Return Styles: Pseud0ch, Terminal, Valhalla, NES, Geocities, Blue Moon. Entire thread

Help with unambiguous grammar

Name: long long 2012-01-13 22:59

I'm creating yet another fail C-family compiler.
Objectives: simple language, generics, type inference, and unambiguous grammar.
I would like to hear your opinions, any help is very appreciated.

bison parser:
%union {
    char *str;
}

%right '=' "+=" "-=" "*=" "/=" "%=" "~=" "&=" "|=" "^=" "<<=" ">>="
%left ':'
%right '?'
%left '|'
%left '^'
%left '&'
%left "or"
%left "and"
%left "==" "!="
%left '<' "<=" '>' ">="
%left "<<" ">>"
%left '+' '-'
%left '*' '/' '%'
%right "not" '~' PREFIX
%left "++" "--"
%left '['
%left '.'
%right '$'

%token <str> ID INT HEX BIN DOUBLE FLOAT COMPLEX STRING CHAR
%token-table

%%

program:
       | program import
       | program decl
       ;

import: "import" module
      | "import" module "as" identifier
      ;

module: identifier
      | module '.' identifier
      ;

decl: declvar
    | declfunc
    | declstruct
    ;

declvar: "var" identifier
       | "var" identifier '=' expr
       | "def" identifier '=' expr
       ;

declstruct: "def" identifier '{' declattrs '}'
          ;

declattrs: declattr
         | declattrs declattr
         ;

declattr: "var" identifier
        | declfunc
        ;

declfunc: "def" identifier '(' declargs ')' block
        ;

declargs: declarg
        | declargs ',' declarg
        ;

declarg: identifier
       | "var" identifier
       | identifier identifier
       ;

block: '{' '}'
     | '{' stmts '}'
     ;

stmts: stmt
     | block
     | stmts stmt
     | stmts block
     ;

stmt: declvar
    | purestmt
    | identifier ':' purestmt
    ;

purestmt: assign
        | call
        | if
        | for
        | while
        | switch
        | "goto" identifier
        | "return" expr
        ;

assign: lvalue '=' expr
      | lvalue "+=" expr
      | lvalue "-=" expr
      | lvalue "*=" expr
      | lvalue "/=" expr
      | lvalue "%=" expr
      | lvalue "~=" expr
      | lvalue "&=" expr
      | lvalue "|=" expr
      | lvalue "^=" expr
      | lvalue "<<=" expr
      | lvalue ">>=" expr
      | lvalue "++"
      | lvalue "--"
//      | "++" lvalue %prec PREFIX
//      | "--" lvalue %prec PREFIX
      ;

lvalue: identifier
      | lvalue '.' identifier
      | lvalue '[' expr ']'
      | '$' lvalue
      ;

identifier: ID
          ;

expr: literal
    | lvalue
    | assign
    | call
    | '~' expr
    | '@' lvalue
    | '-' expr %prec PREFIX
    | expr '+' expr
    | expr '-' expr
    | expr '*' expr
    | expr '/' expr
    | expr '%' expr
    | expr '&' expr
    | expr '|' expr
    | expr '^' expr
    | "not" expr
    | expr "==" expr
    | expr "!=" expr
    | expr '<' expr
    | expr "<=" expr
    | expr '>' expr
    | expr ">=" expr
    | expr "<<" expr
    | expr ">>" expr
    | expr "and" expr
    | expr "or" expr
    | expr '?' expr ':' expr
    | "proc" '(' declargs ')' block
    | '(' expr ')'
    ;

literal: INT
       | HEX
       | BIN
       | FLOAT
       | DOUBLE
       | COMPLEX
       | STRING
       | CHAR
       ;

call: onecall
    | call '.' onecall
    ;

onecall: lvalue '(' args ')'
       ;

args:
    | expr
    | args ',' expr
    ;

if: "if" '(' expr ')' block
  | "if" '(' expr ')' block "else" block
  ;

for : "for" '(' forarg "in" foriter ')' loop
    | "for" '(' forinit ';' expr ';' forincr ')' loop
    ;

forarg: identifier
      | "var" identifier
      ;

forinit: forinitarg
       | forinit ',' forinitarg
       ;

forinitarg: assign
          | "var" identifier '=' expr
          ;

forincr: assign
       | forincr ',' assign
       ;

foriter: lvalue
       | call
       ;

loop: ';'
    | loopblock
    ;

loopblock: '{' '}'
         | '{' loopstmts '}'
         ;

loopstmts: loopstmt
         | loopstmts loopstmt
         ;

loopstmt: "break"
        | "continue"
        | loopblock
        | stmt
        ;

while: "while" '(' expr ')' loop
     | "do" loopblock "while" '(' expr ')'
     ;

switch: "switch" '(' expr ')' '{' cases '}'
      ;

cases: allcases "default" ':' stmts
     ;

allcases: "case" expr ':' casestmts
        | allcases "case" expr ':' casestmts
        ;

casestmts: casestmt
         | casestmts casestmt
         ;

casestmt: "break"
        | caseblock
        | stmt
        ;

caseblock: '{' '}'
         | '{' casestmts '}'
         ;

Name: Anonymous 2012-01-14 13:30

Ok, I added C-like array initializers (var x : list[int] = {1,2,3}; var g : Game = {"loser", 0}) and pointer types (var y : ptr[int]) to the grammar.

But for now I'll take some time (maybe one week?) to mature ideas before writing the AST to LLVM backend.

Also guys, thanks for your time and help.

Newer Posts
Don't change these.
Name: Email:
Entire Thread Thread List