mycc

A toy c compiler for learning purpose, support some C99 features and generate code for x86_64.

References

C

• C99 Gramer
• C standards
• C99 html pdf
• C operator precedence
• Understand integer conversion rules(Integer Promotions, Integer Conversion Rank, Usual Arithmetic Conversions), C99(6.3)
• Implicity prototype C99 vs C90
• alignment

x86

• Guide to x86-64
• x86 and amd64 instruction reference
• gas manual
• x86_64-abi (3.2 Function Calling Sequence)
• call printf in x86_64 using gas
• RPI relative addressing

Concept

• Computer Systems A Programmers Perspective (3rd)
• Compiler
• Compilers Principles, Techniques, and Tools (2rd)
• Recursive descent parser
• LL parser
• Scope wiki C

Some Implements

• chibicc
• lcc ( book )
• acwj

Others

• How to disable GNU C extensions?
• Makefile header dependencies

TODO

• Types
  • integrial types
    • char
    • shrot
    • int
    • long
    • long long
  • floaing types
  • pointer types
  • enumeration types
  • function types
    • returing struct/union or taking sruct/union as paramenter
  • array types
    • variable length array type
  • struct types
  • union types
  • type alignment
• Conversions
  • floating types
• Lexical elements
  • constant
    • integral constant
      • decimal
      • octal / hex
      • suffix
    • floating constant
    • character constant
    • string literials
      • simple escape
      • octal/hex escape
  • comments
• Statements
  • switch statement
  • goto statement
• Expressions
  • operand constrains: null pointer? type qualifer? type compatible?
  • sizeof type-name
  • struct/union related operator
  • cast operator
• Declarations
  • typedef
  • initialization
  • declaration of function without name
  • struct/union with bit-field member
  • struct/union with flexible array member
• Preprocessing
• Others
  • remove print statement by calling library function "printf"
  • for test, compile the source code with gcc, use the output of gcc as target output
  • for debug: add corrsponding source line in output file
  • add source location in error message
    • associate related Token to Node and Type
    • distinguish identifiers referencing to the same var (we can error at current identifier that reference the var, instead of where the var declared)
  • error recovery, (test/085.c)
  • read srouce code from file, write output to file