Compilation Phases - Mini-Compilador Educativo

Overview

The Mini-Compilador Educativo transforms source code through six sequential phases, each building on the output of the previous one. This page provides an in-depth look at how a sample program flows through the entire compilation pipeline.

Example Program

We’ll trace this simple program through all phases:

let a = 5;
let b = 10;
let c = a + b * 2;
print c;

Expected output: 30 (because b * 2 = 20, then a + 20 = 25… wait, let me recalculate: 10 * 2 = 20, 5 + 20 = 25. Actually the output should be 25.)

Phase 1: Lexical Analysis (Scanner)

Input

Raw character stream:

let a = 5;\nlet b = 10;\nlet c = a + b * 2;\nprint c;

Processing

The Scanner class reads character-by-character:

Recognizes keywords by dictionary lookup
Identifies operators as single characters
Accumulates digit sequences into numbers
Collects alphanumeric sequences into identifiers
Tracks line and column positions

Output

List of tokens with metadata:

[
  Token(LET, 'let', línea=1, col=1),
  Token(IDENTIFICADOR, 'a', línea=1, col=5),
  Token(IGUAL, '=', línea=1, col=7),
  Token(NUMERO, '5', línea=1, col=9, valor=5),
  Token(PUNTO_COMA, ';', línea=1, col=10),
  
  Token(LET, 'let', línea=2, col=1),
  Token(IDENTIFICADOR, 'b', línea=2, col=5),
  Token(IGUAL, '=', línea=2, col=7),
  Token(NUMERO, '10', línea=2, col=9, valor=10),
  Token(PUNTO_COMA, ';', línea=2, col=11),
  
  Token(LET, 'let', línea=3, col=1),
  Token(IDENTIFICADOR, 'c', línea=3, col=5),
  Token(IGUAL, '=', línea=3, col=7),
  Token(IDENTIFICADOR, 'a', línea=3, col=9),
  Token(SUMA, '+', línea=3, col=11),
  Token(IDENTIFICADOR, 'b', línea=3, col=13),
  Token(MULTIPLICACION, '*', línea=3, col=15),
  Token(NUMERO, '2', línea=3, col=17, valor=2),
  Token(PUNTO_COMA, ';', línea=3, col=18),
  
  Token(PRINT, 'print', línea=4, col=1),
  Token(IDENTIFICADOR, 'c', línea=4, col=7),
  Token(PUNTO_COMA, ';', línea=4, col=8),
  
  Token(FIN_ARCHIVO, '', línea=4, col=9)
]

Key Features

Position Tracking

Every token stores its line and column for precise error reporting.

Value Extraction

Numeric tokens carry both the original string (lexema) and parsed integer (valor).

Comment Handling

// comments are consumed without generating tokens (see compfinal.py:319-327).

Error Collection

Invalid characters add to Scanner.errores list but don’t stop scanning.

Phase 2: Syntactic Analysis (Parser)

Input

Token stream from Scanner

Processing

The Parser uses recursive descent to match tokens against grammar rules:

parsear() → processes statements until EOF
_sentencia() → dispatches to _declaracion_variable() or _sentencia_print()
_expresion() → delegates to _suma_resta() → _multiplicacion_division() → _primario()
Operator precedence handled by nesting depth

Output

Abstract Syntax Tree (AST)

AST Representation

Tree View
Python Objects

Programa
├── DeclaracionVariable
│   ├── nombre: 'a'
│   └── valor:
│       └── NumeroLiteral(5)
├── DeclaracionVariable
│   ├── nombre: 'b'
│   └── valor:
│       └── NumeroLiteral(10)
├── DeclaracionVariable
│   ├── nombre: 'c'
│   └── valor:
│       └── ExpresionBinaria
│           ├── operador: '+'
│           ├── izquierda: Identificador('a')
│           └── derecha:
│               └── ExpresionBinaria
│                   ├── operador: '*'
│                   ├── izquierda: Identificador('b')
│                   └── derecha: NumeroLiteral(2)
└── SentenciaPrint
    └── expresion:
        └── Identificador('c')

Programa(
  sentencias=[
    DeclaracionVariable(
      token_let=Token(LET, 'let', 1, 1),
      nombre=Token(IDENTIFICADOR, 'a', 1, 5),
      expresion=NumeroLiteral(
        token=Token(NUMERO, '5', 1, 9, 5),
        valor=5
      )
    ),
    
    DeclaracionVariable(
      token_let=Token(LET, 'let', 2, 1),
      nombre=Token(IDENTIFICADOR, 'b', 2, 5),
      expresion=NumeroLiteral(
        token=Token(NUMERO, '10', 2, 9, 10),
        valor=10
      )
    ),
    
    DeclaracionVariable(
      token_let=Token(LET, 'let', 3, 1),
      nombre=Token(IDENTIFICADOR, 'c', 3, 5),
      expresion=ExpresionBinaria(
        izquierda=Identificador(
          token=Token(IDENTIFICADOR, 'a', 3, 9),
          nombre='a'
        ),
        operador=Token(SUMA, '+', 3, 11),
        derecha=ExpresionBinaria(
          izquierda=Identificador(
            token=Token(IDENTIFICADOR, 'b', 3, 13),
            nombre='b'
          ),
          operador=Token(MULTIPLICACION, '*', 3, 15),
          derecha=NumeroLiteral(
            token=Token(NUMERO, '2', 3, 17, 2),
            valor=2
          )
        )
      )
    ),
    
    SentenciaPrint(
      token_print=Token(PRINT, 'print', 4, 1),
      expresion=Identificador(
        token=Token(IDENTIFICADOR, 'c', 4, 7),
        nombre='c'
      )
    )
  ]
)

Precedence in Action

Expression: a + b * 2Why it parses as a + (b * 2) and not (a + b) * 2:

_expresion() calls _suma_resta()
_suma_resta() calls _multiplicacion_division() for left side → gets a
Sees +, so calls _multiplicacion_division() for right side
_multiplicacion_division() processes b * 2 completely (consuming both tokens) before returning
Result: ExpresionBinaria(+, a, ExpresionBinaria(*, b, 2))

Source: compfinal.py:802-843

Phase 3: Semantic Analysis

Input

AST from Parser

Processing

The AnalizadorSemantico traverses the AST with these checks:Variable Tracking:

Maintains variables_declaradas set
On DeclaracionVariable: analyze expression first, then add variable
On Identificador: verify variable exists in set

Division by Zero:

On ExpresionBinaria with / operator
If right side is NumeroLiteral with valor == 0, report error

Output

Boolean success flag + error list

Semantic Trace

# Statement 1: let a = 5;
1. Analyze expression: NumeroLiteral(5) ✓
2. Add 'a' to variables_declaradas
3. variables_declaradas = {'a'}

# Statement 2: let b = 10;
1. Analyze expression: NumeroLiteral(10) ✓
2. Add 'b' to variables_declaradas
3. variables_declaradas = {'a', 'b'}

# Statement 3: let c = a + b * 2;
1. Analyze expression:
   - Left side: Identificador('a')
     → Check: 'a' in variables_declaradas ✓
   - Right side: ExpresionBinaria(*, ...)
     - Left: Identificador('b')
       → Check: 'b' in variables_declaradas ✓
     - Right: NumeroLiteral(2) ✓
     - Operator: * (not division) ✓
2. Add 'c' to variables_declaradas
3. variables_declaradas = {'a', 'b', 'c'}

# Statement 4: print c;
1. Analyze expression: Identificador('c')
   → Check: 'c' in variables_declaradas ✓

# Result: No errors ✓

Semantic analysis only catches literal division by zero. Runtime divisions like 10 / x where x becomes 0 are not detected.

Source: compfinal.py:968-1085

Phase 4: Intermediate Code Generation

Input

Validated AST

Processing

The GeneradorIR creates Three Address Code (TAC):

Each operation becomes one TAC instruction
Temporary variables (t0, t1, …) store intermediate results
Variables use their original names

Output

List of TAC instructions as strings

Generated IR Code

a = 5
b = 10
t0 = b * 2
t1 = a + t0
c = t1
print c

TAC Properties

Explicit Ordering

Operations occur in sequential order, making control flow obvious.

Single Operator

Each instruction has at most one operator, simplifying optimization and code generation.

Platform Independent

TAC is abstract; not tied to any specific CPU architecture.

Optimization Ready

Easy to apply optimizations like constant folding, dead code elimination, etc.

Source: compfinal.py:1090-1149

Phase 5: Interpretation

Input

AST (same as used for IR generation)

Processing

The Interprete directly executes the AST:

Maintains variables dictionary for runtime values
Recursively evaluates expressions
Outputs print statement results to console

Output

Console output from print statements

Execution Trace

# Statement 1: let a = 5;
evaluar(NumeroLiteral(5)) → 5
variables['a'] = 5
# variables = {'a': 5}

# Statement 2: let b = 10;
evaluar(NumeroLiteral(10)) → 10
variables['b'] = 10
# variables = {'a': 5, 'b': 10}

# Statement 3: let c = a + b * 2;
evaluar(ExpresionBinaria(+, ...)):
  izquierda = evaluar(Identificador('a'))
    → variables['a'] → 5
  derecha = evaluar(ExpresionBinaria(*, ...)):
    izquierda = evaluar(Identificador('b'))
      → variables['b'] → 10
    derecha = evaluar(NumeroLiteral(2))
      → 2
    operador.lexema == '*'
      → 10 * 2 → 20
  operador.lexema == '+'
    → 5 + 20 → 25
variables['c'] = 25
# variables = {'a': 5, 'b': 10, 'c': 25}

# Statement 4: print c;
evaluar(Identificador('c')) → variables['c'] → 25
print("     Resultado:")
print("                → 25")

The interpreter uses floor division for / operator: 7 / 2 evaluates to 3, not 3.5.Source: compfinal.py:1209

Source: compfinal.py:1156-1215

Phase 6: Assembly Code Generation

Input

AST

Processing

The GeneradorASM produces x86 assembly for EMU8086:

Data section: Declare all variables as words (dw)
Code section: Translate each statement:
- Use AX register as accumulator
- Use stack for nested expressions
- Call print_num procedure for output
Utility procedures: Include print_num for integer output

Output

Complete .asm file as string

Generated Assembly

.model small
.stack 100h

.data
a dw 0
b dw 0
c dw 0
msg db 13,10,'$'

.code
main proc
mov ax, @data
mov ds, ax

; let a = 5;
mov ax, 5
mov a, ax

; let b = 10;
mov ax, 10
mov b, ax

; let c = a + b * 2;
mov ax, a
push ax
mov ax, b
push ax
mov ax, 2
mov bx, ax
pop ax
imul bx
mov bx, ax
pop ax
add ax, bx
mov c, ax

; print c;
mov ax, c
call print_num
mov ah,09h
lea dx,msg
int 21h

mov ah,4ch
int 21h
main endp

; imprimir numero en AX
print_num proc
push ax
push bx
push cx
push dx
mov cx,0
mov bx,10
cmp ax,0
jne convert
mov dl,'0'
mov ah,02h
int 21h
jmp done
convert:
next_digit:
xor dx,dx
div bx
push dx
inc cx
cmp ax,0
jne next_digit
print_loop:
pop dx
add dl,'0'
mov ah,02h
int 21h
loop print_loop
done:
pop dx
pop cx
pop bx
pop ax
ret
print_num endp

end main

Assembly Generation Strategy

Stack-Based Evaluation
Register Usage
DOS Interrupts
Print Number Logic

Binary expressions use stack to preserve left operand:

Evaluate left → result in AX
Push AX
Evaluate right → result in AX
Move AX to BX
Pop AX (restore left)
Perform operation (AX op BX → AX)

This handles nested expressions naturally through recursion.

AX: Primary accumulator, holds expression results
BX: Right operand for binary operations
CX: Loop counter in print_num
DX: DOS interrupt parameter, digit extraction
DS: Data segment register
Stack: Temporary storage for nested expressions

Source: compfinal.py:1221-1390

Phase Comparison

Here’s how the same expression a + b * 2 is represented at each phase:

Phase	Representation
1. Lexer	`[ID(a), SUMA, ID(b), MULT, NUM(2)]`
2. Parser	`BinExpr(+, ID(a), BinExpr(*, ID(b), Lit(2)))`
3. Semantic	✓ Variables exist, no errors
4. IR	`t0 = b * 2` `t1 = a + t0`
5. Interpreter	Evaluates to: `25`
6. Assembly	`mov ax, b` `imul 2` `add ax, a` (simplified)

Error Handling Across Phases

Lexical Errors (Phase 1)

Detected by: Scanner
Examples:

Invalid characters: let x = 5@;
Unexpected symbols: let x = $100;

Handling:

Error added to Scanner.errores
Token type set to ERROR
Scanning continues to find more errors

Output:

Error léxico en línea 1, columna 10: carácter inesperado '@'

Syntax Errors (Phase 2)

Detected by: Parser
Examples:

Missing semicolon: let x = 5
Wrong token order: = x 5 let;
Unmatched parentheses: let x = (5 + 3;

Handling:

ErrorSintaxis exception raised
Error caught and added to Parser.errores
Synchronization to next statement boundary
Parsing continues

Output:

Error de sintaxis en línea 1, columna 10: Se esperaba ';' al final de la declaración. Se encontró ''

Semantic Errors (Phase 3)

Detected by: AnalizadorSemantico
Examples:

Undeclared variable: print x; (without let x = ...;)
Division by zero: let x = 10 / 0;

Handling:

Error added to AnalizadorSemantico.errores
Analysis continues to find more errors

Output:

Error semántico en línea 2, columna 7: la variable 'x' no ha sido declarada

No Runtime Errors (Phases 4-6)

Once semantic analysis passes, subsequent phases assume correctness:

IR Generation: Always succeeds with valid AST
Interpretation: May encounter runtime division by zero (not caught)
Assembly Generation: Produces syntactically correct code

The compiler does not detect:

Runtime division by zero (variable divisor)
Integer overflow
Stack overflow from deep recursion

Complete Compilation Output

Console Output
Generated Assembly

═══════════════════════════════════════════════════════════════════════════
  MINI-COMPILADOR EDUCATIVO
═══════════════════════════════════════════════════════════════════════════

[FASE 1] Análisis Léxico...
         (Convirtiendo código en tokens)
  ✓ Completado: 24 tokens generados

[FASE 2] Análisis Sintáctico...
         (Verificando gramática y construyendo AST)
  ✓ Completado: 4 sentencias parseadas

[FASE 3] Análisis Semántico...
         (Verificando que el código tenga sentido)
  ✓ Completado: 3 variables verificadas

[FASE 4] Generación de Código Intermedio (IR)
         (Three Address Code)
    a = 5
    b = 10
    t0 = b * 2
    t1 = a + t0
    c = t1
    print c

[FASE 5] Ejecución del programa...
              (Interpretando el AST)
     Resultado:
                → 25
  ✓ Ejecución finalizada

───────────────────────────────────────────────────────────────────────────
  ✓ RESULTADO: ¡Compilación exitosa!

Available via “Exportar ASM” button in GUI, produces .asm file ready for EMU8086.

Performance Characteristics

Time Complexity

Lexer: O(n) where n = character count
Parser: O(n) where n = token count
Semantic: O(n) where n = AST node count
IR/ASM: O(n) where n = AST node count
Overall: O(n) linear in input size

Space Complexity

Token list: O(n) tokens
AST: O(n) nodes
Symbol table: O(v) variables
IR code: O(n) instructions
Overall: O(n) linear in input size

The compiler completes in milliseconds for typical programs (< 100 lines). GUI shows compilation time in status bar.

Next Steps

Try It Yourself

Experiment with the compiler by:

Modifying the example program
Introducing intentional errors to see error messages
Viewing the AST and IR output (enable in GUI checkboxes)
Exporting and running the assembly code in EMU8086

See Installation and Quick Start for setup and usage instructions.

Get Started

Core Concepts

Guides

Compiler Components

API Reference

Examples

​Overview

​Example Program

​Phase 1: Lexical Analysis (Scanner)

​Key Features

Position Tracking

Value Extraction

Comment Handling

Error Collection

​Phase 2: Syntactic Analysis (Parser)

​AST Representation

​Precedence in Action

​Phase 3: Semantic Analysis

​Semantic Trace

​Phase 4: Intermediate Code Generation

​Generated IR Code

​TAC Properties

Explicit Ordering

Single Operator

Platform Independent

Optimization Ready

​Phase 5: Interpretation

​Execution Trace

​Phase 6: Assembly Code Generation

​Generated Assembly

​Assembly Generation Strategy

​Phase Comparison

​Error Handling Across Phases

​Complete Compilation Output

​Performance Characteristics

Time Complexity

Space Complexity

​Next Steps

Try It Yourself

Build docs developers (and LLMs) love

Overview

Example Program

Phase 1: Lexical Analysis (Scanner)

Key Features

Phase 2: Syntactic Analysis (Parser)

AST Representation

Precedence in Action

Phase 3: Semantic Analysis

Semantic Trace

Phase 4: Intermediate Code Generation

Generated IR Code

TAC Properties

Phase 5: Interpretation

Execution Trace

Phase 6: Assembly Code Generation

Generated Assembly

Assembly Generation Strategy

Phase Comparison

Error Handling Across Phases

Complete Compilation Output

Performance Characteristics

Next Steps