Introduction
Nant is the scripting tool
for the Dot Net platform. It is inspired from its java counterpart, <st1:place w:st="on">Ant. <o:p>
NAnt
is an open source project, with a homepage at http://nant.sourceforge.net.
There<o:p>
are
links to many resources from this page. Additionally,
there is an adjunct to NAnt called NAntContrib. It has a homepage at http://nantcontrib.sourceforge.net. This project is for task
contributions to NAnt that do<o:p>
not make it into the core NAnt release. Occasionally, tasks in
NAntContrib will make the<o:p>
switch to NAnt.<o:p>
NAnt Nomenclature<o:p>
Just before we dive in, let us clarify some of the terminology
that we will use when discussing<o:p>
NAnt:<o:p>
NAnt: The physical executable
(and the associated tasks) that form the application.<o:p>
NAnt script/build script/build file: An XML file describing a project comprised of one or<o:p>
more targets and zero or more properties.<o:p>
Project: The root node of a build
script. Each script has one project.<o:p>
Target: A subdivision of a
project. A target is a discrete piece of work that can consist of<o:p>
zero or more tasks. Targets can be dependent on each other.<o:p>
Task: A task is a defined
activity that NAnt will execute; for example, creating a directory<o:p>
or zipping a file. There are many built-in tasks to NAnt, and we
will look at the most valuable<o:p>
of these in Chapter 3.<o:p>
Property: An XML
key/value pair that can be used to configure tasks and targets in a<o:p>
more dynamic fashion.
Property values can be overridden from the command line.
Installation Instructions:<o:p>
<o:p> Download the Nant executable
from the Nant site, http://nant.sourceforge.net. The latest
version was 0.85 at the time of writing this article. Extract the executable to
a local directory, say C:\Nant\. The binary executable for nant will be stored
in C:\Nant\bin, which is Nant.exe. This exe will be responsible for executing
your scripts. So <o:p>
i.
Either you have to dump
your scripts to C:\Nant folder or<o:p>
ii.
Add the path (C:\Nant\bin) to your PATH
environment variable.<o:p>
<o:p>In
this article I will follow the first approach.<o:p>
<o:p> Using the code
A "Hello World" Example<o:p>
Consider the following very simple NAnt script:<o:p>
<?xml version="1.0"
encoding="utf-8" ?><o:p>
<project
name="HelloWorld" default="go"><o:p>
<property
name="message" value="Hello World!"/><o:p>
<target
name="go"><o:p>
<echo
message="${message}"/><o:p>
</target><o:p>
</project><o:p>
<o:p>
<o:p>
Save this script as HelloWorld.build and then do
one of two things. Either navigate to the<o:p>
directory in which the file is saved and type<o:p>
Nant<o:p>
or use an explicit path to the file at the command prompt such
as<o:p>
You will get the following output:<o:p>
<o:p>
---------- NAnt ----------<o:p>
NAnt 0.85<o:p>
Copyright (C) 2001-2003
Gerry Shaw<o:p>
http://nant.sourceforge.net<o:p>
Buildfile:
file:///HelloWorld.build<o:p>
Target(s) specified: go<o:p>
go:<o:p>
[echo] Hello World!<o:p>
BUILD SUCCEEDED<o:p>
Total time: 0 seconds.<o:p>
Output completed (0 sec
consumed) - Normal Termination The Project<o:p>
The project node is the root node of a build file. This node can
contain any number of<o:p>
<property> nodes, <task>
nodes, and <target> nodes.
Generally, the rest of the build file is<o:p>
now defined either as a property, or as a target, or as a task
within a target.<o:p>
<o:p>
Project Node attributes: <o:p>
<o:p>
name The
name of the project.<o:p>
default The default target to use when no target is
supplied.<o:p>
basedir The base directory from
which all path calculations are done. No<o:p>
Current directory is the
default.<o:p>
<o:p>
<o:p>
The Target<o:p>
Targets are used to
"modularize" the build file. A target contains zero or more tasks to be completed<o:p>
in a sequence.<o:p>
Target Node attributes <o:p>
<o:p>
name<o:p>
The name attribute is crucial to the target, as the target is invoked by
name. Simple as that.<o:p>
description<o:p>
The description is shown when the –projecthelp switch is
used at the command line, and so<o:p>
can be of some assistance to a user of the build file.<o:p>
depends<o:p>
Targets can be made to depend on each other by use of the depends attribute. This attribute<o:p>
takes a comma-delimited list of targets that are to be executed
prior to the execution of the<o:p>
target in question. For example:<o:p>
<target name="go"
depends="foo, bar"/><o:p>
This means that target go will not execute until
target foo and then target bar have been executed.<o:p>
Additionally, any targets that foo or bar depend on must be executed. You will be pleased<o:p>
to note that NAnt can figure out circular dependencies so that
the following build file will not<o:p>
execute:<o:p>
<o:p>
<o:p>
Properties :<o:p>
These are similar to the
variables used in a programming language. They are used as a key value pair.<o:p>
<o:p>
<o:p>
Important tasks with NAnt:<o:p>
<o:p>
NUnit: <o:p>
Nant can be used to run unit
test suite written in NUnit framework.
Nant provides the <Nunit2>
element for this purpose. (Note that here 2 stands for version 2 of nunit
framework) . The sample script is as follows:<o:p>
<o:p>
<?xml
version="1.0"?><o:p>
<project
name="testNUnit"><o:p>
<nunit2><o:p>
<formatter<o:p>
type="Xml"<o:p>
usefile="true"<o:p>
extension=".xml"<o:p>
outputdir="C:\NAnt\NUnitResults\"
/><o:p>
<test
assemblyname="E:\testing\TestClassLibrary1\\bin\Debug\TestClassLibrary1.dll"
/><o:p>
</nunit2><o:p>
</project><o:p>
<o:p>
<o:p>
NDoc:<o:p>
<o:p>
NDoc is a tool that can be used to create
extremely presentable<o:p>
MSDN-style documentation in web or compiled HTML (CHM) formats
from the<o:p>
XML documentation
capabilities of the C# language.<o:p>
NAnt comes with a version of the core<o:p>
NDoc assembly, and this task
can be used to perform the same action:<o:p>
<o:p>
<?xml
version="1.0"?><o:p>
<project
name="TestNDoc"><o:p>
<ndoc><o:p>
<assemblies
basedir="E:\testing\ClassLibrary1\ClassLibrary1\bin\Debug\"><o:p>
<include
name="ClassLibrary1.dll" /><o:p>
</assemblies><o:p>
<summaries><o:p>
<include
name="ClassLibrary1.xml" /><o:p>
</summaries><o:p>
<documenters><o:p>
<documenter
name="MSDN"><o:p>
<property
name="OutputDirectory" value="C:\MyDocs\NDOC" /><o:p>
<property
name="HtmlHelpName" value="MyProject" /><o:p>
<property
name="ShowMissingSummaries" value="True" /><o:p>
<property<o:p>
name="HtmlHelpCompilerFilename"
value="hhc.exe" /><o:p>
<property
name="IncludeFavorites" value="False" /><o:p>
<property
name="Title" value="MySystem (NDoc)" /><o:p>
<property
name="SplitTOCs" value="False" /><o:p>
<property
name="DefaulTOC" value="" /><o:p>
<property
name="ShowVisualBasic" value="False" /><o:p>
<property
name="ShowMissingSummaries" value="True" /><o:p>
<property
name="ShowMissingRemarks" value="False" /><o:p>
<property
name="ShowMissingParams" value="True" /><o:p>
<property
name="ShowMissingReturns" value="True" /><o:p>
<property
name="ShowMissingValues" value="True" /><o:p>
<property name="DocumentInternals"
value="True" /><o:p>
<property
name="DocumentProtected" value="True" /><o:p>
<property
name="DocumentPrivates" value="False" /><o:p>
<property
name="DocumentEmptyNamespaces" value="False" /><o:p>
<property
name="IncludeAssemblyVersion" value="True" /><o:p>
<property
name="CopyrightText" value="Etomic 2005" /><o:p>
<property
name="CopyrightHref" value="" /><o:p>
</documenter><o:p>
</documenters><o:p>
</ndoc><o:p>
</project><o:p>
<o:p>
The above script generates
the ndoc for the assembly classlibrary1 which is present in the base directory
mentioned.<o:p>
<o:p>
<o:p>
FxCop:<o:p>
Nant can be used to automate
fxcop command. The following script explains how: <o:p>
<o:p>
<?xml
version="1.0"?><o:p>
<project
name="testFxCop"><o:p>
<exec<o:p>
program="C:\Program
Files\Microsoft FxCop 1.35\FxCopCmd.exe"<o:p>
commandline="/f:E:\testing\ClassLibrary1\ClassLibrary1\bin\Debug\ClassLibrary1.dll /o:fxcop.xml"<o:p>
failonerror="false"
/><o:p>
</project><o:p>
Replace value of program attribute to whatever is the
version of fxcop executable in your machine. <o:p>
Mention the path for assembly in commandline,
along with name of the output file desired.<o:p>
Faileonerror indicates whether to stop
the process once an error is
encountered. <o:p>
<o:p>
<o:p>
NCover: <o:p>
<o:p>
<project
name="testNCover" default="coverage"
failonerror="true"><o:p>
<loadtasks
assembly="NCoverExplorer.NAntTasks.dll" /><o:p>
<target
name="coverage" description="Code coverage test run."><o:p>
<ncover<o:p>
program="C:\Program
Files\NCover\Ncover.console.exe" <o:p>
commandLineExe="C:\Program
Files\NUnit-Net-2.0 2.2.8\bin\nunit-console.exe" <o:p>
commandLineArgs="E:\REL\Test\TestTest\bin\Debug\TestTest.dll"<o:p>
logLevel="Verbose"<o:p>
excludeAttributes="CoverageExcludeAttribute"><o:p>
<assemblies
basedir="E:\REL\Test\Test\bin\Debug"><o:p>
<!-- include this to have
only the test.dll reviewed and not the TestTest.dll --><o:p>
<include
name="Test.dll"/> <o:p>
</assemblies><o:p>
</ncover>
<o:p>
To exclude a selected class
or method from being reviewed by ncoverage, use the CoverageExclude attribute.<o:p>
public class
CoverageExcludeAttribute : Attribute { }<o:p>
namespace MyApp<o:p>
{<o:p>
[CoverageExclude]<o:p>
Partial class MyClass<o:p>
{<o:p>
<o:p>
}<o:p>
}<o:p>
Add the CoverageExclude attribute to
your file , either on top of class or method. <o:p>
This will exclude the
designated method or class from being reviewed.<o:p>
Summary:
I hope this article has given you a basic idea about writing Nant scripts for cruise control integration. You can get more documenation and online help at http://nant.sourceforge.net..
Thill then happy scripting !!!.
