Search Results

Search found 5623 results on 225 pages for 'inline assembly'.

Page 1/225 | 1 2 3 4 5 6 7 8 9 10 11 12  | Next Page >

  • c# - can you make a "weak" assembly reference to a strong named assembly

    - by Tim
    hi, for various reasons i would rather not use strong named (signed) assemblies in my project. however, one of the projects is referenced by a sharepoint web part which means it must be signed. is it possible to have this assembly signed but when I reference it from other projects, to do so using a non-strong reference. this would give me the advantages of having a non-signed assembly for the rest of my code but still allow it to be loaded by sharepoint Tim

    Read the article

  • How to Specify AssemblyKeyFile Attribute in .NET Assembly and Issues

    How to specify strong key file in assembly? Answer: You can specify snk file information using following line [assembly: AssemblyKeyFile(@"c:\Key2.snk")] Where to specify an strong key file (snk file)? Answer: You have two options to specify the AssemblyKeyFile infromation. 1. In class 2. In AssemblyInfo.cs [assembly: AssemblyKeyFile(@"c:\Key2.snk")] 1. In Class you must specify above line before defining namespace of the class and after all the imports or usings Example: See Line 7 in bellow sample class using System;using System.Collections.Generic;using System.Linq;using System.Text;using System.Reflection;[assembly: AssemblyKeyFile(@"c:\Key1.snk")]namespace Csharp3Part1{ class Person { public string GetName() { return "Smith"; } }}2. In AssemblyInfo.cs You can aslo specify assembly information in AssemblyInfo.cs Example: See Line 16 in bellow sample AssemblyInfo.csusing System.Reflection;using System.Runtime.CompilerServices;using System.Runtime.InteropServices;// General Information about an assembly is controlled through the following// set of attributes. Change these attribute values to modify the information// associated with an assembly.[assembly: AssemblyTitle("Csharp3Part1")][assembly: AssemblyDescription("")][assembly: AssemblyConfiguration("")][assembly: AssemblyCompany("Deloitte")][assembly: AssemblyProduct("Csharp3Part1")][assembly: AssemblyCopyright("Copyright © Deloitte 2009")][assembly: AssemblyTrademark("")][assembly: AssemblyCulture("")][assembly: AssemblyKeyFile(@"c:\Key1.snk")]// Setting ComVisible to false makes the types in this assembly not visible// to COM components. If you need to access a type in this assembly from// COM, set the ComVisible attribute to true on that type.[assembly: ComVisible(false)]// The following GUID is for the ID of the typelib if this project is exposed to COM[assembly: Guid("4350396f-1a5c-4598-a79f-2e1f219654f3")]// 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")]Issues:You should not sepcify this in following ways. 1. In multiple classes. 2. In both class and AssemblyInfo.cs If you did wrong in either one of the above ways, Visual Studio or C#/VB.NET compilers shows following Error Duplicate 'AssemblyKeyFile' attribute and warning Use command line option '/keyfile' or appropriate project settings instead of 'AssemblyKeyFile' To avoid this, Please specity your keyfile information only one time either only in one class or in AssemblyInfo.cs file. It is suggested to specify this at AssemblyInfo.cs file You might also encounter the errors like Error: type or namespace name 'AssemblyKeyFileAttribute' and 'AssemblyKeyFile' could not be found. Solution. Please find herespan.fullpost {display:none;} span.fullpost {display:none;}

    Read the article

  • More information wanted on error: CREATE ASSEMBLY for assembly failed because assembly failed verif

    - by turnip.cyberveggie
    I have a small application that uses SQL Server 2005 Express with CLR stored procedures. It has been successfully installed and runs on many computers running XP and Vista. To create the assembly the following SQL is executed (names changed to protect the innocent): CREATE ASSEMBLY myAssemblyName FROM 'c:\pathtoAssembly\myAssembly.dll' On one computer (a test machine that reflects other computers targeted for installation) that is running Vista and has some very aggressive security policy restrictions I receive the following error: << Start Error Message Msg 6218, Level 16, State 2, Server domain\servername, Line 2 CREATE ASSEMBLY for assembly 'myAssembly' failed because assembly 'myAssembly' failed verification. Check if the referenced assemblies are up-to-date and trusted (for external_access or unsafe) to execute in the database. CLR Verifier error messages if any will follow this message [ : myProcSupport.Axis::Proc1][mdToken=0x6000004] [HRESULT 0x8007000E] - Not enough storage is available to complete this operation. [ : myProcSupport.Axis::Proc2][mdToken=0x6000005] [HRESULT 0x8007000E] - Not enough storage is available to complete this operation. [ : myProcSupport.Axis::Proc3][mdToken=0x6000006] [HRESULT 0x8007000E] - Not enough storage is available to complete this operation. [ : myProcSupport.Axis::.ctor][mdToken=0x600000a] [HRESULT 0x8007000E] - Not enough storage is available to complete this operation. [ : myProcSupport.Axis::Proc4][mdToken=0x6000001] [HRESULT 0x8007000E] - Not enough storage is available to complete this operation. [ : myProcSupport.Axis::Proc5][mdToken=0x6000002] [HRESULT 0x8007000E] - Not enough storage is available to complete this operation. [ : myProcSupport.Axis::Proc6][mdToken=0x6000007] [HRESULT 0x8007000E] - Not enough storage is available to complete this operation. [ : myProcSupport.Axis::Proc7][mdToken=0x6000008] [HRESULT 0x8007000E] - Not enough storage is available to complete this operation. [ : myProcSupport.Axis::Proc8][mdToken=0x6000009] [HRESULT 0x8007000E] - Not enough storage is available to complete this operation. [ : myProcSupport.Axis::Proc8][mdToken=0x600000b] [HRESULT 0x8007000E] - Not enough storage is available to complete this operation. [ : myProcSupport.Axis::Proc9][mdToken=0x600000c] [HRESULT 0x8007000E] - Not enough storage is available to complete this operation.... << End Error Message The C# DLL is defined as “Safe” as it only uses data contained in the database. The DLL is not normally signed, but I provided a signed version to test and received the same results. The installation is being done by someone else, and I don’t have access to the box, but they are executing scripts that I provided and work on other computers. I have tried to find information about this error beyond what the results of the script provide, but I haven’t found anything helpful. The person executing the script to create the assembly is logged in with an Admin account, is running CMD as admin, is connecting to the DB via Windows Authentication, has been added to the dbo_owner role, and added to the server role SysAdmin with the hopes that it is a permissions issue. This hasn't changed anything. Do I need to configure SQL Server 2005 Express differently for this environment? Is this error logged anywhere other than just the output from SQLCMD? What could cause this error? Could Vista security policies cause this? I don’t have access to the computer (the customer is doing the testing) so I can’t examine the box myself. TIA

    Read the article

  • T4 Template error - Assembly Directive cannot locate referenced assembly in Visual Studio 2010 proje

    - by CodeSniper
    I ran into the following error recently in Visual Studio 2010 while trying to port Phil Haack’s excellent T4CSS template which was originally built for Visual Studio 2008.   The Problem Error Compiling transformation: Metadata file 'dotless.Core' could not be found In “T4 speak”, this simply means that you have an Assembly directive in your T4 template but the T4 engine was not able to locate or load the referenced assembly. In the case of the T4CSS Template, this was a showstopper for making it work in Visual Studio 2010. On a side note: The T4CSS template is a sweet little wrapper to allow you to use DotLessCss to generate static .css files from .less files rather than using their default HttpHandler or command-line tool.    If you haven't tried DotLessCSS yet, go check it out now!  In short, it is a tool that allows you to templatize and program your CSS files so that you can use variables, expressions, and mixins within your CSS which enables rapid changes and a lot of developer-flexibility as you evolve your CSS and UI. Back to our regularly scheduled program… Anyhow, this post isn't about DotLessCss, its about the T4 Templates and the errors I ran into when converting them from Visual Studio 2008 to Visual Studio 2010. In VS2010, there were quite a few changes to the T4 Template Engine; most were excellent changes, but this one bit me with T4CSS: “Project assemblies are no longer used to resolve template assembly directives.” In VS2008, if you wanted to reference a custom assembly in your T4 Template (.tt file) you would simply right click on your project, choose Add Reference and select that assembly.  Afterwards you were allowed to use the following syntax in your T4 template to tell it to look at the local references: <#@ assembly name="dotless.Core.dll" #> This told the engine to look in the “usual place” for the assembly, which is your project references. However, this is exactly what they changed in VS2010.  They now basically sandbox the T4 Engine to keep your T4 assemblies separate from your project assemblies.  This can come in handy if you want to support different versions of an assembly referenced both by your T4 templates and your project. Who broke the build?  Oh, Microsoft Did! In our case, this change causes a problem since the templates are no longer compatible when upgrading to VS 2010 – thus its a breaking change.  So, how do we make this work in VS 2010? Luckily, Microsoft now offers several options for referencing assemblies from T4 Templates: GAC your assemblies and use Namespace Reference or Fully Qualified Type Name Use a hard-coded Fully Qualified UNC path Copy assembly to Visual Studio "Public Assemblies Folder" and use Namespace Reference or Fully Qualified Type Name.  Use or Define a Windows Environment Variable to build a Fully Qualified UNC path. Use a Visual Studio Macro to build a Fully Qualified UNC path. Option #1 & 2 were already supported in Visual Studio 2008, so if you want to keep your templates compatible with both Visual Studio versions, then you would have to adopt one of these approaches. Yakkety Yak, use the GAC! Option #1 requires an additional pre-build step to GAC the referenced assembly, which could be a pain.  But, if you go that route, then after you GAC, all you need is a simple type name or namespace reference such as: <#@ assembly name="dotless.Core" #> Hard Coding aint that hard! The other option of using hard-coded paths in Option #2 is pretty impractical in most situations since each developer would have to use the same local project folder paths, or modify this setting each time for their local machines as well as for production deployment.  However, if you want to go that route, simply use the following assembly directive style: <#@ assembly name="C:\Code\Lib\dotless.Core.dll" #> Lets go Public! Option #3, the Visual Studio Public Assemblies Folder, is the recommended place to put commonly used tools and libraries that are only needed for Visual Studio.  Think of it like a VS-only GAC.  This is likely the best place for something like dotLessCSS and is my preferred solution.  However, you will need to either use an installer or a pre-build action to copy the assembly to the right folder location.   Normally this is located at:  C:\Program Files (x86)\Microsoft Visual Studio 10.0\Common7\IDE\PublicAssemblies Once you have copied your assembly there, you use the type name or namespace syntax again: <#@ assembly name="dotless.Core" #> Save the Environment! Option #4, using a Windows Environment Variable, is interesting for enterprise use where you may have standard locations for files, but less useful for demo-code, frameworks, and products where you don't have control over the local system.  The syntax for including a environment variable in your assembly directive looks like the following, just as you would expect: <#@ assembly name="%mypath%\dotless.Core.dll" #> “mypath” is a Windows environment variable you setup that points to some fully qualified UNC path on your system.  In the right situation this can be a great solution such as one where you use a msi installer for deployment, or where you have a pre-existing environment variable you can re-use. OMG Macros! Finally, Option #5 is a very nice option if you want to keep your T4 template’s assembly reference local and relative to the project or solution without muddying-up your dev environment or GAC with extra deployments.  An example looks like this: <#@ assembly name="$(SolutionDir)lib\dotless.Core.dll" #> In this example, I’m using the “SolutionDir” VS macro so I can reference an assembly in a “/lib” folder at the root of the solution.   This is just one of the many macros you can use.  If you are familiar with creating Pre/Post-build Event scripts, you can use its dialog to look at all of the different VS macros available. This option gives the best solution for local assemblies without the hassle of extra installers or other setup before the build.   However, its still not compatible with Visual Studio 2008, so if you have a T4 Template you want to use with both, then you may have to create multiple .tt files, one for each IDE version, or require the developer to set a value in the .tt file manually.   I’m not sure if T4 Templates support any form of compiler switches like “#if (VS2010)”  statements, but it would definitely be nice in this case to switch between this option and one of the ones more compatible with VS 2008. Conclusion As you can see, we went from 3 options with Visual Studio 2008, to 5 options (plus one problem) with Visual Studio 2010.  As a whole, I think the changes are great, but the short-term growing pains during the migration may be annoying until we get used to our new found power. Hopefully this all made sense and was helpful to you.  If nothing else, I’ll just use it as a reference the next time I need to port a T4 template to Visual Studio 2010.  Happy T4 templating, and “May the fourth be with you!”

    Read the article

  • Installed SQL Server 2008 and now TFS is broken.

    - by johnnycakes
    Hi, My W2K3 server was running TFS 2008 SP1, SQL Server 2005 Development edition. I installed SQL Server 2008 Standard. I installed it while leaving SQL Server 2005 alone. Upgrading was not possible due to the differences in editions of the SQL Servers. Now TFS is broken. On a client computer, if I go Team - Connect to Team Foundation Server, I get this error: Team Foundation services are not available from server myserver. Technical information (for administrator): TF30059: Fatal error while initializing web service. So I head on over to my event viewer on the server. Under Application, I see one warning and two errors. First, the warning: Source: SQLSERVERAGENT Event ID: 208 Description: SQL Server Scheduled Job 'TfsWorkItemTracking Process Identities Job' (0x21F395C1F444CA499A63EBF05D717749) - Status: Failed - Invoked on: 2010-04-26 13:30:00 - Message: The job failed. The Job was invoked by Schedule 9 (ProcessIdentitiesSchedule). The last step to run was step 1 (Process Identities). Then the first error: Source: TFS Services Event ID: 3017 Description: TF53010: The following error has occurred in a Team Foundation component or extension: Date (UTC): 4/26/2010 5:36:29 PM Machine: myserver Application Domain: /LM/W3SVC/799623628/Root/Services-2-129167769888923968 Assembly: Microsoft.TeamFoundation.Server, Version=9.0.0.0, Culture=neutral, PublicKeyToken=b03f5f7f11d50a3a; v2.0.50727 Process Details: Process Name: w3wp Process Id: 4008 Thread Id: 224 Account name: DOMAIN\TFSService Detailed Message: TF53013: A crash report is being prepared for Microsoft. The following information is included in that report: System Values OS Version Information=Microsoft Windows NT 5.2.3790 Service Pack 2 CLR Version Information=2.0.50727.3053 Machine Name=myserver Processor Count=1 Working Set=34897920 System Directory=C:\WINDOWS\system32 Process Values ExitCode=0 Interactive=False Has Shutdown Started=False Process Environment Variables Path = C:\WINDOWS\system32;C:\WINDOWS;C:\WINDOWS\System32\Wbem;C:\Program Files\Microsoft SQL Server\80\Tools\Binn\;C:\Program Files\Microsoft SQL Server\90\Tools\binn\;C:\Program Files\Microsoft SQL Server\90\DTS\Binn\;C:\Program Files\Microsoft SQL Server\90\Tools\Binn\VSShell\Common7\IDE\;C:\Program Files\Microsoft Visual Studio 8\Common7\IDE\PrivateAssemblies\;C:\Program Files\Microsoft SQL Server\100\Tools\Binn\;C:\Program Files\Microsoft SQL Server\100\DTS\Binn\;C:\Program Files\Microsoft SQL Server\100\Tools\Binn\VSShell\Common7\IDE\;C:\Program Files\Microsoft Visual Studio 9.0\Common7\IDE\PrivateAssemblies\;C:\WINDOWS\system32\WindowsPowerShell\v1.0 PATHEXT = .COM;.EXE;.BAT;.CMD;.VBS;.VBE;.JS;.JSE;.WSF;.WSH;.PSC1 PROCESSOR_ARCHITECTURE = x86 SystemDrive = C: windir = C:\WINDOWS TMP = C:\WINDOWS\TEMP USERPROFILE = C:\Documents and Settings\Default User ProgramFiles = C:\Program Files FP_NO_HOST_CHECK = NO COMPUTERNAME = myserver APP_POOL_ID = Microsoft Team Foundation Server Application Pool NUMBER_OF_PROCESSORS = 1 PROCESSOR_IDENTIFIER = x86 Family 16 Model 5 Stepping 2, AuthenticAMD ClusterLog = C:\WINDOWS\Cluster\cluster.log SystemRoot = C:\WINDOWS ComSpec = C:\WINDOWS\system32\cmd.exe CommonProgramFiles = C:\Program Files\Common Files PROCESSOR_LEVEL = 16 PROCESSOR_REVISION = 0502 lib = C:\Program Files\SQLXML 4.0\bin\ ALLUSERSPROFILE = C:\Documents and Settings\All Users TEMP = C:\WINDOWS\TEMP OS = Windows_NT Request Details Url=http://myserver.domain.local:8080/Services/v1.0/Registration.asmx [method = POST] User Agent=Team Foundation (devenv.exe, 10.0.30128.1) Headers=Content-Length=390&Content-Type=text%2fxml%3b+charset%3dutf-8&Accept-Encoding=gzip%2cgzip%2cgzip&Accept-Language=en-US&Authorization=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%2f6h5U30CEXgoAEAAAAAAAAAAAAAAAAAAAAAAACQAyAEgAVABUAFAALwB0AGkAdABhAG4ALgBoAHkAcABlAHIAaQBvAG4ALgBsAG8AYwBhAGwAAAAAAAAAAAA%3d&Expect=100-continue&Host=myserver.domain.local%3a8080&User-Agent=Team+Foundation+(devenv.exe%2c+10.0.30128.1)&X-TFS-Version=1.0.0.0&X-TFS-Session=b7e7fdec-e7ee-48fc-92e8-537d1cd87ea4&SOAPAction=%22http%3a%2f%2fschemas.microsoft.com%2fTeamFoundation%2f2005%2f06%2fServices%2fRegistration%2f03%2fGetRegistrationEntries%22 Path=/Services/v1.0/Registration.asmx Local Request=False User Host Address=10.0.5.78 User=DOMAIN\Johnny [auth = NTLM] Application Provided Information Team Foundation Application Information Event Log Source = TFS Services Configured Team Foundation Server = http://myserver:8080 License Type = WorkgroupLicense Server Culture = en-US Activity Logging Name = Integration Component Name = CS Initialized = No Requests Processed = 0 Exception: TypeInitializationException Message: The type initializer for 'Microsoft.TeamFoundation.Server.IntegrationResourceComponent' threw an exception. Stack Trace: at Microsoft.TeamFoundation.Server.IntegrationResourceComponent.RegisterExceptions() at Microsoft.TeamFoundation.Server.Global.Initialize() at Microsoft.TeamFoundation.Server.TeamFoundationApplication.Init() Inner Exception Details Exception: ReflectionTypeLoadException Message: Unable to load one or more of the requested types. Retrieve the LoaderExceptions property for more information. Stack Trace: at System.Reflection.Module._GetTypesInternal(StackCrawlMark& stackMark) at System.Reflection.Assembly.GetTypes() at Microsoft.TeamFoundation.Server.SqlResourceComponent.RegisterExceptions(Assembly assembly) at Microsoft.TeamFoundation.Server.IntegrationResourceComponent.RegisterExceptions() at Microsoft.TeamFoundation.Server.IntegrationResourceComponent..cctor() Application Domain Information Assembly Name=mscorlib, Version=2.0.0.0, Culture=neutral, PublicKeyToken=b77a5c561934e089 Assembly CLR Version=v2.0.50727 Assembly Version=2.0.0.0 Assembly Location=C:\WINDOWS\Microsoft.NET\Framework\v2.0.50727\mscorlib.dll Assembly File Version: File: C:\WINDOWS\Microsoft.NET\Framework\v2.0.50727\mscorlib.dll InternalName: mscorlib.dll OriginalFilename: mscorlib.dll FileVersion: 2.0.50727.3053 (netfxsp.050727-3000) FileDescription: Microsoft Common Language Runtime Class Library Product: Microsoft® .NET Framework ProductVersion: 2.0.50727.3053 Debug: False Patched: False PreRelease: False PrivateBuild: False SpecialBuild: False Language: English (United States) Assembly Name=System.Web, Version=2.0.0.0, Culture=neutral, PublicKeyToken=b03f5f7f11d50a3a Assembly CLR Version=v2.0.50727 Assembly Version=2.0.0.0 Assembly Location=C:\WINDOWS\assembly\GAC_32\System.Web\2.0.0.0__b03f5f7f11d50a3a\System.Web.dll Assembly File Version: File: C:\WINDOWS\assembly\GAC_32\System.Web\2.0.0.0__b03f5f7f11d50a3a\System.Web.dll InternalName: System.Web.dll OriginalFilename: System.Web.dll FileVersion: 2.0.50727.3053 (netfxsp.050727-3000) FileDescription: System.Web.dll Product: Microsoft® .NET Framework ProductVersion: 2.0.50727.3053 Debug: False Patched: False PreRelease: False PrivateBuild: False SpecialBuild: False Language: English (United States) Assembly Name=System, Version=2.0.0.0, Culture=neutral, PublicKeyToken=b77a5c561934e089 Assembly CLR Version=v2.0.50727 Assembly Version=2.0.0.0 Assembly Location=C:\WINDOWS\assembly\GAC_MSIL\System\2.0.0.0__b77a5c561934e089\System.dll Assembly File Version: File: C:\WINDOWS\assembly\GAC_MSIL\System\2.0.0.0__b77a5c561934e089\System.dll InternalName: System.dll OriginalFilename: System.dll FileVersion: 2.0.50727.3053 (netfxsp.050727-3000) FileDescription: .NET Framework Product: Microsoft® .NET Framework ProductVersion: 2.0.50727.3053 Debug: False Patched: False PreRelease: False PrivateBuild: False SpecialBuild: False Language: English (United States) Assembly Name=System.Xml, Version=2.0.0.0, Culture=neutral, PublicKeyToken=b77a5c561934e089 Assembly CLR Version=v2.0.50727 Assembly Version=2.0.0.0 Assembly Location=C:\WINDOWS\assembly\GAC_MSIL\System.Xml\2.0.0.0__b77a5c561934e089\System.Xml.dll Assembly File Version: File: C:\WINDOWS\assembly\GAC_MSIL\System.Xml\2.0.0.0__b77a5c561934e089\System.Xml.dll InternalName: System.Xml.dll OriginalFilename: System.Xml.dll FileVersion: 2.0.50727.3053 (netfxsp.050727-3000) FileDescription: .NET Framework Product: Microsoft® .NET Framework ProductVersion: 2.0.50727.3053 Debug: False Patched: False PreRelease: False PrivateBuild: False SpecialBuild: False Language: English (United States) Assembly Name=System.Configuration, Version=2.0.0.0, Culture=neutral, PublicKeyToken=b03f5f7f11d50a3a Assembly CLR Version=v2.0.50727 Assembly Version=2.0.0.0 Assembly Location=C:\WINDOWS\assembly\GAC_MSIL\System.Configuration\2.0.0.0__b03f5f7f11d50a3a\System.Configuration.dll Assembly File Version: File: C:\WINDOWS\assembly\GAC_MSIL\System.Configuration\2.0.0.0__b03f5f7f11d50a3a\System.Configuration.dll InternalName: System.Configuration.dll OriginalFilename: System.Configuration.dll FileVersion: 2.0.50727.3053 (netfxsp.050727-3000) FileDescription: System.Configuration.dll Product: Microsoft® .NET Framework ProductVersion: 2.0.50727.3053 Debug: False Patched: False PreRelease: False PrivateBuild: False SpecialBuild: False Language: English (United States) Assembly Name=Microsoft.JScript, Version=8.0.0.0, Culture=neutral, PublicKeyToken=b03f5f7f11d50a3a Assembly CLR Version=v2.0.50727 Assembly Version=8.0.0.0 Assembly Location=C:\WINDOWS\assembly\GAC_MSIL\Microsoft.JScript\8.0.0.0__b03f5f7f11d50a3a\Microsoft.JScript.dll Assembly File Version: File: C:\WINDOWS\assembly\GAC_MSIL\Microsoft.JScript\8.0.0.0__b03f5f7f11d50a3a\Microsoft.JScript.dll InternalName: Microsoft.JScript.dll OriginalFilename: Microsoft.JScript.dll FileVersion: 8.0.50727.3053 FileDescription: Microsoft.JScript.dll Product: Microsoft (R) Visual Studio (R) 2005 ProductVersion: 8.0.50727.3053 Debug: False Patched: False PreRelease: False PrivateBuild: False SpecialBuild: False Language: Language Neutral Assembly Name=App_global.asax.4nq_g1xi, Version=0.0.0.0, Culture=neutral, PublicKeyToken=null Assembly CLR Version=v2.0.50727 Assembly Version=0.0.0.0 Assembly Location=C:\WINDOWS\Microsoft.NET\Framework\v2.0.50727\Temporary ASP.NET Files\services\87e24ff8\921625fe\App_global.asax.4nq_g1xi.dll Assembly File Version: File: C:\WINDOWS\Microsoft.NET\Framework\v2.0.50727\Temporary ASP.NET Files\services\87e24ff8\921625fe\App_global.asax.4nq_g1xi.dll InternalName: App_global.asax.4nq_g1xi.dll OriginalFilename: App_global.asax.4nq_g1xi.dll FileVersion: 0.0.0.0 FileDescription: Product: ProductVersion: 0.0.0.0 Debug: False Patched: False PreRelease: False PrivateBuild: False SpecialBuild: False Language: Language Neutral Assembly Name=Microsoft.TeamFoundation.Server, Version=9.0.0.0, Culture=neutral, PublicKeyToken=b03f5f7f11d50a3a Assembly CLR Version=v2.0.50727 Assembly Version=9.0.0.0 Assembly Location=C:\WINDOWS\Microsoft.NET\Framework\v2.0.50727\Temporary ASP.NET Files\services\87e24ff8\921625fe\assembly\dl3\9051eeb6\603ea9a2_d822c801\Microsoft.TeamFoundation.Server.DLL Assembly File Version: File: C:\WINDOWS\Microsoft.NET\Framework\v2.0.50727\Temporary ASP.NET Files\services\87e24ff8\921625fe\assembly\dl3\9051eeb6\603ea9a2_d822c801\Microsoft.TeamFoundation.Server.DLL InternalName: Microsoft.TeamFoundation.Server.dll OriginalFilename: Microsoft.TeamFoundation.Server.dll FileVersion: 9.0.21022.8 FileDescription: Microsoft.TeamFoundation.Server.dll Product: Microsoft (R) Visual Studio (R) 2008 ProductVersion: 9.0.21022.8 Debug: False Patched: False PreRelease: False PrivateBuild: False SpecialBuild: False Language: Language Neutral Assembly Name=Microsoft.TeamFoundation.Common, Version=9.0.0.0, Culture=neutral, PublicKeyToken=b03f5f7f11d50a3a Assembly CLR Version=v2.0.50727 Assembly Version=9.0.0.0 Assembly Location=C:\WINDOWS\assembly\GAC_32\Microsoft.TeamFoundation.Common\9.0.0.0__b03f5f7f11d50a3a\Microsoft.TeamFoundation.Common.dll Assembly File Version: File: C:\WINDOWS\assembly\GAC_32\Microsoft.TeamFoundation.Common\9.0.0.0__b03f5f7f11d50a3a\Microsoft.TeamFoundation.Common.dll InternalName: Microsoft.TeamFoundation.Common.dll OriginalFilename: Microsoft.TeamFoundation.Common.dll FileVersion: 9.0.30729.1 FileDescription: Microsoft.TeamFoundation.Common.dll Product: Microsoft (R) Visual Studio (R) 2008 ProductVersion: 9.0.30729.1 Debug: False Patched: False PreRelease: False PrivateBuild: False SpecialBuild: False Language: Language Neutral Assembly Name=Microsoft.TeamFoundation, Version=9.0.0.0, Culture=neutral, PublicKeyToken=b03f5f7f11d50a3a Assembly CLR Version=v2.0.50727 Assembly Version=9.0.0.0 Assembly Location=C:\WINDOWS\assembly\GAC_32\Microsoft.TeamFoundation\9.0.0.0__b03f5f7f11d50a3a\Microsoft.TeamFoundation.dll Assembly File Version: File: C:\WINDOWS\assembly\GAC_32\Microsoft.TeamFoundation\9.0.0.0__b03f5f7f11d50a3a\Microsoft.TeamFoundation.dll InternalName: Microsoft.TeamFoundation.dll OriginalFilename: Microsoft.TeamFoundation.dll FileVersion: 9.0.30729.1 FileDescription: Microsoft.TeamFoundation.dll Product: Microsoft (R) Visual Studio (R) 2008 ProductVersion: 9.0.30729.1 Debug: False Patched: False PreRelease: False PrivateBuild: False SpecialBuild: False Language: Language Neutral Assembly Name=System.Security, Version=2.0.0.0, Culture=neutral, PublicKeyToken=b03f5f7f11d50a3a Assembly CLR Version=v2.0.50727 Assembly Version=2.0.0.0 Assembly Location=C:\WINDOWS\assembly\GAC_MSIL\System.Security\2.0.0.0__b03f5f7f11d50a3a\System.Security.dll Assembly File Version: File: C:\WINDOWS\assembly\GAC_MSIL\System.Security\2.0.0.0__b03f5f7f11d50a3a\System.Security.dll InternalName: System.Security.dll OriginalFilename: System.Security.dll FileVersion: 2.0.50727.3053 (netfxsp.050727-3000) FileDescription: System.Security.dll Product: Microsoft® .NET Framework ProductVersion: 2.0.50727.3053 Debug: False Patched: False PreRelease: False PrivateBuild: False SpecialBuild: False Language: English (United States) Assembly Name=System.Data, Version=2.0.0.0, Culture=neutral, PublicKeyToken=b77a5c561934e089 Assembly CLR Version=v2.0.50727 Assembly Version=2.0.0.0 Assembly Location=C:\WINDOWS\assembly\GAC_32\System.Data\2.0.0.0__b77a5c561934e089\System.Data.dll Assembly File Version: File: C:\WINDOWS\assembly\GAC_32\System.Data\2.0.0.0__b77a5c561934e089\System.Data.dll InternalName: system.data.dll OriginalFilename: system.data.dll FileVersion: 2.0.50727.3053 (netfxsp.050727-3000) FileDescription: .NET Framework Product: Microsoft® .NET Framework ProductVersion: 2.0.50727.3053 Debug: False Patched: False PreRelease: False PrivateBuild: False SpecialBuild: False Language: English (United States) Assembly Name=Microsoft.TeamFoundation.Common.Library, Version=9.0.0.0, Culture=neutral, PublicKeyToken=b03f5f7f11d50a3a Assembly CLR Version=v2.0.50727 Assembly Version=9.0.0.0 Assembly Location=C:\WINDOWS\assembly\GAC_32\Microsoft.TeamFoundation.Common.Library\9.0.0.0__b03f5f7f11d50a3a\Microsoft.TeamFoundation.Common.Library.dll Assembly File Version: File: C:\WINDOWS\assembly\GAC_32\Microsoft.TeamFoundation.Common.Library\9.0.0.0__b03f5f7f11d50a3a\Microsoft.TeamFoundation.Common.Library.dll InternalName: Microsoft.TeamFoundation.Common.Library.dll OriginalFilename: Microsoft.TeamFoundation.Common.Library.dll FileVersion: 9.0.30729.1 FileDescription: Microsoft.TeamFoundation.Common.Library.dll Product: Microsoft (R) Visual Studio (R) 2008 ProductVersion: 9.0.30729.1 Debug: False Patched: False PreRelease: False PrivateBuild: False SpecialBuild: False Language: Language Neutral Assembly Name=System.Web.Mobile, Version=2.0.0.0, Culture=neutral, PublicKeyToken=b03f5f7f11d50a3a Assembly CLR Version=v2.0.50727 Assembly Version=2.0.0.0 Assembly Location=C:\WINDOWS\assembly\GAC_MSIL\System.Web.Mobile\2.0.0.0__b03f5f7f11d50a3a\System.Web.Mobile.dll As And finally, the second error: Source: Team Foundation Error Reporting Event ID: 5000 Description: EventType teamfoundationue, P1 1.0.0.0, P2 tfs, P3 9.0.30729.1, P4 9.0.0.0, P5 general, P6 typeinitializationexcept, P7 4758b22a940fe6d9, P8 d15c14bb, P9 NIL, P10 NIL. Any ideas? Thanks.

    Read the article

  • Anatomy of a .NET Assembly - CLR metadata 1

    - by Simon Cooper
    Before we look at the bytes comprising the CLR-specific data inside an assembly, we first need to understand the logical format of the metadata (For this post I only be looking at simple pure-IL assemblies; mixed-mode assemblies & other things complicates things quite a bit). Metadata streams Most of the CLR-specific data inside an assembly is inside one of 5 streams, which are analogous to the sections in a PE file. The name of each section in a PE file starts with a ., and the name of each stream in the CLR metadata starts with a #. All but one of the streams are heaps, which store unstructured binary data. The predefined streams are: #~ Also called the metadata stream, this stream stores all the information on the types, methods, fields, properties and events in the assembly. Unlike the other streams, the metadata stream has predefined contents & structure. #Strings This heap is where all the namespace, type & member names are stored. It is referenced extensively from the #~ stream, as we'll be looking at later. #US Also known as the user string heap, this stream stores all the strings used in code directly. All the strings you embed in your source code end up in here. This stream is only referenced from method bodies. #GUID This heap exclusively stores GUIDs used throughout the assembly. #Blob This heap is for storing pure binary data - method signatures, generic instantiations, that sort of thing. Items inside the heaps (#Strings, #US, #GUID and #Blob) are indexed using a simple binary offset from the start of the heap. At that offset is a coded integer giving the length of that item, then the item's bytes immediately follow. The #GUID stream is slightly different, in that GUIDs are all 16 bytes long, so a length isn't required. Metadata tables The #~ stream contains all the assembly metadata. The metadata is organised into 45 tables, which are binary arrays of predefined structures containing information on various aspects of the metadata. Each entry in a table is called a row, and the rows are simply concatentated together in the file on disk. For example, each row in the TypeRef table contains: A reference to where the type is defined (most of the time, a row in the AssemblyRef table). An offset into the #Strings heap with the name of the type An offset into the #Strings heap with the namespace of the type. in that order. The important tables are (with their table number in hex): 0x2: TypeDef 0x4: FieldDef 0x6: MethodDef 0x14: EventDef 0x17: PropertyDef Contains basic information on all the types, fields, methods, events and properties defined in the assembly. 0x1: TypeRef The details of all the referenced types defined in other assemblies. 0xa: MemberRef The details of all the referenced members of types defined in other assemblies. 0x9: InterfaceImpl Links the types defined in the assembly with the interfaces that type implements. 0xc: CustomAttribute Contains information on all the attributes applied to elements in this assembly, from method parameters to the assembly itself. 0x18: MethodSemantics Links properties and events with the methods that comprise the get/set or add/remove methods of the property or method. 0x1b: TypeSpec 0x2b: MethodSpec These tables provide instantiations of generic types and methods for each usage within the assembly. There are several ways to reference a single row within a table. The simplest is to simply specify the 1-based row index (RID). The indexes are 1-based so a value of 0 can represent 'null'. In this case, which table the row index refers to is inferred from the context. If the table can't be determined from the context, then a particular row is specified using a token. This is a 4-byte value with the most significant byte specifying the table, and the other 3 specifying the 1-based RID within that table. This is generally how a metadata table row is referenced from the instruction stream in method bodies. The third way is to use a coded token, which we will look at in the next post. So, back to the bytes Now we've got a rough idea of how the metadata is logically arranged, we can now look at the bytes comprising the start of the CLR data within an assembly: The first 8 bytes of the .text section are used by the CLR loader stub. After that, the CLR-specific data starts with the CLI header. I've highlighted the important bytes in the diagram. In order, they are: The size of the header. As the header is a fixed size, this is always 0x48. The CLR major version. This is always 2, even for .NET 4 assemblies. The CLR minor version. This is always 5, even for .NET 4 assemblies, and seems to be ignored by the runtime. The RVA and size of the metadata header. In the diagram, the RVA 0x20e4 corresponds to the file offset 0x2e4 Various flags specifying if this assembly is pure-IL, whether it is strong name signed, and whether it should be run as 32-bit (this is how the CLR differentiates between x86 and AnyCPU assemblies). A token pointing to the entrypoint of the assembly. In this case, 06 (the last byte) refers to the MethodDef table, and 01 00 00 refers to to the first row in that table. (after a gap) RVA of the strong name signature hash, which comes straight after the CLI header. The RVA 0x2050 corresponds to file offset 0x250. The rest of the CLI header is mainly used in mixed-mode assemblies, and so is zeroed in this pure-IL assembly. After the CLI header comes the strong name hash, which is a SHA-1 hash of the assembly using the strong name key. After that comes the bodies of all the methods in the assembly concatentated together. Each method body starts off with a header, which I'll be looking at later. As you can see, this is a very small assembly with only 2 methods (an instance constructor and a Main method). After that, near the end of the .text section, comes the metadata, containing a metadata header and the 5 streams discussed above. We'll be looking at this in the next post. Conclusion The CLI header data doesn't have much to it, but we've covered some concepts that will be important in later posts - the logical structure of the CLR metadata and the overall layout of CLR data within the .text section. Next, I'll have a look at the contents of the #~ stream, and how the table data is arranged on disk.

    Read the article

  • Display Img and Div inline - it's not rendered inline

    - by user359372
    In order to follow correct web standards, I've tried to layout image and div inline. In orde to achieve that - I've used display:inline property. But then I experienced the following issue: image renders from the center line, and div doesn't respect height parameter set to it. I've tried using line-height parameter, but that didn't give any useful results. I've also tried various combinations with setting margin/padding to some values or to auto, or replacing div with span, or wrapping img and div with additional divs. I've managed to achieve desired result by using position:absolute, but that doesn't help in cases where I want to use centered/relative positioning of the whole component... Any clues or ideas or troubleshooting hints? Please find the html example below: Test Page Some text that should be displayed in the center/middle of the div 123   Some text that should be displayed in the center/middle of the div

    Read the article

  • 32bit to 64bit inline assembly porting

    - by Simone Margaritelli
    I have a piece of C++ code (compiled with g++ under a GNU/Linux environment) that load a function pointer (how it does that doesn't matter), pushes some arguments onto the stack with some inline assembly and then calls that function, the code is like : unsigned long stack[] = { 1, 23, 33, 43 }; /* save all the registers and the stack pointer */ unsigned long esp; asm __volatile__ ( "pusha" ); asm __volatile__ ( "mov %%esp, %0" :"=m" (esp)); for( i = 0; i < sizeof(stack); i++ ){ unsigned long val = stack[i]; asm __volatile__ ( "push %0" :: "m"(val) ); } unsigned long ret = function_pointer(); /* restore registers and stack pointer */ asm __volatile__ ( "mov %0, %%esp" :: "m" (esp) ); asm __volatile__ ( "popa" ); I'd like to add some sort of #ifdef _LP64 // 64bit inline assembly #else // 32bit version as above example #endif But i don't know inline assembly for 64bit machines, anyone could help me? Thanks

    Read the article

  • Anatomy of a .NET Assembly - Custom attribute encoding

    - by Simon Cooper
    In my previous post, I covered how field, method, and other types of signatures are encoded in a .NET assembly. Custom attribute signatures differ quite a bit from these, which consequently affects attribute specifications in C#. Custom attribute specifications In C#, you can apply a custom attribute to a type or type member, specifying a constructor as well as the values of fields or properties on the attribute type: public class ExampleAttribute : Attribute { public ExampleAttribute(int ctorArg1, string ctorArg2) { ... } public Type ExampleType { get; set; } } [Example(5, "6", ExampleType = typeof(string))] public class C { ... } How does this specification actually get encoded and stored in an assembly? Specification blob values Custom attribute specification signatures use the same building blocks as other types of signatures; the ELEMENT_TYPE structure. However, they significantly differ from other types of signatures, in that the actual parameter values need to be stored along with type information. There are two types of specification arguments in a signature blob; fixed args and named args. Fixed args are the arguments to the attribute type constructor, named arguments are specified after the constructor arguments to provide a value to a field or property on the constructed attribute type (PropertyName = propValue) Values in an attribute blob are limited to one of the basic types (one of the number types, character, or boolean), a reference to a type, an enum (which, in .NET, has to use one of the integer types as a base representation), or arrays of any of those. Enums and the basic types are easy to store in a blob - you simply store the binary representation. Strings are stored starting with a compressed integer indicating the length of the string, followed by the UTF8 characters. Array values start with an integer indicating the number of elements in the array, then the item values concatentated together. Rather than using a coded token, Type values are stored using a string representing the type name and fully qualified assembly name (for example, MyNs.MyType, MyAssembly, Version=1.0.0.0, Culture=neutral, PublicKeyToken=0123456789abcdef). If the type is in the current assembly or mscorlib then just the type name can be used. This is probably done to prevent direct references between assemblies solely because of attribute specification arguments; assemblies can be loaded in the reflection-only context and attribute arguments still processed, without loading the entire assembly. Fixed and named arguments Each entry in the CustomAttribute metadata table contains a reference to the object the attribute is applied to, the attribute constructor, and the specification blob. The number and type of arguments to the constructor (the fixed args) can be worked out by the method signature referenced by the attribute constructor, and so the fixed args can simply be concatenated together in the blob without any extra type information. Named args are different. These specify the value to assign to a field or property once the attribute type has been constructed. In the CLR, fields and properties can be overloaded just on their type; different fields and properties can have the same name. Therefore, to uniquely identify a field or property you need: Whether it's a field or property (indicated using byte values 0x53 and 0x54, respectively) The field or property type The field or property name After the fixed arg values is a 2-byte number specifying the number of named args in the blob. Each named argument has the above information concatenated together, mostly using the basic ELEMENT_TYPE values, in the same way as a method or field signature. A Type argument is represented using the byte 0x50, and an enum argument is represented using the byte 0x55 followed by a string specifying the name and assembly of the enum type. The named argument property information is followed by the argument value, using the same encoding as fixed args. Boxed objects This would be all very well, were it not for object and object[]. Arguments and properties of type object allow a value of any allowed argument type to be specified. As a result, more information needs to be specified in the blob to interpret the argument bytes as the correct type. So, the argument value is simple prepended with the type of the value by specifying the ELEMENT_TYPE or name of the enum the value represents. For named arguments, a field or property of type object is represented using the byte 0x51, with the actual type specified in the argument value. Some examples... All property signatures start with the 2-byte value 0x0001. Similar to my previous post in the series, names in capitals correspond to a particular byte value in the ELEMENT_TYPE structure. For strings, I'll simply give the string value, rather than the length and UTF8 encoding in the actual blob. I'll be using the following enum and attribute types to demonstrate specification encodings: class AttrAttribute : Attribute { public AttrAttribute() {} public AttrAttribute(Type[] tArray) {} public AttrAttribute(object o) {} public AttrAttribute(MyEnum e) {} public AttrAttribute(ushort x, int y) {} public AttrAttribute(string str, Type type1, Type type2) {} public int Prop1 { get; set; } public object Prop2 { get; set; } public object[] ObjectArray; } enum MyEnum : int { Val1 = 1, Val2 = 2 } Now, some examples: Here, the the specification binds to the (ushort, int) attribute constructor, with fixed args only. The specification blob starts off with a prolog, followed by the two constructor arguments, then the number of named arguments (zero): [Attr(42, 84)] 0x0001 0x002a 0x00000054 0x0000 An example of string and type encoding: [Attr("MyString", typeof(Array), typeof(System.Windows.Forms.Form))] 0x0001 "MyString" "System.Array" "System.Windows.Forms.Form, System.Windows.Forms, Version=4.0.0.0, Culture=neutral, PublicKeyToken=b77a5c561934e089" 0x0000 As you can see, the full assembly specification of a type is only needed if the type isn't in the current assembly or mscorlib. Note, however, that the C# compiler currently chooses to fully-qualify mscorlib types anyway. An object argument (this binds to the object attribute constructor), and two named arguments (a null string is represented by 0xff and the empty string by 0x00) [Attr((ushort)40, Prop1 = 12, Prop2 = "")] 0x0001 U2 0x0028 0x0002 0x54 I4 "Prop1" 0x0000000c 0x54 0x51 "Prop2" STRING 0x00 Right, more complicated now. A type array as a fixed argument: [Attr(new[] { typeof(string), typeof(object) })] 0x0001 0x00000002 // the number of elements "System.String" "System.Object" 0x0000 An enum value, which is simply represented using the underlying value. The CLR works out that it's an enum using information in the attribute constructor signature: [Attr(MyEnum.Val1)] 0x0001 0x00000001 0x0000 And finally, a null array, and an object array as a named argument: [Attr((Type[])null, ObjectArray = new object[] { (byte)2, typeof(decimal), null, MyEnum.Val2 })] 0x0001 0xffffffff 0x0001 0x53 SZARRAY 0x51 "ObjectArray" 0x00000004 U1 0x02 0x50 "System.Decimal" STRING 0xff 0x55 "MyEnum" 0x00000002 As you'll notice, a null object is encoded as a null string value, and a null array is represented using a length of -1 (0xffffffff). How does this affect C#? So, we can now explain why the limits on attribute arguments are so strict in C#. Attribute specification blobs are limited to basic numbers, enums, types, and arrays. As you can see, this is because the raw CLR encoding can only accommodate those types. Special byte patterns have to be used to indicate object, string, Type, or enum values in named arguments; you can't specify an arbitary object type, as there isn't a generalised way of encoding the resulting value in the specification blob. In particular, decimal values can't be encoded, as it isn't a 'built-in' CLR type that has a native representation (you'll notice that decimal constants in C# programs are compiled as several integer arguments to DecimalConstantAttribute). Jagged arrays also aren't natively supported, although you can get around it by using an array as a value to an object argument: [Attr(new object[] { new object[] { new Type[] { typeof(string) } }, 42 })] Finally... Phew! That was a bit longer than I thought it would be. Custom attribute encodings are complicated! Hopefully this series has been an informative look at what exactly goes on inside a .NET assembly. In the next blog posts, I'll be carrying on with the 'Inside Red Gate' series.

    Read the article

  • Fastest inline-assembly spinlock

    - by sigvardsen
    I'm writing a multithreaded application in c++, where performance is critical. I need to use a lot of locking while copying small structures between threads, for this I have chosen to use spinlocks. I have done some research and speed testing on this and I found that most implementations are roughly equally fast: Microsofts CRITICAL_SECTION, with SpinCount set to 1000, scores about 140 time units Implementing this algorithm with Microsofts InterlockedCompareExchange scores about 95 time units Ive also tried to use some inline assembly with __asm {} using something like this code and it scores about 70 time units, but I am not sure that a proper memory barrier has been created. Edit: The times given here are the time it takes for 2 threads to lock and unlock the spinlock 1,000,000 times. I know this isn't a lot of difference but as a spinlock is a heavily used object, one would think that programmers would have agreed on the fastest possible way to make a spinlock. Googling it leads to many different approaches however. I would think this aforementioned method would be the fastest if implemented using inline assembly and using the instruction CMPXCHG8B instead of comparing 32bit registers. Furthermore memory barriers must be taken into account, this could be done by LOCK CMPXHG8B (I think?), which guarantees "exclusive rights" to the shared memory between cores. At last [some suggests] that for busy waits should be accompanied by NOP:REP that would enable Hyper-threading processors to switch to another thread, but I am not sure whether this is true or not? From my performance-test of different spinlocks, it is seen that there is not much difference, but for purely academic purpose I would like to know which one is fastest. However as I have extremely limited experience in the assembly-language and with memory barriers, I would be happy if someone could write the assembly code for the last example I provided with LOCK CMPXCHG8B and proper memory barriers in the following template: __asm { spin_lock: ;locking code. spin_unlock: ;unlocking code. }

    Read the article

  • How to: Inline assembler in C++ (under Visual Studio 2010)

    - by toxic shock
    I'm writing a performance-critical, number-crunching C++ project where 70% of the time is used by the 200 line core module. I'd like to optimize the core using inline assembly, but I'm completely new to this. I do, however, know some x86 assembly languages including the one used by GCC and NASM. All I know: I have to put the assembler instructions in _asm{} where I want them to be. Problem: I have no clue where to start. What is in which register at the moment my inline assembly comes into play?

    Read the article

  • Creating a dynamic proxy generator – Part 1 – Creating the Assembly builder, Module builder and cach

    - by SeanMcAlinden
    I’ve recently started a project with a few mates to learn the ins and outs of Dependency Injection, AOP and a number of other pretty crucial patterns of development as we’ve all been using these patterns for a while but have relied totally on third part solutions to do the magic. We thought it would be interesting to really get into the details by rolling our own IoC container and hopefully learn a lot on the way, and you never know, we might even create an excellent framework. The open source project is called Rapid IoC and is hosted at http://rapidioc.codeplex.com/ One of the most interesting tasks for me is creating the dynamic proxy generator for enabling Aspect Orientated Programming (AOP). In this series of articles, I’m going to track each step I take for creating the dynamic proxy generator and I’ll try my best to explain what everything means - mainly as I’ll be using Reflection.Emit to emit a fair amount of intermediate language code (IL) to create the proxy types at runtime which can be a little taxing to read. It’s worth noting that building the proxy is without a doubt going to be slightly painful so I imagine there will be plenty of areas I’ll need to change along the way. Anyway lets get started…   Part 1 - Creating the Assembly builder, Module builder and caching mechanism Part 1 is going to be a really nice simple start, I’m just going to start by creating the assembly, module and type caches. The reason we need to create caches for the assembly, module and types is simply to save the overhead of recreating proxy types that have already been generated, this will be one of the important steps to ensure that the framework is fast… kind of important as we’re calling the IoC container ‘Rapid’ – will be a little bit embarrassing if we manage to create the slowest framework. The Assembly builder The assembly builder is what is used to create an assembly at runtime, we’re going to have two overloads, one will be for the actual use of the proxy generator, the other will be mainly for testing purposes as it will also save the assembly so we can use Reflector to examine the code that has been created. Here’s the code: DynamicAssemblyBuilder using System; using System.Reflection; using System.Reflection.Emit; namespace Rapid.DynamicProxy.Assembly {     /// <summary>     /// Class for creating an assembly builder.     /// </summary>     internal static class DynamicAssemblyBuilder     {         #region Create           /// <summary>         /// Creates an assembly builder.         /// </summary>         /// <param name="assemblyName">Name of the assembly.</param>         public static AssemblyBuilder Create(string assemblyName)         {             AssemblyName name = new AssemblyName(assemblyName);               AssemblyBuilder assembly = AppDomain.CurrentDomain.DefineDynamicAssembly(                     name, AssemblyBuilderAccess.Run);               DynamicAssemblyCache.Add(assembly);               return assembly;         }           /// <summary>         /// Creates an assembly builder and saves the assembly to the passed in location.         /// </summary>         /// <param name="assemblyName">Name of the assembly.</param>         /// <param name="filePath">The file path.</param>         public static AssemblyBuilder Create(string assemblyName, string filePath)         {             AssemblyName name = new AssemblyName(assemblyName);               AssemblyBuilder assembly = AppDomain.CurrentDomain.DefineDynamicAssembly(                     name, AssemblyBuilderAccess.RunAndSave, filePath);               DynamicAssemblyCache.Add(assembly);               return assembly;         }           #endregion     } }   So hopefully the above class is fairly explanatory, an AssemblyName is created using the passed in string for the actual name of the assembly. An AssemblyBuilder is then constructed with the current AppDomain and depending on the overload used, it is either just run in the current context or it is set up ready for saving. It is then added to the cache.   DynamicAssemblyCache using System.Reflection.Emit; using Rapid.DynamicProxy.Exceptions; using Rapid.DynamicProxy.Resources.Exceptions;   namespace Rapid.DynamicProxy.Assembly {     /// <summary>     /// Cache for storing the dynamic assembly builder.     /// </summary>     internal static class DynamicAssemblyCache     {         #region Declarations           private static object syncRoot = new object();         internal static AssemblyBuilder Cache = null;           #endregion           #region Adds a dynamic assembly to the cache.           /// <summary>         /// Adds a dynamic assembly builder to the cache.         /// </summary>         /// <param name="assemblyBuilder">The assembly builder.</param>         public static void Add(AssemblyBuilder assemblyBuilder)         {             lock (syncRoot)             {                 Cache = assemblyBuilder;             }         }           #endregion           #region Gets the cached assembly                  /// <summary>         /// Gets the cached assembly builder.         /// </summary>         /// <returns></returns>         public static AssemblyBuilder Get         {             get             {                 lock (syncRoot)                 {                     if (Cache != null)                     {                         return Cache;                     }                 }                   throw new RapidDynamicProxyAssertionException(AssertionResources.NoAssemblyInCache);             }         }           #endregion     } } The cache is simply a static property that will store the AssemblyBuilder (I know it’s a little weird that I’ve made it public, this is for testing purposes, I know that’s a bad excuse but hey…) There are two methods for using the cache – Add and Get, these just provide thread safe access to the cache.   The Module Builder The module builder is required as the create proxy classes will need to live inside a module within the assembly. Here’s the code: DynamicModuleBuilder using System.Reflection.Emit; using Rapid.DynamicProxy.Assembly; namespace Rapid.DynamicProxy.Module {     /// <summary>     /// Class for creating a module builder.     /// </summary>     internal static class DynamicModuleBuilder     {         /// <summary>         /// Creates a module builder using the cached assembly.         /// </summary>         public static ModuleBuilder Create()         {             string assemblyName = DynamicAssemblyCache.Get.GetName().Name;               ModuleBuilder moduleBuilder = DynamicAssemblyCache.Get.DefineDynamicModule                 (assemblyName, string.Format("{0}.dll", assemblyName));               DynamicModuleCache.Add(moduleBuilder);               return moduleBuilder;         }     } } As you can see, the module builder is created on the assembly that lives in the DynamicAssemblyCache, the module is given the assembly name and also a string representing the filename if the assembly is to be saved. It is then added to the DynamicModuleCache. DynamicModuleCache using System.Reflection.Emit; using Rapid.DynamicProxy.Exceptions; using Rapid.DynamicProxy.Resources.Exceptions; namespace Rapid.DynamicProxy.Module {     /// <summary>     /// Class for storing the module builder.     /// </summary>     internal static class DynamicModuleCache     {         #region Declarations           private static object syncRoot = new object();         internal static ModuleBuilder Cache = null;           #endregion           #region Add           /// <summary>         /// Adds a dynamic module builder to the cache.         /// </summary>         /// <param name="moduleBuilder">The module builder.</param>         public static void Add(ModuleBuilder moduleBuilder)         {             lock (syncRoot)             {                 Cache = moduleBuilder;             }         }           #endregion           #region Get           /// <summary>         /// Gets the cached module builder.         /// </summary>         /// <returns></returns>         public static ModuleBuilder Get         {             get             {                 lock (syncRoot)                 {                     if (Cache != null)                     {                         return Cache;                     }                 }                   throw new RapidDynamicProxyAssertionException(AssertionResources.NoModuleInCache);             }         }           #endregion     } }   The DynamicModuleCache is very similar to the assembly cache, it is simply a statically stored module with thread safe Add and Get methods.   The DynamicTypeCache To end off this post, I’m going to create the cache for storing the generated proxy classes. I’ve spent a fair amount of time thinking about the type of collection I should use to store the types and have finally decided that for the time being I’m going to use a generic dictionary. This may change when I can actually performance test the proxy generator but the time being I think it makes good sense in theory, mainly as it pretty much maintains it’s performance with varying numbers of items – almost constant (0)1. Plus I won’t ever need to loop through the items which is not the dictionaries strong point. Here’s the code as it currently stands: DynamicTypeCache using System; using System.Collections.Generic; using System.Security.Cryptography; using System.Text; namespace Rapid.DynamicProxy.Types {     /// <summary>     /// Cache for storing proxy types.     /// </summary>     internal static class DynamicTypeCache     {         #region Declarations           static object syncRoot = new object();         public static Dictionary<string, Type> Cache = new Dictionary<string, Type>();           #endregion           /// <summary>         /// Adds a proxy to the type cache.         /// </summary>         /// <param name="type">The type.</param>         /// <param name="proxy">The proxy.</param>         public static void AddProxyForType(Type type, Type proxy)         {             lock (syncRoot)             {                 Cache.Add(GetHashCode(type.AssemblyQualifiedName), proxy);             }         }           /// <summary>         /// Tries the type of the get proxy for.         /// </summary>         /// <param name="type">The type.</param>         /// <returns></returns>         public static Type TryGetProxyForType(Type type)         {             lock (syncRoot)             {                 Type proxyType;                 Cache.TryGetValue(GetHashCode(type.AssemblyQualifiedName), out proxyType);                 return proxyType;             }         }           #region Private Methods           private static string GetHashCode(string fullName)         {             SHA1CryptoServiceProvider provider = new SHA1CryptoServiceProvider();             Byte[] buffer = Encoding.UTF8.GetBytes(fullName);             Byte[] hash = provider.ComputeHash(buffer, 0, buffer.Length);             return Convert.ToBase64String(hash);         }           #endregion     } } As you can see, there are two public methods, one for adding to the cache and one for getting from the cache. Hopefully they should be clear enough, the Get is a TryGet as I do not want the dictionary to throw an exception if a proxy doesn’t exist within the cache. Other than that I’ve decided to create a key using the SHA1CryptoServiceProvider, this may change but my initial though is the SHA1 algorithm is pretty fast to put together using the provider and it is also very unlikely to have any hashing collisions. (there are some maths behind how unlikely this is – here’s the wiki if you’re interested http://en.wikipedia.org/wiki/SHA_hash_functions)   Anyway, that’s the end of part 1 – although I haven’t started any of the fun stuff (by fun I mean hairpulling, teeth grating Relfection.Emit style fun), I’ve got the basis of the DynamicProxy in place so all we have to worry about now is creating the types, interceptor classes, method invocation information classes and finally a really nice fluent interface that will abstract all of the hard-core craziness away and leave us with a lightning fast, easy to use AOP framework. Hope you find the series interesting. All of the source code can be viewed and/or downloaded at our codeplex site - http://rapidioc.codeplex.com/ Kind Regards, Sean.

    Read the article

  • Creating a Hello World library function in assembly and calling it from C#

    - by Filip Ekberg
    Let's say we use NASM as they do in this answer: how to write hellow world in assembly under windows. I got a couple of thoughts and questions regarding assembly combined with c# or any other .net languages for that matter. First of all I want to be able to create a library that has the following function HelloWorld that takes this parameter: Name In C# the method signature would looke like this: void HelloWorld(string name) and it would print out something like Hello World from name I've searched around a bit but can't find that much good and clean material for this to get me started. I know some basic assembly from before mostly gasthough. So any pointers in the right direction is very much apprechiated. To sum it up Create a function in ASM ( NASM ) that takes one or more parameters Compile and create a library of the above functionality Include the library in any .net language Call the included library function Bonus features How does one handle returned values? Is it possible to write the ASM-method inline? When creating libraries in assembly or c, you do follow a certain "pre defined" way, the c calling convetion, correct?

    Read the article

  • Maven Assembly Plugin - install the created assembly

    - by Walter White
    I have a project that simply consists of files. I want to package those files into a zip and store them in a maven repository. I have the assembly plugin configured to build the zip file and that part works just fine, but I cannot seem to figure out how to install the zip file? Also, if I want to use this assembly in another artifact, how would I do that? I am intending on calling dependency:unpack, but I don't have an artifact in the repository to unpack. How can I get a zip file to be in my repository so that I may re-use it in another artifact? parent pom <build> <plugins> <plugin> <!--<groupId>org.apache.maven.plugins</groupId>--> <artifactId>maven-assembly-plugin</artifactId> <version>2.2-beta-5</version> <configuration> <filters> <filter></filter> </filters> <descriptors> <descriptor>../packaging.xml</descriptor> </descriptors> </configuration> </plugin> </plugins> </build> Child POM <parent> <groupId>com. ... .virtualHost</groupId> <artifactId>pom</artifactId> <version>0.0.1</version> <relativePath>../pom.xml</relativePath> </parent> <name>Virtual Host - ***</name> <groupId>com. ... .virtualHost</groupId> <artifactId>***</artifactId> <version>0.0.1</version> <packaging>pom</packaging> I filtered the name out. Is this POM correct? I just want to bundle files for a particular virtual host together. Thanks, Walter

    Read the article

  • C2244 when trying to call the pow function from inline assembly

    - by schrödingers cat
    I would like to call the pow function from inline assembly. The problem is i'm getting error C2244: 'pow' : unable to match function definition to an existing declaration. I'm new to assembly so this may be a trivial question but how do i resolve this? I guess it has something to do with the compiler not beeing able to properly resolve the overload of pow. The following code fragment is causing the error: do_POW: // push first argument to the stack sub esp, size value_type fld qword ptr [ecx] fstp qword ptr [esp] // push second argument to the stack sub esp, size value_type fld qword ptr [ecx - size value_type] fstp qword ptr [esp]and pop fpu stack // call the pow function call pow sub ecx, size value_type fstp qword ptr [ecx] add esp, 2 * size value_type jmp loop_start

    Read the article

  • maven: Including module assembly into assembly of parent project

    - by Ilya Izhovkin
    I have a set of projects, each has been configured to build with maven, and each has assembly plugin enabled. So, I have something like that: project1/ target/ lib/ ... project1-1.0.jar start.sh ... project2/ target/ lib/ ... project2-1.0.jar start.sh ... I need to create parent project, and configure assembly for it, which must include all subproject's assemblies. parent_project/ target/ project1/ lib/ ... project1-1.0.jar project2/ lib/ ... project2-1.0.jar start.sh ... How can I achieve this?

    Read the article

  • Anatomy of a .NET Assembly - PE Headers

    - by Simon Cooper
    Today, I'll be starting a look at what exactly is inside a .NET assembly - how the metadata and IL is stored, how Windows knows how to load it, and what all those bytes are actually doing. First of all, we need to understand the PE file format. PE files .NET assemblies are built on top of the PE (Portable Executable) file format that is used for all Windows executables and dlls, which itself is built on top of the MSDOS executable file format. The reason for this is that when .NET 1 was released, it wasn't a built-in part of the operating system like it is nowadays. Prior to Windows XP, .NET executables had to load like any other executable, had to execute native code to start the CLR to read & execute the rest of the file. However, starting with Windows XP, the operating system loader knows natively how to deal with .NET assemblies, rendering most of this legacy code & structure unnecessary. It still is part of the spec, and so is part of every .NET assembly. The result of this is that there are a lot of structure values in the assembly that simply aren't meaningful in a .NET assembly, as they refer to features that aren't needed. These are either set to zero or to certain pre-defined values, specified in the CLR spec. There are also several fields that specify the size of other datastructures in the file, which I will generally be glossing over in this initial post. Structure of a PE file Most of a PE file is split up into separate sections; each section stores different types of data. For instance, the .text section stores all the executable code; .rsrc stores unmanaged resources, .debug contains debugging information, and so on. Each section has a section header associated with it; this specifies whether the section is executable, read-only or read/write, whether it can be cached... When an exe or dll is loaded, each section can be mapped into a different location in memory as the OS loader sees fit. In order to reliably address a particular location within a file, most file offsets are specified using a Relative Virtual Address (RVA). This specifies the offset from the start of each section, rather than the offset within the executable file on disk, so the various sections can be moved around in memory without breaking anything. The mapping from RVA to file offset is done using the section headers, which specify the range of RVAs which are valid within that section. For example, if the .rsrc section header specifies that the base RVA is 0x4000, and the section starts at file offset 0xa00, then an RVA of 0x401d (offset 0x1d within the .rsrc section) corresponds to a file offset of 0xa1d. Because each section has its own base RVA, each valid RVA has a one-to-one mapping with a particular file offset. PE headers As I said above, most of the header information isn't relevant to .NET assemblies. To help show what's going on, I've created a diagram identifying all the various parts of the first 512 bytes of a .NET executable assembly. I've highlighted the relevant bytes that I will refer to in this post: Bear in mind that all numbers are stored in the assembly in little-endian format; the hex number 0x0123 will appear as 23 01 in the diagram. The first 64 bytes of every file is the DOS header. This starts with the magic number 'MZ' (0x4D, 0x5A in hex), identifying this file as an executable file of some sort (an .exe or .dll). Most of the rest of this header is zeroed out. The important part of this header is at offset 0x3C - this contains the file offset of the PE signature (0x80). Between the DOS header & PE signature is the DOS stub - this is a stub program that simply prints out 'This program cannot be run in DOS mode.\r\n' to the console. I will be having a closer look at this stub later on. The PE signature starts at offset 0x80, with the magic number 'PE\0\0' (0x50, 0x45, 0x00, 0x00), identifying this file as a PE executable, followed by the PE file header (also known as the COFF header). The relevant field in this header is in the last two bytes, and it specifies whether the file is an executable or a dll; bit 0x2000 is set for a dll. Next up is the PE standard fields, which start with a magic number of 0x010b for x86 and AnyCPU assemblies, and 0x20b for x64 assemblies. Most of the rest of the fields are to do with the CLR loader stub, which I will be covering in a later post. After the PE standard fields comes the NT-specific fields; again, most of these are not relevant for .NET assemblies. The one that is is the highlighted Subsystem field, and specifies if this is a GUI or console app - 0x20 for a GUI app, 0x30 for a console app. Data directories & section headers After the PE and COFF headers come the data directories; each directory specifies the RVA (first 4 bytes) and size (next 4 bytes) of various important parts of the executable. The only relevant ones are the 2nd (Import table), 13th (Import Address table), and 15th (CLI header). The Import and Import Address table are only used by the startup stub, so we will look at those later on. The 15th points to the CLI header, where the CLR-specific metadata begins. After the data directories comes the section headers; one for each section in the file. Each header starts with the section's ASCII name, null-padded to 8 bytes. Again, most of each header is irrelevant, but I've highlighted the base RVA and file offset in each header. In the diagram, you can see the following sections: .text: base RVA 0x2000, file offset 0x200 .rsrc: base RVA 0x4000, file offset 0xa00 .reloc: base RVA 0x6000, file offset 0x1000 The .text section contains all the CLR metadata and code, and so is by far the largest in .NET assemblies. The .rsrc section contains the data you see in the Details page in the right-click file properties page, but is otherwise unused. The .reloc section contains address relocations, which we will look at when we study the CLR startup stub. What about the CLR? As you can see, most of the first 512 bytes of an assembly are largely irrelevant to the CLR, and only a few bytes specify needed things like the bitness (AnyCPU/x86 or x64), whether this is an exe or dll, and the type of app this is. There are some bytes that I haven't covered that affect the layout of the file (eg. the file alignment, which determines where in a file each section can start). These values are pretty much constant in most .NET assemblies, and don't affect the CLR data directly. Conclusion To summarize, the important data in the first 512 bytes of a file is: DOS header. This contains a pointer to the PE signature. DOS stub, which we'll be looking at in a later post. PE signature PE file header (aka COFF header). This specifies whether the file is an exe or a dll. PE standard fields. This specifies whether the file is AnyCPU/32bit or 64bit. PE NT-specific fields. This specifies what type of app this is, if it is an app. Data directories. The 15th entry (at offset 0x168) contains the RVA and size of the CLI header inside the .text section. Section headers. These are used to map between RVA and file offset. The important one is .text, which is where all the CLR data is stored. In my next post, we'll start looking at the metadata used by the CLR directly, which is all inside the .text section.

    Read the article

  • Assembly Language being used in Aircraft System

    - by caramel23
    Today my lecturer mentioned the reason why the aircraft system is programmed in assembly language is due to the program being written have less error . Is this statement true ? Because when he asked about our opinion I said assembly can create faster program thus it is a good language for real-time oriented aircraft system program . I search around google but can't seem to find an article clarifying my lecturer's statement .

    Read the article

  • Why would I use Assembly.LoadFile in lieu of Assembly.LoadFrom?

    - by Cheeso
    It's my impression that Assembly.LoadFrom uses the ApplicationBase and PrivateBinPath. It also my impression that Assembly.LoadFile does not. Why would anyone want to use LoadFile? In other words, if my understanding is correct, why would anyone want to NOT use the ApplicationBase and PrivateBinPath? I'm working with some existing code, which uses LoadFile, and I don't understand why it would do so. LoadFile apparently does not load dependencies from the same directory. The LoadFrom method does load dependencies (From the doc: The load-from context...allows dependencies on that path to be found and loaded because the path information is maintained by the context.) I'd like to convert it from using LoadFile, to use LoadFrom. What is likely to break, if anything, if I replace LoadFile with LoadFrom? Even if it iss benign, it may be that I cannot do the replacement, just based on project schedules. If I cannot replace LoadFile with LoadFrom, is there a way to convince assemblies loaded with LoadFile to load dependencies? Is there a packaging trick I can use (embedded assembly, ILMerge, an AssemblyResolve event, something like that) that can allow an assembly loaded with LoadFile to also load its dependencies?

    Read the article

  • C inline assembly of x96 fbstp instruction

    - by David HUnter
    Was wondering how to inline a usage of fbstp on a 32 bit I86 architecture. I tried something like int main( ) { double foo = 100.0; long bar = 0; asm( "pushl %1; fbstp %0" : "=m"(bar) : "r"(foo) ); ... But bar is unchanged. I have tried reading everything I can find on this but most example simply do things like add two integers together. I can’t find any that talk about pushing operands onto the stack and what I should be doing when an instruction like fbstp writes 80 bits of data back to memory ( i.e. what C type to use ) and how to specify it in the asm syntax. Also on x86-64 there seems to be a pushq and no pushl but fbstp still exists whereas fbstq does not. Is there some other magic for 64 bit.

    Read the article

  • Mac OS ? Assembly Language Esoteria

    - by veryfoolish
    I've been playing around with assembly and object files in general on Mac OS ? and was wondering if somebody could provide some edification. Specifically, I'm wondering what the extra code GCC generates when compiling the C file in the following example does. I have a toy C program so I can comprehend the assembly output. int main() { int a = 5; int b = 5; int c = a + b; } Running this through gcc -S creates the following assembly: .text .globl _main _main: LFB2: pushq %rbp LCFI0: movq %rsp, %rbp LCFI1: movl $5, -4(%rbp) movl $5, -8(%rbp) movl -8(%rbp), %eax addl -4(%rbp), %eax movl %eax, -12(%rbp) leave ret LFE2: .section __TEXT,__eh_frame,coalesced,no_toc+strip_static_syms+live_support EH_frame1: .set L$set$0,LECIE1-LSCIE1 .long L$set$0 LSCIE1: .long 0x0 .byte 0x1 .ascii "zR\0" .byte 0x1 .byte 0x78 .byte 0x10 .byte 0x1 .byte 0x10 .byte 0xc .byte 0x7 .byte 0x8 .byte 0x90 .byte 0x1 .align 3 LECIE1: .globl _main.eh _main.eh: LSFDE1: .set L$set$1,LEFDE1-LASFDE1 .long L$set$1 LASFDE1: .long LASFDE1-EH_frame1 .quad LFB2-. .set L$set$2,LFE2-LFB2 .quad L$set$2 .byte 0x0 .byte 0x4 .set L$set$3,LCFI0-LFB2 .long L$set$3 .byte 0xe .byte 0x10 .byte 0x86 .byte 0x2 .byte 0x4 .set L$set$4,LCFI1-LCFI0 .long L$set$4 .byte 0xd .byte 0x6 .align 3 LEFDE1: .subsections_via_symbols The LCFI1 section seems to contain the actual logic for the program, but I'm not sure what the misc. other stuff is for... also, is there any scheme these labels are following? I'm sorry this is such a vague question. I'd appreciate anything, including being pointed to a resource where I can find out more about this. Thanks!

    Read the article

  • Mac OS ? Assembly Language Esoteria

    - by veryfoolish
    I've been playing around with assembly and object files in general on Mac OS ? and was wondering if somebody could provide some edification. Specifically, I'm wondering what the extra code GCC generates when compiling the C file in the following example does. I have a toy C program so I can comprehend the assembly output. int main() { int a = 5; int b = 5; int c = a + b; } Running this through gcc -S creates the following assembly: .text .globl _main _main: LFB2: pushq %rbp LCFI0: movq %rsp, %rbp LCFI1: movl $5, -4(%rbp) movl $5, -8(%rbp) movl -8(%rbp), %eax addl -4(%rbp), %eax movl %eax, -12(%rbp) leave ret LFE2: .section __TEXT,__eh_frame,coalesced,no_toc+strip_static_syms+live_support EH_frame1: .set L$set$0,LECIE1-LSCIE1 .long L$set$0 LSCIE1: .long 0x0 .byte 0x1 .ascii "zR\0" .byte 0x1 .byte 0x78 .byte 0x10 .byte 0x1 .byte 0x10 .byte 0xc .byte 0x7 .byte 0x8 .byte 0x90 .byte 0x1 .align 3 LECIE1: .globl _main.eh _main.eh: LSFDE1: .set L$set$1,LEFDE1-LASFDE1 .long L$set$1 LASFDE1: .long LASFDE1-EH_frame1 .quad LFB2-. .set L$set$2,LFE2-LFB2 .quad L$set$2 .byte 0x0 .byte 0x4 .set L$set$3,LCFI0-LFB2 .long L$set$3 .byte 0xe .byte 0x10 .byte 0x86 .byte 0x2 .byte 0x4 .set L$set$4,LCFI1-LCFI0 .long L$set$4 .byte 0xd .byte 0x6 .align 3 LEFDE1: .subsections_via_symbols The LCFI1 section seems to contain the actual logic for the program, but I'm not sure what the misc. other stuff is for... also, is there any scheme these labels are following? I'm sorry this is such a vague question. I'd appreciate anything, including being pointed to a resource where I can find out more about this. Thanks!

    Read the article

  • Back to Basics: When does a .NET Assembly Dependency get loaded

    - by Rick Strahl
    When we work on typical day to day applications, it's easy to forget some of the core features of the .NET framework. For me personally it's been a long time since I've learned about some of the underlying CLR system level services even though I rely on them on a daily basis. I often think only about high level application constructs and/or high level framework functionality, but the low level stuff is often just taken for granted. Over the last week at DevConnections I had all sorts of low level discussions with other developers about the inner workings of this or that technology (especially in light of my Low Level ASP.NET Architecture talk and the Razor Hosting talk). One topic that came up a couple of times and ended up a point of confusion even amongst some seasoned developers (including some folks from Microsoft <snicker>) is when assemblies actually load into a .NET process. There are a number of different ways that assemblies are loaded in .NET. When you create a typical project assemblies usually come from: The Assembly reference list of the top level 'executable' project The Assembly references of referenced projects Dynamically loaded at runtime via AppDomain/Reflection loading In addition .NET automatically loads mscorlib (most of the System namespace) the boot process that hosts the .NET runtime in EXE apps, or some other kind of runtime hosting environment (runtime hosting in servers like IIS, SQL Server or COM Interop). In hosting environments the runtime host may also pre-load a bunch of assemblies on its own (for example the ASP.NET host requires all sorts of assemblies just to run itself, before ever routing into your user specific code). Assembly Loading The most obvious source of loaded assemblies is the top level application's assembly reference list. You can add assembly references to a top level application and those assembly references are then available to the application. In a nutshell, referenced assemblies are not immediately loaded - they are loaded on the fly as needed. So regardless of whether you have an assembly reference in a top level project, or a dependent assembly assemblies typically load on an as needed basis, unless explicitly loaded by user code. The same is true of dependent assemblies. To check this out I ran a simple test: I have a utility assembly Westwind.Utilities which is a general purpose library that can work in any type of project. Due to a couple of small requirements for encoding and a logging piece that allows logging Web content (dependency on HttpContext.Current) this utility library has a dependency on System.Web. Now System.Web is a pretty large assembly and generally you'd want to avoid adding it to a non-Web project if it can be helped. So I created a Console Application that loads my utility library: You can see that the top level Console app a reference to Westwind.Utilities and System.Data (beyond the core .NET libs). The Westwind.Utilities project on the other hand has quite a few dependencies including System.Web. I then add a main program that accesses only a simple utillity method in the Westwind.Utilities library that doesn't require any of the classes that access System.Web: static void Main(string[] args) { Console.WriteLine(StringUtils.NewStringId()); Console.ReadLine(); } StringUtils.NewStringId() calls into Westwind.Utilities, but it doesn't rely on System.Web. Any guesses what the assembly list looks like when I stop the code on the ReadLine() command? I'll wait here while you think about it… … … So, when I stop on ReadLine() and then fire up Process Explorer and check the assembly list I get: We can see here that .NET has not actually loaded any of the dependencies of the Westwind.Utilities assembly. Also not loaded is the top level System.Data reference even though it's in the dependent assembly list of the top level project. Since this particular function I called only uses core System functionality (contained in mscorlib) there's in fact nothing else loaded beyond the main application and my Westwind.Utilities assembly that contains the method accessed. None of the dependencies of Westwind.Utilities loaded. If you were to open the assembly in a disassembler like Reflector or ILSpy, you would however see all the compiled in dependencies. The referenced assemblies are in the dependency list and they are loadable, but they are not immediately loaded by the application. In other words the C# compiler and .NET linker are smart enough to figure out the dependencies based on the code that actually is referenced from your application and any dependencies cascading down into the dependencies from your top level application into the referenced assemblies. In the example above the usage requirement is pretty obvious since I'm only calling a single static method and then exiting the app, but in more complex applications these dependency relationships become very complicated - however it's all taken care of by the compiler and linker figuring out what types and members are actually referenced and including only those assemblies that are in fact referenced in your code or required by any of your dependencies. The good news here is: That if you are referencing an assembly that has a dependency on something like System.Web in a few places that are not actually accessed by any of your code or any dependent assembly code that you are calling, that assembly is never loaded into memory! Some Hosting Environments pre-load Assemblies The load behavior can vary however. In Console and desktop applications we have full control over assembly loading so we see the core CLR behavior. However other environments like ASP.NET for example will preload referenced assemblies explicitly as part of the startup process - primarily to minimize load conflicts. Specifically ASP.NET pre-loads all assemblies referenced in the assembly list and the /bin folder. So in Web applications it definitely pays to minimize your top level assemblies if they are not used. Understanding when Assemblies Load To clarify and see it actually happen what I described in the first example , let's look at a couple of other scenarios. To see assemblies loading at runtime in real time lets create a utility function to print out loaded assemblies to the console: public static void PrintAssemblies() { var assemblies = AppDomain.CurrentDomain.GetAssemblies(); foreach (var assembly in assemblies) { Console.WriteLine(assembly.GetName()); } } Now let's look at the first scenario where I have class method that references internally uses System.Web. In the first scenario lets add a method to my main program like this: static void Main(string[] args) { Console.WriteLine(StringUtils.NewStringId()); Console.ReadLine(); PrintAssemblies(); } public static void WebLogEntry() { var entry = new WebLogEntry(); entry.UpdateFromRequest(); Console.WriteLine(entry.QueryString); } UpdateFromWebRequest() internally accesses HttpContext.Current to read some information of the ASP.NET Request object so it clearly needs a reference System.Web to work. In this first example, the method that holds the calling code is never called, but exists as a static method that can potentially be called externally at some point. What do you think will happen here with the assembly loading? Will System.Web load in this example? No - it doesn't. Because the WebLogEntry() method is never called by the mainline application (or anywhere else) System.Web is not loaded. .NET dynamically loads assemblies as code that needs it is called. No code references the WebLogEntry() method and so System.Web is never loaded. Next, let's add the call to this method, which should trigger System.Web to be loaded because a dependency exists. Let's change the code to: static void Main(string[] args) { Console.WriteLine(StringUtils.NewStringId()); Console.WriteLine("--- Before:"); PrintAssemblies(); WebLogEntry(); Console.WriteLine("--- After:"); PrintAssemblies(); Console.ReadLine(); } public static void WebLogEntry() { var entry = new WebLogEntry(); entry.UpdateFromRequest(); Console.WriteLine(entry.QueryString); } Looking at the code now, when do you think System.Web will be loaded? Will the before list include it? Yup System.Web gets loaded, but only after it's actually referenced. In fact, just until before the call to UpdateFromRequest() System.Web is not loaded - it only loads when the method is actually called and requires the reference in the executing code. Moral of the Story So what have we learned - or maybe remembered again? Dependent Assembly References are not pre-loaded when an application starts (by default) Dependent Assemblies that are not referenced by executing code are never loaded Dependent Assemblies are just in time loaded when first referenced in code All of this is nothing new - .NET has always worked like this. But it's good to have a refresher now and then and go through the exercise of seeing it work in action. It's not one of those things we think about everyday, and as I found out last week, I couldn't remember exactly how it worked since it's been so long since I've learned about this. And apparently I'm not the only one as several other people I had discussions with in relation to loaded assemblies also didn't recall exactly what should happen or assumed incorrectly that just having a reference automatically loads the assembly. The moral of the story for me is: Trying at all costs to eliminate an assembly reference from a component is not quite as important as it's often made out to be. For example, the Westwind.Utilities module described above has a logging component, including a Web specific logging entry that supports pulling information from the active HTTP Context. Adding that feature requires a reference to System.Web. Should I worry about this in the scope of this library? Probably not, because if I don't use that one class of nearly a hundred, System.Web never gets pulled into the parent process. IOW, System.Web only loads when I use that specific feature and if I am, well I clearly have to be running in a Web environment anyway to use it realistically. The alternative would be considerably uglier: Pulling out the WebLogEntry class and sticking it into another assembly and breaking up the logging code. In this case - definitely not worth it. So, .NET definitely goes through some pretty nifty optimizations to ensure that it loads only what it needs and in most cases you can just rely on .NET to do the right thing. Sometimes though assembly loading can go wrong (especially when signed and versioned local assemblies are involved), but that's subject for a whole other post…© Rick Strahl, West Wind Technologies, 2005-2012Posted in .NET  CSharp   Tweet !function(d,s,id){var js,fjs=d.getElementsByTagName(s)[0];if(!d.getElementById(id)){js=d.createElement(s);js.id=id;js.src="//platform.twitter.com/widgets.js";fjs.parentNode.insertBefore(js,fjs);}}(document,"script","twitter-wjs"); (function() { var po = document.createElement('script'); po.type = 'text/javascript'; po.async = true; po.src = 'https://apis.google.com/js/plusone.js'; var s = document.getElementsByTagName('script')[0]; s.parentNode.insertBefore(po, s); })();

    Read the article

1 2 3 4 5 6 7 8 9 10 11 12  | Next Page >