1 Introduction 1
1.1 History of Compilation 2
1.2 What Compilers Do 4
1.2.1 Machine CodeGenerated byCompilers 4
1.2.2 TargetCode Formats 7
1.3 Interpreters 9
1.4 Syntax and Semantics 10
1.4.1 Static Semantics 11
1.4.2 Runtime Semantics 12
1.5 Organization of aCompiler 14
1.5.1 TheScanner 16
1.5.2 TheParser 16
1.5.3 TheTypeChecker(Semantic Analysis) 17
1.5.4 Translator(Program Synthesis) 17
1.5.5 Symbol Tables 18
1.5.6 TheOptimizer 18
1.5.7 TheCode Generator 19
1.5.8 Compiler Writing Tools 19
1.6 Programming Language and Compiler Design 20
1.7 Computer Architecture and Compiler Design 21
1.8 Compiler DesignConsiderations 22
1.8.1 Debugging(Development)Compilers 22
1.8.2 Optimizing Compilers 23
1.8.3 Retargetable Compilers 23
1.9 Integrated DevelopmentEnvironments 24
Exercises 26
2 A Simple Compiler 31
2.1 AnInformalDe.nition of the acLanguage 32
2.2 FormalDe.nition of ac 33
2.2.1 SyntaxSpeci.cation 33
2.2.2 TokenSpeci.cation 36
2.3 Phasesof aSimpleCompiler 37
2.4 Scanning 38
2.5 Parsing 39
2.5.1 Predicting aParsingProcedure 41
2.5.2 Implementing theProduction 43
2.6 AbstractSyntaxTrees 45
2.7 Semantic Analysis 46
2.7.1 Symbol Tables 47
2.7.2 TypeChecking 48
2.8 Code Generation 51
Exercises 54
3 Scanning—Theory and Practice 57
3.1 Overviewof aScanner 58
3.2 Regular Expressions 60
3.3 Examples 62
3.4 Finite Automata and Scanners 64
3.4.1 Deterministic Finite Automata 65
3.5 The LexScannerGenerator 69
3.5.1 De.ning Tokensin Lex 70
3.5.2 TheCharacterClass 71
3.5.3 UsingRegular Expressions to De.neTokens 73
3.5.4 CharacterProcessingUsingLex 76
3.6 OtherScannerGenerators 77
3.7 PracticalConsiderations ofBuildingScanners 79
3.7.1 ProcessingIdenti.ers andLiterals 79
3.7.2 UsingCompiler Directives and Listing SourceLines 83
3.7.3 Terminating theScanner 85
3.7.4 Multicharacter Lookahead 86
3.7.5 PerformanceConsiderations 87
3.7.6 LexicalErrorRecovery 89
3.8 Regular Expressions and Finite Automata 92
3.8.1 Transforming aRegularExpressioninto anNFA 93
3.8.2 Creating theDFA 94
3.8.3 Optimizing Finite Automata 97
3.8.4 Translating Finite Automata into Regular Expressions 100
3.9 Summary 103 Exercises 106
4 Grammars and Parsing 113
4.1 Context-FreeGrammars 114
4.1.1 Leftmost Derivations 116
4.1.2 Rightmost Derivations 116
4.1.3 ParseTrees 117
4.1.4 OtherTypes ofGrammars 118
4.2 Properties of CFGs 120
4.2.1 Reduced Grammars 120
4.2.2 Ambiguity 121
4.2.3 FaultyLanguage De.nition 122
4.3 TransformingExtended Grammars 122
4.4 Parsers and Recognizers 123
4.5 GrammarAnalysisAlgorithms 127
4.5.1 GrammarRepresentation 127
4.5.2 Deriving theEmpty String 128
4.5.3 FirstSets 130
4.5.4 FollowSets 134
Exercises 138
5 Top-Down Parsing 143
5.1 Overview 144
5.2 LL(k)Grammars 145
5.3 Recursive-DescentLL(1)Parsers 149
5.4 Table-Driven LL(1)Parsers 150
5.5 Obtaining LL(1)Grammars 154
5.5.1 Common Pre.xes 156
5.5.2 Left Recursion 157
5.6 A Non-LL(1)Language 159
5.7 Properties ofLL(1)Parsers 161
5.8 ParseTable Representation 163
5.8.1 Compaction 164
5.8.2 Compression 165
5.9 SyntacticErrorRecovery andRepair 168
5.9.1 ErrorRecovery 169
5.9.2 ErrorRepair 169
5.9.3 ErrorDetectionin LL(1)Parsers 171
5.9.4 ErrorRecoveryinLL(1)Parsers 171
Exercises 173
6 Bottom-Up Parsing 179
6.1 Overview 180
6.2 Shift-Reduce Parsers 181
6.2.1 LRParsers and RightmostDerivations 182
6.2.2 LRParsing asKnitting 182
6.2.3 LRParsing Engine 184
6.2.4 TheLRParseTable 185
6.2.5 LR(k)Parsing 187
6.3 LR(0)Table Construction 191
6.4 Con.ictDiagnosis 197
6.4.1 Ambiguous Grammars 199
6.4.2 Grammars that are not LR(k) 202
6.5 Con.ictResolution andTableConstruction 204
6.5.1 SLR(k)Table Construction 204
6.5.2 LALR(k)Table Construction 209
6.5.3 LALR Propagation Graph 211
6.5.4 LR(k)Table Construction 219 Exercises 224
7 Syntax-Directed Translation 235
7.1 Overview 235
7.1.1 SemanticActions andValues 236
7.1.2 Synthesized and Inherited Attributes 237
7.2 Bottom-Up Syntax-DirectedTranslation 239
7.2.1 Example 239
7.2.2 Rule Cloning 243
7.2.3 ForcingSemantic Actions 244
7.2.4 AggressiveGrammarRestructuring 246
7.3 Top-DownSyntax-DirectedTranslation 247
7.4 AbstractSyntaxTrees 250
7.4.1 Concrete and AbstractTrees 250
7.4.2 An Ef.cientAST DataStructure 251
7.4.3 InfrastructureforCreating ASTs 252
7.5 ASTDesign andConstruction 254
7.5.1 Design 256
7.5.2 Construction 258
7.6 ASTStructuresforLeft andRightValues 261
7.7 Design PatternsforASTs 264
7.7.1 Node ClassHierarchy 264
7.7.2 Visitor Pattern 265
7.7.3 Re.ectiveVisitorPattern 268 Exercises 272
8 Symbol Tables and Declaration Processing 279
8.1 Constructing aSymbolTable 280
8.1.1 Static Scoping 282
8.1.2 A Symbol Table Interface 282
8.2 Block-StructuredLanguages andScopes 284
8.2.1 Handling Scopes 284
8.2.2 OneSymbolTable orMany? 285
8.3 Basic Implementation Techniques 286
8.3.1 Entering andFindingNames 286
8.3.2 TheName Space 289
8.3.3 An Ef.cientSymbol Table Implementation 290
8.4 Advanced Features 293
8.4.1 Records and Typenames 294
8.4.2 Overloading andTypeHierarchies 294
8.4.3 Implicit Declarations 296
8.4.4 Export andImportDirectives 296
8.4.5 Altered SearchRules 297
8.5 Declaration ProcessingFundamentals 298
8.5.1 Attributes in the Symbol Table 298
8.5.2 TypeDescriptorStructures 299
8.5.3 TypeCheckingUsing anAbstractSyntaxTree 300
8.6 Variable andTypeDeclarations 303
8.6.1 Simple Variable Declarations 303
8.6.2 Handling TypeNames 304
8.6.3 TypeDeclarations 305
8.6.4 Variable DeclarationsRevisited 308
8.6.5 Static ArrayTypes 311
8.6.6 Struct and RecordTypes 312
8.6.7 Enumeration Types 313
8.7 Class and Method Declarations 316
8.7.1 ProcessingClassDeclarations 317
8.7.2 ProcessingMethod Declarations 321
8.8 An Introduction toTypeChecking 323
8.8.1 Simple Identi.ers and Literals 327
8.8.2 Assignment Statements 328
8.8.3 Checking Expressions 328
8.8.4 Checking ComplexNames 329
8.9 Summary 334 Exercises 336
9 Semantic Analysis 343
9.1 Semantic AnalysisforControl Structures 343
9.1.1 Reachability and Termination Analysis 345
9.1.2 IfStatements 348
9.1.3 While, Do, andRepeat Loops 350
9.1.4 ForLoops 353
9.1.5 Break,Continue, Return, andGoto Statements 356
9.1.6 Switch andCaseStatements 364
9.1.7 Exception Handling 369
9.2 Semantic Analysis ofCalls 376
9.3 Summary 384 Exercises 385
10 Intermediate Representations 391
10.1 Overview 392
10.1.1 Examples 393
10.1.2 TheMiddle-End 395
10.2 Java Virtual Machine 397
10.2.1 Introduction andDesignPrinciples 398
10.2.2 Contents of aClassFile 399
10.2.3 JVMInstructions 401
10.3 Static Single Assignment Form 410
10.3.1 Renaming and φ-functions 411
Exercises 414
11 Code Generation for a Virtual Machine 417
11.1 Visitors forCode Generation 418
11.2 Class and Method Declarations 420
11.2.1 ClassDeclarations 422
11.2.2 Method Declarations 424
11.3 The MethodBodyVisitor 425
11.3.1 Constants 425
11.3.2 References to LocalStorage 426
11.3.3 Static References 427
11.3.4 Expressions 427
11.3.5 Assignment 429
11.3.6 Method Calls 430
11.3.7 Field References 432
11.3.8 ArrayReferences 433
11.3.9 Conditional Execution 435 11.3.10Loops 436
11.4 The LHSVisitor 437
11.4.1 Local References 437
11.4.2 Static References 438
11.4.3 Field References 439
11.4.4 ArrayReferences 439 Exercises 441
12 Runtime Support 445
12.1 Static Allocation 446
12.2 Stack Allocation 447
12.2.1 Field AccessinClasses andStructs 449
12.2.2 AccessingFrames at Runtime 450
12.2.3 Handling Classes and Objects 451
12.2.4 Handling Multiple Scopes 453
12.2.5 Block-LevelAllocation 455
12.2.6 MoreAbout Frames 457
12.3 Arrays 460
12.3.1 Static One-Dimensional Arrays 460
12.3.2 Multidimensional Arrays 465
12.4 Heap Management 468
12.4.1 Allocation Mechanisms 468
12.4.2 Deallocation Mechanisms 471
12.4.3 Automatic GarbageCollection 472
12.5 Region-Based MemoryManagement 479 Exercises 482
13 Target Code Generation 489
13.1 Translating Bytecodes 490
13.1.1 Allocating memory addresses 493
13.1.2 Allocating Arrays andObjects 493
13.1.3 Method Calls 496
13.1.4 Example ofBytecodeTranslation 498
13.2 Translating ExpressionTrees 501
13.3 Register Allocation 505
13.3.1 On-the-FlyRegister Allocation 506
13.3.2 RegisterAllocation Using GraphColoring 508
13.3.3 Priority-Based RegisterAllocation 516
13.3.4 Interprocedural RegisterAllocation 517
13.4 Code Scheduling 519
13.4.1 ImprovingCode Scheduling 523
13.4.2 Global andDynamicCodeScheduling 524
13.5 Automatic Instruction Selection 526
13.5.1 InstructionSelection UsingBURS 529
13.5.2 InstructionSelection UsingTwig 531
13.5.3 OtherApproaches 532
13.6 Peephole Optimization 532
13.6.1 Levels ofPeepholeOptimization 533
13.6.2 AutomaticGeneration ofPeepholeOptimizers 536
Exercises 538
14 Program Optimization 547
14.1 Overview 548
14.1.1 WhyOptimize? 549
14.2 Control FlowAnalysis 555
14.2.1 Control FlowGraphs 556
14.2.2 Program andControlFlowStructure 559
14.2.3 DirectProcedureCall Graphs 560
14.2.4 Depth-FirstSpanning Tree 560
14.2.5 Dominance 565
14.2.6 Simple Dominance Algorithm 567
14.2.7 FastDominanceAlgorithm 571
14.2.8 Dominance Frontiers 581
14.2.9 Intervals 585
14.3 Introduction to DataFlowAnalysis 598
14.3.1 Available Expressions 598
14.3.2 LiveVariables 601
14.4 Data FlowFrameworks 604
14.4.1 Data FlowEvaluation Graph 604
14.4.2 Meet Lattice 606
14.4.3 TransferFunctions 608
14.5 Evaluation 611
14.5.1 Iteration 611
14.5.2 Initialization 615
14.5.3 Termination andRapidFrameworks 616
14.5.4 Distributive Frameworks 620
14.6 Constant Propagation 623
14.7 SSA Form 627
14.7.1 Placing φ-Functions 629
14.7.2 Renaming 631
Exercises 636
Bibliography 651
Abbreviations 661
Pseudocode Guide 663
Index
鄂公网安备 42010302001612号 