X-Git-Url: https://git.kianting.info/?a=blobdiff_plain;f=src%2Findex.ts;h=26a690eb3b71659a119668c7e4c5f202faf04a39;hb=d4dfd2e99f564ca880d40172687986ea3ce757f0;hp=60e74871dc96261bf3ddcb2eb4d3b332952f8f57;hpb=fed01e9044149985b3df74a9225ce68805b8f478;p=clo

diff --git a/src/index.ts b/src/index.ts
index 60e7487..26a690e 100644
--- a/src/index.ts
+++ b/src/index.ts
@@ -1,7 +1,9 @@
 var fs = require('fs');
 import jsTokens from "js-tokens";
 import * as util from 'util';
-
+import * as p from 'typescript-parsec';
+import { Token } from 'typescript-parsec';
+import { TokenType } from "./tokenize";
 /**
  * 
  * # REPRESENTATION
@@ -21,7 +23,7 @@ export function tkTreeToSExp(t: tkTree): string{
         if (t=== undefined){
             str = "%undefined"
         }else{
-            str = t.value;
+            str = t;
         }
     }
 
@@ -35,156 +37,229 @@ let repr = (x : any)=>{return util.inspect(x, {depth: null})};
  * # TYPES
  */
 
+
+type tkTree = string | tkTree[];
+
+enum TokenKind {
+    Seperator, // ---
+    Semicolon, // ;
+    Number,
+    Op,
+    ExprMark, // @
+    ExcapeAt, // \@
+    Paren,
+    SpaceNL, // \s\t\n\r
+    Id,
+    Str,
+    Comment, // /* ooo */
+}
+
 /**
- * TokenPair for tokens' parser combinator
- *
- * matched: the matched (now and before) tokens
- * 
- * remained: tokens to be matched
- * 
- * ast: abstract syntax tree
+ * Parsing
  */
-export interface TokenPair {
-    matched: jsTokens.Token[]
-    remained: jsTokens.Token[]
-    ast : tkTree[]
-}
-export type Some<T> = { _tag: "Some"; value: T };
-export type None = { _tag: "None" };
-export type Maybe<T> = Some<T> | None;
+const lexer = p.buildLexer([
+    [true, /^\d+(\.\d+)?/g, TokenKind.Number],
+    [true, /^\\\@/g, TokenKind.ExcapeAt],
+    [true, /^\/\*([^/]|\/[^*])*\*\//g, TokenKind.Comment],
+    [true, /^\;/g, TokenKind.Semicolon],
+    [true, /^[-][-][-]/g, TokenKind.Seperator],
+    [true, /^[\+\-\*\/\&\|\!\^\<\>\~\=\?]+/g, TokenKind.Op],
+    [true, /^\@/g, TokenKind.ExprMark],
+    [true, /^[()\[\]{}]/g, TokenKind.Paren],
+    [true, /^["]([\"]|[\\].)*["]/g, TokenKind.Str],
+    [true, /^[']([\']|[\\].)*[']/g, TokenKind.Str],
+    [true, /^[()\[\]{}]/g, TokenKind.Paren],
+    [true, /^[^\/\\\@\s\n\t\r;]+/g, TokenKind.Id],
+    [true, /^(\s|\n|\r|\t)+/g, TokenKind.SpaceNL],
 
-type Token = jsTokens.Token;
-type tkTree = Token | tkTree[];
+]);
 
 /**
  * 
- * # PARSER UNITS
+ * # TEST
  */
-function toSome<T>(x:T): Maybe<T>{
-    return {_tag: "Some", value: x};
+const inputTxt=
+`import a as b; /*bacourt*/
+/* ba choir 
+ipsum lorem*/
+
+import you as john;
+---
+
+臺中市\\\@
+
+公園
+@1+2==3;
+
+console.log("122");@
+
+山頂
+`;
+
+
+const PROG = p.rule<TokenKind, tkTree>();
+const SEGMENT = p.rule<TokenKind, tkTree>();
+const IMPORT = p.rule<TokenKind, tkTree>();
+const IMPORTS = p.rule<TokenKind, tkTree>();
+const SEMICOLON = p.rule<TokenKind, tkTree>();
+const EXCAPE_AT = p.rule<TokenKind, tkTree>();
+const NOT_AT_TEXT = p.rule<TokenKind, tkTree>();
+const CONTENT = p.rule<TokenKind, tkTree>();
+
+let doubleMinus = { type: 'Punctuator', value: '--' };
+let doubleMinus2 = p.str('--');
+const TERM = p.rule<TokenKind, tkTree>();
+
+function applySegment(input: [Token<TokenKind>, Token<TokenKind>[],
+        Token<TokenKind>]): tkTree[]{
+    let unpackedInnerExprs = input[1].map((x)=>{return x.text});
+    return ["%exprs", unpackedInnerExprs];
 }
 
-/**
- * like `m ==> f` in ocaml
- * @param m matchee wrapped
- * @param f matching function
- * @returns wrapped result
- */
-function thenDo(m : Maybe<TokenPair>, f : Function){
-    if (m._tag == "None"){
-        return m;
-    }else{
-        var a : Maybe<TokenPair> = f(m.value);
-        if (a._tag == "Some"){
-            a.value.ast = m.value.ast.concat(a.value.ast);
-        }
+function applySemiColon(value: Token<TokenKind.Semicolon>): tkTree{
+    return value.text;
+}
+
+function applyParts(first: tkTree,
+                    second: [Token<TokenKind>, tkTree]):tkTree {
+    return ["%clo", first , second[1]];
+}
+
+
+function applyComment(value: Token<TokenKind.Comment>): tkTree[]{
+    return [value.text];
+}
+
+
+function applyImport(input: [Token<TokenKind>,Token<TokenKind>[], tkTree]) : tkTree{
+    let importTail = input[1].map(x=>x.text);
+    return ["import"].concat(importTail);
+};
+
+
+/*
+function applyImportComment(input: [Token<TokenKind>,Token<TokenKind>[],
+    tkTree, Token<TokenKind.Comment>]) : tkTree{
+    let importTail = input[1].map(x=>x.text);
+    let comment = [input[3].text];
+    return ["import"].concat(importTail).concat(comment);
+};*/
+
+function applyImports(input : [tkTree, tkTree[]]): tkTree{
+    let resultBody = [input[0]].concat(input[1]);
+    let resultWrapper = ["%import", resultBody];
+    return resultWrapper;
+};
 
-        return a;
+
+
+
+function applyNotAtText(value : Token<TokenKind>): tkTree{
+    if (value.text == "\\\@"){
+        return '@';
     }
+    else{return value.text;}
+};
+
+function applyText (input : tkTree): tkTree[]{
+    return ["%text", input];
+};
+
+function applyContent(input : tkTree[]): tkTree[]{
+    return ["%content", input];
+};
+
+function applySpaceNL(value : Token<TokenKind.SpaceNL>): tkTree{
+    return value.text;
 }
+
 /**
- * 
- * @param m : the `TokenPair` to be consumed.
- * @returns if the length of `m.remained` >= 1; consumes the matchee by 1 token
- *  and wraps it in `Some`,
- * otherwise, returns `None`.
+ * IMPORTEE:  Number, Op, Paren, Id, Str, Comment,
  */
-export function matchAny(m: TokenPair): Maybe<TokenPair> {
-    if (m.remained.length >= 1) {
-        return {
-            _tag: "Some", value: {
-                matched: m.matched.concat(m.remained[0]),
-                remained: m.remained.slice(1),
-                ast :  [m.remained[0]],
-            }
-        };
-    } else {
-        return { _tag: "None" };
-    }
-}
+let IMPORTEE = p.alt(p.tok(TokenKind.Number),
+    p.tok(TokenKind.Op),
+    p.tok(TokenKind.Paren),
+    p.tok(TokenKind.Id),
+    p.tok(TokenKind.Str),
+    p.tok(TokenKind.SpaceNL),
+    p.tok(TokenKind.Comment));
+
+let NOT_AT = p.alt(p.tok(TokenKind.Seperator),
+    p.tok(TokenKind.Semicolon),
+    p.tok(TokenKind.Number),
+    p.tok(TokenKind.ExcapeAt),
+    p.tok(TokenKind.Op),
+    p.tok(TokenKind.Paren),
+    p.tok(TokenKind.SpaceNL),
+    p.tok(TokenKind.Id),
+    p.tok(TokenKind.Str),
+    p.tok(TokenKind.Comment),
+    );
 
 /**
- * like `f1 | f2` in regex
- * @param f1 the first tried function
- * @param f2 the second tried function
- * @returns wrapped result
+ * PROG : IMPORTS '---' CONTENT;
  */
-function orDo(f1 : Function, f2 : Function){
-    return (x : TokenPair) =>{
-        let res1 : Maybe<TokenPair> = f1(x);
-        if (res1._tag == "Some"){
-            return res1;
-        }else{
-            let res2 : Maybe<TokenPair> = f2(x);
-            return res2;
-        }
-    } 
-}
+PROG.setPattern(
+    p.lrec_sc(IMPORTS, p.seq(p.str('---'), CONTENT), applyParts)
+
+)
 
 /**
- * like regex [^c]
- * @param f input token function. one token only.
- * @returns combined finction
+ * NOT_AT_TEXT : NOT_AT
  */
-function notDo(f : Function){
-    return (x : TokenPair) =>{
-        let res1 : Maybe<TokenPair> = f(x);
-        if (res1._tag == "Some"){
-            return {_tag:"None"};
-        }else{
-            let res2 = matchAny(x);
-            return res2;
-        }
-    } 
-}
+NOT_AT_TEXT.setPattern(
+    p.apply(NOT_AT, applyNotAtText)
+);
 
-function matchToken(typeName : string, value? : string):
- (t : TokenPair) => Maybe<TokenPair>{
-    return (t)=>{
-        let headToken = t.remained[0];
-        if (headToken.type != typeName){
-            return {_tag:"None"};
-        }else{
-            if (value === undefined || value == headToken.value){
-                let newTokenPair = {
-                    matched: t.matched.concat(headToken),
-                    remained: t.remained.slice(1),
-                    ast : [headToken]
-                };
-                return {_tag : "Some", value : newTokenPair};
-            }else{
-                return {_tag:"None"};
-            }
-        };
-    }
-};
+IMPORTS.setPattern(
+    p.apply( p.seq(IMPORT, p.rep(IMPORT)), applyImports)
+);
+
+/**
+ * IMPORT :
+ * 'import' IMPORTEE* SEMICOLON |
+ * COMMENT |
+ */
+IMPORT.setPattern(
+    p.alt(
+        p.apply(p.seq(p.str('import'), p.rep_sc(IMPORTEE), SEMICOLON),
+            applyImport),
+        p.apply(p.tok(TokenKind.Comment), applyComment),
+        p.apply(p.tok(TokenKind.SpaceNL), applySpaceNL)
 
+    )
+);
 
 /**
- * 
- * # TEST
+ * SEMICOLON : ';';
  */
-const tokens = Array.from(jsTokens(
-`import foo from\t 'bar';
-import * as util  from 'util';
+SEMICOLON.setPattern(
+    p.apply(p.tok(TokenKind.Semicolon), applySemiColon)
+);
 
 
-花非花，霧\\{非霧 。{{foo();}}下
-一句`));
 
-console.log("RESULT="+repr(tokens));
+/**
+ * SEGMENT : '@' NOT_AT* '@' |
+ * (NOT_AT_TEXT | EXCAPE_AT)*
+ */
+SEGMENT.setPattern(
+    p.alt(
+        p.apply(p.rep_sc(NOT_AT_TEXT), applyText),
+        p.apply(p.seq(p.str('@'), p.rep(NOT_AT), p.str('@')), applySegment),
+    )
+);
+
+/**
+ * CONTENT : SEGMENT*
+ */
+CONTENT.setPattern(
+    p.apply(p.rep(SEGMENT), applyContent)
+);
 
 
-var mainTokenPair : TokenPair = {
-    matched : [] ,
-    remained : tokens,
-    ast : []};
+let tree = p.expectSingleResult(p.expectEOF(PROG.parse(lexer.parse(inputTxt))));
 
-let a = thenDo(thenDo(toSome(mainTokenPair), matchToken('IdentifierName')),
-        notDo(matchToken('Punctuator', ';')));
 
 
-console.log("RESULT="+repr(a));
-if (a._tag == "Some"){
-    console.log("SEXP="+tkTreeToSExp(a.value.ast));
-}
+console.log("RESULT="+tkTreeToSExp(tree));