Structure of a compiler

Any large software is easier to understand and implement if it is divided into well-defined modules.

Figure 6: The structure of a compiler.

• In a compiler,
  • linear analysis
    • is called LEXICAL ANALYSIS or SCANNING and
    • is performed by the LEXICAL ANALYZER or LEXER,
  • hierarchical analysis
    • is called SYNTAX ANALYSIS or PARSING and
    • is performed by the SYNTAX ANALYZER or PARSER.
• During the analysis, the compiler manages a SYMBOL TABLE by
  • recording the identifiers of the source program
  • collecting information (called ATTRIBUTES) about them: storage allocation, type, scope, and (for functions) signature.
• When the identifier x is found by the lexical analyzer
  • generates the token id
  • enters the lexeme x in the symbol-table (if it is not already there)
  • associates to the generated token a pointer to the symbol-table entry x. This pointer is called the LEXICAL VALUE of the token.
• During the analysis or synthesis, the compiler may DETECT ERRORS and report on them.
  • However, after detecting an error, the compilation should proceed allowing further errors to be detected.
  • The syntax and semantic phases usually handle a large fraction of the errors detectable by the compiler.