doc/org.eclipse.xpand.doc/help/backend.html - m2t/org.eclipse.xpand - Git at Google

 <html>
 <head>
 <title>Improve performance using the backend</title>
 <link href="book.css" rel="stylesheet" type="text/css"/>
 <meta content="DocBook XSL Stylesheets V1.75.1" name="generator"/>
 <link rel="home" href="index.html" title="Xpand Documentation"/>
 <link rel="up" href="Reference.html" title="Part I. Reference"/>
 <link rel="prev" href="incrementalGeneration_notes.html" title="Additional Notes"/>
 <link rel="next" href="backend_usage.html" title="Using the backend"/>
 </head>
 <body bgcolor="white" text="black" link="#0000FF" vlink="#840084" alink="#0000FF">
 <h1 xmlns:l="http://docbook.sourceforge.net/xmlns/l10n/1.0">Improve performance using the backend</h1>
 <div class="chapter" title="Improve performance using the backend">
 <div class="titlepage">
 <div>
 <div>
 <h2 class="title">
 <a name="backend"/>Improve performance using the backend</h2>
 </div>
 </div>
 </div>
 <div class="toc">
 <dl>
 <dt>
 <span class="section">
 <a href="backend.html#backend_background">Technical background</a>
 </span>
 </dt>
 <dt>
 <span class="section">
 <a href="backend_usage.html">Using the backend</a>
 </span>
 </dt>
 <dt>
 <span class="section">
 <a href="call_functions_backend.html">Call functions registered at the backend from Java</a>
 </span>
 </dt>
 </dl>
 </div>
 <p>
 		Another approach to improve generation and model transformation
 		performance, is to use the M2T Backend. The M2T Backend is a new
 		execution engine for model to text (M2T) and model to model (M2M)
 		transformations that provides:
 		</p>
 <div class="itemizedlist">
 <ul class="itemizedlist" type="disc">
 <li class="listitem">
 				compilation of M2T and M2M languages
 				supported on the backend by providing a middleend that translates the respective
 				language into a Backend abstract syntax tree. This AST is in turn compiled to Java. The
 				compiler is still experimental.
       		</li>
 <li class="listitem">
 				Backeend interpreted execution of M2T and M2M languages
 				supported on the backend.
       		</li>
 <li class="listitem">
 				performance improvement employing lazy string concatenation
 				and aggressive caching
       		</li>
 <li class="listitem">
 				language interoperability, e.g. mix Xpand with Xtend2
       		</li>
 </ul>
 </div>
 <p>
 	</p>
 <p>
 		The M2T Backend provides a compiler that compiles resources written in Xpand / Xtend / Check and other
 		languages supported on the backend to Java. The backend is used as a common platform just
 		like the JVM is a common platform for multiple languages. Languages supported on the backend are interoperable
 		in same way as languages supported on the JVM are interoperable.
 		This enables you to import templates, extensions and transformations written
 		in other languages than Xpand/Xtend/Check and call functions declared
 		therein, e.g. call an QVT transformation from Xpand. As soon as a middleend for Xtend2 is available,
 		interoperability with old fashioned Xpand/Xtend/Check will be supported.
 	</p>
 <p>
 		Beginning with the Eclipse Helios release (Eclipse 3.6) the M2T
 		Backend is part	of M2T Xpand. As of the Eclipse Indigo release
 		the Backend Compiler has been added.
 	</p>
 <div class="section" title="Technical background">
 <div class="titlepage">
 <div>
 <div>
 <h2 class="title" style="clear: both">
 <a name="backend_background"/>Technical background</h2>
 </div>
 </div>
 </div>
 <p>
 			The basic idea behind the <span class="emphasis">
 <em>M2T Backend</em>
 </span>
 			is to not build a compiler for every language but use an intermediary
 			"language", i.e. the backend abstract syntax tree, that is compiled
 			into executable Java code. Xpand, Xtend and Check are translated into a
 			backend AST by so called <span class="emphasis">
 <em>middleends</em>
 </span>.
 			Middleends are specific for each language. Hence, there exists a
 			<span class="emphasis">
 <em>language specific	middleend</em>
 </span>
 			for each supported language. If a middleend
 			for a certain language is available, it is compilable into Java code and finally to bytecode
 			using a plain Java Compiler or automatically during the build Eclipse JDT Compiler.
 			In contrast to the backend language, Xpand, Xtend and Check are called <span class="emphasis">
 <em>frontend languages</em>
 </span>.
 		</p>
 <p>
 			The M2T Backend supports functional and aspect oriented
 			programming. Amongst others it supports language features such as:
 			</p>
 <div class="itemizedlist">
 <ul class="itemizedlist" type="disc">
 <li class="listitem">
 					closures
 			</li>
 <li class="listitem">
 					higher order functions - use any expression as parameters,
 					even function calls
 			</li>
 <li class="listitem">
 					currying of functions
 			</li>
 <li class="listitem">
 					multi dispatch - call a function based on the runtime
 					parameter types as in Xpand / Xtend / Check instead of the
 					statically assigned	parameter types as in Java
 				</li>
 <li class="listitem">
 					guards on functions - constrain the execution of a
 					function to certain conditions defined in the guard
 			</li>
 <li class="listitem">
 					deferred evaluation of expressions - use the function call
 					as a placeholder and evaluate the expression later
 			</li>
 <li class="listitem">
 					AOP
 			</li>
 <li class="listitem">
 					declare functions in Java using several annotations
 				</li>
 <li class="listitem">
 					Support for multiple typesystems a.k.a metamodels.
 					Currently JavaBeans, EMF, Eclipse UML2 and XSD are supported.
 				</li>
 </ul>
 </div>
 <p>
 		</p>
 <div class="section" title="How it works in interpreted mode">
 <div class="titlepage">
 <div>
 <div>
 <h3 class="title">
 <a name="backend_interpretation_background"/>How it works in interpreted mode</h3>
 </div>
 </div>
 </div>
 <p>
 				In interpreted mode the following steps
 				are executed:

 				</p>
 <div class="itemizedlist">
 <ul class="itemizedlist" type="disc">
 <li class="listitem">
 						Read a resource, e.g. an Xpand template.
 		      	</li>
 <li class="listitem">
 						Find the responsible <span class="emphasis">
 <em>middleend</em>
 </span>,
 						that is capable to convert the <span class="emphasis">
 <em>frontend language</em>
 </span>
 						AST (e.g. Xpand) into the <span class="emphasis">
 <em>backend AST</em>
 </span>. Whether
 						a middleend can handle the transformation of a certain
 						frontend language AST into the backend AST, is determined based on the
 						file ending of the resource, e.g ".xpt" and by trying to parse the resource
 						with the middleends	frontend parser.
 					</li>
 <li class="listitem">
 						Finally, execute the backend AST using the backend.
 					</li>
 </ul>
 </div>
 <p>
 			</p>
 </div>
 </div>
 </div>
 </body>
 </html>
	<html>
	<head>
	<title>Improve performance using the backend</title>
	<link href="book.css" rel="stylesheet" type="text/css"/>
	<meta content="DocBook XSL Stylesheets V1.75.1" name="generator"/>
	<link rel="home" href="index.html" title="Xpand Documentation"/>
	<link rel="up" href="Reference.html" title="Part I. Reference"/>
	<link rel="prev" href="incrementalGeneration_notes.html" title="Additional Notes"/>
	<link rel="next" href="backend_usage.html" title="Using the backend"/>
	</head>
	<body bgcolor="white" text="black" link="#0000FF" vlink="#840084" alink="#0000FF">
	<h1 xmlns:l="http://docbook.sourceforge.net/xmlns/l10n/1.0">Improve performance using the backend</h1>
	<div class="chapter" title="Improve performance using the backend">
	<div class="titlepage">
	<div>
	<div>
	<h2 class="title">
	<a name="backend"/>Improve performance using the backend</h2>
	</div>
	</div>
	</div>
	<div class="toc">
	<dl>
	<dt>
	<span class="section">
	<a href="backend.html#backend_background">Technical background</a>
	</span>
	</dt>
	<dt>
	<span class="section">
	<a href="backend_usage.html">Using the backend</a>
	</span>
	</dt>
	<dt>
	<span class="section">
	<a href="call_functions_backend.html">Call functions registered at the backend from Java</a>
	</span>
	</dt>
	</dl>
	</div>
	<p>
	Another approach to improve generation and model transformation
	performance, is to use the M2T Backend. The M2T Backend is a new
	execution engine for model to text (M2T) and model to model (M2M)
	transformations that provides:
	</p>
	<div class="itemizedlist">
	<ul class="itemizedlist" type="disc">
	<li class="listitem">
	compilation of M2T and M2M languages
	supported on the backend by providing a middleend that translates the respective
	language into a Backend abstract syntax tree. This AST is in turn compiled to Java. The
	compiler is still experimental.
	</li>
	<li class="listitem">
	Backeend interpreted execution of M2T and M2M languages
	supported on the backend.
	</li>
	<li class="listitem">
	performance improvement employing lazy string concatenation
	and aggressive caching
	</li>
	<li class="listitem">
	language interoperability, e.g. mix Xpand with Xtend2
	</li>
	</ul>
	</div>
	<p>
	</p>
	<p>
	The M2T Backend provides a compiler that compiles resources written in Xpand / Xtend / Check and other
	languages supported on the backend to Java. The backend is used as a common platform just
	like the JVM is a common platform for multiple languages. Languages supported on the backend are interoperable
	in same way as languages supported on the JVM are interoperable.
	This enables you to import templates, extensions and transformations written
	in other languages than Xpand/Xtend/Check and call functions declared
	therein, e.g. call an QVT transformation from Xpand. As soon as a middleend for Xtend2 is available,
	interoperability with old fashioned Xpand/Xtend/Check will be supported.
	</p>
	<p>
	Beginning with the Eclipse Helios release (Eclipse 3.6) the M2T
	Backend is part of M2T Xpand. As of the Eclipse Indigo release
	the Backend Compiler has been added.
	</p>
	<div class="section" title="Technical background">
	<div class="titlepage">
	<div>
	<div>
	<h2 class="title" style="clear: both">
	<a name="backend_background"/>Technical background</h2>
	</div>
	</div>
	</div>
	<p>
	The basic idea behind the <span class="emphasis">
	<em>M2T Backend</em>
	</span>
	is to not build a compiler for every language but use an intermediary
	"language", i.e. the backend abstract syntax tree, that is compiled
	into executable Java code. Xpand, Xtend and Check are translated into a
	backend AST by so called <span class="emphasis">
	<em>middleends</em>
	</span>.
	Middleends are specific for each language. Hence, there exists a
	<span class="emphasis">
	<em>language specific middleend</em>
	</span>
	for each supported language. If a middleend
	for a certain language is available, it is compilable into Java code and finally to bytecode
	using a plain Java Compiler or automatically during the build Eclipse JDT Compiler.
	In contrast to the backend language, Xpand, Xtend and Check are called <span class="emphasis">
	<em>frontend languages</em>
	</span>.
	</p>
	<p>
	The M2T Backend supports functional and aspect oriented
	programming. Amongst others it supports language features such as:
	</p>
	<div class="itemizedlist">
	<ul class="itemizedlist" type="disc">
	<li class="listitem">
	closures
	</li>
	<li class="listitem">
	higher order functions - use any expression as parameters,
	even function calls
	</li>
	<li class="listitem">
	currying of functions
	</li>
	<li class="listitem">
	multi dispatch - call a function based on the runtime
	parameter types as in Xpand / Xtend / Check instead of the
	statically assigned parameter types as in Java
	</li>
	<li class="listitem">
	guards on functions - constrain the execution of a
	function to certain conditions defined in the guard
	</li>
	<li class="listitem">
	deferred evaluation of expressions - use the function call
	as a placeholder and evaluate the expression later
	</li>
	<li class="listitem">
	AOP
	</li>
	<li class="listitem">
	declare functions in Java using several annotations
	</li>
	<li class="listitem">
	Support for multiple typesystems a.k.a metamodels.
	Currently JavaBeans, EMF, Eclipse UML2 and XSD are supported.
	</li>
	</ul>
	</div>
	<p>
	</p>
	<div class="section" title="How it works in interpreted mode">
	<div class="titlepage">
	<div>
	<div>
	<h3 class="title">
	<a name="backend_interpretation_background"/>How it works in interpreted mode</h3>
	</div>
	</div>
	</div>
	<p>
	In interpreted mode the following steps
	are executed:

	</p>
	<div class="itemizedlist">
	<ul class="itemizedlist" type="disc">
	<li class="listitem">
	Read a resource, e.g. an Xpand template.
	</li>
	<li class="listitem">
	Find the responsible <span class="emphasis">
	<em>middleend</em>
	</span>,
	that is capable to convert the <span class="emphasis">
	<em>frontend language</em>
	</span>
	AST (e.g. Xpand) into the <span class="emphasis">
	<em>backend AST</em>
	</span>. Whether
	a middleend can handle the transformation of a certain
	frontend language AST into the backend AST, is determined based on the
	file ending of the resource, e.g ".xpt" and by trying to parse the resource
	with the middleends frontend parser.
	</li>
	<li class="listitem">
	Finally, execute the backend AST using the backend.
	</li>
	</ul>
	</div>
	<p>
	</p>
	</div>
	</div>
	</div>
	</body>
	</html>