Compilation principle reading notes (1/12)

The first 1 Chapter Introduction

1.1 What is a compiler

Compiler is the translator of programming language , Translate the code written in high-level programming language into machine code that the machine can execute .

1.2 The compilation process and the structure of the compiler

1.2.1 Compilation process overview

The typical compilation process is divided into lexical analysis 、 Syntax analysis 、 Semantic analysis 、 Intermediate code generation 、 Code optimization and object code generation 6 Stages .

1. Lexical analysis
   Read the source program from left to right , Recognize the words one by one , Such as identifier 、 Reserved words 、 operator 、 Boundary symbol, etc .
2. Syntax analysis
   On the basis of lexical analysis, the word sequence is decomposed into various grammatical phrases , Like the procedure 、 sentence 、 Expression etc. . You can generate a syntax tree in the form of a table .
3. Semantic analysis
   Semantic analysis is to check whether the source program has semantic errors , Collect type information for the code generation phase .
4. Intermediate code generation
   Intermediate code is a simple structure 、 A well-defined system of notation , There are two important design principles ： Easy to generate and translate into object code . Many translation programs use an approximation “ Three address instructions ” Of “ Quaternion ” In the middle of the code , In the form of （ Operator , Operands 1, Operands 2, result ）.
5. Code optimization
   Change and transform the intermediate code , The purpose is to make the generated object code more efficient . Such as the deletion of common subexpressions 、 The intensity is weakened 、 Cycle optimization, etc .
6. Target code generation
   Transform the intermediate code into absolute instructions or relocatable instruction code or assembly instruction code on a specific machine . This stage is related to hardware system structure and instruction meaning .

The division of the above stages is a typical logical division , Not all of them are needed in the actual program .

1.2.2 The structure of the compiler

In addition to the above six components of the compiler , There are also forms management and error handling procedures , These two procedures are similar to the above 6 All stages are related .

The table processing program is equivalent to a small database , Store all kinds of data in the whole compiler processing process of the program .

1.2.3 Combination of compilation stages

The compilation process can be divided into front-end and back-end , The work of the front end mainly depends on the source language and has nothing to do with the target machine . The back end refers to the work that depends on the target machine and generally does not depend on the source language .

It is similar to the separation of front end and back end in software development , So the same front end , Equipped with multiple backend , You can generate multiple machine languages for one language .

1.3 Interpreter and some software tools

1.3.1 Interpretation procedure

The interpreter does not need to translate the source program into object code before running . Its working mode is to acquire one by one 、 Analyze and execute source program statements , There will be results immediately after the implementation , Suitable for interactive working .

python It's interpretive language , and java、C All compiled languages .

The interpretation of the program is very slow , Sometimes the interpretation of a high-level language source program is slower than running an equivalent machine code program 100 times . When running speed is very important , It cannot be explained . in addition , The space cost of interpreting language is also great .

The advantage of interpretation language is that the program written is easy to debug .

1.3.2 Software tools for processing source programs

Structured editor of language
Debugging tools for language programs
Program formatter
Language program testing tools
Program understanding tools
Conversion tools between high-level languages

These tools are found in many integrated development environments .

1.4 PL/0 Language compiler system

PL/0 The language compilation system will PL/0 The source program language is transformed into class P-code Target language . The programming language of this compiler given in this book is C Language .

Yes PL/0 And the class P-code I don't know much about the language . Baidu said PL/0 yes pascal A subset of .

1.4.1 PL/0 Language compilation system composition

PL/0 The language compilation system consists of two parts: compiler and interpreter , Known as PL/0 Compilers and classes P-code Interpretation procedure .PL/0 Language program is PL/0 The compiler converts to an equivalent class P-code Program , When the compiler ends normally ,PL/0 The language compilation system will call the interpreter , Explain and execute the generated target program .

PL/0 The compiler of is not directly programming machine code , Instead, it is compiled into a class P-code Code , This class P-code The code needs to be interpreted by the interpreter before it can be executed .

1.4.2 PL/0 Language

PL/0 What's important about language Pascal A subset of .
In practice , The syntax description of program languages often adopts an extended barks paradigm （EBNF） To describe .

ENBF It describes program syntax , It is also the rule that the language constructor follows .

1.4.3 class P-code Language

class P-code Languages can be regarded as classes P-code Assembly language of virtual machine .
class P-code Virtual machine is a simple machine with pure Trestle Structure , There is a stack memory , Yes 4 Control registers .
class P-code Data storage and arithmetic and logic operations during program operation are carried out at the top of the stack . class P-code The instruction format of the virtual machine is as follows ：
F L A
F Represents the opcode of an instruction ;L： The function represents the hierarchical difference between the reference layer and the declaration layer , If it doesn't work, set it to 0; A： Different instructions have different meanings .

class P-code Similar to assembly language . Operate directly on the register .

1.4.4 PL/0 compiler

PL/0 The compiler adopts the compilation process of single pass scanning , By lexical analysis program 、 Parser and code generator 3 An independent process .PL/0 The compiler takes the semantic analysis program as the core , When the grammar analysis needs to read words, it calls the lexical analysis program , When grammatical and semantic analysis correctly generates the target code of the corresponding language component , Just call the code generator .

1.4.5 PL/0 Driver code of language compilation system

The so-called driver code , Is the main function of the program , Connect the various parts of the compiler in series in the main function .