【MyBatis源码分析】Configuration加载(上篇)

config.xml解析为org.w3c.dom.Document

本文首先来简单看一下MyBatis中将config.xml解析为org.w3c.dom.Document的流程,代码为上文的这部分:

1 static {
2     try {
3         reader = Resources.getResourceAsReader("mybatis/config.xml");
4         ssf = new SqlSessionFactoryBuilder().build(reader);
5     } 
6     catch (IOException e) {
7         e.printStackTrace();
8     }
9 }

第3行的代码实现为:

1 public static Reader getResourceAsReader(String resource) throws IOException {
2     Reader reader;
3     if (charset == null) {
4       reader = new InputStreamReader(getResourceAsStream(resource));
5     } else {
6       reader = new InputStreamReader(getResourceAsStream(resource), charset);
7     }
8     return reader;
9 }

相当于就是将输入的路径转换为一个字符输入流并返回。

接着继续看静态块第4行的代码,new SqlSessionFactoryBuilder().build(reader),把代码定位到SqlSessionFactoryBuilder类的builder方法,这里使用了多态,直接跟到build方法:

 1 public SqlSessionFactory build(Reader reader, String environment, Properties properties) {
 2     try {
 3       XMLConfigBuilder parser = new XMLConfigBuilder(reader, environment, properties);
 4       return build(parser.parse());
 5     } catch (Exception e) {
 6       throw ExceptionFactory.wrapException("Error building SqlSession.", e);
 7     } finally {
 8       ErrorContext.instance().reset();
 9       try {
10         reader.close();
11       } catch (IOException e) {
12         // Intentionally ignore. Prefer previous error.
13       }
14     }
15 }

解析config.xml的代码在第3行XMLConfigBuilder类的构造方法中,看一下XMLConfigBuilder类的构造方法做了什么:

 1 public XMLConfigBuilder(Reader reader, String environment, Properties props) {
 2     this(new XPathParser(reader, true, props, new XMLMapperEntityResolver()), environment, props);
 3 }

这里的关键是第二行代码的第一个参数XPathParser,看一下实例化XPathParser类的代码:

1 public XPathParser(Reader reader, boolean validation, Properties variables, EntityResolver entityResolver) {
2     commonConstructor(validation, variables, entityResolver);
3     this.document = createDocument(new InputSource(reader));
4 }

第2行的代码commonConstructor方法没什么好看的,将validation、variables、entityResolver设置到XPathParser类的参数中而已,顺便再实例化一个javax.xml.xpath.XPath出来,XPath用于在XML文档中通过元素和属性进行导航,并对元素和属性进行遍历。

接着看第3行的createDocument方法:

 1 private Document createDocument(InputSource inputSource) {
 2     // important: this must only be called AFTER common constructor
 3     try {
 4       DocumentBuilderFactory factory = DocumentBuilderFactory.newInstance();
 5       factory.setValidating(validation);
 6 
 7       factory.setNamespaceAware(false);
 8       factory.setIgnoringComments(true);
 9       factory.setIgnoringElementContentWhitespace(false);
10       factory.setCoalescing(false);
11       factory.setExpandEntityReferences(true);
12 
13       DocumentBuilder builder = factory.newDocumentBuilder();
14       builder.setEntityResolver(entityResolver);
15       builder.setErrorHandler(new ErrorHandler() {
16         @Override
17         public void error(SAXParseException exception) throws SAXException {
18           throw exception;
19         }
20 
21         @Override
22         public void fatalError(SAXParseException exception) throws SAXException {
23           throw exception;
24         }
25 
26         @Override
27         public void warning(SAXParseException exception) throws SAXException {
28         }
29       });
30       return builder.parse(inputSource);
31     } catch (Exception e) {
32       throw new BuilderException("Error creating document instance.  Cause: " + e, e);
33     }
34 }

看一下第5行~第11行的代码设置DocumentBuilderFactory中参数的含义:

  • setValidating表示是否验证xml文件,这个验证是DTD验证
  • setNamespaceAware表示是否支持xml命名空间
  • setIgnoringComments表示是否忽略注释
  • setIgnoringElementContentWhitespace表示是否忽略元素中的空白
  • setCoalescing表示是否将CDATA节点转换为Text节点,并将其附加到相邻(如果有)的Text节点
  • setExpandEntityReferences表示是否扩展实体引用节点

第13行的代码由设置的参数从DocumentBuilderFactory中获取一个DocumentBuilder实例DocumentBuilderImpl,并由第14行的代码设置一个实体解析器,由第15行~第29行的代码设置一个错误处理器。

最后看一下第30行的代码parse方法:

 1 public Document parse(InputSource is) throws SAXException, IOException {
 2         if (is == null) {
 3             throw new IllegalArgumentException(
 4                 DOMMessageFormatter.formatMessage(DOMMessageFormatter.DOM_DOMAIN,
 5                 "jaxp-null-input-source", null));
 6         }
 7         if (fSchemaValidator != null) {
 8             if (fSchemaValidationManager != null) {
 9                 fSchemaValidationManager.reset();
10                 fUnparsedEntityHandler.reset();
11             }
12             resetSchemaValidator();
13         }
14         domParser.parse(is);
15         Document doc = domParser.getDocument();
16         domParser.dropDocumentReferences();
17         return doc;
18 }

看过Spring配置文件解析源码的朋友应该对这一段代码比较熟悉,一样的,使用DocumentBuilder将解析InputSource成org.w3c.dom.Document并将Document存储到XPathParser中。

 

Document转换为Configuration

前面的代码将config.xml转换为了org.w3c.dom.Document,下一步就是将org.w3c.dom.Document中的内容转换为Java对象了,其中最主要的一个对象就是org.apache.ibatis.session.Configuration,还是回到之前的SqlSessionFactoryBuilder的build方法:

 1 public SqlSessionFactory build(Reader reader, String environment, Properties properties) {
 2     try {
 3       XMLConfigBuilder parser = new XMLConfigBuilder(reader, environment, properties);
 4       return build(parser.parse());
 5     } catch (Exception e) {
 6       throw ExceptionFactory.wrapException("Error building SqlSession.", e);
 7     } finally {
 8       ErrorContext.instance().reset();
 9       try {
10         reader.close();
11       } catch (IOException e) {
12         // Intentionally ignore. Prefer previous error.
13       }
14     }
15 }

先看一下第4行的parse方法,parse方法是XMLConfigBuilder中的,之前重点分析了它的属性XPathParser,看一下XMLConfigBuilder的parse方法是如何实现的:

1 public Configuration parse() {
2     if (parsed) {
3       throw new BuilderException("Each XMLConfigBuilder can only be used once.");
4     }
5     parsed = true;
6     parseConfiguration(parser.evalNode("/configuration"));
7     return configuration;
8 }

这里看一下第6行,可以使用XPathParser的evalNode方法解析标签,后面解析标签会大量用到此方法,此方法将标签解析为XNode,像config.xml(可见上一篇文章的示例)解析完之后的XNode,toString()方法输出的内容是这样的:

<configuration>
<properties resource="properties/db.properties"/>
<settings>
<setting name="cacheEnabled" value="true"/>
<setting name="lazyLoadingEnabled" value="true"/>
<setting name="useGeneratedKeys" value="true"/>
</settings>
<typeAliases>
<typeAlias alias="Mail" type="org.xrq.mybatis.pojo.Mail"/>
</typeAliases>
<environments default="development">
<environment id="development">
<transactionManager type="JDBC"/>
<dataSource type="POOLED">
<property name="driver" value="${driveClass}"/>
<property name="url" value="${url}"/>
<property name="username" value="${userName}"/>
<property name="password" value="${password}"/>
</dataSource>
</environment>
</environments>
<mappers>
<mapper resource="mybatis/mail.xml"/>
</mappers>
</configuration>

可见xml文件中<configuration>中所有内容都已经被成功解析并放在XNode中了,剩下的只要调用XNode的方法获取自己想要的内容即可。

最后扫一眼parseConfiguration方法,之所以说扫一眼,因为之后要分析里面的一些常用的和重点的内容,这里只是列一下代码而已:

 1 private void parseConfiguration(XNode root) {
 2     try {
 3       Properties settings = settingsAsPropertiess(root.evalNode("settings"));
 4       //issue #117 read properties first
 5       propertiesElement(root.evalNode("properties"));
 6       loadCustomVfs(settings);
 7       typeAliasesElement(root.evalNode("typeAliases"));
 8       pluginElement(root.evalNode("plugins"));
 9       objectFactoryElement(root.evalNode("objectFactory"));
10       objectWrapperFactoryElement(root.evalNode("objectWrapperFactory"));
11       reflectorFactoryElement(root.evalNode("reflectorFactory"));
12       settingsElement(settings);
13       // read it after objectFactory and objectWrapperFactory issue #631
14       environmentsElement(root.evalNode("environments"));
15       databaseIdProviderElement(root.evalNode("databaseIdProvider"));
16       typeHandlerElement(root.evalNode("typeHandlers"));
17       mapperElement(root.evalNode("mappers"));
18     } catch (Exception e) {
19       throw new BuilderException("Error parsing SQL Mapper Configuration. Cause: " + e, e);
20     }
21 }

这里就是逐个解析<configuration>标签下的子标签,并将数据设置到对应的属性中,这里要一个一个看一下。

 

settings解析

首先看settingsAsPropertiess(root.evalNode("settings"))这句代码,显而易见这句话获取了<configuration>下的<settings>节点。跟一下代码的实现:

 1 private Properties settingsAsPropertiess(XNode context) {
 2     if (context == null) {
 3       return new Properties();
 4     }
 5     Properties props = context.getChildrenAsProperties();
 6     // Check that all settings are known to the configuration class
 7     MetaClass metaConfig = MetaClass.forClass(Configuration.class, localReflectorFactory);
 8     for (Object key : props.keySet()) {
 9       if (!metaConfig.hasSetter(String.valueOf(key))) {
10         throw new BuilderException("The setting " + key + " is not known.  Make sure you spelled it correctly (case sensitive).");
11       }
12     }
13     return props;
14 }

第5行将节点解析成键值对的形式(Properties是Hashtable的子类),看一下props的toString方法打印的内容:

{useGeneratedKeys=true, lazyLoadingEnabled=true, cacheEnabled=true}

可见settings里面的数据已经被解析成了Properties了。之后还有一步,<settings>标签下的每个<setting>中的name属性不是随便填写的,都是MyBatis支持的配置,因此需要对Properties里面的Key做一个校验,校验的代码就是第7行至第12行的代码,其中有一个name不是MyBatis支持的就会抛出异常,MyBatis初始化整体失败。

至于具体校验的是哪些Key,这就要跟一下第7行的代码了,首先是MetaClass.forClass(Configuration.class, localReflectorFactory),第二个实参是XMLConfigBuilder里面直接new出来的,它的实际类型为DefaultReflectorFactory,看一下forClass方法实现:

 1 public static MetaClass forClass(Class<?> type, ReflectorFactory reflectorFactory) {
 2     return new MetaClass(type, reflectorFactory);
 3 }

看一下new MetaClass做了什么事:

1 private MetaClass(Class<?> type, ReflectorFactory reflectorFactory) {
2     this.reflectorFactory = reflectorFactory;
3     this.reflector = reflectorFactory.findForClass(type);
4 }

显而易见,继续跟一下第3行的代码DefaultRelectorFactory的findForClass方法:

 1 public Reflector findForClass(Class<?> type) {
 2     if (classCacheEnabled) {
 3             // synchronized (type) removed see issue #461
 4       Reflector cached = reflectorMap.get(type);
 5       if (cached == null) {
 6         cached = new Reflector(type);
 7         reflectorMap.put(type, cached);
 8       }
 9       return cached;
10     } else {
11       return new Reflector(type);
12     }
13 }

不管怎么样都会执行new Reflector(type)这一句代码,看一下此时做了什么事,注意传入的参数是Configuration的class对象:

 1 public Reflector(Class<?> clazz) {
 2     type = clazz;
 3     addDefaultConstructor(clazz);
 4     addGetMethods(clazz);
 5     addSetMethods(clazz);
 6     addFields(clazz);
 7     readablePropertyNames = getMethods.keySet().toArray(new String[getMethods.keySet().size()]);
 8     writeablePropertyNames = setMethods.keySet().toArray(new String[setMethods.keySet().size()]);
 9     for (String propName : readablePropertyNames) {
10       caseInsensitivePropertyMap.put(propName.toUpperCase(Locale.ENGLISH), propName);
11     }
12     for (String propName : writeablePropertyNames) {
13       caseInsensitivePropertyMap.put(propName.toUpperCase(Locale.ENGLISH), propName);
14     }
15 }

这么多方法至于具体要看哪个,要注意的是之前XMLConfigBuilder里面对于Key的判断是"!metaConfig.hasSetter(String.valueOf(key))",代码的意思是判断的是否Key有set方法,那么显而易见这里要继续跟第5行的addSetMethods方法:

 1 private void addSetMethods(Class<?> cls) {
 2     Map<String, List<Method>> conflictingSetters = new HashMap<String, List<Method>>();
 3     Method[] methods = getClassMethods(cls);
 4     for (Method method : methods) {
 5       String name = method.getName();
 6       if (name.startsWith("set") && name.length() > 3) {
 7         if (method.getParameterTypes().length == 1) {
 8           name = PropertyNamer.methodToProperty(name);
 9           addMethodConflict(conflictingSetters, name, method);
10         }
11       }
12     }
13     resolveSetterConflicts(conflictingSetters);
14 }

到这里应该很明显了,结论就是:<setting>的name属性对应的值,必须在Configuration类有相应的Setter,比如设置了一个属性useGenerateKeys方法,那么必须在Configuration类中有setUseGenerateKeys方法才行

顺便说一下第13行有一个resolveSetterConflicts方法,其作用是:Setter有可能在类中被重载导致有多个,此时取Setter中方法参数只有一个且参数类型与Getter一致的Setter

 

properties解析

接着看一下propertiesElement(root.evalNode("properties"))方法,这句读取的是<configuration>下的<properties>节点,代码实现为:

 1 private void propertiesElement(XNode context) throws Exception {
 2     if (context != null) {
 3       Properties defaults = context.getChildrenAsProperties();
 4       String resource = context.getStringAttribute("resource");
 5       String url = context.getStringAttribute("url");
 6       if (resource != null && url != null) {
 7         throw new BuilderException("The properties element cannot specify both a URL and a resource based property file reference.  Please specify one or the other.");
 8       }
 9       if (resource != null) {
10         defaults.putAll(Resources.getResourceAsProperties(resource));
11       } else if (url != null) {
12         defaults.putAll(Resources.getUrlAsProperties(url));
13       }
14       Properties vars = configuration.getVariables();
15       if (vars != null) {
16         defaults.putAll(vars);
17       }
18       parser.setVariables(defaults);
19       configuration.setVariables(defaults);
20     }
21 }

看到第4行~第7行的代码指定了MyBatis的<properties>标签下不能同时指定"resource"属性和"url"属性。

接着第9行~第13行的代码将.properties资源解析为Properties类,最后将Properties类设置到XPathParser和Configuration的variables属性中,variables是一个Propreties变量。

 

类型别名解析

跳过loadCustomVfs(settings)直接看typeAliasesElement(root.evalNode("typeAliases"))这行,因为前者我也没看懂干什么用的,后者是用于定义类型的别名的,解析的是<configuration>下的<typeAliases>标签,用过MyBatis的应该很熟悉。看一下源码实现:

 1 private void typeAliasesElement(XNode parent) {
 2     if (parent != null) {
 3       for (XNode child : parent.getChildren()) {
 4         if ("package".equals(child.getName())) {
 5           String typeAliasPackage = child.getStringAttribute("name");
 6           configuration.getTypeAliasRegistry().registerAliases(typeAliasPackage);
 7         } else {
 8           String alias = child.getStringAttribute("alias");
 9           String type = child.getStringAttribute("type");
10           try {
11             Class<?> clazz = Resources.classForName(type);
12             if (alias == null) {
13               typeAliasRegistry.registerAlias(clazz);
14             } else {
15               typeAliasRegistry.registerAlias(alias, clazz);
16             }
17           } catch (ClassNotFoundException e) {
18             throw new BuilderException("Error registering typeAlias for '" + alias + "'. Cause: " + e, e);
19           }
20         }
21       }
22     }
23 }

从源码实现中我们可以知道两点,<typeAliases>标签下可以定义<package>和<typeAlias>两种标签,但是看第4行和第7行的判断,这是一段if...else...,因此可以知道<package>标签和<typeAlias>标签只能定义其中的一种。首先看一下解析<package>标签的代码,第6行的registerAliases方法:

 1 public void registerAliases(String packageName, Class<?> superType){
 2     ResolverUtil<Class<?>> resolverUtil = new ResolverUtil<Class<?>>();
 3     resolverUtil.find(new ResolverUtil.IsA(superType), packageName);
 4     Set<Class<? extends Class<?>>> typeSet = resolverUtil.getClasses();
 5     for(Class<?> type : typeSet){
 6       // Ignore inner classes and interfaces (including package-info.java)
 7       // Skip also inner classes. See issue #6
 8       if (!type.isAnonymousClass() && !type.isInterface() && !type.isMemberClass()) {
 9         registerAlias(type);
10       }
11     }
12 }

第3行根据路径packageName寻找它下面的".class"文件拿到所有的".class"文件对应的类的Class,然后遍历所有的Class,做了三层判断

  • 必须不是匿名类
  • 必须不是接口
  • 必须不是成员类

此时此Class对应的类符合条件,会进行注册,通过registerAlias方法进行注册,看一下方法实现:

1 public void registerAlias(Class<?> type) {
2     String alias = type.getSimpleName();
3     Alias aliasAnnotation = type.getAnnotation(Alias.class);
4     if (aliasAnnotation != null) {
5       alias = aliasAnnotation.value();
6     } 
7     registerAlias(alias, type);
8 }

第2行获取Class的simpleName,simpleName指的是移除了包名的名称,比如aa.bb.cc.Mail,getSimpleName()获取的就是Mail。

第3行获取类上面的注解Alias,如果Alias注解中有定义value属性且指定了值,那么第4行~第6行的判断优先取这个值作为Class的别名。

第7行注册别名:

 1 public void registerAlias(String alias, Class<?> value) {
 2     if (alias == null) {
 3       throw new TypeException("The parameter alias cannot be null");
 4     }
 5     // issue #748
 6     String key = alias.toLowerCase(Locale.ENGLISH);
 7     if (TYPE_ALIASES.containsKey(key) && TYPE_ALIASES.get(key) != null && !TYPE_ALIASES.get(key).equals(value)) {
 8       throw new TypeException("The alias '" + alias + "' is already mapped to the value '" + TYPE_ALIASES.get(key).getName() + "'.");
 9     }
10     TYPE_ALIASES.put(key, value);
11 }

其实就做了两步操作:

  1. 将alias全部小写
  2. 将alias以及Class对象放到TYPE_ALIASES中,TYPE_ALIASES是一个HashMap

这样一个流程,就将<package>标签name属性路径下的Class(如果符合要求),全部放到了HashMap中以供使用。

接着看一下<typeAlias>标签的解析,也就是前面说的else部分:

 1 String alias = child.getStringAttribute("alias");
 2 String type = child.getStringAttribute("type");
 3 try {
 4   Class<?> clazz = Resources.classForName(type);
 5   if (alias == null) {
 6     typeAliasRegistry.registerAlias(clazz);
 7   } else {
 8     typeAliasRegistry.registerAlias(alias, clazz);
 9   }
10 } catch (ClassNotFoundException e) {
11   throw new BuilderException("Error registering typeAlias for '" + alias + "'. Cause: " + e, e);
12 }

这里先解析<typeAlias>中的alias属性,再解析<typeAlias>中的type属性,当然alias也可以不定义,不定义走的就是第6行的registerAlias方法,定义走的就是第8行的registerAlias方法,这两个重载的registerAlias方法前面也都说过了,就不说了。

 

默认typeAlias

上面说的是自定义typeAlias,MyBatis本身也默认提供给开发者了一些typeAlias定义,在两处地方。第一处地方在Configuration的构造方法中:

public Configuration() {
    typeAliasRegistry.registerAlias("JDBC", JdbcTransactionFactory.class);
    typeAliasRegistry.registerAlias("MANAGED", ManagedTransactionFactory.class);

    typeAliasRegistry.registerAlias("JNDI", JndiDataSourceFactory.class);
    typeAliasRegistry.registerAlias("POOLED", PooledDataSourceFactory.class);
    typeAliasRegistry.registerAlias("UNPOOLED", UnpooledDataSourceFactory.class);

    typeAliasRegistry.registerAlias("PERPETUAL", PerpetualCache.class);
    typeAliasRegistry.registerAlias("FIFO", FifoCache.class);
    typeAliasRegistry.registerAlias("LRU", LruCache.class);
    typeAliasRegistry.registerAlias("SOFT", SoftCache.class);
    typeAliasRegistry.registerAlias("WEAK", WeakCache.class);

    typeAliasRegistry.registerAlias("DB_VENDOR", VendorDatabaseIdProvider.class);

    typeAliasRegistry.registerAlias("XML", XMLLanguageDriver.class);
    typeAliasRegistry.registerAlias("RAW", RawLanguageDriver.class);

    typeAliasRegistry.registerAlias("SLF4J", Slf4jImpl.class);
    typeAliasRegistry.registerAlias("COMMONS_LOGGING", JakartaCommonsLoggingImpl.class);
    typeAliasRegistry.registerAlias("LOG4J", Log4jImpl.class);
    typeAliasRegistry.registerAlias("LOG4J2", Log4j2Impl.class);
    typeAliasRegistry.registerAlias("JDK_LOGGING", Jdk14LoggingImpl.class);
    typeAliasRegistry.registerAlias("STDOUT_LOGGING", StdOutImpl.class);
    typeAliasRegistry.registerAlias("NO_LOGGING", NoLoggingImpl.class);

    typeAliasRegistry.registerAlias("CGLIB", CglibProxyFactory.class);
    typeAliasRegistry.registerAlias("JAVASSIST", JavassistProxyFactory.class);

    ...
  }

第二处地方是在TypeAliasRegistry的构造方法中:

 1 public TypeAliasRegistry() {
 2     registerAlias("string", String.class);
 3 
 4     registerAlias("byte", Byte.class);
 5     registerAlias("long", Long.class);
 6     registerAlias("short", Short.class);
 7     registerAlias("int", Integer.class);
 8     registerAlias("integer", Integer.class);
 9     registerAlias("double", Double.class);
10     registerAlias("float", Float.class);
11     registerAlias("boolean", Boolean.class);
12 
13     registerAlias("byte[]", Byte[].class);
14     registerAlias("long[]", Long[].class);
15     registerAlias("short[]", Short[].class);
16     registerAlias("int[]", Integer[].class);
17     registerAlias("integer[]", Integer[].class);
18     registerAlias("double[]", Double[].class);
19     registerAlias("float[]", Float[].class);
20     registerAlias("boolean[]", Boolean[].class);
21 
22     registerAlias("_byte", byte.class);
23     registerAlias("_long", long.class);
24     registerAlias("_short", short.class);
25     registerAlias("_int", int.class);
26     registerAlias("_integer", int.class);
27     registerAlias("_double", double.class);
28     registerAlias("_float", float.class);
29     registerAlias("_boolean", boolean.class);
30 
31     registerAlias("_byte[]", byte[].class);
32     registerAlias("_long[]", long[].class);
33     registerAlias("_short[]", short[].class);
34     registerAlias("_int[]", int[].class);
35     registerAlias("_integer[]", int[].class);
36     registerAlias("_double[]", double[].class);
37     registerAlias("_float[]", float[].class);
38     registerAlias("_boolean[]", boolean[].class);
39 
40     registerAlias("date", Date.class);
41     registerAlias("decimal", BigDecimal.class);
42     registerAlias("bigdecimal", BigDecimal.class);
43     registerAlias("biginteger", BigInteger.class);
44     registerAlias("object", Object.class);
45 
46     registerAlias("date[]", Date[].class);
47     registerAlias("decimal[]", BigDecimal[].class);
48     registerAlias("bigdecimal[]", BigDecimal[].class);
49     registerAlias("biginteger[]", BigInteger[].class);
50     registerAlias("object[]", Object[].class);
51 
52     registerAlias("map", Map.class);
53     registerAlias("hashmap", HashMap.class);
54     registerAlias("list", List.class);
55     registerAlias("arraylist", ArrayList.class);
56     registerAlias("collection", Collection.class);
57     registerAlias("iterator", Iterator.class);
58 
59     registerAlias("ResultSet", ResultSet.class);
60 }

对于这些数据,我们可以直接使用registerAlias方法的第一个参数对应的字符串而不需要定义这些typeAlias。

 

posted @ 2017-05-06 10:36  五月的仓颉  阅读(6402)  评论(0编辑  收藏  举报