JAVACC语法分析器 java语法树分析

转载

小蝌蚪 2023-07-20 20:36:57

文章标签 JAVACC语法分析器 java AST 抽象语法树 Express 文章分类 Java 后端开发

一、AST

AST abstract syntax tree （抽象语法结构树），是对java语言的一种抽象，每个节点都能对应到一种java语法，最终一个java文件就是由棵节点树构成

二、Tree

Tree是一个接口，是AST节点的抽象，内部有一个最重要的枚举Kind，定义了java中的每条语句的格式，也是对应着AST的一个节点

以下均为子接口，每个结构都定义了语法节点类型

Tree

WildcardTree
TypeParameterType
StatementTree (一个完整的语句，比如简单的"int a = 1;"，注意最后的“;”)

ThrowTree
ReturnTree
TryTree
ContinueTree
BreakTree
BlockTree
DoWhileLoopTree
ClassTree
VariableTree
WhileLoopTree
ForLoopTree
SwitchTree
SynchronizedTree
EnhancedForLoopTree
LabeedStatementTree
AssertTree
EmptyStatementTree
IfTree
ExpressionStatementTree

CatchTree
MethodTree
ArrayTypeTree
UnionTypeTree
ModifiersTree
ParameterizedTypeTree
CaseTree
ExpressionTree (一个非完整的语句，比如简单的"a = 1"，注意最后没有“;”)

MemberReferenceTree
LambdaExpressionTree
UnaryTree
MethodInvocationTree
AnnotatedTypeTree
CompoundAssignmentTree
BinaryTree
ErroneousTree
MemberSelectTree
ArrayAccessTree
TypeCastTree
NewClassTree
ConditionalExpressionTree
ParenthesizedTree
AnnotationTree
NewArrayType
InstanceOfTree
IdentifierTree
AssignmentTree
LiteralTree

CompilationUnitTree
IntersectionTypeTree
PrimitiveTypeTree
Import

三、Kind

AST 节点类型，对应着java的语法

枚举类型	Tree节点（就是上面列出的接口）	定义	示例
ANNOTATED_TYPE	AnnotatedTypeTree	注解类型，可以用在任何类型上注解的Target为：TYPE_PARAMETER TYPE_USE	@NotNull String
ANNOTATION	AnnotationTree	普通的注解	@Data @JsonProperties(value="sdf")
TYPE_ANNOTATION	AnnotationTree	跟ANNOTATION，待验证有什么区别
ARRAY_ACCESS	ArrayAccessTree	数组访问	array[3]
ARRAY_TYPE	ArrayTypeTree	数组类型	int[]
ASSIGNMENT	AssignmentTree	赋值	a = 1
BLOCK	BlockTree	语句块	{} { statements } static { statements }
BREAK	BreakTree	break	break; break label;
CASE	CaseTree	case 语句	case expression：statements default: statements
CATCH	CatchTree	catch 语句	catch(parameter) block 注：这些语句会有嵌套问题
CLASS INTERFACE ENUM ANNOTATION_TYPE	ClassTree	class 定义	modifiers class simpleName typeParameters extends extendsClause implements implementsClause { <em>members</em> }
COMPILATION_UNIT	CompilationUnitTree	一个编译单元包括源文件和package-info.java
CONDITIONAL_EXPRESSION	ConditionalExpressionTree	三元运算符	a == 1 ? 1 : 0
CONTINUE	ContinueTree	continue 语句	continue; continue label;
DO_WHILE_LOOP	DoWhileLoopTree	do...while
ENHANCED_FOR_LOOP	EnhancedForLoopTree	增强for	for(int i: intlist) { statement }
EXPRESSION_STATEMENT	ExpressionStatementTree	语句
MEMBER_SELECT	MemberSelectTree	对象.操作	aClass.aparams
MEMBER_REFERENCE	MemberReferenceTree	成员引用表达式
FOR_LOOP	ForLoopTree	for循环
IDENTIFIER	IdentifierTree	识别符，与literal区分开	a = 1 a 就是一个identifier
IF	IfTree
IMPORT	ImportTree
INSTANCE_OF	InstanceOfTree	instanceof 表达式
LABELED_STATEMENT	LabeledStatementTree	label语法	label : statement
METHOD	MethodTree	方法定义	方法 modifiers typeParameters type name (parameters ) body 注解 modifiers type name () default defaultValue
METHOD_INVOCATION	MethodInvocationTree	method invoke语法
MODIFIERS	ModifiersTree	修饰符（Modifier）包括注解
NEW_ARRAY	NewArrayTree
NEW_CLASS	NewClassTree
LAMBDA_EXPRESSION	LambdaExpressionTree	lambda表达式
PRIMITIVE_TYPE	PrimitiveTypeTree	基础类型
RETURN	ReturnTree	return
EMPTY_STATEMENT	EmptyStatementTree	空的";"
SWITCH	SwitchTree
SYNCHRONIZED	SynchronizedTree
THROW	ThrowTree
TRY	TryTree
PARAMETERIZED_TYPE	ParameterizedTypeTree	泛型	List<Integer>
UNION_TYPE	UnionTypeTree	多变量声明	try { }catch(Exception e1 \| Exception e2){ }
INTERSECTION_TYPE	IntersectionTypeTree		T extends MyA & MyB
TYPE_CAST	TypeCastTree	强转语句	(int)a
TYPE_PARAMETER	TypeParameterTree
VARIABLE	VariableTree	变量	modifiers type name initializer ; modifiers type qualified-name.this
WHILE_LOOP	WhileLoopTree
POSTFIX_INCREMENT POSTFIX_DECREMENT PREFIX_INCREMENT PREFIX_DECREMENT UNARY_PLUS UNARY_MINUS BITWISE_COMPLEMENT LOGICAL_COMPLEMENT	UnaryTree	运算符	++、--、+、-、~、！、*、/、%
MULTIPLY DIVIDE REMAINDER PLUS MINUS LEFT_SHIFT RIGHT_SHIFT UNSIGNED_RIGHT_SHIFT LESS_THAN GREATER_THAN LESS_THAN_EQUAL GREATER_THAN_EQUAL EQUAL_TO NOT_EQUAL_TO AND XOR OR CONDITIONAL_AND CONDITIONAL_OR	BinaryTree	两元运算
MULTIPLY_ASSIGNMENT DIVIDE_ASSIGNMENT REMAINDER_ASSIGNMENT PLUS_ASSIGNMENT MINUS_ASSIGNMENT LEFT_SHIFT_ASSIGNMENT RIGHT_SHIFT_ASSIGNMENT UNSIGNED_RIGHT_SHIFT_ASSIGNMENT AND_ASSIGNMENT XOR_ASSIGNMENT OR_ASSIGNMENT	CompoundAssignmentTree	赋值运算符	+=、-=、*=
INT_LITERAL LONG_LITERAL FLOAT_LITERAL DOUBLE_LITERAL BOOLEAN_LITERAL CHAR_LITERAL STRING_LITERAL NULL_LITERAL	LiteralTree	字面值	a = 1 (1就是)
OTHER	null

四、JCTree

com.sun.tools.javac.tree.JCTree 是Tree的一个抽象实现类，有各种实现类，代表着一个个的AST节点

JC具体类	实现Tree	Tag	Kind	备注
LetExpr		LETEXPR	不支持
JCErroneous	ErroneousTree	ERRONEOUS	ERRONEOUS
JCAnnotatedType	AnnotatedTypeTree	ANNOTATED_TYPE	ANNOTATED_TYPE
JCModifiers	ModifiersTree	MODIFIERS	MODIFIERS	✅
JCAnnotation	AnnotationTree			✅
JCTypeParameter	TypeParameterTree	TYPEPARAMETER	TYPE_PARAMETER
JCTypeIntersection	IntersectionTypeTree	TYPEINTERSECTION	INTERSECTION_TYPE
JCTypeUnion	UnionTypeTree	TYPEUNION	UNION_TYPE
JCTypeApply	ParameterizedTypeTree	TYPEAPPLY	PARAMETERIZED_TYPE
JCArrayTypeTree	ArrayTypeTree	TYPEARRAY	ARRAY_TYPE
JCPrimitiveTypeTree	PrimitiveTypeTree	TYPEIDENT	PRIMITIVE_TYPE	✅
JCLiteral	LiteralTree	LITERAL	INT_LITERAL LONG_LITERAL ... 以LITERAL结尾的	✅
JCIdent	IdentifierTree	IDENT	IDENTIFIER	✅
JCMemberReference	MemberReferenceTree	REFERENCE	MEMBER_REFERENCE
JCFieldAccess	MemberSelectTree	SELECT	MEMBER_SELECT	✅
JCArrayAccess	ArrayAccessTree	INDEXED	ARRAY_ACCESS	✅
JCInstanceOf	InstanceOfTree	TYPETEST	INSTANCE_OF
JCTypeCast	TypeCastTree	TYPE_CAST	TYPECAST
JCBinary	BinaryTree	根据操作符定	根据tag映射
JCAssignOp	CompoundAssignmentTree	赋值，根据操作符定支持的是 “+=”这种	根据tag映射
JCAssign	AssignmentTree	ASSIGN 支持的是直接赋值	ASSIGNMENT	✅
JCParens	ParenthesizedTree	PARENS	PARENTHESIZED
JCLambda	LambdaExpressionTree	LAMBDA	LAMBDA_EXPRESSION
JCNewArray	NewArrayTree	NEWARRAY	NEW_ARRAY
JCNewClass	NewClassTree	NEWCLASS	NEW_CLASS
JCMethodInvocation	MethodInvocationTree	APPLY	METHOD_INVOCATION
JCAssert	AssertTree	ASSERT	ASSERT
JCThrow	ThrowTree	THROW	THROW	✅
JCReturn	ReturnTree	RETURN	RETURN	✅
JCContinue	ContinueTree	CONTINUE	CONTINUE
JCBreak	BreakTree	BREAK	BREAK
JCExpressionStatement	ExpressionStatementTree	EXEC	EXPRESSION_STATEMENT
JCIf	IfTree	IF	IF
JCConditional	ConditionalExpressionTree	CONDEXPR	CONDITIONAL_EXPRESSION
JCCatch	CatchTree	CATCH	CATCH
JCTry	TryTree	TRY	TRY
JCSynchronized	SynchronizedTree	SYNCHRONIZED	SYNCHRONIZED
JCCase	CaseTree	CASE	CASE
JCSwitch	SwitchTree	SWITCH	SWITCH
JCLabeledStatement	LabeledStatementTree	LABELLED	LABELED_STATEMENT
JCEnhancedForLoop	EnhancedForLoopTree	FOREACHLOOP	ENHANCED_FOR_LOOP
JCForLoop	ForLoopTree	FORLOOP	FOR_LOOP
JCWhileLoop	WhileLoopTree	WHILELOOP	WHILE_LOOP
JCDoWhileLoop	DoWhileLoopTree	DOLOOP	DO_WHILE_LOOP
JCBlock	BlockTree	BLOCK	BLOCK
JCSkip	EmptyStatementTree	SKIP	EMPTY_STATEMENT
JCVariableDecl	VariableTree	VARDEF	VARIABLE	✅
JCMethodDecl	MethodTree	METHODDEF	METHOD	✅
JCClassDecl	ClassTree	CLASSDEF	ANNOTATION_TYPE INTERFACE ENUM CLASS	✅
JCImport	ImportTree	IMPORT	IMPORT	✅
JCCompilationUnit	CompilationUnitTree	TOPLEVEL	COMPILATION_UNIT

五、JC类的Factory

细心阅读会发现，各种JC类的构造函数都是protected，也就是说，并不能直接进行初始化，所以现在要使用Factory，Factory是各种JC类的工厂定义

JAVACC语法分析器 java语法树分析_JAVACC语法分析器

六、TreeMaker

TreeMaker实现了Factory接口，所以我们它会是我们经常使用的一个类

还需要关注的一个Names，这个基本是创建一个节点的开始

Names: 用来存储和表示解析后的词法，每个字符集合都会是一个Name对象，所有的对象都存储在Name.Table内部类中。

在学习以下内容时，一定要熟练掌握JCIdent和Names。

Javac编译原理 - wade&luffy - 博客园

javac源码笔记与简单的编译原理 | Ren's Time Overflow Skyline ^_^

深入分析 Javac 编译原理 - 知乎

6.1、添加 import

// import com.leisure.lombok.thrift.AstTreeClass
// 获取标识符
JCTree.JCIdent jcIdent = treeMaker.Ident(names.fromString("com.leisure.lombok.thrift"));
// 获取标识符
Name className = names.fromString("AstTreeClass");
// 定义访问
JCTree.JCFieldAccess jcFieldAccess = treeMaker.Select(jcIdent, className);
// 添加import节点
JCTree.JCImport anImport = treeMaker.Import(jcFieldAccess, false);
// 加入AST树种
jcCompilationUnit.defs = jcCompilationUnit.defs.append(anImport);

6.2、定义属性

默认构造函数

//private AstTreeClass variable = new AstTreeClass()
        // new AstTreeClass()
        JCTree.JCNewClass newClass = treeMaker.NewClass(null, null, treeMaker.Ident(names.fromString("AstTreeClass")), List.nil(), null);

        // private AstTreeClass variable = ...
        JCTree.JCVariableDecl variableDecl = treeMaker.VarDef(treeMaker.Modifiers(Modifier.PRIVATE.ordinal()),
                names.fromString("variable"), treeMaker.Ident(names.fromString("AstTreeClass")), newClass);

带参数的构造函数

使用这个我们就可以实现@Slf4j了（当然了，还需要有插件，才能真正像@Slf4j一样使用）

//public static final AstTreeClass staticVariable = new AstTreeClass(1)
        JCTree.JCLiteral constantOne = treeMaker.Literal(1);
        JCTree.JCNewClass newClass1 = treeMaker.NewClass(null, null, treeMaker.Ident(names.fromString("AstTreeClass")), List.of(constantOne), null);

        JCTree.JCVariableDecl staticVariableDecl = treeMaker.VarDef(
                treeMaker.Modifiers(Flags.PUBLIC + Flags.STATIC + Flags.FINAL),
                names.fromString("staticVariable"),
                treeMaker.Ident(names.fromString("AstTreeClass")),
                newClass1
        );

6.4、添加方法

String name = variableDecl.name.toString();
            JCTree.JCExpression varType = variableDecl.vartype;
            // getter，假设name长度大于1
            String getterMethodName = "get" + name.substring(0, 1).toUpperCase() + name.substring(1);
            // this.name
            JCTree.JCExpression thisName = treeMaker.Select(treeMaker.This(classDecl.sym.type), names.fromString(name));
            JCTree.JCReturn jcReturn = treeMaker.Return(thisName);
            // public type getName() {return this.name;}
            JCTree.JCMethodDecl getter = treeMaker.MethodDef(
                    treeMaker.Modifiers(Flags.PUBLIC), // 访问标识
                    names.fromString(getterMethodName), // 方法名字
                    varType,                            // 返回类型
                    List.nil(),                          // 泛型形参列表
                    List.nil(),                        // 参数列表
                    List.nil(),                        // 异常列表
                    treeMaker.Block(0, List.of(jcReturn)), // block
                    null                   // 默认方法
            );

// setter
            // public void setName(nameType name) {this.name = name;}
            String setterMethodName = "set" + name.substring(0, 1).toUpperCase() + name.substring(1);
            // 这里只是生成了this.name = name表达式
            JCTree.JCAssign assign = treeMaker.Assign(thisName, treeMaker.Ident(names.fromString(name)));
            // 生成一条语句 this.name = name;
            JCTree.JCExpressionStatement assignStat = treeMaker.Exec(assign);
            JCTree.JCMethodDecl setterMethodDecl = treeMaker.MethodDef(
                    treeMaker.Modifiers(Flags.PUBLIC),     // 访问标识
                    names.fromString(setterMethodName),    // 方法名
                    treeMaker.TypeIdent(TypeTag.VOID),                     // 返回类型
                    List.nil(),                    // 泛型形参列表
                    List.of(treeMaker.VarDef(
                            treeMaker.Modifiers(Flags.PARAMETER),   // 普通参数
                            variableDecl.name,
                            varType,
                            null   // 初始化为空，方法入参时为空即可
                            )
                    ), // 参数列表
                    List.nil(), // 异常列表
                    treeMaker.Block(0, List.of(assignStat)), // 方法体
                    null //默认方法
            );

TreeMaker -> JCTree -> Tree ,接下来的一系列就是应用了

本文章为转载内容，我们尊重原作者对文章享有的著作权。如有内容错误或侵权问题，欢迎原作者联系我们进行内容更正或删除文章。