyData = new Byte[stream.Length];
stream.Read(assemblyData, 0, assemblyData.Length);
return Assembly.Load(assemblyData);
}
}
I name the vbs script to match the .cs filename (e.g. ssh.vbs looks for ssh.cs); this makes running the script numerous times a lot easier, but if you aren't an idiot like me then a generic script could pick up the target .cs file from a drag-and-drop:
Dim name_,oShell,fso
Set oShell = CreateObject("Shell.Application")
Set fso = CreateObject("Scripting.fileSystemObject")
'TAKE THE VBS SCRIPT NAME AS THE TARGET FILE NAME
'################################################
name_ = Split(wscript.ScriptName, ".")(0)
'GET THE EXTERNAL DLL's AND ICON NAMES FROM THE .CS FILE
'#######################################################
Const OPEN_FILE_FOR_READING = 1
Set objInputFile = fso.OpenTextFile(name_ & ".cs", 1)
'READ EVERYTHING INTO AN ARRAY
'#############################
inputData = Split(objInputFile.ReadAll, vbNewline)
For each strData In inputData
if left(strData,7)="//+ref>" then
csc_references = csc_references & " /reference:" & trim(replace(strData,"//+ref>","")) & " "
end if
if left(strData,7)="//+res>" then
csc_resources = csc_resources & " /resource:" & trim(replace(strData,"//+res>","")) & " "
end if
if left(strData,7)="//+ico>" then
csc_icon = " /win32icon:" & trim(replace(strData,"//+ico>","")) & " "
end if
Next
objInputFile.Close
'COMPILE THE FILE
'################
oShell.ShellExecute "c:\windows\microsoft.net\framework\v3.5\csc.exe", "/warn:1 /target:exe " & csc_references & csc_resources & csc_icon & " " & name_ & ".cs", "", "runas", 2
WScript.Quit(0)

If you are using .NET Core 3.0
You can do this with the dotnet publish command with PublishSingleFile property:
dotnet publish -r win-x64 -c Release /p:PublishSingleFile=true
The only downside is you end up with a single EXE file with a huge size.

It's possible but not all that easy, to create a hybrid native/managed assembly in C#. Were you using C++ instead it'd be a lot easier, as the Visual C++ compiler can create hybrid assemblies as easily as anything else.
Unless you have a strict requirement to produce a hybrid assembly, I'd agree with MusiGenesis that this isn't really worth the trouble to do with C#. If you need to do it, perhaps look at moving to C++/CLI instead.

Generally you would need some form of post build tool to perform an assembly merge like you are describing. There is a free tool called Eazfuscator (eazfuscator.blogspot.com/) which is designed for bytecode mangling that also handles assembly merging. You can add this into a post build command line with Visual Studio to merge your assemblies, but your mileage will vary due to issues that will arise in any non trival assembly merging scenarios.
You could also check to see if the build make untility NANT has the ability to merge assemblies after building, but I am not familiar enough with NANT myself to say whether the functionality is built in or not.
There are also many many Visual Studio plugins that will perform assembly merging as part of building the application.
Alternatively if you don't need this to be done automatically, there are a number of tools like ILMerge that will merge .net assemblies into a single file.
The biggest issue I've had with merging assemblies is if they use any similar namespaces. Or worse, reference different versions of the same dll (my problems were generally with the NUnit dll files).

Try this:
https://github.com/ytk2128/dll-merger
here you can merge all 32 bit dlls/exe - even its not ".net" dlls - so for me better then ilmerge for example ...

How to check if an assembly was built using Debug or Release configuration?

I'm starting deployment of my web application and I need to guarantee that all the assemblies that are going to be deployed were built using Release configuration. Our system was developed using C#/.Net 3.5.
Is there any way to achieve this?

Check this. The idea is that you get the list of assembly attributes using Assembly.GetCustomAttributes() and search for DebuggableAttribute and then find if such attribute has IsJITTrackingEnabled property set.
public bool IsAssemblyDebugBuild(Assembly assembly)
{
return assembly.GetCustomAttributes(false).OfType<DebuggableAttribute>().Any(da => da.IsJITTrackingEnabled);
}

I loved that David suggestion, but you could also go this way (AssemblyInfo.cs):
#if DEBUG
[assembly: AssemblyDescription("Your application assembly (DEBUG version)")]
#else if RELEASE
[assembly: AssemblyDescription("Your application assembly (RELEASE version)")]
#endif
This is more human friendly, as anyone can right-click that assembly, to select Properties and go to Details tab.

If it is your assembly I believe using the AssemblyConfiguration attribute is the best approach. It is documented as "Specifies the build configuration, such as retail or debug, for an assembly."
Depending on your build configurations you might have code like this:
#if DEBUG
[assembly: AssemblyConfiguration("Debug")]
#else
[assembly: AssemblyConfiguration("Release")]
#endif
Then check the assembly attribute:
public static bool IsAssemblyConfiguration(Assembly assembly, string configuration)
{
var attributes = assembly.GetCustomAttributes(typeof(AssemblyConfigurationAttribute), false);
if (attributes.Length == 1)
{
var assemblyConfiguration = attributes[0] as AssemblyConfigurationAttribute;
if (assemblyConfiguration != null)
{
return assemblyConfiguration.Configuration.Equals(configuration, StringComparison.InvariantCultureIgnoreCase);
}
}
return true;
}
(I know R. Schreurs comment at Rubens Farias says the same, but I've find this information somewhere else before seeing the comment so I believe this requires a more important entry like a full response instead of a comment)

If you have Reflector installed you can also click on the assembly and look for the debuggable attribute ([assembly: Debuggable()]) in the Disassembler pane.

Assuming only Debug and Release configuration, DEBUG symbol is by default defined with Debug configuration, so the code below in AssemblyInfo.cs (under Properties folder).
#if DEBUG
[assembly: AssemblyTitle("Debug")]
#else
[assembly: AssemblyTitle("Release")]
#endif
I use AssemblyTitle over AssemblyDescription as it will show up on my Windows 7 file explorer properties:
For those who like David and stevieg's answer, here is a LINQPad script written in C#. To use the script, you need to download LINQPad 5 and make sure C# Program is selected as shown in screenshot below.
Simply replace DLL_FOLDER_PATH to point to folder containing the DLLs to be inspected.
// TODO - Specify your folder containing DLLs to inspect
static string DLL_FOLDER_PATH = #"C:\Program Files (x86)\Reference Assemblies\Microsoft\Framework\.NETFramework\v4.0";
void Main()
{
(from dllPath in Directory.GetFiles(DLL_FOLDER_PATH, "*.dll")
let assembly = dllPath.SafeLoad()
let build = assembly == null ? "Error" : (dllPath.SafeLoad().IsAssemblyDebugBuild() ? "Debug" : "Release")
select new {
Assembly_Path = dllPath,
Build = build,
}).Dump();
}
static class Extensions {
public static bool IsAssemblyDebugBuild(this Assembly assembly)
{
return assembly.GetCustomAttributes(false).OfType<DebuggableAttribute>().Select(da => da.IsJITTrackingEnabled).FirstOrDefault();
}
public static Assembly SafeLoad(this string path){
try{
return Assembly.LoadFrom(path);
}
catch {
return null;
}
}
}
LINQPAD 5 can be downloaded here.

Don't deploy to production via Visual Studio. Look into Continuous Integration and scripted builds (such as with NAnt, or perhaps something more legible like FAKE).
The F5 Key Is Not a Build Process
To detractors who believe that this does not answer the question, the OP wrote:
...I need to guarantee that all the assemblies that are going to be
deployed were built using Release configuration.
To guarantee that, use a build server such as TeamCity and possibly a release management tool like Octopus Deploy. Lock down your production systems so that developers must go through the official build process.

Automatically update version number

I would like the version property of my application to be incremented for each build but I'm not sure on how to enable this functionality in Visual Studio (2005/2008). I have tried to specify the AssemblyVersion as 1.0.* but it doesn't get me exactly what I want.
I'm also using a settings file and in earlier attempts when the assembly version changed my settings got reset to the default since the application looked for the settings file in another directory.
I would like to be able to display a version number in the form of 1.1.38 so when a user finds a problem I can log the version they are using as well as tell them to upgrade if they have an old release.
A short explanation of how the versioning works would also be appreciated. When does the build and revision number get incremented?

With the "Built in" stuff, you can't, as using 1.0.* or 1.0.0.* will replace the revision and build numbers with a coded date/timestamp, which is usually also a good way.
For more info, see the Assembly Linker Documentation in the /v tag.
As for automatically incrementing numbers, use the AssemblyInfo Task:
AssemblyInfo Task
This can be configured to automatically increment the build number.
There are 2 Gotchas:
Each of the 4 numbers in the Version string is limited to 65535. This is a Windows Limitation and unlikely to get fixed.
Why are build numbers limited to 65535?
Using with with Subversion requires a small change:
Using MSBuild to generate assembly version info at build time (including SubVersion fix)
Retrieving the Version number is then quite easy:
Version v = Assembly.GetExecutingAssembly().GetName().Version;
string About = string.Format(CultureInfo.InvariantCulture, #"YourApp Version {0}.{1}.{2} (r{3})", v.Major, v.Minor, v.Build, v.Revision);
And, to clarify: In .net or at least in C#, the build is actually the THIRD number, not the fourth one as some people (for example Delphi Developers who are used to Major.Minor.Release.Build) might expect.
In .net, it's Major.Minor.Build.Revision.

VS.NET defaults the Assembly version to 1.0.* and uses the following logic when auto-incrementing: it sets the build part to the number of days since January 1st, 2000, and sets the revision part to the number of seconds since midnight, local time, divided by two. See this MSDN article.
Assembly version is located in an assemblyinfo.vb or assemblyinfo.cs file. From the file:
' Version information for an assembly consists of the following four values:
'
' Major Version
' Minor Version
' Build Number
' Revision
'
' You can specify all the values or you can default the Build and Revision Numbers
' by using the '*' as shown below:
' <Assembly: AssemblyVersion("1.0.*")>
<Assembly: AssemblyVersion("1.0.0.0")>
<Assembly: AssemblyFileVersion("1.0.0.0")>

I have found that it works well to simply display the date of the last build using the following wherever a product version is needed:
System.IO.File.GetLastWriteTime(System.Reflection.Assembly.GetExecutingAssembly().Location).ToString("yyyy.MM.dd.HH.mm.ss")
Rather than attempting to get the version from something like the following:
System.Reflection.Assembly assembly = System.Reflection.Assembly.GetExecutingAssembly();
object[] attributes = assembly.GetCustomAttributes(typeof(System.Reflection.AssemblyFileVersionAttribute), false);
object attribute = null;
if (attributes.Length > 0)
{
attribute = attributes[0] as System.Reflection.AssemblyFileVersionAttribute;
}

[Visual Studio 2017, .csproj properties]
To automatically update your PackageVersion/Version/AssemblyVersion property (or any other property), first, create a new Microsoft.Build.Utilities.Task class that will get your current build number and send back the updated number (I recommend to create a separate project just for that class).
I manually update the major.minor numbers, but let MSBuild to automatically update the build number (1.1.1, 1.1.2, 1.1.3, etc. :)
using Microsoft.Build.Framework;
using System;
using System.Collections.Generic;
using System.Text;
public class RefreshVersion : Microsoft.Build.Utilities.Task
{
[Output]
public string NewVersionString { get; set; }
public string CurrentVersionString { get; set; }
public override bool Execute()
{
Version currentVersion = new Version(CurrentVersionString ?? "1.0.0");
DateTime d = DateTime.Now;
NewVersionString = new Version(currentVersion.Major,
currentVersion.Minor, currentVersion.Build+1).ToString();
return true;
}
}
Then call your recently created Task on MSBuild process adding the next code on your .csproj file:
<Project Sdk="Microsoft.NET.Sdk">
...
<UsingTask TaskName="RefreshVersion" AssemblyFile="$(MSBuildThisFileFullPath)\..\..\<dll path>\BuildTasks.dll" />
<Target Name="RefreshVersionBuildTask" BeforeTargets="Pack" Condition="'$(Configuration)|$(Platform)'=='Release|AnyCPU'">
<RefreshVersion CurrentVersionString="$(PackageVersion)">
<Output TaskParameter="NewVersionString" PropertyName="NewVersionString" />
</RefreshVersion>
<Message Text="Updating package version number to $(NewVersionString)..." Importance="high" />
<XmlPoke XmlInputPath="$(MSBuildProjectDirectory)\mustache.website.sdk.dotNET.csproj" Query="/Project/PropertyGroup/PackageVersion" Value="$(NewVersionString)" />
</Target>
...
<PropertyGroup>
..
<PackageVersion>1.1.4</PackageVersion>
..
When picking Visual Studio Pack project option (just change to BeforeTargets="Build" for executing the task before Build) the RefreshVersion code will be triggered to calculate the new version number, and XmlPoke task will update your .csproj property accordingly (yes, it will modify the file).
When working with NuGet libraries, I also send the package to NuGet repository by just adding the next build task to the previous example.
<Message Text="Uploading package to NuGet..." Importance="high" />
<Exec WorkingDirectory="$(MSBuildProjectDirectory)\bin\release" Command="c:\nuget\nuget push *.nupkg -Source https://www.nuget.org/api/v2/package" IgnoreExitCode="true" />
c:\nuget\nuget is where I have the NuGet client (remember to save your NuGet API key by calling nuget SetApiKey <my-api-key> or to include the key on the NuGet push call).
Just in case it helps someone ^_^.

What source control system are you using?
Almost all of them have some form of $ Id $ tag that gets expanded when the file is checked in.
I usually use some form of hackery to display this as the version number.
The other alternative is use to use the date as the build number: 080803-1448

Some time ago I wrote a quick and dirty exe that would update the version #'s in an assemblyinfo.{cs/vb} - I also have used rxfind.exe (a simple and powerful regex-based search replace tool) to do the update from a command line as part of the build process. A couple of other helpfule hints:
separate the assemblyinfo into product parts (company name, version, etc.) and assembly specific parts (assembly name etc.). See here
Also - i use subversion, so I found it helpful to set the build number to subversion revision number thereby making it really easy to always get back to the codebase that generated the assembly (e.g. 1.4.100.1502 was built from revision 1502).

If you want an auto incrementing number that updates each time a compilation is done, you can use VersionUpdater from a pre-build event. Your pre-build event can check the build configuration if you prefer so that the version number will only increment for a Release build (for example).

Here is a handcranked alternative option: This is a quick-and-dirty PowerShell snippet I wrote that gets called from a pre-build step on our Jenkins build system.
It sets the last digit of the AssemblyVersion and AssemblyFileVersion to the value of the BUILD_NUMBER environment variable which is automatically set by the build system.
if (Test-Path env:BUILD_NUMBER) {
Write-Host "Updating AssemblyVersion to $env:BUILD_NUMBER"
# Get the AssemblyInfo.cs
$assemblyInfo = Get-Content -Path .\MyShinyApplication\Properties\AssemblyInfo.cs
# Replace last digit of AssemblyVersion
$assemblyInfo = $assemblyInfo -replace
"^\[assembly: AssemblyVersion\(`"([0-9]+)\.([0-9]+)\.([0-9]+)\.[0-9]+`"\)]",
('[assembly: AssemblyVersion("$1.$2.$3.' + $env:BUILD_NUMBER + '")]')
Write-Host ($assemblyInfo -match '^\[assembly: AssemblyVersion')
# Replace last digit of AssemblyFileVersion
$assemblyInfo = $assemblyInfo -replace
"^\[assembly: AssemblyFileVersion\(`"([0-9]+)\.([0-9]+)\.([0-9]+)\.[0-9]+`"\)]",
('[assembly: AssemblyFileVersion("$1.$2.$3.' + $env:BUILD_NUMBER + '")]')
Write-Host ($assemblyInfo -match '^\[assembly: AssemblyFileVersion')
$assemblyInfo | Set-Content -Path .\MyShinyApplication\Properties\AssemblyInfo.cs -Encoding UTF8
} else {
Write-Warning "BUILD_NUMBER is not set."
}