CSC 338: Compiler design and implementation

1
CSC 338 Compiler design and implementation

• Dr. Mohamed Ben Othman

2
Goals

• Compilers are used everyday in most computers.
• Allow students to implement big programming
  projects.
• Parts of most projects (specially those
  containing command language) can be built in the
  same way compilers are built.

3
Introduction

• Terminology
• Compilers implementation languages
• Compiler structure
• Example

4
Terminology

• Compilers translate programs written in high
  level languages. The entire program has to be
  translated before execution (compared to a book
  translation).
• Interpreters translate statement by statement
  (or line by line) and then this statement is
  given to the CPU to be executed before the next
  statement in being translated (compared to a man
  translator).
• High Level Language
• Source Language
• Source Code
• Implementation language
• Target machine the machine to which the
  translation will be done.

5
How to choose the compiler implementation language

• The first compiler was written in assembly
  language
• The compiler of Fortran77 programming language is
  Pascal.
• It is preferable to write the compiler in the
  source language from which it will translate to
  machine language.

6
Translation operation

• The are some compilers do not respect this
  pattern.
• Load-and-go Compilers gives programs ready for
  execution.
• Cross-language compilers compilers between high
  level languages.

7
Compiler structure

• Compiler is divided into two parts Front End
  and Back End.
• The Front End translates a program from source
  language to an intermediate language.
• The back End translates from the intermediate
  language to the machine language.

8
The Front End
9
The Back End
10
Lexical analysis

• The lexical analyzer reads the input program as
  a character stream and produces a stream of
  lexemes (or token strings) as output.
• The lexical analyzer reads the input program
  character by character until it reads a word
  (symbol).
• The lexical analyzer searches the current work in
  a table (called symbol table) and adds it if not
  found. The lexical analyzer produces an output
  for each symbol called token. Tokens are
  generally integer numbers.

11
Syntactic Analysis

• Syntactic Analyzer (or parser) takes as input
  the Token stream produced by the lexical
  analyzer.
• The parser produces a Parse Tree

12
Semantic Analysis

• The semantic analyzer determines if the meaning
  is respected in the user program.
• The syntax rules may be respected without
  respecting the meaning.
• Semantic analysis is mainly to be sure that data
  types are used properly.
• Semantic analysis is part of syntactic analysis.

13
Detecting errors in the source program

• In all phases above the main goal is to determine
  if the program source respects the source
  language rules.
• Example
• for (int i 1 iltn i) x
• if x gt N Y - 3 else Y 3

This error may be detected at lexical analysis
This is a syntax error
14
Intermediate code generation

• The intermediate code generator produces a code
  that is not related to the target machine. The
  intermediate code has to be very close to the and
  very easy to translate to machine language.

15
Object Code

• The target code is the machine code.
• The machine code generation is not the same as
  the intermediate code generation.
• A assembly language code generation may be done
  in the same time

16
Lexical analysis example (Pascal program)

• PROGRAM AverageNumbers(Input, Output)
• CONST Amount 3
• VAR Average Real
• x ARRAY1..Amount OF Integer
• i, Sum Integer
• BEGIN
• x13
• x26
• x310
• Sum 0
• FOR i 1 TO Amount DO Sum Sum xi
• Average Sum/Amount
• END. AverageNumbers

17
Lexical analysis result

• PROGRAM ID (ID , ID )
• CONST ID NUMLITERAL
• VAR ID ID
• ID ARRAY1 .. ID OF ID
• ID, ID ID
• BEGIN
• and so on

18
Miscellaneous

• Symbol Table contains all keywords and symbols.
• Symbol Table Handler manages the symbol table.
• Error Handling gives a clear description of
  errors.

19
Resume