Antony Lee的技術之路
步步為營 |
Table of Contents
本文翻譯自Lars Vogel的DocBook with Eclipse教程, http://www.vogella.de/articles/DocBook/article.html
在翻譯過程中,本人保留原文的一切鏈接。
原文一切權利歸原作者所有,譯文一切權利歸本人所有。如欲轉載原文,請征得作者授權。如欲轉載譯文,請注明本文原始鏈接。
本文的翻譯已獲得原作者授權。
DocBook是一種文檔標準,用以創建結構化的純文本文檔。用DocBook創建的文檔能夠方便的在不同的操作系統之間以及不同的文本處理 工具之間進行移植,并可以通過XSLT 轉為其他的輸出格式。XSLT是EXtensible Stylesheet Language Transformation的縮寫。 由于DocBook是使用純文本編輯的,因此你可以使用任何一個文本 編輯器來編寫DocBook,并納入版本控制系統的管理。
目前,有多種不同的樣式表,能夠把DocBook文件轉換為不同的輸出格式,例如轉換為HTML,pdf,java help以及Unix man pages.
DocBook有兩種主要的文檔,一種是book,另一種是article。其中
Article: 用來寫一些技術文章,下文都以article為例。article通常是由一系列的section組成。
Book: 用來寫一些更長的描述。book比article多了一種結構:chapter。
下面就是一個DocBook文檔的例子。
<?xml version="1.0" encoding="UTF-8" standalone="no"?> <!DOCTYPE article PUBLIC "-//OASIS//DTD DocBook XML V4.5//EN" "../docbook-xml-4.5/docbookx.dtd"> <article> <articleinfo> <title>DocBook Intro</title> <author> <firstname>Lars</firstname> <surname>Vogel</surname> </author> </articleinfo> <sect1 label="1.0"> <title>An introduction to DocBook</title> <para> This is text. </para> </sect1> </article>
請注意,在上面的例子中,DTD文件的位置:當前目錄的上一級目錄,其中的docbook-xml-4.5文件夾內。
想要創建DocBook文件并轉換成其他格式,你需要以下工具:
DocBook的DTD文件。這個文件定義了DocBook文檔的格式。
XSLT樣式表,用來把DocBook文檔轉換成其他格式。
XSLT解析器
我們使用Eclipse 作為XML編輯器,Xalan作為XSLT解析器,并使用 Apache Ant 來進行XSLT的轉換。
你需要安裝Eclipse,可以參看這篇文章 Eclipse IDE 來學習Eclipse的安裝和使用。我們需要用到的Ant已經被集成到Eclipse里面了,因此關于Ant我們不需要安裝任何額外的東西。
你還需要下載Docbook的DTD,以及用來轉換的Docbook文檔的樣式表。Docbook的DTD可以在 http://www.oasis-open.org/docbook/xml/4.5下載, 而XSLT樣式表可以在 http://docbook.sourceforge.net/ 下載。在寫這篇文檔時,最新的版本是“1.75.2”(譯者注:在翻譯這篇文檔時,最新的版本是1.76.1)。你需要下載的docbook-xsl的發布文件, 例如“docbook-xsl-1.75.2.zip”。
要命的是JVM自帶的XSL處理器在處理XSLT樣式表的時候有問題……所以我們需要從 http://xml.apache.org/xalan-j/ 下載Xalan,用來作為我們的XSL處理器。
可以在Eclipse中,創建一個"de.vogella.docbook.first"的新工程,方法是File -> New -> Project,并從彈出的 窗口中選擇General -> Projects.
在工程中創建如下的目錄結構:
output : Docbook轉換成其他格式時的輸出目錄
docbook-xml-4.5: 用來放Docbook的DTD定義文件
docbook-xsl: 用來放進行Docbook轉換的樣式表文件
lib: 用來包含你需要的庫文件(用來創建pdf)
documents: 用來放你的DocBook文件
把DocBook的DTD和XSLT的樣式表放入相應的文件夾中。
在lib文件夾下創建xalan文件夾,并把xalan相關的jar包拷入這個文件夾中。 結果應該看起來是這樣的:
在“documents”文件夾里面,創建一個“book.xml”文件,并輸入下面的xml文件
<?xml version="1.0" encoding="UTF-8" standalone="no"?> <!DOCTYPE article PUBLIC "-//OASIS//DTD DocBook XML V4.5//EN" "../docbook-xml-4.5/docbookx.dtd"> <article> <articleinfo> <title>DocBook Intro</title> <author> <firstname>Lars</firstname> <surname>Vogel</surname> </author> </articleinfo> <sect1 label="1.0"> <title>An introduction to DocBook</title> <para> This is text. </para> </sect1> </article>
在xml文件中,“../docbook-xml-4.5/docbook.dtd”對應于我們剛剛創建的文件夾以及放入的DTD文件。
接下來我們配置ANT。在工程目錄中,創建build.xml文件如下:
<?xml version="1.0"?> <!-- - Author: Lars Vogel --> <project name="docbook-src" default="build-html"> <description> This Ant buildhtml.xml file is used to transform DocBook XML to html output </description> <!-- - Configure basic properties that will be used in the file. --> <property name="docbook.xsl.dir" value="docbook-xsl" /> <property name="doc.dir" value="output" /> <property name="documents" value="documents" /> <property name="html.stylesheet" value="${docbook.xsl.dir}/html/docbook.xsl" /> <!-- Making xalan available --> <path id="xalan.class.path"> <pathelement location="lib/xalan/serializer.jar" /> <pathelement location="lib/xalan/xalan.jar" /> <pathelement location="lib/xalan/xercesImpl.jar" /> <pathelement location="lib/xalan/xml-apis.jar" /> </path> <!-- - target: usage --> <target name="usage" description="Prints the Ant build.xml usage"> <echo message="Use -projecthelp to get a list of the available targets." /> </target> <!-- - target: clean --> <target name="clean" description="Cleans up generated files."> <delete dir="${doc.dir}" /> </target> <!-- - target: depends --> <target name="depends"> <mkdir dir="${doc.dir}" /> </target> <!-- - target: build-html - description: Iterates through a directory and transforms - .xml files into .html files using the DocBook XSL. --> <target name="build-html" depends="depends" description="Generates HTML files from DocBook XML"> <xslt style="${html.stylesheet}" extension=".html" basedir="${documents}" destdir="${doc.dir}"> <include name="**/*book.xml" /> <include name="**/*article.xml" /> <param name="html.stylesheet" expression="style.css" /> <classpath refid="xalan.class.path" /> </xslt> <!-- Copy the stylesheet to the same directory as the HTML files --> <copy todir="${doc.dir}"> <fileset dir="lib"> <include name="style.css" /> </fileset> </copy> </target> </project>
運行build.xml文件(右鍵 -> Run As -> Ant Build)。運行之后,在你的output文件夾里面, 應該已經有一個“book.html”了。
恭喜你完成了第一個Docbook文檔,并順利的轉成了HTML格式!
下面是一些使用Docbook標簽的概覽。
Table 1. Docbook一些重要的標簽
Tag | 說明 |
---|---|
<![CDATA[ 此處可輸入特殊字符,e.g. & ]]> | 在標簽中可以輸入某些特殊字符,例如某些xml以及Docbook的特殊字符。 |
<programlisting> </programlisting> | 表示該文本是程序代碼 |
<emphasis> </emphasis> | 表示用強調(Highlight)該文本 |
<xi:include xmlns:xi="http://www.w3.org/2001/XInclude" parse="text" href="example1.txt" /> | 包含example1.xml的內容。該文件可以是一個獨立的xml文件 |
<ulink url="http://www.heise.de/newsticker">German IT News</ulink> [a] | 在文檔中創建一個超鏈接 |
& | 在文檔中插入“&”符號 |
[a] 譯者注:Docbook5改變了超鏈接的格式。關于Docbook5以及更詳細的Eclipse配置Docbook教程,會很快放出。 不過,即使是在Docbook5的環境下,使用ulink一樣可以順利的用XSLT轉換,只不過Eclipse的xml編輯器會提示有錯誤。 |
下面是一個創建表格的例子
<table frame='all'> <title>Sample Table</title> <tgroup cols='2' align='left' colsep='1' rowsep='1'> <colspec colname='c1' /> <colspec colname='c2' /> <thead> <row> <entry>a4</entry> <entry>a5</entry> </row> </thead> <tfoot> <row> <entry>f4</entry> <entry>f5</entry> </row> </tfoot> <tbody> <row> <entry>b1</entry> <entry>b2</entry> </row> <row> <entry>d1</entry> <entry>d5</entry> </row> </tbody> </tgroup> </table>
生成的表格看起來是這樣的
沒有序號的列表可以這樣創建:
<itemizedlist> <listitem> <para>Item1</para> </listitem> <listitem> <para>Item2</para> </listitem> <listitem> <para>Item3</para> </listitem> <listitem> <para>Item4</para> </listitem> </itemizedlist>
輸出結果如下:
Item1
Item2
Item3
Item4
而帶編號的列表可以這樣寫:
<orderedlist> <listitem> <para>This is a list entry</para> </listitem> <listitem> <para>This is another list entry</para> </listitem> </orderedlist>
輸出結果如下:
This is a list entry
This is another list entry
鏈接可以用下面的方法來創建:[a]
<para> We use the Ant integrated into Eclipse. See <ulink url="http://www.vogella.de/articles/ApacheAnt/article.html"> Apache Ant Tutorial</ulink> for an introduction into Apache Ant. </para>
Docbook轉成pdf的過程是:先由docbook轉成XSL-FO格式,再利用Apache FOP把FO轉成 pdf。因此,我們首先需要Apache FOP相關的庫。
XML FO,是XML Formating Object的意思,FO格式是一種處理打印、印刷介質的XML標準。
可以從http://xmlgraphics.apache.org/fop/下載FOP的最新版本。
從下載的FOP發行版中,把所有的jar文件都拷貝到你的lib文件夾中,并把這些庫都加入到ant 的build path中。可以參考 Apach Ant Tutorial來修改ant的build path。
要在ant中使用fop,我們首先應當定義一個fop相關的ant task,然后在后面的腳本中使用這個任務。 下面的這個例子演示了怎樣定義一個ant task并怎樣調用。第二個例子是一個完整的build.xml文件的例子。
<!-- - Defines the ant task for fop --> <taskdef name="fop" classname="org.apache.fop.tools.anttasks.Fop" /> <!-- Transformation into pdf - Two steps - 1.) First create the FO files - 2.) Then transform the FO files into pdf files --> <!-- - target: build-pdf - description: Iterates through a directory and transforms - .xml files into .fo files using the DocBook XSL. --> <target name="build-pdf" depends="depends, xinclude" description="Generates HTML files from DocBook XML"> <!-- Convert DocBook Files into FO --> <xslt style="${fo.stylesheet}" extension=".fo" basedir="${src.tmp}" destdir="${src.tmp}"> <include name="**/*book.xml" /> <include name="**/*article.xml" /> <param name="section.autolabel" expression="1" /> </xslt> <!-- Convert FO Files into pdf --> <fop format="application/pdf" outdir="${doc.dir}"> <fileset dir="${src.tmp}"> <include name="**/*.fo" /> </fileset> </fop> </target>
<?xml version="1.0"?> <!-- - Author: Lars Vogel --> <project name="docbook-src" default="all"> <description> This Ant build.xml file is used to transform DocBook XML to various output formats </description> <!-- - Defines the ant task for xinclude --> <taskdef name="xinclude" classname="de.vogella.xinclude.XIncludeTask" /> <!-- - Defines the ant task for xinclude --> <taskdef name="fop" classname="org.apache.fop.tools.anttasks.Fop" /> <!-- - Configure basic properties that will be used in the file. --> <property name="javahelp.dir" value="${basedir}/../Documentation/output/vogella/javahelp" /> <property name="src" value="${basedir}/documentation" /> <property name="output.dir" value="${basedir}/../Documentation/output/vogella/articles" /> <property name="output.tmp" value="${basedir}/output.tmp" /> <property name="lib" value="${basedir}/lib/" /> <property name="docbook.xsl.dir" value="${basedir}/docbook-xsl-1.72.0" /> <property name="xinclude.lib.dir" value="${basedir}/lib/" /> <!-- - Usage of the differect style sheets which will be used for the transformation --> <property name="eclipse.stylesheet" value="${docbook.xsl.dir}/eclipse/eclipse.xsl" /> <property name="html.stylesheet" value="${docbook.xsl.dir}/html/docbook.xsl" /> <property name="fo.stylesheet" value="${docbook.xsl.dir}/fo/docbook.xsl" /> <property name="javahelp.stylesheet" value="${docbook.xsl.dir}/javahelp/javahelp.xsl" /> <property name="chunk-html.stylesheet" value="${docbook.xsl.dir}/html/chunk.xsl" /> <!-- - target: usage --> <target name="usage" description="Prints the Ant build.xml usage"> <echo message="Use -projecthelp to get a list of the available targets." /> </target> <!-- - target: clean --> <target name="clean" description="Cleans up generated files."> <delete dir="${output.dir}" /> </target> <!-- - target: depends --> <target name="depends"> <mkdir dir="${output.dir}" /> </target> <!-- - target: copy - Copies the images from the subdirectories to the target folder --> <target name="copy"> <echo message="Copy the images" /> <copy todir="${output.dir}"> <fileset dir="${src}"> <include name="**/images/*.*" /> </fileset> </copy> </target> <!-- - target: xinclude - description: Creates one combined temporary files for the different inputs files. - The combined file will then be processed via different ant tasks --> <target name="xinclude"> <xinclude in="${src}/DocBook/article.xml" out="${output.tmp}/DocBook/article.xml" /> <xinclude in="${src}/JavaConventions/article.xml" out="${output.tmp}/JavaConventions/article.xml" /> <xinclude in="${src}/JUnit/article.xml" out="${output.tmp}/JUnit/article.xml" /> <xinclude in="${src}/EclipseReview/article.xml" out="${output.tmp}/EclipseReview/article.xml" /> <xinclude in="${src}/HTML/article.xml" out="${output.tmp}/HTML/article.xml" /> <xinclude in="${src}/Eclipse/article.xml" out="${output.tmp}/Eclipse/article.xml" /> <xinclude in="${src}/Logging/article.xml" out="${output.tmp}/Logging/article.xml" /> <!-- <xinclude in="${src}/ant/article.xml" out="${src.tmp}/ant/article.xml" /> --> </target> <!-- - target: build-html - description: Iterates through a directory and transforms - .xml files into .html files using the DocBook XSL. --> <target name="build-html" depends="depends, xinclude" description="Generates HTML files from DocBook XML"> <xslt style="${html.stylesheet}" extension=".html" basedir="${output.tmp}" destdir="${output.dir}"> <include name="**/*book.xml" /> <include name="**/*article.xml" /> <param name="html.stylesheet" expression="styles.css" /> <param name="section.autolabel" expression="1" /> <param name="html.cleanup" expression="1" /> <outputproperty name="indent" value="yes" /> </xslt> <!-- Copy the stylesheet to the same directory as the HTML files --> <copy todir="${output.dir}"> <fileset dir="lib"> <include name="styles.css" /> </fileset> </copy> </target> <!-- - target: build-javahelp - description: Iterates through a directory and transforms - .xml files into .html files using the DocBook XSL. --> <target name="build-javahelp" depends="depends, xinclude" description="Generates HTML files from DocBook XML"> <xslt style="${javahelp.stylesheet}" extension=".html" basedir="${output.tmp}" destdir="${javahelp.dir}"> <include name="**/*book.xml" /> <include name="**/*article.xml" /> <outputproperty name="indent" value="yes" /> </xslt> </target> <!-- - target: chunks-html - description: Iterates through a directory and transforms - .xml files into seperate .html files using the DocBook XSL. --> <target name="build-chunks" depends="depends, xinclude" description="Generates chunk HTML files from DocBook XML"> <xslt style="${html.stylesheet}" extension=".html" basedir="${output.tmp}" destdir="${output.dir}"> <include name="**/*book.xml" /> <include name="**/*article.xml" /> <param name="html.stylesheet" expression="styles.css" /> <param name="section.autolabel" expression="1" /> <param name="html.cleanup" expression="1" /> <param name="chunk.first.selection" expression="1" /> </xslt> <!-- Copy the stylesheet to the same directory as the HTML files --> <copy todir="${output.dir}"> <fileset dir="lib"> <include name="styles.css" /> </fileset> </copy> </target> <!-- Transformation into pdf - Two steps - 1.) First create the FO files - 2.) Then transform the FO files into pdf files --> <!-- - target: build-pdf - description: Iterates through a directory and transforms - .xml files into .fo files using the DocBook XSL. - Relativebase is set to true to enable FOP to find the graphics which are included - in the images directory --> <target name="build-pdf" depends="depends, xinclude" description="Generates HTML files from DocBook XML"> <!-- Convert DocBook Files into FO --> <xslt style="${fo.stylesheet}" extension=".fo" basedir="${output.tmp}" destdir="${output.tmp}"> <include name="**/*book.xml" /> <include name="**/*article.xml" /> <param name="section.autolabel" expression="1" /> </xslt> <!-- Convert FO Files into pdf --> <fop format="application/pdf" outdir="${output.dir}" relativebase="true"> <fileset dir="${output.tmp}"> <include name="**/*.fo" /> </fileset> </fop> </target> <!-- - target: chunks-html - description: Iterates through a directory and transforms - .xml files into seperate .html files using the DocBook XSL. --> <target name="build-eclipse" depends="depends, xinclude" description="Generates Eclipse help files from DocBook XML"> <xslt style="${eclipse.stylesheet}" basedir="${output.tmp}" destdir="${output.dir}"> <include name="**/*book.xml" /> <include name="**/*article.xml" /> </xslt> </target> <target name="all" depends="copy, build-html, build-pdf, build-chunks, build-eclipse"> </target> </project>
我們可以通過修改XSLT樣式表的參數來影響輸出的結果。下面其中一些參數的介紹。
Docbook允許在轉換成html的時候,導入并包含一個外部的html文件。你可以使用這種技術,在生成 html的時候向其中插入javascript代碼。
下面是一個包含html文件的例子。
<?dbhtml-include href="../../myadditonalcontent.html"?>
Inserting external HTML code 有更多的描述。
XInclude技術能幫你重新組織你的docbook文件。你可以在書寫每一個章節的時候,都使用 一個單獨的xml文件,然后用一個總的xml文件把這些章節都組合起來。簡單的說,XInclude 能把不同的xml文件組合成為一個大的xml文件。
例如,假設你要引入一個“foo.xml”文件,則可以寫成:
<xi:include xmlns:xi="http://www.w3.org/2001/XInclude" href="foo.xml" />
下面這個例子是要把導入的文件當做文本: [1]
<xi:include xmlns:xi="http://www.w3.org/2001/XInclude" parse="text" href="bar.xml" />Eclipse XSL project提供一個XInclude的ant task。在此,我很自豪的告訴各位:這個ant task是我提供給XSL project 的:)
Eclipse XSL工具提供了對XSLT的支持,包括XSL的編輯以及debug的支持。雖然我們這里僅僅使用其中的ant task, 但還是得完整的安裝整個包。
安裝XSL工具可以通過Eclipse的update manager完成。 [2] 你可以通過 Using the Eclipse update manager 來獲得更多信息。
在你的Eclipse安裝路徑中找到“org.eclipse.wst.xsl.core.jar”并把這個jar包加入到ANT的classpath中。 這樣,你應該就可以創建和運行xinclude tast了。下面是一個build.xml文件的例子:
<?xml version="1.0"?> <!-- - Author: Lars Vogel --> <project name="docbook-src" default="usage"> <description> This Ant build.xml file is used to transform DocBook XML to various output formats </description> <!-- - Configure basic properties that will be used in the file. --> <property name="doc.dir" value="${basedir}/output" /> <property name="src" value="${basedir}/src" /> <property name="src.tmp" value="${basedir}/src.tmp" /> <property name="lib" value="${basedir}/lib/" /> <property name="docbook.xsl.dir" value="${basedir}/docbook-xsl-1.72.0" /> <property name="html.stylesheet" value="${docbook.xsl.dir}/html/docbook.xsl" /> <property name="xinclude.lib.dir" value="${basedir}/lib/" /> <!-- - target: usage --> <target name="usage" description="Prints the Ant build.xml usage"> <echo message="Use -projecthelp to get a list of the available targets." /> </target> <!-- - target: clean --> <target name="clean" description="Cleans up generated files."> <delete dir="${doc.dir}" /> </target> <!-- - target: depends --> <target name="depends"> <mkdir dir="${doc.dir}" /> </target> <!-- - target: xinclude - description: Creates one combined temporary files for the different inputs files. - The combined file will then be processed via different ant tasks --> <target name="xinclude"> <xsl.xinclude in="${src}/DocBook/article.xml" out="${src.tmp}/DocBook/article.xml" /> </target> <!-- - target: build-html - description: Iterates through a directory and transforms - .xml files into .html files using the DocBook XSL. --> <target name="build-html" depends="depends, xinclude" description="Generates HTML files from DocBook XML"> <xslt style="${html.stylesheet}" extension=".html" basedir="${src.tmp}" destdir="${doc.dir}"> <include name="**/*book.xml" /> <include name="**/*article.xml" /> <param name="html.stylesheet" expression="styles.css" /> </xslt> <!-- Copy the stylesheet to the same directory as the HTML files --> <copy todir="${doc.dir}"> <fileset dir="lib"> <include name="styles.css" /> </fileset> </copy> </target> </project>
http://www.sagehill.net/docbookxsl/index.html Docbook XSL Online Book
http://sourceforge.net/projects/docbook/ Docbook XSLT樣式表項目
http://docbook.sourceforge.net/release/xsl/current/doc/ XSLT樣式表的參數手冊
http://sourceforge.net/apps/mediawiki/xslthl/index.php?title=Main_Page 如何使用XSLT實現語法高亮
http://www.docbook.org/tdg/en/html/docbook.html Docbook參數手冊
http://xmlgraphics.apache.org/fop/ Apache FOP項目
我會爭取在近期用Python寫個圖形界面。之所以現在不寫,是因為現在我還不會……
原理很簡單,就是用程序去獲得騰訊星座網站的源碼,然后簡單解析一下就可以了。原來試圖用dom解析,結果發現騰訊和新浪的網站都會解析出錯,一狠心干脆直接用字符串的替換。
Java那個程序的功能多一些,寫了讀取“明天”、“本周”、“下周”、“本月”、“下月”的功能,主函數中提供了讀取“明天”的功能,默認會在當前工作目錄下生成“yuncheng_XXX.txt”文件。實在懶得寫客戶端了,甚至于命令行客戶端都不想寫了,因為昨天寫這個破東西寫到了11點半,原以為很簡單的呢,結果郁悶壞了。真是水平大幅度下滑啊。
今天晚上用python寫了一個程序,這是我用Python寫的第一個程序喲~~(如果不算helloworld的話。。。)這個程序功能比較簡單,只能獲得明日運程。
剛剛寫出來而已,注釋很亂,代碼很亂。寫這篇博客只是因為母雞心態:好歹有了個新東西了,怎么也得讓我叫喚兩聲吧。
下載地址:
http://www.rayfile.com/files/bc9485dc-88e8-11de-b777-0014221f469f/一、基本概念
所謂動態代理,基本上是如下場景:假設我有個接口IHelloWorld
問題是,現在,我打算為HelloWorldImpl增強功能,需要在調用sayHello方法前后各執行一些操作。在有些情況下,你無法修改HelloWorldImpl的源代碼,那怎么辦呢?
從道理上來說,我們可以攔截對HelloWorldImpl對象里sayHello()函數的調用。也就是說,每當有代碼調用sayHello函數時,我們都把這種調用請求攔截下來之后,做自己想做的事情。
那怎么攔截呢?
首先,需要開發一個InvocationHandler。這個東東表示的是,你攔截下函數調用之后,究竟想干什么。InvocationHandler是一個接口,里面的聲明的函數只有一個:
Object invoke(Object proxy, Method method, Object[] args) throws Throwable
這個函數表示一次被攔截的函數調用。因此,proxy表示這個被攔截的調用,原本是對哪個對象調用的;method表示這個被攔截的調用,究竟是調用什么方法;args表示這個被攔截的調用里,參數分別是什么。
我們下面寫一個攔截器,讓他在函數調用之前和之后分別輸出一句話。
有了這個Handler之后,下面要做的,就是把這個Handler和一個IHelloWorld類型的對象裝配起來。重點的函數只有一個,那就是java.lang.reflect.Proxy類中的一個靜態工廠方法:
public static Object newProxyInstance(ClassLoader loader, Class<?>[] interfaces, InvocationHandler h) throws IllegalArgumentException
這個方法返回一個對象,我們稱返回的對象為代理對象(proxy)。
而后,我們就不把真正的原對象暴露給外接,而使用這個代理對象。這個代理對象接受對源對象的一切函數調用(也就是把所有調用都攔截了),然后根據我們寫的InvocationHandler,來對函數進行處理。
產生代理對象的過程,我把它理解成一個裝配的過程:由源對象、源對象實現的接口、InvocationHandler裝配產生一個代理對象。
相應的測試代碼如下:
利用ant編譯運行的結果:
[java] ################################
[java] method name : sayHello
[java] Do Before
[java] Hello, World
[java] Do After
[java] ################################
二、更多理解
我們看產生代理對象的newProxyInstance函數的聲明:
public static Object newProxyInstance(ClassLoader loader, Class<?>[] interfaces, InvocationHandler h) throws IllegalArgumentException
這個函數的第一個參數是ClassLoader,第三個參數是InvocationHandler,基本都沒什么問題。
第二個參數是一個Class類型的數組,名字叫interfaces,表示的是產生的動態代理對象實現的接口。
仔細想想,有兩個問題。第一,產生一個代理對象,需要源對象么?第二,我能不能產生一個動態代理對象,來實現源對象沒有實現的接口?
第一個問題和第二個問題其實是一致的。我們完全可以脫離源對象,而直接產生一個代理對象,也可以利用動態代理,讓源對象實現更多的接口,為源對象增強功能。
例如,假設我們希望讓源對象實現java.io.Closeable接口,則首先修改一下我們的Handler的invoke方法,讓他在獲取colse方法時,不要傳遞給源對象(因為源對象沒有實現該方法):
然后,我們在裝配的過程中,改變一下參數,并強轉之后調用一下close方法:
ant運行結果:
[java] ################################
[java] method name : close
[java] Do Before
[java] I got the close() method!
[java] Do After
[java] ################################
三、更多的代理~
我們現在能夠讓sayHello()函數執行之前和之后,輸出一些內容了。那如果我還想在裝配一個Handler呢?
最簡單的方法:
和人討論設計模式的時候,看到這樣一句話:
大阿亮<yighter@qq.com> 22:24:40
java擴展功能就是繼承和組合。肯定結構都很相似。模式思想都是從解決問題背景和目的來區分的。
恍然大悟。原來,很多情況下,所謂設計模式,是對同一種技術、實現的不同角度的理解。
所以,設計無所謂好壞,只要能解決問題的,就是好設計。至于所謂“強耦合”,“Bad Smell”,本質上是因為采用這些設計無法解決問題(就是無法快速應對需求變化)。
“不管黑貓白貓,只要能抓住耗子,就是好貓”,這句話蘊含著深刻的設計思想。
在啃《The Java Programming Language 4th Edition》時看到的一個小知識點。先描述一下問題。
一個類中,靜態初始代碼塊中的代碼會在類加載時自動運行。考慮下面這種情況:
ClassA定義了靜態初始代碼塊,其中調用了ClassB的一個方法m(靜態非靜態均可)。而在ClassB的m方法中,又使用了ClassA類的信息。則,當虛擬機在沒有ClassB類的情況下,加載ClassA類時,會遇到這樣一條線索:
加載ClassA --> 調用ClassA的靜態初始化代碼塊 --> 調用ClassB的m方法 --> 加載ClassB --> 使用ClassA的信息
注意這條線索的一頭一尾,我們要在對ClassA還沒完成加載時,使用ClassA的信息!
示例代碼:
首先,編譯器無法解決這個問題,因為在編譯ClassA類時,無法找到ClassB的代碼,也就無法檢查是否存在靜態初始化代碼塊循環問題。事實上,上述程序在java中是能夠編譯通過的。
其次,運行時的結果。當程序運行到第3行時,JVM加載ClassA類,此時,會執行ClassA類中的靜態初始化代碼塊。當程序執行到第12行時,調用ClassB的print方法,此時,程序跳轉到18行。
關鍵在這兒:此時的print方法需要調用ClassA的信息,并打印其靜態屬性。而ClassA的信息正在加載過程中。此時,JVM采用的策略是:在print方法中使用ClassA不完整的信息。在print方法中ClassA的信息,是在第12行對ClassB.print方法之前的信息。此時ClassA.a1已經被賦值為10,而ClassA.a2還未被賦值,它的值為默認值。因此,最后打印出的是10、0。
| |||||||||
日 | 一 | 二 | 三 | 四 | 五 | 六 | |||
---|---|---|---|---|---|---|---|---|---|
28 | 29 | 30 | 31 | 1 | 2 | 3 | |||
4 | 5 | 6 | 7 | 8 | 9 | 10 | |||
11 | 12 | 13 | 14 | 15 | 16 | 17 | |||
18 | 19 | 20 | 21 | 22 | 23 | 24 | |||
25 | 26 | 27 | 28 | 29 | 30 | 31 | |||
1 | 2 | 3 | 4 | 5 | 6 | 7 |