ICS 142 Klefstad

Compilers and Interpreters

Outline Introduction Compiler Relatives Example Compilers and Interpreters Languages Lexical Analysis Syntax Analysis Consistency Checks Syntax-Directed Translation Intermediate Code Code Optimization Run-time Environments Code Generation

Introduction

source language
- the language being compiling
target language
- the language being compiled to
compiler
- a program which translates programs in the source language to equivalent programs in the target language
interpreter
- a program which executes the instructions of the source language
the division between compiler and interpreter is arbitrary

Compiler Relatives

preprocessors
- support for symbolic constants
- parameterized macro expansion
- conditional compilation
- file inclusion
assemblers
- translates from assembly source code to machine code
loaders and link-editors
- facilitates separate compilation and relocatable object code
- resolves external references via persistent symbol table information

Example Compilers

language processing tools
- EG FLEX and BISON
text formatters
- EG NROFF, TeX, LATEX, TBL, EQN, PIC
source-to-source translators
- EG AT&T's C++ translator -- generates C
source-to-object compilers
- EG Verdix Ada, GNU C, Turbo Pascal

Example Interpreters

programming languages
- Basic
- Interlisp
- Prolog
- Snobol
- APL
editor languages
- AWK
- SED
- TECO
- EMACS Lisp extension language
command-interpreters
- CSH and DOS

Languages

a language is a formal notation which has a syntax and semantics
syntax may be described by a grammar
a grammar contains for components
- terminals, non-terminals, rewrite rules, and a start symbol
a string a sequence of terminals derived from a the start symbol of a grammar
a parse tree shows how a string is derived, and has:
- a root
- leaves - labeled by terminals
- internal nodes - labeled by a non-terminal

Lexical Analysis

useful for many applications, not just programming language compilers
regular languages may be described by regular expressions
scanning is the process of recognizing a regular language
used in compilers to convert input characters into lexical tokens
scanners may be hand-written or generated by tool

Syntax analysis

context free languages may be described by context free grammars
parsing is the process of determining if a string of tokens can be generated from a grammar
two approaches
- top-down
- bottom-up
parsers may be hand-written or generated by tool

Consistency checks

checks done by a language processor as an attempt to validate some of the semantics of a program
consistent use of identifiers
consistent matching of data types
correct number of parameters passed to a subroutine
check values to ensure they are within the correct range of values

Syntax-directed translation

a formalism for specifying the translation of a construct in terms of its attributes such as type, offset, and kind of entity
attributes may be either
- synthesized --> delivered by children
- inherited --> provided by parent

if <expr> then <stmt> --> {
    require_boolean_type(<expr>.type);
    generate_code(
        test <expr>.code
        branch_to_label_if_zero AFTER
        <stmt>.code
      AFTER:)
}

Intermediate Code

generation of intermediate code allows machine oriented optimizations to be done machine independently
common intermediate forms:
- expression trees or graphs
- postfix
```
(a+b)*(c-d) --> ab+cd-*
```
- three address code
```
T1 := a + b
T2 := c - d
T3 := T1 * T2
```

Code Optimization

space and time improvements which maintain program meaning
two kinds
- source level
- intermediate code level

Run-time Environments

storage organization
- program space
- static data
- activation stack
- heap
run-time support
- exception handling
- tasking
- input/output
- dynamic-storage allocation

Code generation

mapping intermediate code into machine-specific instructions
typically deals with
- register allocation
- instruction selection