Computer Languages - minijava - Concrete & Abstract...

Parsing Syntactic categories Concrete syntax Abstract syntax Composite pattern

Computer Languagesminijava - Concrete & Abstract syntax

1 Parsing2 Syntactic categories3 Concrete syntax4 Abstract syntax5 Composite pattern

February 4

Overview of a compiler

The compiler has to

Analyze the source code to understandwhat it means!

in the process of doing so it has toreject meaningless sources!

Generate code in a language for whicha machine exists.

The compiler has to

Compilersourcecode

machinecode

errors

Has to distinguish correct fromincorrect programs (has tounderstand!)

Has to generate correct machinecode!

Has to organize memory forvariables and instructions!

Has to agree with OS on theform of object code!

Compilersourcecode

machinecode

errors

Compilersourcecode

machinecode

errors

Compilersourcecode

machinecode

errors

Compilersourcecode

machinecode

errors

The Front End (analysis phase)

Scanner Parser Types Trans.sourcecode

tokens AS AS IR

errors

The Scanner (lexical analyzer) transforms a sequence ofcharacters (source code) into a sequence of tokens: arepresentation of the lexemes of the language.

The Parser (syntactical analyzer) takes the sequence of tokensand generates a tree representation, the Abstract Syntax.

This tree is analyzed by the type checker and is then used togenerate the intermediate representation.

tokens AS AS IR

errors

tokens AS AS IR

errors

tokens AS AS IR

errors

Plan for today

Syntactic categories

What kind of phrases are used in aprogramming language? Whatcomponents do they have?

Concrete syntax

How do the phrases look like? Whatwords, separators, punctuations areused?

Abstract syntax

What data structure can be used tostore the program so that thestructure is evident?

Composite pattern

How can we implement thisdata structure in Java?

Plan for today

Concrete syntax

Abstract syntax

Composite pattern

Plan for today

Concrete syntax

Abstract syntax

Composite pattern

Plan for today

Concrete syntax

Abstract syntax

Composite pattern

Plan for today

Concrete syntax

Abstract syntax

Composite pattern

Syntactic categories in programming languages

Example

class Small{public static void main(String args[]){

System.out.println("Hello "+args[0]);}

Example

resource small()string[10] usergetarg(1,user)write("Hello",user)

Example

import System

main =do (user:_) <- getArgs

print("Hello " ++ user)

Example

import System

Example

import System

Example

import System

What is a Program?

In Java and minijava: a class with a static main method!

and some class declarations!

In C and C++: a main function (a static method!)!

and some function or class declarations!

In other languages it could be something else.

What is a Program?

Declarations

What we use to buildabstractions: we give names toentities.

Expressions

Entities that have a value.Expressions usually have anassociated type.

Statements

Entities that change state.

Some programming languageshave no statements (Haskell). Insome languages statements andexpressions are mixed up (C).Different programming paradigmssupport different kinds of valuesand abstractions.

Declarations

Expressions

Statements

Declarations

Expressions

Statements

Declarations

Expressions

Statements

Declarations

Expressions

Statements

Syntactic categories in minijava - declarations

class A {int x;...

public int f(){int y;...

classes

A class can be used as thetype of an object.

fields

A field is a global variableinside a class, stands for aplace in memory.

methods

A method can be used insteadof a statement or anexpression.

local variables

Local to a method, stands fora register or a place inmemory.

class A {int x;...

classes

fields

methods

local variables

class A {int x;...

classes

fields

methods

local variables

class A {int x;...

classes

fields

methods

local variables

class A {int x;...

classes

fields

methods

local variables

Syntactic categories in minijava - expressions

3 true falsexx+98*y<x*y*ya.f()new A()arr[39]

constants

identifiers

using operators

calling a method

object creation

array element

What are the values of these expressions?

What are the values of identifiers? Are theexpressions meaningful?

constants

identifiers

using operators

calling a method

object creation

array element

constants

identifiers

using operators

calling a method

object creation

array element

constants

identifiers

using operators

calling a method

object creation

array element

constants

identifiers

using operators

calling a method

object creation

array element

constants

identifiers

using operators

calling a method

object creation

array element

constants

identifiers

using operators

calling a method

object creation

array element

constants

identifiers

using operators

calling a method

object creation

array element

constants

identifiers

using operators

calling a method

object creation

array element

Syntactic categories in minijava - statements

x = a.f();arr[38]=x;System.out.println(x);if (x<y) max=y;

else max=x;while(x<y)

x=x+y;

assignment

array assign

output

conditional

x=x+y;

assignment

array assign

output

conditional

x=x+y;

assignment

array assign

output

conditional

x=x+y;

assignment

array assign

output

conditional

x=x+y;

assignment

array assign

output

conditional

x=x+y;

assignment

array assign

output

conditional

Overview of the grammar for minijava

goal → mainClass classDeclarationList

mainClass → CLASS ID ’{’ PUBLIC STATIC VOID MAIN ’(’STRING ’[’ ’]’ ID ’)’ ’{’ statement ’}’

classDeclarationList → classDeclarationList classDeclaration| ∈

classDeclaration → CLASS ID ’{’ varDeclarationListmetDeclarationList ’}’

Overview of the grammar

metDeclaration → PUBLIC type ID ’(’ parameters ’)’ ’{’varDeclarationList statementListRETURN exp ’}’

statement → ’{’ statementList ’}’| IF ’(’ exp ’)’ statement ELSE statement| WHILE ’(’ exp ’)’ statement| SYSTEMOUT ’(’ exp ’)’ ’;’| ID ’=’ exp ’;’| ID ’[’ exp ’]’ ’=’ exp ’;’

Remarks

The grammar in the appendix ofthe book describes a small subsetof Java called minijava. Do notadd things to it! We will compileminijava!

When understanding thegrammar, start from the goal.But when implementing it injacc start with expressions, thenstatements and then thedeclarations!

Read in the jacc manual how todo lists, lists with separators, etc.

The grammar will have someconflicts. Test it a lot in order todecide whether you can let jaccsolve them or whether you willhave to modify the grammar!

Remarks

Storing the structure

While the parser recognizes aprogram, we cannot calculate a valuebut we can calculate a representationof the structure!

We forget about connectingterminals and about the derivationand concentrate in the othersyntactic categories that are used!

Storing the structure

While the parser recognizes aprogram, we cannot calculate a valuebut we can calculate a representationof the structure!

We forget about connectingterminals and about the derivationand concentrate in the othersyntactic categories that are used!

What do we need?

Main class

We only need the name and the statement!

Class Declaration

We only need the name, the list of variable declarations and thelist of method declarations.

Method declaration

We only need the result type, the name, the list of parameters, thelist of variable declarations, the statement list and the returnedexpression.

For a statement, it depends . . .

What do we need?

Main class

Class Declaration

Method declaration

What do we need?

Main class

Class Declaration

Method declaration

What do we need?

Main class

Class Declaration

Method declaration

What do we need?

Main class

Class Declaration

Method declaration

Statements

Assign

We need the identifier and theexpression.

Array element assign

We need the identifier, theindexing expression and the righthand side expression.

We need the expression and twostatements!

We need and expression and astatement!

Statements

Assign

Statements

Assign

Statements

Assign

Statements

Assign

All this looks very much like trees with diferent type and numberof children for different types of nodes!

Abstract syntax in Java

We need a datastructure inJava for the abstract syntaxtrees.

Problem

There will be a class for eachtype of node. But we wantthat all types of statementnodes can be used where astatement is needed! (and thesame for expressions by theway.)

Containers and components

Bricks are used as components,contained in a construction.But the construction can beused as a component in alarger construction!

Problem

Statements

Node for IF

public class If {public Exp e;public Statement s1,s2;

public If(Exp ae, Statement as1, Statement as2) {e=ae; s1=as1; s2=as2;

How do we see that this builds a statement, that can be usedwhere a statement is required?

Statements

Node for IF

public class If {public Exp e;public Statement s1,s2;

How do we see that this builds a statement, that can be usedwhere a statement is required?

Statements

Node for IF

public class If extends Statement{public Exp e;public Statement s1,s2;

Statements

public abstract class Statement {}

Statements

Node for IF

public class If extends Statement{public Exp e;public Statement s1,s2;

Statements

public abstract class Statement {}

Building syntax trees

As the parser recognizes sentences of the different forms, it willbuild the abstract syntax trees.

Example

statement : IF ’(’ exp ’)’ statement ELSE statement{ $$ = new If($3, $5, $7); }

There is an abstract class for thesyntactic category statement andone class inheriting from it foreach production for statement!

And the same idea is used forexpressions!

Example

Parsing minijava

tokens AS AS IR

errors

The project in the coursecontinues by using jacc to builda parser for minijava that, in theprocess of parsing a minijavaprogram, creates an abstractsyntax tree for the program!

You get the datastructure for theabstract syntax! (many classes,described in chapter 4 of thecourse book).

Parsing minijava

tokens AS AS IR

errors

Parsing minijava

tokens AS AS IR

errors

Computer Languages - minijava - Concrete & Abstract...

Documents

Transcript of Computer Languages - minijava - Concrete & Abstract...