Kompilatorer och interpretatorer: Lecture 1

Information about the course

• Me
• http://www.aass.oru.se/~tpy/koi/
• Subject and objectives of the course
• Lectures
• Labs
• The scheduled lab on Wednesday
• Exam
• What you're supposed to know already

Today: Aho et al, Chapter 1, Introduction to Compiling,
KP kapitel 1, Kompilatorns beståndsdelar

1.1 Compilers

input: source program
output: target program, error messages, warnings

the target program: machine code for physical or abstract machine, assembler, C, silicon design...

Systematic techniques: The first Fortran compiler took 18 staff-years in the 1950's. In the 1980's, "a substantial compiler can be implemented even as a student project in a one-semester compiler design course".

Main idea:

analysis (Swedish: "analys")
(one or more) intermediate representation(s) (Sw: "internform"): usually a tree, syntax tree
synthesis (Sw: "syntes")

Examples of compiler-like tools that perform analysis:

• Structure editors (ex: Emacs' C mode)
• Interpreters (execute the tree, or generate byte-code or similar, compare: abstract machine) "script languages", or "command languages", are often interpreted
• Silicon compilers
• HTML or XML processors
• Programs that read configuration files

The context (Sw: "omgivning", "kontext") of a compiler:

Plus:

• Sometimes a preprocessor (C: "#include" etc) before the actual compiler
• Sometimes the compiler generates assembly code, which is translated into machine code by by an assembler
• Relocatable machine code = object code, absolute machine code
• Static and dynamic linking

1.2 Analysis of the source program, 1.3 The Phases of a Compiler

1. Lexical analysis ("scanning")
2. Syntactical analysis ("parsing")
3. Semantic analysis
4. Intermediate code generation
5. Code optimization
6. Code generation
7. (The symbol table)
8. (Error handling: detection and reporting)

Lexical analysis

Tokens

Syntax analysis

• expression -> identifier
• expression -> number
• expression -> expression₁ "+" expression₂
• expression -> expression₁ "*" expression₂
• expression -> ( expression₁ )
• assignment statement -> identifier₁ ":=" expression₁

This means: An expression can consist of an identifier. An expression can also consist of a number.

Again: position := initial + rate * 60

Aho fig 1.4: "concrete syntax tree", or "parse tree"

Aho fig 1.2: "abstract syntax tree", or just "syntax tree"

All the phases

1.5 The Grouping of Phases

• Front end
• Back end
• Pass

1.4 Cousins of the Compiler

1.6 Compiler-Construction Tools

• Parser generators (ex: Yacc, Bison)
• Scanner generators (ex: Lex, Flex)