Description Given as grammatical rules States what strings are - - PowerPoint PPT Presentation

78

Description

• f a

language

• Syntax

– Describes the structure of a program – Given as grammatical rules – States what strings are legitimate programs of the given language

• Syntax check (parsing)

– Generating a parse tree – Does the input string correspond to the grammar of the language?

• Semantics

– What is the meaning a given legal program – What kind of computation does a legal program produce – Static semantics (compile time), "contextual" checking (types, scopes) – Run-time semantics (program operation)

79

Phases of compilation

• Compilation usually divided to

separate phases: easier, simpler, clearer

• Output of previous phase is input of

next

• Grammar of each phase defines the

programming language

• Symbol table collects info on user-

defined constructs (variables, functions, types, ...)

80

Phases of code analysis

• If applied to natural language:

– lexical analysis: ”Dog cha5es the c? a” – syntax analysis: ”Man blue drive car.” – semantic analysis: ”Car hides moon”

• Phases of compilation:

type compatability etc. code generation parsing scanning Process contextual semantic

• syntactic

lexical Analysis

81

Source code Lexical analysis Syntactic analysis Optimisation Code generation

• Semantic analysis

tokens target language(assembly, machine)

• Intermediate codegen

characters

Symbol table Parse tree Intermediate code Executable

(maybe same)

(maybe)

Compilation

82

Source code Interpreter Inputs Results

Interpretation

Symbol table

83

Source Lexical analysis Syntax analysis Intermediate code gen. interpreter

• utput

tokens Parse tree Byte code

Input

Hybrid process

84

Precompiler

#define MAX_LOOP 100 #define INCR (a) ( a )++ #define FOR_LOOP ( var, from, to ) \\ for ( var = from; var <= to; INCR ( a ) ) { #define END_FOR } #define NULL FOR_LOOP ( n, 1, MAX_LOOP ) NULL; END_FOR; Example on macros

C-code:

85

Lexical analysis

Lexer

• grammar: regular
• format: regular expressions
• (implementation:

finite state machine) characters (source code) list of tokens

86

Lexical analysis

• Grouping of input strings

– lexeme

• Some item in the input text

– token

• Classification of lexemes, output of lexical analysis
• A "name/type" given to a lexeme
• Tokens have a type (keyword IF, number literal,
• perator, etc.) and value (IF, 123, +, i.e. interpreted

lexeme)

• Token

– A terminal symbol in the language syntax, further syntactic structures are built on tokens

index = 2 * count;

Lexeme Token index

identifier: index

=

assignment

2

integer: 2

*

• mult. operator

count

identifier: count

;

semicolon

87

Tokens

• Keywords: Reserved words
• Identifiers: Names chosen by the

programmer

• Literals: Values for constants
• Operators: Arithmetic and similar operations
• Separators: Symbols and strings that

separate language constructs

• Other things to consider in lexical analysis:

comments, white space, indentation

• Grammar for tokens: regular expressions

– [0-9][0-9]* – [A-Za-z][A-Za-z0-9]* – ".*"

88

Regular expressions ("regexes")

• "normal" characters: keyword
• . = any (1) char: k..word
• [abc] = set of chars: k[aeu]yword
• [a-z] = range of chars: k[a-e]yword
• Above can be combined: k[a-e0-9]y
• * = any number of previous (incl. zero):

ke*yw[o-u]*rd.*

• () = group chars together: key(word)*
• ? = 0 or 1 of previous: keyword[0-9]?
• + = at least 1 of prev: key(word)+
• | = alternative regexes: key(word|phrase)

89

program gcd ( input, output ); var i, j: integer; begin read ( i, j ); while i <> j do if i > j then i := i – j else j := j – i; writeln ( i ) end.

Identifying the lexemes

Pascal-code:

Lexical analysis

kw-program, ident(gcd), lparen, ident(input), comma, ident(output), rparen, semicolon, kw-var, ident(i), comma, ident(j), colon, ident(integer), semicolon, kw-begin, ident(read), lparen, ident(i), comma, ident(j), rparen, semicolon, kw-while, ident(i), noteq, ident(j), kw-do, kw-if, ident(i), greater, ident(j), kw-then, ident(i), assign, ident(i), minus, ident(j), kw-else, ident(j), assign, ident(j), minus, ident(i), semicolon, ident(writeln), lparen, ident(i), rparen, kw-end, fullstop