Develop Reference

Read an XML standard config file selected from console

I created a CLI application which actually uses the standard config app.Config file.
In this file I put some subsections, like
<typicsTable>
<mainSettings>
<add key="sheetNumber" value="1"/>
<add key="firstDataRow" value="2"/>
</mainSettings>
</typicsTable>
I actually read these settings with
NameValueCollection TypicsConversionTableSettings = (NameValueCollection)ConfigurationManager.GetSection("typicsTable/mainSettings");
int ctSheetNumber = Int32.Parse(TypicsConversionTableSettings["sheetNumber"]);
int ctFirstDataRow = Int32.Parse(TypicsConversionTableSettings["firstDataRow"]);
Everything works fine in this way.
What I want to do now is
1) I want different config files with custom names (i.e. test1.config , test2.config) and take via CLI the right config file;
2) switch to a less ".net config file", and take data from a standard XML file.
I'm now focusing on the point 1, I tried different attempts, I used
ExeConfigurationFileMap configFileMap = new ExeConfigurationFileMap();
configFileMap.ExeConfigFilename = #"C:\folderTest\conf1.config";
Configuration config = ConfigurationManager.OpenMappedExeConfiguration(configFileMap, ConfigurationUserLevel.None);
But I absolutely don't get how to read sections AND subsections in the file. How can I do that?

The class that is going to help you, I believe, is System.Xml.Linq.
using System.Xml.Linq;
So Part 1 would be load the file into an XElement:
XElement xConfig = XElement.Load("app.simulated.config");
Here's a quick demo of how you can iterate through everything and also find a single element using a matching condition.
class Program
{
static void Main(string[] args)
{
Console.WriteLine("Iterating the config file values and attributes...");
Console.WriteLine("==================================================");
XElement xConfig = XElement.Load("app.simulated.config");
foreach (var element in xConfig.DescendantsAndSelf())
{
Console.WriteLine(element.Name);
foreach (var attribute in element.Attributes())
{
Console.WriteLine("\t" + attribute.Name + "," + attribute.Value);
}
}
Console.WriteLine();
Console.WriteLine("Finding a value using matching conditions.");
Console.WriteLine("==========================================");
XElement xel =
xConfig
.DescendantsAndSelf()
.FirstOrDefault(match =>
(match.Attribute("key") != null) &&
(match.Attribute("key").Value == "sheetNumber"));
Console.WriteLine(
"The value of 'sheetNumber' is " +
xel.Attribute("value").Value
);
// Pause
Console.ReadKey();
}
}
Clone or Download this example from GitHub.

C# - How to set a Service Ref. Client Endpoint [duplicate]

I'd like to programmatically modify my app.config file to set which service file endpoint should be used. What is the best way to do this at runtime? For reference:
<endpoint address="http://mydomain/MyService.svc"
binding="wsHttpBinding" bindingConfiguration="WSHttpBinding_IASRService"
contract="ASRService.IASRService" name="WSHttpBinding_IASRService">
<identity>
<dns value="localhost" />
</identity>
</endpoint>

Is this on the client side of things??
If so, you need to create an instance of WsHttpBinding, and an EndpointAddress, and then pass those two to the proxy client constructor that takes these two as parameters.
// using System.ServiceModel;
WSHttpBinding binding = new WSHttpBinding();
EndpointAddress endpoint = new EndpointAddress(new Uri("http://localhost:9000/MyService"));
MyServiceClient client = new MyServiceClient(binding, endpoint);
If it's on the server side of things, you'll need to programmatically create your own instance of ServiceHost, and add the appropriate service endpoints to it.
ServiceHost svcHost = new ServiceHost(typeof(MyService), null);
svcHost.AddServiceEndpoint(typeof(IMyService),
new WSHttpBinding(),
"http://localhost:9000/MyService");
Of course you can have multiple of those service endpoints added to your service host. Once you're done, you need to open the service host by calling the .Open() method.
If you want to be able to dynamically - at runtime - pick which configuration to use, you could define multiple configurations, each with a unique name, and then call the appropriate constructor (for your service host, or your proxy client) with the configuration name you wish to use.
E.g. you could easily have:
<endpoint address="http://mydomain/MyService.svc"
binding="wsHttpBinding" bindingConfiguration="WSHttpBinding_IASRService"
contract="ASRService.IASRService"
name="WSHttpBinding_IASRService">
<identity>
<dns value="localhost" />
</identity>
</endpoint>
<endpoint address="https://mydomain/MyService2.svc"
binding="wsHttpBinding" bindingConfiguration="SecureHttpBinding_IASRService"
contract="ASRService.IASRService"
name="SecureWSHttpBinding_IASRService">
<identity>
<dns value="localhost" />
</identity>
</endpoint>
<endpoint address="net.tcp://mydomain/MyService3.svc"
binding="netTcpBinding" bindingConfiguration="NetTcpBinding_IASRService"
contract="ASRService.IASRService"
name="NetTcpBinding_IASRService">
<identity>
<dns value="localhost" />
</identity>
</endpoint>
(three different names, different parameters by specifying different bindingConfigurations) and then just pick the right one to instantiate your server (or client proxy).
But in both cases - server and client - you have to pick before actually creating the service host or the proxy client. Once created, these are immutable - you cannot tweak them once they're up and running.
Marc

I use the following code to change the endpoint address in the App.Config file. You may want to modify or remove the namespace before usage.
using System;
using System.Xml;
using System.Configuration;
using System.Reflection;
//...
namespace Glenlough.Generations.SupervisorII
{
public class ConfigSettings
{
private static string NodePath = "//system.serviceModel//client//endpoint";
private ConfigSettings() { }
public static string GetEndpointAddress()
{
return ConfigSettings.loadConfigDocument().SelectSingleNode(NodePath).Attributes["address"].Value;
}
public static void SaveEndpointAddress(string endpointAddress)
{
// load config document for current assembly
XmlDocument doc = loadConfigDocument();
// retrieve appSettings node
XmlNode node = doc.SelectSingleNode(NodePath);
if (node == null)
throw new InvalidOperationException("Error. Could not find endpoint node in config file.");
try
{
// select the 'add' element that contains the key
//XmlElement elem = (XmlElement)node.SelectSingleNode(string.Format("//add[#key='{0}']", key));
node.Attributes["address"].Value = endpointAddress;
doc.Save(getConfigFilePath());
}
catch( Exception e )
{
throw e;
}
}
public static XmlDocument loadConfigDocument()
{
XmlDocument doc = null;
try
{
doc = new XmlDocument();
doc.Load(getConfigFilePath());
return doc;
}
catch (System.IO.FileNotFoundException e)
{
throw new Exception("No configuration file found.", e);
}
}
private static string getConfigFilePath()
{
return Assembly.GetExecutingAssembly().Location + ".config";
}
}
}

SomeServiceClient client = new SomeServiceClient();
var endpointAddress = client.Endpoint.Address; //gets the default endpoint address
EndpointAddressBuilder newEndpointAddress = new EndpointAddressBuilder(endpointAddress);
newEndpointAddress.Uri = new Uri("net.tcp://serverName:8000/SomeServiceName/");
client = new SomeServiceClient("EndpointConfigurationName", newEndpointAddress.ToEndpointAddress());
I did it like this. The good thing is it still picks up the rest of your endpoint binding settings from the config and just replaces the URI.

this short code worked for me:
Configuration wConfig = ConfigurationManager.OpenExeConfiguration(ConfigurationUserLevel.None);
ServiceModelSectionGroup wServiceSection = ServiceModelSectionGroup.GetSectionGroup(wConfig);
ClientSection wClientSection = wServiceSection.Client;
wClientSection.Endpoints[0].Address = <your address>;
wConfig.Save();
Of course you have to create the ServiceClient proxy AFTER the config has changed.
You also need to reference the System.Configuration and System.ServiceModel assemblies to make this work.
Cheers

This is the shortest code that you can use to update the app config file even if don't have a config section defined:
void UpdateAppConfig(string param)
{
var doc = new XmlDocument();
doc.Load("YourExeName.exe.config");
XmlNodeList endpoints = doc.GetElementsByTagName("endpoint");
foreach (XmlNode item in endpoints)
{
var adressAttribute = item.Attributes["address"];
if (!ReferenceEquals(null, adressAttribute))
{
adressAttribute.Value = string.Format("http://mydomain/{0}", param);
}
}
doc.Save("YourExeName.exe.config");
}

I have modified and extended Malcolm Swaine's code to modify a specific node by it's name attribute, and to also modify an external config file. Hope it helps.
using System;
using System.Collections.Generic;
using System.Linq;
using System.Text;
using System.Xml;
using System.Reflection;
namespace LobbyGuard.UI.Registration
{
public class ConfigSettings
{
private static string NodePath = "//system.serviceModel//client//endpoint";
private ConfigSettings() { }
public static string GetEndpointAddress()
{
return ConfigSettings.loadConfigDocument().SelectSingleNode(NodePath).Attributes["address"].Value;
}
public static void SaveEndpointAddress(string endpointAddress)
{
// load config document for current assembly
XmlDocument doc = loadConfigDocument();
// retrieve appSettings node
XmlNodeList nodes = doc.SelectNodes(NodePath);
foreach (XmlNode node in nodes)
{
if (node == null)
throw new InvalidOperationException("Error. Could not find endpoint node in config file.");
//If this isnt the node I want to change, look at the next one
//Change this string to the name attribute of the node you want to change
if (node.Attributes["name"].Value != "DataLocal_Endpoint1")
{
continue;
}
try
{
// select the 'add' element that contains the key
//XmlElement elem = (XmlElement)node.SelectSingleNode(string.Format("//add[#key='{0}']", key));
node.Attributes["address"].Value = endpointAddress;
doc.Save(getConfigFilePath());
break;
}
catch (Exception e)
{
throw e;
}
}
}
public static void SaveEndpointAddress(string endpointAddress, string ConfigPath, string endpointName)
{
// load config document for current assembly
XmlDocument doc = loadConfigDocument(ConfigPath);
// retrieve appSettings node
XmlNodeList nodes = doc.SelectNodes(NodePath);
foreach (XmlNode node in nodes)
{
if (node == null)
throw new InvalidOperationException("Error. Could not find endpoint node in config file.");
//If this isnt the node I want to change, look at the next one
if (node.Attributes["name"].Value != endpointName)
{
continue;
}
try
{
// select the 'add' element that contains the key
//XmlElement elem = (XmlElement)node.SelectSingleNode(string.Format("//add[#key='{0}']", key));
node.Attributes["address"].Value = endpointAddress;
doc.Save(ConfigPath);
break;
}
catch (Exception e)
{
throw e;
}
}
}
public static XmlDocument loadConfigDocument()
{
XmlDocument doc = null;
try
{
doc = new XmlDocument();
doc.Load(getConfigFilePath());
return doc;
}
catch (System.IO.FileNotFoundException e)
{
throw new Exception("No configuration file found.", e);
}
}
public static XmlDocument loadConfigDocument(string Path)
{
XmlDocument doc = null;
try
{
doc = new XmlDocument();
doc.Load(Path);
return doc;
}
catch (System.IO.FileNotFoundException e)
{
throw new Exception("No configuration file found.", e);
}
}
private static string getConfigFilePath()
{
return Assembly.GetExecutingAssembly().Location + ".config";
}
}
}

MyServiceClient client = new MyServiceClient(binding, endpoint);
client.Endpoint.Address = new EndpointAddress("net.tcp://localhost/webSrvHost/service.svc");
client.Endpoint.Binding = new NetTcpBinding()
{
Name = "yourTcpBindConfig",
ReaderQuotas = XmlDictionaryReaderQuotas.Max,
ListenBacklog = 40 }
It's very easy to modify the uri in config or binding info in config.
Is this what you want?

For what it's worth, I needed to update the port and scheme for SSL for my RESTFul service. This is what I did. Apologies that it is a bit more that the original question, but hopefully useful to someone.
// Don't forget to add references to System.ServiceModel and System.ServiceModel.Web
using System.ServiceModel;
using System.ServiceModel.Configuration;
var port = 1234;
var isSsl = true;
var scheme = isSsl ? "https" : "http";
var currAssembly = System.Reflection.Assembly.GetExecutingAssembly().CodeBase;
Configuration config = ConfigurationManager.OpenExeConfiguration(currAssembly);
ServiceModelSectionGroup serviceModel = ServiceModelSectionGroup.GetSectionGroup(config);
// Get the first endpoint in services. This is my RESTful service.
var endp = serviceModel.Services.Services[0].Endpoints[0];
// Assign new values for endpoint
UriBuilder b = new UriBuilder(endp.Address);
b.Port = port;
b.Scheme = scheme;
endp.Address = b.Uri;
// Adjust design time baseaddress endpoint
var baseAddress = serviceModel.Services.Services[0].Host.BaseAddresses[0].BaseAddress;
b = new UriBuilder(baseAddress);
b.Port = port;
b.Scheme = scheme;
serviceModel.Services.Services[0].Host.BaseAddresses[0].BaseAddress = b.Uri.ToString();
// Setup the Transport security
BindingsSection bindings = serviceModel.Bindings;
WebHttpBindingCollectionElement x =(WebHttpBindingCollectionElement)bindings["webHttpBinding"];
WebHttpBindingElement y = (WebHttpBindingElement)x.ConfiguredBindings[0];
var e = y.Security;
e.Mode = isSsl ? WebHttpSecurityMode.Transport : WebHttpSecurityMode.None;
e.Transport.ClientCredentialType = HttpClientCredentialType.None;
// Save changes
config.Save();

I think what you want is to swap out at runtime a version of your config file, if so create a copy of your config file (also give it the relevant extension like .Debug or .Release) that has the correct addresses (which gives you a debug version and a runtime version ) and create a postbuild step that copies the correct file depending on the build type.
Here is an example of a postbuild event that I've used in the past which overrides the output file with the correct version (debug/runtime)
copy "$(ProjectDir)ServiceReferences.ClientConfig.$(ConfigurationName)" "$(ProjectDir)ServiceReferences.ClientConfig" /Y
where :
$(ProjectDir) is the project directory where the config files are located
$(ConfigurationName) is the active configuration build type
EDIT:
Please see Marc's answer for a detailed explanation on how to do this programmatically.

You can do it like this:
Keep your settings in a separate xml file and read through it when you create a proxy for your service.
For example , i want to modify my service endpoint address at runtime so i have the following ServiceEndpoint.xml file.
<?xml version="1.0" encoding="utf-8" ?>
<Services>
<Service name="FileTransferService">
<Endpoints>
<Endpoint name="ep1" address="http://localhost:8080/FileTransferService.svc" />
</Endpoints>
</Service>
</Services>
For reading your xml :
var doc = new XmlDocument();
doc.Load(FileTransferConstants.Constants.SERVICE_ENDPOINTS_XMLPATH);
XmlNodeList endPoints = doc.SelectNodes("/Services/Service/Endpoints");
foreach (XmlNode endPoint in endPoints)
{
foreach (XmlNode child in endPoint)
{
if (child.Attributes["name"].Value.Equals("ep1"))
{
var adressAttribute = child.Attributes["address"];
if (!ReferenceEquals(null, adressAttribute))
{
address = adressAttribute.Value;
}
}
}
}
Then get your web.config file of your client at runtime and assign the service endpoint address as:
Configuration wConfig = ConfigurationManager.OpenMappedExeConfiguration(new ExeConfigurationFileMap { ExeConfigFilename = #"C:\FileTransferWebsite\web.config" }, ConfigurationUserLevel.None);
ServiceModelSectionGroup wServiceSection = ServiceModelSectionGroup.GetSectionGroup(wConfig);
ClientSection wClientSection = wServiceSection.Client;
wClientSection.Endpoints[0].Address = new Uri(address);
wConfig.Save();

see if you are placing the client section in the correct web.config file. SharePoint has around 6 to 7 config files.
http://msdn.microsoft.com/en-us/library/office/ms460914(v=office.14).aspx (http://msdn.microsoft.com/en-us/library/office/ms460914%28v=office.14%29.aspx)
Post this you can simply try
ServiceClient client = new ServiceClient("ServiceSOAP");

Modifying app.config via custom msi installer

I need to create an address string in app.config as:
<client>
<endpoint address="http://ServerName/xxx/yyy.svc"
binding="basicHttpBinding" bindingConfiguration="BasicHttpBinding_IClientIInfoService"
contract="DocuHealthLinkSvcRef.IClientIInfoService" name="BasicHttpBinding_IClientIInfoService" />
</client>
The ServerName need to be entered by the user during installation.
For that i have created a new UI dialog in the Installer. I have also written an Installer.cs class and overrided the install () as:
public override void Install(System.Collections.IDictionary stateSaver)
{
base.Install(stateSaver);
string targetDirectory = Context.Parameters["targetdir"];
string ServerName = Context.Parameters["ServerName"];
System.Diagnostics.Debugger.Break();
string exePath = string.Format("{0}myapp.exe", targetDirectory);
Configuration config = ConfigurationManager.OpenExeConfiguration(exePath);
config.AppSettings.Settings["ServerName"].Value = ServerName;
config.Save();
}
}
But how do i use this ServerName in my app.config to create the specified string.
I'm working on VS2010.

You could use WiX (Windows Installer XML toolset) to build your MSI, in which case you can use the XmlFile utility tag to update the server name:
<util:XmlFile Id="UpdateServerName" File="[INSTALLLOCATION]AppName.exe.config" Action="setValue" ElementPath="/client/endpoint" Name="address" Value="http://[SERVERNAME]/xxx/yyy.svc" />
You can capture the server name during installation using a WixUI extension form.
Advantages of WiX: WiX is msbuild compliant (unlike .vdproj files), and gives you much finer-grained control over your installer, among other things

Assuming you are using the full ServiceModel section group in the app.config
Essentially you follow these steps:
Load ServiceModel config section
Get Client Section
Get ChannelEndpoint Element
Change Address value by replacing string "ServerName" with entered value
Set Address attribute to new value
Save config
public override void Install(System.Collections.IDictionary stateSaver)
{
base.Install(stateSaver);
string targetDirectory = Context.Parameters["targetdir"];
string ServerName = Context.Parameters["ServerName"];
System.Diagnostics.Debugger.Break();
string exePath = string.Format("{0}myapp.exe", targetDirectory);
Configuration config = ConfigurationManager.OpenExeConfiguration(exePath);
config.AppSettings.Settings["ServerName"].Value = ServerName;
//Get ServiceModelSectionGroup from config
ServiceModelSectionGroup group = ServiceModelSectionGroup.GetSectionGroup (config);
//get the client section
ClientSection clientSection = group.Client;
//get the first endpoint
ChannelEndpointElement channelEndpointElement = clientSection.Endpoints[0];
//get the address attribute and replace servername in the string.
string address = channelEndpointElement.Address.ToString().Replace("ServerName", ServerName);
//set the Address attribute to the new value
channelEndpointElement.Address = new Uri(address);
config.Save();
}

At the end of the day, app.config is xml file. You can use Linq To XML or XPathNavigator to replace the address attribute of endpoint element.
Below code uses Linq to Xml
using System.Xml.Linq;
public override void Install(System.Collections.IDictionary stateSaver)
{
base.Install(stateSaver);
string targetDirectory = Context.Parameters["targetdir"];
string ServerName = Context.Parameters["ServerName"];
System.Diagnostics.Debugger.Break();
string configPath = string.Format("{0myapp.exe.config", targetDirectory);
XElement root = XElement.Load(configPath);
var endPointElements = root.Descendants("endpoint");
foreach(var element in endPointElements)
{
element.Attribute("address").Value = ServerName;
}
root.Save(configPath);
}
}

Since you have a windows-installer tag, I assume you either have an MSI package, or can create one...
Then:
You can create a public MSI property like ENDPOINTSERVER you require
during installation.
Add a custom action that modifies app.config to
run after "InstallFinalize" with the value of ENDPOINTSERVER
A silent installation will be possible using:
msiexec /i app.msi ENDPOINTSERVER=www.MyServer.com /qb-

try to use below two lines before saving the config file changes:
config.Save(ConfigurationSaveMode.Modified);
ConfigurationManager.RefreshSection("Section Name");
root.Save(configPath);
P.S: it doesn't update the solution item 'app.config', but the '.exe.config' one in the bin/ folder if you run it with F5.

Modifying App.Config file at the time of installation using c#

<configuration>
<configSections>
<section name="ADMIN" type="System.Configuration.DictionarySectionHandler"/>
</configSections>
<User>
<add key="ExtendTime" value="20"/>
<add key="Name" value="sss"/>
</User>
<configuration>
i have to remove first child element in user config section i.e . Reply me if you have any idea for this.
i am using
Configuration config = ConfigurationManager.OpenExeConfiguration(Context.Parameters["assemblypath"]);
ConfigurationSection section = config.GetSection("USER");

This article may have what you're looking for : http://raquila.com/software/configure-app-config-application-settings-during-msi-install/
Excerpt from article:
string exePath = string.Format("{0}MyWindowsFormsApplication.exe", targetDirectory);
Configuration config = ConfigurationManager.OpenExeConfiguration(exePath);
config.AppSettings.Settings["Param1"].Value = param1;
config.AppSettings.Settings["Param2"].Value = param2;
config.AppSettings.Settings["Param3"].Value = param3;
config.Save();
EDIT: Adding additional code sample and blog reference: http://ryanfarley.com/blog/archive/2004/07/13/879.aspx
using System;
using System.Xml;
using System.Configuration;
using System.Reflection;
//...
public class ConfigSettings
{
private ConfigSettings() {}
public static string ReadSetting(string key)
{
return ConfigurationSettings.AppSettings[key];
}
public static void WriteSetting(string key, string value)
{
// load config document for current assembly
XmlDocument doc = loadConfigDocument();
// retrieve appSettings node
XmlNode node = doc.SelectSingleNode("//appSettings");
if (node == null)
throw new InvalidOperationException("appSettings section not found in config file.");
try
{
// select the 'add' element that contains the key
XmlElement elem = (XmlElement)node.SelectSingleNode(string.Format("//add[#key='{0}']", key));
if (elem != null)
{
// add value for key
elem.SetAttribute("value", value);
}
else
{
// key was not found so create the 'add' element
// and set it's key/value attributes
elem = doc.CreateElement("add");
elem.SetAttribute("key", key);
elem.SetAttribute("value", value);
node.AppendChild(elem);
}
doc.Save(getConfigFilePath());
}
catch
{
throw;
}
}
public static void RemoveSetting(string key)
{
// load config document for current assembly
XmlDocument doc = loadConfigDocument();
// retrieve appSettings node
XmlNode node = doc.SelectSingleNode("//appSettings");
try
{
if (node == null)
throw new InvalidOperationException("appSettings section not found in config file.");
else
{
// remove 'add' element with coresponding key
node.RemoveChild(node.SelectSingleNode(string.Format("//add[#key='{0}']", key)));
doc.Save(getConfigFilePath());
}
}
catch (NullReferenceException e)
{
throw new Exception(string.Format("The key {0} does not exist.", key), e);
}
}
private static XmlDocument loadConfigDocument()
{
XmlDocument doc = null;
try
{
doc = new XmlDocument();
doc.Load(getConfigFilePath());
return doc;
}
catch (System.IO.FileNotFoundException e)
{
throw new Exception("No configuration file found.", e);
}
}
private static string getConfigFilePath()
{
return Assembly.GetExecutingAssembly().Location + ".config";
}
}
Then you would use it like this:
// read the Test1 value from the config file
string test1 = ConfigSettings.ReadSetting("Test1");
// write a new value for the Test1 setting
ConfigSettings.WriteSetting("Test1", "This is my new value");
// remove the Test1 setting from the config file
ConfigSettings.RemoveSetting("Test1");

I have come to the conclusion that it is not possible to access a custom configuration section during installation using:
MyCustomConfigurationSection section = (MyCustomConfigurationSection)config.GetSection("MyCustomConfigurationSection");
When the MSI package is installed, the program executed is the Windows Install(MsiExec), not the program which contains the installer class.
'%windir%\system32\msiexec.exe
In order to access the config we need to workaround this issue either by using the context:
Configuration config = ConfigurationManager.OpenExeConfiguration(this.Context.Parameters["assemblypath"]);
Or, by using reflection retrieve the location of the executing assembly:
Configuration config = ConfigurationManager.OpenExeConfiguration(System.Reflection.Assembly.GetExecutingAssembly().Location);
As Chuck suggested, you can access the AppSettings and modify them:
AppSettingsSection appSettings = (AppSettingsSection)config.GetSection("appSettings");
appSettings.Settings["Environment"].Value = _Environment;
config.Save();
This is fine, because the installer knows exactly how to deal with
System.Configuration.AppSettingsSection
because that library is part of .NET. However, when it comes to a custom section, the installer needs to know how to deal with that custom config. More than likely, you have it in a class library (in a DLL) that is referenced by your application and that DLL has now been installed in the install directory.
The problem is, as we know from above, MSIExec.exe isn't running in the context of that directory, so the install fails, when it can't find the appropriate DLL in system32, it throws the error:
An error occurred creating the
configuration section handler for
'XXX': Could not load file or assembly
'XXX.dll' or one of its dependencies.
The system cannot find the file
specified.
The only way therefore, to access the custom config, is to treat the config file as an XML document, and edit it using traditional XML management tools:
// load the doc
XmlDocument doc = new XmlDocument();
doc.Load(Assembly.GetExecutingAssembly().Location + ".config");
// Get the node
XmlNode node = doc.SelectSingleNode("//MyCustomConfigurationSection");
// edit node here
// ...
// Save
doc.Save(Assembly.GetExecutingAssembly().Location + ".config");
This technique is described on Ryan Farley's blog, as Chuck pointed out in the comments to his original answer.

Good news! I have found a way how to work-around this problem.
Solution is to intercept loading of assembly and return one we have. To do so
ResolveEventHandler handler = new ResolveEventHandler(CurrentDomain_AssemblyResolve);
AppDomain.CurrentDomain.AssemblyResolve += handler;
try
{
section = config.GetSection("mySection") as MySection;
}
catch(Exception)
{
}
AppDomain.CurrentDomain.AssemblyResolve -= handler;
and
Assembly CurrentDomain_AssemblyResolve(object sender, ResolveEventArgs args)
{
if (args.Name == "Dead.Beef.Rocks")
{
return typeof(MySection).Assembly;
}
return null;
}

WCF Can't Return XmlElement from XmlDataDocument.DocumentElement

I'm building a WCF web service which returns a composite object that looks similar to the following:
[DataContract]
public class WebServiceReturn
{
...
[DataMember]
public XmlElement Results { get; set; }
...
}
When I return a WebServiceReturn object with the following code, everything is fine:
XElement cities = new XElement("Cities",
from r in results
select new XElement("City", r));
using (XmlReader xmlReader = cities.CreateReader())
{
XmlDocument xmlDoc = new XmlDocument();
xmlDoc.Load(xmlReader);
WebServiceReturn response = new WebServiceReturn();
response.Results = xmlDoc.DocumentElement;
}
However, when I use the code below, which takes an XmlElement from the results of a stored procedure call that returns an XmlDataDocument, a CommunicationException is thrown (which has no inner exceptions).
XmlDataDocument xdd = DataAccess.ExecuteXML("MyStoredProc", parameter);
response.Results = xdd.DocumentElement;
The confusing part is if I convert the XmlDataDocument.DocumentElement (which is an XmlElement) into an XElement and then back into an XmlElement, there are no problems (wow that was a mouthful) - so the following code returns with no problem.
XmlElement xe = DataAccess.ExecuteXML("MyStoredProc", parameter).DocumentElement;
XDocument xDoc = new XDocument();
using (XmlWriter xmlWriter = xDoc.CreateWriter()){
xe.WriteTo(xmlWriter);
}
using (XmlReader xmlReader = xDoc.Root.CreateReader())
{
XmlDocument xmlDoc = new XmlDocument();
xmlDoc.Load(xmlReader);
response.Results = xmlDoc.DocumentElement;
}
The communicationexception details are:
[CommunicationException: The server did not provide a meaningful reply; this might be caused by a contract mismatch, a premature session shutdown or an internal server error.]
I have also updated the service reference in my test application multiple times which has had no effect.
Is the problem with my test code that is calling the web service? Why would converting an XmlElement into an XElement and then back into an XmlElement fix the issue? Any information at all would be much appreciated! :)

I don't know anything odd about XmlDataDocument, but you don't necessarily need the XDocument - try:
XmlDocument newDoc = new XmlDocument();
newDoc.Load(new XmlNodeReader(doc.DocumentElement));
return newDoc.DocumentElement;
Still not ideal, but it looks cleaner to me...

Well, in order to get more error information, you'll need to enable the debugging details in your server's fault - the message you're getting right now is the generic, reveal-nothing-to-possible-attackers WCF error message, basically saying: something went wrong.
In order to do that, you need to tweak your service config - add this section (if you don't already have one):
<behaviors>
<serviceBehaviors>
<behavior name="MEXandDebug">
<serviceMetadata />
<serviceDebug includeExceptionDetailInFaults="true"/>
</behavior>
</serviceBehaviors>
</behaviors>
and then reference that section from your service definition:
<services>
<service behaviorConfiguration="MEXandDebug" name="WCFService.MyWCFService">
That should give you a more meaningful error, which hopefully gives you an idea what goes wrong.
Otherwise you'll need to debug into your server-side code and find out what's happening there.