Check类之duplicate declaration checking/Class name generation/Type Checking
1、duplicate declaration checking
/** Check that variable does not hide variable with same name in * immediately enclosing local scope. * * e.g * public void m1(boolean a){ * int a = 3; * boolean a = true; // 不调用checkTransparentVar()方法,因为调用checkUnique()方法时能测出冲突 * } * * @param pos Position for error reporting. * @param v The symbol. * @param s The scope. */ void checkTransparentVar(DiagnosticPosition pos, VarSymbol v, Scope s) { if (s.next != null) { for (Scope.Entry e = s.next.lookup(v.name); e.scope != null && e.sym.owner == v.owner; e = e.next()) { if (e.sym.kind == VAR && (e.sym.owner.kind & (VAR | MTH)) != 0 && v.name != names.error) { duplicateError(pos, e.sym); return; } } } }
e.g
public void m1(boolean a){ int a = 3; for(int x = 0;x<2;x++){ int x = 3; } }
其中int a = 3与int x= 3将调用checkTransparentVar()方法检测出重复声明。
/** Check that a class or interface does not hide a class or * interface with same name in immediately enclosing local scope. * @param pos Position for error reporting. * @param c The symbol. * @param s The scope. */ void checkTransparentClass(DiagnosticPosition pos, ClassSymbol c, Scope s) { if (s.next != null) { for (Scope.Entry e = s.next.lookup(c.name); e.scope != null && e.sym.owner == c.owner; e = e.next()) { if (e.sym.kind == TYP && e.sym.type.tag != TYPEVAR && (e.sym.owner.kind & (VAR | MTH)) != 0 && c.name != names.error) { duplicateError(pos, e.sym); return; } } } }
e.g
public void m1() { class A{} { class A{} // err Duplicate nested type A }
而对于e.sym.owner.kind为VAR的情况还没有举出例子。
/** Check that class does not have the same name as one of * its enclosing classes, or as a class defined in its enclosing scope. * return true if class is unique in its enclosing scope. * @param pos Position for error reporting. * @param name The class name. * @param s The enclosing scope. */ boolean checkUniqueClassName(DiagnosticPosition pos, Name name, Scope s) { for (Scope.Entry e = s.lookup(name); e.scope == s; e = e.next()) { if (e.sym.kind == TYP && e.sym.name != names.error) { duplicateError(pos, e.sym); return false; } } for (Symbol sym = s.owner; sym != null; sym = sym.owner) { if (sym.kind == TYP && sym.name == name && sym.name != names.error) { duplicateError(pos, sym); return true; } } return true; }
e.g
{ class A {} interface A{} // A类型的冲突将由checkUniqueClassName()方法来检测 } { class B { class B { } } }
2、Class name generation
/** Return name of local class. * This is of the form <enclClass> $ n <classname> * where * enclClass is the flat name of the enclosing class, * classname is the simple name of the local class */ Name localClassName(ClassSymbol c) { for (int i=1; ; i++) { Name flatname = names. fromString("" + c.owner.enclClass().flatname + syntheticNameChar + i + c.name); if (compiled.get(flatname) == null) return flatname; } }
e.g
package com.test07; public class Test3{ interface I { } // com.test07.Test3$1 final I i1 = new I() { }; // com.test07.Test3$2 final I i2 = new I() { }; { // com.test07.Test3$1B class B { } } public void m1() { // com.test07.Test3$2B class B { } // com.test07.Test3$1C class C { } // com.test07.Test3$3 final I i3 = new I() { }; } public void m2() { // com.test07.Test3$3B class B { } // com.test07.Test3$1D class D { } } }
3、Type Checking
/** Check that a given type is assignable to a given proto-type. * If it is, return the type, otherwise return errType. * @param pos Position to be used for error reporting. * @param found The type that was found. * @param req The type that was required. */ Type checkType(DiagnosticPosition pos, Type found, Type req) { // 不兼容的类型 return checkType(pos, found, req, "incompatible.types"); } Type checkType(DiagnosticPosition pos, Type found, Type req, String errKey) { if (req.tag == ERROR) return req; if (found.tag == FORALL) return instantiatePoly(pos, (ForAll)found, req, convertWarner(pos, found, req)); if (req.tag == NONE) return found; if (types.isAssignable(found, req, convertWarner(pos, found, req))) return found; if (found.tag <= DOUBLE && req.tag <= DOUBLE) // 可能损失精度 return typeError(pos, diags.fragment("possible.loss.of.precision"), found, req); if (found.isSuperBound()) { // 从 super-bound 类型{0}进行分配 log.error(pos, "assignment.from.super-bound", found); return types.createErrorType(found); } if (req.isExtendsBound()) { log.error(pos, "assignment.to.extends-bound", req); return types.createErrorType(found); } return typeError(pos, diags.fragment(errKey), found, req); }
checkType方法主要是在Attr类中的check()方法调用的,这个方法的代码如下:
/** Check kind and type of given tree against protokind and prototype. * If check succeeds, store type in tree and return it. * If check fails, store errType in tree and return it. * No checks are performed if the prototype is a method type. * It is not necessary in this case since we know that kind and type * are correct. * * @param tree The tree whose kind and type is checked * @param owntype The computed type of the tree * @param ownkind The computed kind of the tree * @param protokind The expected kind (or: protokind) of the tree * @param prototype The expected type (or: prototype) of the tree */ Type check(JCTree tree, Type owntype, int ownkind, int protokind, Type prototype) { // prototype不允许为method type if (owntype.tag != ERROR && prototype.tag != METHOD && prototype.tag != FORALL) { if ((ownkind & ~protokind) == 0) { owntype = chk.checkType(tree.pos(), owntype, prototype, errKey); } else { log.error(tree.pos(), "unexpected.type",kindNames(protokind),kindName(ownkind)); owntype = types.createErrorType(owntype); } } tree.type = owntype; return owntype; }
owntype与prototype都是Type类型,其tag取值是TypeTags类中的预先定义好的值。
ownkind与protokind都是int类型,其取值是Kinds预先定义好的值。
通过比较如上的两类型值,其实也就是比较了语法节点的Type类型与Symbol类型。
/** Check that a type is within some bounds. * * Used in TypeApply to verify that, e.g., X in V<X> is a valid type argument. * @param pos Position to be used for error reporting. * @param a The type that should be bounded by bs. * @param bs The bound. */ private boolean checkExtends(Type a, Type bound) { if (a.isUnbound()) { return true; } else if (a.tag != WILDCARD) { a = types.upperBound(a); return types.isSubtype(a, bound); } else if (a.isExtendsBound()) { Type ut = types.upperBound(a); boolean result = types.isCastable(bound, ut , Warner.noWarnings); return result; } else if (a.isSuperBound()) { Type lt = types.lowerBound(a); boolean result = !types.notSoftSubtype(lt, bound); return result; } return true; }
e.g
class BC<T extends Closeable> { public void test() { BC<InputStream> a = null; BC<? extends InputStream> b = null; BC<? super InputStream> c = null; } }
当a.isExtendsBound()时调用了isCastable()方法。
当a.isSuperBound()时调用了notSoftSubtype()方法。
Check类之duplicate declaration checking/Class name generation/Type Checking的更多相关文章
- 类型检查和鸭子类型 Duck typing in computer programming is an application of the duck test 鸭子测试 鸭子类型 指示编译器将类的类型检查安排在运行时而不是编译时 type checking can be specified to occur at run time rather than compile time.
Go所提供的面向对象功能十分简洁,但却兼具了类型检查和鸭子类型两者的有点,这是何等优秀的设计啊! Duck typing in computer programming is an applicati ...
- qt ISO C++ forbids declaration of 'XXXX' with no type
error: ISO C++ forbids declaration of 'XXXX' with no type 出现这个错误,一般是由于两个CPP相互都相互包含了对方的头文件造成的,比如: 当 ...
- Refused to execute script from '....js' because its MIME type ('text/html') is not executable, and strict MIME type checking is enabled.md
目录 问题描述 解决过程 总结 问题描述 在整合 Spring Boot.Spring Security.Thymeleaf 的练习中,对页面进行调试时,发现如下错误提示: Refused to ex ...
- Dynamic type checking and runtime type information
动态类型的关键是将动态对象与实际类型信息绑定. See also: Dynamic programming language and Interpreted language Dynamic type ...
- 类型检查 Type Checking 一些编程语言并不安全 类型化语言的优点 定型环境 (符号表) 断言的种类
Compiler http://staff.ustc.edu.cn/~bjhua/courses/compiler/2014/ http://staff.ustc.edu.cn/~bjhua/cour ...
- Check类中的incl、union,excl,diff,intersect
定义一些类,这些类之间有父子关系,如下: class Father{} class Son1 extends Father{} class Son2 extends Father{} class To ...
- Check类的validate方法解读
此方法的实现如下: public void validate(JCTree tree, Env<AttrContext> env, boolean checkRaw) { Validato ...
- Check类之TypeValidation
(1)Validator类的visitTypeApply()方法 实例1: class TestTypeVal<T extends InputStream>{ TestTypeVal< ...
- 怎么在eclipse中查到这个类用的是哪个jar的类和Eclipse 编译错误 Access restriction:The type *** is not accessible due to restriction on... 解决方案
找到了一个办法,你先按F3,然后点击Change Attached Source..按钮,在弹出的框里有个路径,我的路径是D:/SNFWorkSpace/JAR/~importu9.jar,然后你去引 ...
随机推荐
- (转)EF Power tool用法
转自:http://msdn.microsoft.com/zh-cn/data/jj593170.aspx 命令摘要 安装 EF Power Tools 后,将提供以下上下文菜单项.本演练将详细讨论这 ...
- Before an Exam
http://acm.hust.edu.cn/vjudge/contest/view.action?cid=93241#problem/B (654123) http://codeforces.com ...
- Lucene.net 性能《第八篇》
一.IndexReader性能分析 IndexReader完成了打开所有索引文件和提供底层reader API等繁重的工作,而IndexSearch则要简单得多.由于打开一个IndexReader需要 ...
- poj2462
看八戒在做这个题,我也做了做.. 坑很多,还是要注意细节.不得不吐槽,难道又到了计算几何只能套模板否则就一串WA的情况了么! 要不是八戒做出来了,这题我估计我也就扔到这里了..哥不服啊~所以得做出来! ...
- Hibernate 的复杂用法HibernateCallback
HibernateTemplate还提供了一种更加灵活的方式来操作数据库,通过这种方式可以完全使用Hibernate的操作方式.HibernateTemplate的灵活访问方式可通过如下两个方法完成: ...
- Python: pyinstaller打包exe(含file version信息)
最近项目上一直都是用Spyder直接运行.py文件的方式来执行每日的自动化程序,每天都要手动去点击Run来执行一次,所以考虑把.py文件直接打包成exe,然后用windows的task schedul ...
- leetcode 第三大的数
给定一个非空数组,返回此数组中第三大的数.如果不存在,则返回数组中最大的数.要求算法时间复杂度必须是O(n). 示例 1: 输入: [3, 2, 1] 输出: 1 解释: 第三大的数是 1. 示例 2 ...
- 构建NetCore应用框架之实战篇(五):BitAdminCore框架1.0登录功能设计实现及源码
本篇承接上篇内容,如果你不小心点击进来,建议从第一篇开始完整阅读,文章内容继承性连贯性. 构建NetCore应用框架之实战篇系列 一.设计原则 1.继承前面框架架构思维,设计以可读性作为首要目标. 2 ...
- 使用FluentScheduler实现定时任务管理
之前定时任务一直用的Windows服务,前段时间发现FluentScheduler这个框架,他跟Quarz.Net,Hangfire一样都是任务调度框架,但是相对使用而言我觉得FluentSchedu ...
- 2D Circular Geometry Kernel ( Geometry Kernels) CGAL 4.13 -User Manual
1 Introduction The goal of the circular kernel is to offer to the user a large set of functionalitie ...