I'm trying to use xml configuration file in my project. Now it looks like:
<?xml version="1.0" encoding="utf-8" ?>
<configuration>
<configSections>
<section name="replication" type="Project.Replication.ReplicationConfigSection, Project.Replication" />
<section name="processing" type="Project.Processing.ProcessingConfigSection, Project.Processing" />
</configSections>
<replication>
<streams>
<stream name="STREAM_DATA_14360" />
</streams>
</replication>
<processing dataStream="STREAM_DATA_14360" />
</configuration>
It works OK, but I'm confused with duplicates in it ("STREAM_DATA_14360").
Can you remind me, how to create variables in XML or something for data reusing to be acceptable in application configuration?
UPDATE:
In real life my configuration has much more sections. There is a value, which apeears in many of this sections: STREAM_DATA_14360. So I want to be able to change this value only in one place of config file, and in other places to use reference to it.
Speed of changing configuration - is the first reason for it.
Size of a file is a second, because values can be huge: STREAM_INFO_FUTURE_SESSION_CONTENTS_12421 (that is third-party names)
You can simply add this value in <appSettings> and access it as you are saying.
You can do this as below:
<appSettings>
<add key="StreamName" value="STREAM_DATA_14360"/>
</appSettings>
In the code, you can access it as below:
string streamName = ConfigurationManager.AppSettings["StreamName"];
Make sure to add reference to System.Configuration assembly before using this.
XML doesn't have any native expansion macros or templating - any scenario would require that you do a preprocess step or have the code that reads the config involved in substituting the value.
If those aren't redacted names though, it seems a simple search/replace would solve the problem without much of a concern on false positives.
You could put something together with T4 templates as a preprocessor, whether that's worth it really depends on how often you expect to modify this file.
It should also be possible to shoehorn the web.config transformation engine into doing the replacements, but you may have to write some hosting code for the XDT engine depending on how your config file is setup.
Apart from using external code that might (or might not) facilitate your life, you can define your own classes that inherit from ConfigurationSection, wherein you define and encapsulate your key/value pairs and use the ConfigurationProperty attribute.
Have look at http://msdn.microsoft.com/en-us/library/2tw134k3.aspx for more info on How to: Create Custom Configuration Sections Using ConfigurationSection.
EDIT: you can make references in xsd (check here)
Thanks for your answers. I agree with Mark, there's no support of variables or references in XML. But, in my case there's much simpler solution. I feel stupid now, but hope that it will help another slowpoke too.
<?xml version="1.0" encoding="utf-8" ?>
<configuration>
<configSections>
<section name="global" type="Project.GlobalConfigSection, Project" />
<section name="replication" type="Project.Replication.ReplicationConfigSection, Project.Replication" />
<section name="processing" type="Project.Processing.ProcessingConfigSection, Project.Processing" />
</configSections>
<global>
<streamNames>
<streamName name="STREAM_DATA_14360" id="1"/>
</streamNames>
</global>
<replication>
<streams>
<stream nameId="1" />
</streams>
</replication>
<processing dataStreamId="1" />
</configuration>
Consequence: need to edit code to use global section as a source of all long names
Advantage: fast renaming, reusability of values
Related
Intro
I'm developing a WebApp built on C# ASP.NET.
I've been researching creating a "Custom Configuration" section with child elements in the Web.config file, and I've hit a bit of a snag when it comes to consuming the keys/values in the data.
I seem to be going round in circles and I don't know how to tackle the issue I'm having.
Situation
I have a few different Connection Strings defined in the Web.Config file, in the <connectionStrings> section. They are for dev, test, and live databases.
<connectionStrings>
<add name="connectionOne" connectionString="..." providerName="..." />
<add name="connectionTwo" connectionString="..." providerName="..." />
<add name="connectionThree" connectionString="..." providerName="..." />
</connectionStrings>
The WebApp is currently hard-coded to use one of these connection strings - if I need to change which one to use, I need to re-compile.
Desired Functionality
I'd like to define a section in the Web.config, let's say DbSettings.
In that, I'd then like to be able to define some child elements for, let's say DbSettings, in which I could define dbConnectionName, foo, bar etc. as attributes.
For example:
<dbSettings>
<dbSetting key="DbSetting1"
dbConnectionName="connectionOne"
foo="fooOne"
bar="barOne" />
... and so on
</dbSettings>
Then, perhaps in the <appSettings> section, define which of these DbSettings elements I want to use to get the settings from:
<appSettings>
<add name="dbSettingsKey" value="DbSetting1" />
</appSettings>
Desired Web.config section
Here is a fuller example of what I'd imagine my Web.config file to look like:
Connection Strings
<connectionStrings>
<add name="connectionOne" connectionString="..." providerName="..." />
<add name="connectionTwo" connectionString="..." providerName="..." />
<add name="connectionThree" connectionString="..." providerName="..." />
</connectionStrings>
App Settings
<add key="dbSettingsKey" value="DbSetting1" /> // or 2, or 3 etc.
DbSettings (custom section)
<dbSettings>
<dbSetting key="DbSetting1"
dbConnectionName="connectionOne"
foo="fooOne"
bar="barOne" />
<dbSetting key="DbSetting2"
dbConnectionName="connectionTwo"
foo="fooTwo"
bar="barTwo" />
<dbSetting key="DbSetting3"
dbConnectionName="connectionThree"
foo="fooThree"
bar="barThree" />
</dbSettings>
My question...
How the devil am I going to get this desired functionality in the C# code?
I've read loads on "creating your own custom section", and similarly "creating a custom config collection". But, I just can't seem to glue it all together to apply for my situation.
I'd like to be able to have a class (like the one I'm using at the moment with the hard-coded strings), which I can reference necessary properties (as I am doing, at the moment) - and then the code can dynamically load the correct settings at run-time from the sections I've described above.
As always, thank you in advance for your suggestions and help.
I agree with the comments. The way this is usually done is you deploy a different web.config to each environment. When your deployment group (or you) deploys, you deploy everything EXCEPT the web.config unless you have changes to push.
In answer to your other question, adding a custom section is not trivial. It's quite a bit of work. Custom section handler which requires a whole bunch of configuration element classes and a bunch of configuration element collection classes... and then, if you want it to "work" correctly, you also need to create a schema and register that with the IDE, etc.
For your particular case, I'd just do it the "normal" way :).
I am trying to create a simple tool for a service person to update a few entries in the App.Config of a different program. The App.Config file contains custom parameters used upon initialization of our program.
Since the App.Config contains many sensitive items a tool is needed to ensure only certain parameters are changed. Thus, the reason not to allow them to edit the App.Config directly.
My questions:
How can I access the name-value pairs from the config sections of an App.config from a separate program?
Which is better suited for the UI: Winforms or WPF? Are their controls that make it easy to add more entries in the future?
The tool should allow the user to set either a String, int, double or Boolean.
Here is the structure of the App.Config:
<?xml version="1.0" encoding="utf-8" ?>
<configuration>
<configSections>
<sectionGroup name="Settings">
<section name="Section1" type="System.Configuration.NameValueSectionHandler"/>
<section name="Section2" type="System.Configuration.NameValueSectionHandler"/>
<section name="Section3" type="System.Configuration.NameValueSectionHandler"/>
<section name="Section4" type="System.Configuration.NameValueSectionHandler"/>
</sectionGroup>
</configSections>
<Settings>
<Section1>
<add key="NAME_STRING" value="Some String"/>
</Section1>
<Section2>
<add key="NAME_INTEGER" value="10"/>
</Section2>
<Section3>
<add key="NAME_DOUBLE" value="10.5"/>
</Section3>
<Section4>
<add key="NAME_BOOLEAN" value="true"/>
</Section4>
</Settings>
... Omitted ...
</configuration>
In the program which uses the App.Config itself, I can easily change the values like so:
NameValueCollection nvc = (NameValueCollection)ConfigurationManager.GetSection("Settings/Section1");
Is there a similar way to do this from a separate program after loading the App.Config?
An answer to Question 1: An app.config file is an XML file. It might be easiest to load it as an XML document and modify that programmatically, followed by a save, than to use System.Configuration classes.
ETA: I believe it can be done with ConfigurationManager. Look at the OpenMappedExeConfiguration method. There's a good example there.
You could treat the app.config file as a normal XML file. Use either XDocument or XmlDocument to load the file.
Then use XPath or Linq to XML to find the name-value pairs.
As for Windows.Forms vs. WPF, its a design decision. Both have good and bad points.
[Update]
If you still want to use System.Configuration, you can use the ConfigurationManager.OpenExeConfiguration to get access to the other program's app.config file. This returns a Configuration object, which has a GetSection method.
Is it possible to add an custom configuration element at runtime.
Here is my app.config file
<?xml version="1.0" encoding="utf-8" ?>
<configuration>
<configSections>
<section name="NodeList"
type="Configuration.NodeListSection, NodeListConfiguration"
requirePermission="false" />
</configSections>
<NodeList>
<nodes>
<add name="Dev1" isdefault="false" description ="Dev server" />
<add name="QA1" isdefault="true" description="QA server"/>
<add name="Prod1" isdefault="false" description="Production" />
</nodes>
</NodeList>
</configuration>
Can we add more nodes at runtime using C# code.
This doesn't appear to be from a built-in configuration section. You will find that "NodesList" is an section/element that is custom written. To determine where in your codebase it is coming from look for "NodesList" at the top of your config file in the configSections element. That will point you at the class to look into.
After that, you need the class to support write operations properly.
To learn a lot more about customising configuration files there is a great series at CodeProject on the topic. In particular, the section on Saving Configuration Changes should be helpful to you.
Edit (after more info added to question):
Try something like (of course it all depends on what's in NodeListSection codebase):
using Configuration;
var nodeListSection = ConfigurationManager.GetSection("NodeList") as Configuration.NodeListSection;
var newNode = new NodeElement() { Name = "xyz", IsDefault = false, Description = "New Guy" };
nodeListSection.Nodes.Add(newNode);
Configuration.Save(ConfigurationSaveMode.Modified);
The file you have posted does not look like a normal .NET config file, but a custom XML file.
In either case - .config files are just XML files - you can open, manipulate and save them using any of the XML libraries within the BCL, such as XDocument.
However, if you want to make changes to configuration during runtime, you will need to decide whether the application should apply these changes at runtime as well and code for this, as normally a configuration file will only be read at startup.
private void AddNewKey_Config(string key, string value, string fileName)
{
var configFile = ConfigurationManager.OpenExeConfiguration(fileName);
configFile.AppSettings.Settings.Add(key, value);
configFile.Save();
}
I have a config file that is used in several projects, general.config, looks like:
<?xml version="1.0" encoding="utf-8" ?>
<appSettings>
<add key="mykey1" value="myvalue1"/>
<add key="mykey2" value="myvalue2"/>
</appSettings>
In one of the projects, I need to override one of the two settings. So the app.config of this project looks like:
<?xml version="1.0"?>
<configuration>
<appSettings file="general.config">
<remove key="mykey1"/>
<add key="mykey1" value="anothervalue"/>
<add key="mykey3" value="myvalue3"/>
</appSettings>
</configuration>
But remove is not working here. How can I override mykey1 without breaking mykey2? add works in this case. I can get myvalue3 from ConfigurationManager.
EDIT: general.config is copied to output folder automatically when compiling. Don't worry about the path issue. Currently I got:
ConfigurationManager.AppSettings["mykey1"]
//I got "myvalue1", but I want "anothervalue" here
//that is, this item is "overrided", just like virtual methods in C#
ConfigurationManager.AppSettings["mykey2"]
//this setting will not be modified, currently it works fine
ConfigurationManager.AppSettings["mykey3"] //good
A friend of mine answered this question. From MSDN:
You can use the file attribute to
specify a configuration file that
provides additional settings or
overrides the settings that are
specified in the appSettings element.
You can use the file attribute in
source control team development
scenarios, such as when a user wants
to override the project settings that
are specified in an application
configuration file. Configuration
files that are specified in a file
attribute must have the appSettings
element rather than configuration
element as the root node.
So in this question, the settings in general.config overrides items in app.config, different from that I think(want) app.config items overrides items in general.config. Now I think I have to resolve this issue in C# code(it will inevitable looks ugly).
Your use of the file attribute to load common settings with an expectation that keys added directly to the <appSettings> element would override those common settings is understandable, but unfortunately that is not how it works.
Microsoft's intention was for the file attribute to load common settings that override the individual application's settings.
This is discussed in some detail in the Microsoft Documentation
To overcome this problem, we very occasionally declare base settings in the common file, and then appropriately named overrides in the application configuration. However this does require additional code which is a bit ugly. e.g.
var config = ConfigurationManager.AppSettings["MSG_QUEUE_PROVIDER_OVERRIDE"]
?? ConfigurationManager.AppSettings["MSG_QUEUE_PROVIDER"]
?? "ActiveMQ";
<appSettings file="common.config">
<!-- Override the common values -->
<add key="MSG_QUEUE_PROVIDER_OVERRIDE" value="RabbitMQ"/>
</appSettings>
The elements are changed from the child and what i mean by that is currently your app.config is the parent file and the values are replaced by the ones existing in General.config
Since you are using remove in the parent file what its effectively doing is removing the element you specify in app.config but after that the elements from general.config are pushed in. Now say here in General.config you say remove mykey3 which is on your app.config you will see that the final collection has no key as mykey3.
In short this is not going to work. Hope this helped you.
You can add another config file say Test.config.
<appSettings>
<add key="mykey1" value="New value"/>
</appSettings>
and in the app.config appsettings section will look like this
<appSettings file="Test.config">
<add key="mykey1" value="myvalue1"/>
</appSettings>
I'm creating a Custom tag in my web.config. I first wrote the following entry under the configSections section.
<section name="castle"
type="Castle.Windsor.Configuration.AppDomain.CastleSectionHandler,
Castle.Windsor" />
But, when I try to create a castle node inside the configuration node as below
<castle>
<components>
</components>
</castle>
I get the following error message:"*Could not find schema information for the element '**castle'*." "***Could not find schema information for the element '**components'***."
Am I missing something? I can't find why. And, if I run the application anyway, I get the following error "Could not find section 'Castle' in the configuration file associated with this domain."
Ps.// The sample comes from "Pro ASP.NET MVC Framework"/Steven Sanderson/APress ISBN-13 (pbk): 978-1-4302-1007-8" on page 99.
Thank you for the help
============================================================
Since I believe to have done exactly what's said in the book and did not succed, I ask the same question in different terms. How do I add a new node using the above information?
=============================================================================
Thank you. I did what you said and do not have the two warnings. However, I've go a big new warning:
"The element 'configuration' in namespace 'MyWindsorSchema' has invalid child element 'configSections' in namespace 'MyWindsorSchema'. List of possible elements expected: 'include, properties, facilities, components' in namespace 'MyWindsorSchema'."
What you get is not an error that will prevent you from running your application. It is just a warning that Visual Studio emits because it does not know the castle node in a config file. You could use a schema to enable intellisense. Download the Castle Windsor Schema file and take a look at the readme.txt inside. It tells you to put windsor.xsd somewhere on your hard drive and then reference it in the config file:
<?xml version="1.0" encoding="utf-8" ?>
<configuration xmlns="MyWindsorSchema"
xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
xsi:schemaLocation="MyWindsorSchema file://S:\Common\Windsor\windsor.xsd">
<configSections>
<section name="castle" type="Castle.Windsor.Configuration.AppDomain.CastleSectionHandler, Castle.Windsor" />
</configSections>
<castle>
<components>
</components>
</castle>
</configuration>