simplified pseq hacks

satyr · Jan 18, 2011 · 1ad4648 · 1ad4648
1 parent 89d6d63
commit 1ad4648
Show file tree

Hide file tree

Showing 9 changed files with 155 additions and 216 deletions.
diff --git a/extras/coco.js b/extras/coco.js
diff --git a/lib/grammar.js b/lib/grammar.js
@@ -10,8 +10,10 @@ bnf = {
   Value: [
     o('Assignable', function(){
       return Value($1);
-    }), o('Primitive', function(){
-      return Value($1);
+    }), o('STRNUM', function(){
+      return Value(L(Literal($1)));
+    }), o('LITERAL', function(){
+      return Value(L(Literal($1, true)));
     }), o('Parenthetical', function(){
       return Value($1);
     }), o('THIS', function(){
@@ -20,13 +22,6 @@ bnf = {
       return Value(L(Call($1, $3, $2)));
     })
   ],
-  Primitive: [
-    o('STRNUM', function(){
-      return L(Literal($1));
-    }), o('LITERAL', function(){
-      return L(Literal($1, true));
-    })
-  ],
   SimpleAssignable: [
     o('IDENTIFIER', function(){
       return L(Literal($1));
@@ -166,17 +161,6 @@ bnf = {
       return $1.concat($4);
     }), o('ArgList OptComma INDENT ArgList OptComma DEDENT', function(){
       return $1.concat($4);
-    }), o('ArgSeq'), o('ArgList , ArgSeq', function(){
-      return $1.concat($3);
-    }), o('ArgList OptComma TERMINATOR ArgSeq', function(){
-      return $1.concat($4);
-    })
-  ],
-  ArgSeq: [
-    o('Primitive Primitive', function(){
-      return [$1, $2];
-    }), o('ArgSeq    Primitive', function(){
-      return $1.concat($2);
     })
   ],
   ThisProperty: [
@@ -243,17 +227,6 @@ bnf = {
       return $1.concat($4);
     }), o('Properties OptComma INDENT Properties OptComma DEDENT', function(){
       return $1.concat($4);
-    }), o('KeySeq'), o('Properties , KeySeq', function(){
-      return $1.concat($3);
-    }), o('Properties OptComma TERMINATOR KeySeq', function(){
-      return $1.concat($4);
-    })
-  ],
-  KeySeq: [
-    o('STRNUM STRNUM', function(){
-      return [L(Literal($1)), L(Literal($2))];
-    }), o('KeySeq STRNUM', function(){
-      return $1.concat(L(Literal($2)));
     })
   ],
   Object: [

diff --git a/lib/lexer.js b/lib/lexer.js
@@ -360,17 +360,18 @@ exports.heregexToken = function(it){
   return heregex.length;
 };
 exports.wordsToken = function(it){
-  var end, line, re, that, word, _i, _ref, _len, _j, _len2;
+  var end, tokens, line, re, that, word, _i, _ref, _len, _j, _len2, _ref2;
   if (!~(end = it.indexOf(']>', 2))) {
     this.carp('unterminated words');
   }
   this.token('[', '[');
+  tokens = this.tokens;
   for (_i = 0, _len = (_ref = it.slice(2, end).split('\n')).length; _i < _len; ++_i) {
     line = _ref[_i];
     if (that = line.match(re || (re = /\S+/g))) {
       for (_j = 0, _len2 = that.length; _j < _len2; ++_j) {
         word = that[_j];
-        this.tokens.push(['STRNUM', string(word, '\''), this.line]);
+        tokens.push((_ref2 = ['STRNUM', string(word, '\''), this.line], _ref2.spaced = true, _ref2));
       }
     }
     ++this.line;

diff --git a/lib/parser.js b/lib/parser.js
diff --git a/lib/rewriter.js b/lib/rewriter.js
@@ -225,7 +225,7 @@ function addImplicitParentheses(tokens){
       token.pair[0] = ')CALL';
       continue;
     }
-    if (!(__indexOf.call(ARG, tag) >= 0 || (tag === 'STRNUM' || tag === 'LITERAL') || tag === '+-' && !(token.spaced || token.eol))) {
+    if (!(__indexOf.call(ARG, tag) >= 0 || tag === '+-' && !(token.spaced || token.eol))) {
       continue;
     }
     skipBlock = seenSwitch = false;
@@ -341,7 +341,7 @@ function rewriteClosingParens(tokens){
   }
 }
 function expandNumbers(tokens){
-  var i, token, num, sig, ts, lno, to, n, _ref, _step;
+  var i, token, num, sig, ts, lno, to, n, _step, _ref;
   i = -1;
   while (token = tokens[++i]) {
     switch (token[0]) {
@@ -350,24 +350,31 @@ function expandNumbers(tokens){
         token[1] = num.slice(1);
         tokens.splice(i++, 0, ['+-', sig, token[2]]);
       }
-      /* fallthrough */
+      break;
     case 'LITERAL':
-      if (_ref = tokens[i + 1][0], __indexOf.call(ARG, _ref) >= 0) {
-        tokens.splice(++i, 0, [',', ',', token[2]]);
-      }
       break;
     case 'RANGE':
       ts = [];
       lno = token[2];
       to = token.to - (token.op === 'to' ? 0 : 1e-15);
       for (n = +token[1], _step = +token.by || 1; _step < 0 ? n >= to : n <= to; n += _step) {
-        if (1024 < ts.push(['STRNUM', n, lno])) {
+        if (2048 < ts.push(['STRNUM', n, lno], [',', ',', lno])) {
           carp('range limit exceeded', lno);
         }
       }
-      ts.length || carp('empty range', lno);
+      if (ts.length) {
+        ts.pop();
+      } else {
+        carp('empty range', lno);
+      }
       tokens.splice.apply(tokens, [i, 1].concat(__slice.call(ts)));
       i += ts.length - 1;
+      break;
+    default:
+      continue;
+    }
+    if (token.spaced && (_ref = tokens[i + 1][0], __indexOf.call(ARG, _ref) >= 0)) {
+      tokens.splice(++i, 0, [',', ',', token[2]]);
     }
   }
 }
@@ -428,4 +435,4 @@ for (_i = 0, _len = BALANCED_PAIRS.length; _i < _len; ++_i) {
   EXPR_START.push(INVERSES[rite] = left);
   EXPR_END.push(INVERSES[left] = rite);
 }
-ARG = ['(', '[', '{', '...', 'IDENTIFIER', 'THISPROP', 'THIS', 'RANGE', 'SUPER', 'PARAM(', 'FUNC_ARROW', 'FUNCTION', 'UNARY', 'CREMENT', 'IF', 'TRY', 'CLASS', 'SWITCH'];
+ARG = ['(', '[', '{', '...', 'IDENTIFIER', 'THISPROP', 'THIS', 'STRNUM', 'LITERAL', 'RANGE', 'SUPER', 'PARAM(', 'FUNC_ARROW', 'FUNCTION', 'UNARY', 'CREMENT', 'IF', 'TRY', 'CLASS', 'SWITCH'];
diff --git a/src/grammar.co b/src/grammar.co
@@ -58,18 +58,14 @@ bnf =
   # The types of things that can be accessed or indexed into.
   Value: [
     o \Assignable    ,-> Value $1
-    o \Primitive     ,-> Value $1
+    o \STRNUM        ,-> Value L Literal $1
+    o \LITERAL       ,-> Value L Literal $1, true
     o \Parenthetical ,-> Value $1
     o \THIS          ,-> Value L Literal \this, true
 
     o 'Value CALL( ArgList OptComma )CALL' ,-> Value L Call $1, $3, $2
   ]
 
-  Primitive: [
-    o \STRNUM  ,-> L Literal $1
-    o \LITERAL ,-> L Literal $1, true
-  ]
-
   # Variables and properties that can be assigned to.
   SimpleAssignable: [
     o \IDENTIFIER ,-> L Literal $1
@@ -198,15 +194,6 @@ bnf =
     o 'ArgList , Arg'                                   ,-> $1.concat $3
     o 'ArgList OptComma TERMINATOR Arg'                 ,-> $1.concat $4
     o 'ArgList OptComma INDENT ArgList OptComma DEDENT' ,-> $1.concat $4
-
-    o \ArgSeq
-    o 'ArgList , ArgSeq'                   ,-> $1.concat $3
-    o 'ArgList OptComma TERMINATOR ArgSeq' ,-> $1.concat $4
-  ]
-  # Sequence of primitive values without explicit separators.
-  ArgSeq: [
-    o 'Primitive Primitive' ,-> [$1, $2]
-    o 'ArgSeq    Primitive' ,-> $1.concat $2
   ]
 
   # A reference to a property on `this`.
@@ -264,14 +251,6 @@ bnf =
     o 'Properties , Property'                                 ,-> $1.concat $3
     o 'Properties OptComma TERMINATOR Property'               ,-> $1.concat $4
     o 'Properties OptComma INDENT Properties OptComma DEDENT' ,-> $1.concat $4
-
-    o \KeySeq
-    o 'Properties , KeySeq'                   ,-> $1.concat $3
-    o 'Properties OptComma TERMINATOR KeySeq' ,-> $1.concat $4
-  ]
-  KeySeq: [
-    o 'STRNUM STRNUM' ,-> [L Literal $1; L Literal $2]
-    o 'KeySeq STRNUM' ,-> $1.concat L Literal $2
   ]
   Object: [
     o '{ Properties OptComma }' ,-> L Obj $2

diff --git a/src/lexer.co b/src/lexer.co
@@ -284,9 +284,10 @@ exports import
   wordsToken: ->
     @carp 'unterminated words' unless ~end = it.indexOf \]> 2
     @token \[ \[
+    {tokens} = this
     for line of it.slice(2, end).split \\n
-      if line.match re ||= /\S+/g
-        @tokens.push [\STRNUM; string word, \'; @line] for word of that
+      if line.match re ||= /\S+/g then for word of that
+        tokens.push [\STRNUM; string word, \'; @line] <<< {+spaced}
       ++@line
     --@line
     @token \STRNUM "''" unless word

diff --git a/src/rewriter.co b/src/rewriter.co
@@ -80,7 +80,7 @@ function addImplicitIndentation (tokens) ->
     continue if next is \THEN
     continue unless tag of <[ THEN FUNC_ARROW DEFAULT TRY CATCH FINALLY ]> or
                     tag is \ELSE and next is not \IF
-    indent = [\INDENT 0, token.2]; dedent = [\DEDENT 0]
+    indent = [\INDENT 0 token.2]; dedent = [\DEDENT 0]
     if tag is \THEN
       tokens.splice --i, 1 if tokens[i-1]?.0 is \TERMINATOR
       tokens[i] = indent
@@ -137,7 +137,7 @@ function addImplicitBraces (tokens) ->
     stack.push [\{]
     idx  = if paren then start.1 else i-1
     idx -= 2 while tokens[idx-2]?.0 is \COMMENT
-    tokens.splice idx, 0, [\{ \{, token.2]
+    tokens.splice idx, 0, [\{ \{ token.2]
     detectEnd tokens, ++i+1, ok, go
   void
 
@@ -159,11 +159,10 @@ function addImplicitParentheses (tokens) ->
       token.     0 = \CALL(
       token.pair.0 = \)CALL
       continue
-    continue unless tag of ARG or tag of <[ STRNUM LITERAL ]> or
-                    tag is \+- and not (token.spaced or token.eol)
+    continue unless tag of ARG or tag is \+- and !(token.spaced or token.eol)
     skipBlock = seenSwitch = false
     tokens.splice --i, 1 if soak = prev.0 is \?
-    tokens.splice i++, 0, [\CALL(, if soak then \?( else \(, token.2]
+    tokens.splice i++, 0, [\CALL( if soak then \?( else \(, token.2]
     detectEnd tokens, i, ok, go
   function ok (token, i) ->
     return true if token.alias and token.1 of <[ && || ]>
@@ -183,7 +182,7 @@ function addImplicitParentheses (tokens) ->
     false
   function go (token, i) ->
     ++i if token.0 is \DEDENT
-    tokens.splice i, 0, [\)CALL \), token.2]
+    tokens.splice i, 0, [\)CALL \) token.2]
   void
 
 # Ensure that all listed pairs of tokens are correctly balanced throughout
@@ -248,19 +247,20 @@ function expandNumbers (tokens) ->
     case \STRNUM
       if ~'-+'.indexOf sig = (num = '' + token.1).charAt 0
         token.1 = num.slice 1
-        tokens.splice i++, 0, [\+-, sig, token.2]
-      fallthrough
-    case \LITERAL
-      tokens.splice ++i, 0, [\, \,; token.2] if tokens[i+1].0 of ARG
+        tokens.splice i++, 0, [\+- sig, token.2]
+    case \LITERAL then break
     case \RANGE
       ts = []; lno = token.2
       to = token.to - if token.op is \to then 0 else 1e-15
       for n from +token.1 to to by +token.by or 1
-        if 32r100 < ts.push [\STRNUM, n, lno]
+        if 32r200 < ts.push [\STRNUM n, lno], [\, \, lno]
           carp 'range limit exceeded', lno
-      ts.length or carp 'empty range', lno
+      if ts.length then ts.pop() else carp 'empty range', lno
       tokens.splice i, 1, ...ts
       i += ts.length - 1
+    default continue
+    if token.spaced and tokens[i+1].0 of ARG
+      tokens.splice ++i, 0, [\, \, token.2]
   void
 
 ### Helpers
@@ -317,6 +317,6 @@ for [left, rite] of BALANCED_PAIRS
   EXPR_START.push INVERSES[rite] = left
   EXPR_END  .push INVERSES[left] = rite
 
-# Tokens that can start an argument list, sans-primitives.
-ARG = <[ ( [ { ... IDENTIFIER THISPROP THIS RANGE SUPER
+# Tokens that can start an argument list.
+ARG = <[ ( [ { ... IDENTIFIER THISPROP THIS STRNUM LITERAL RANGE SUPER
          PARAM( FUNC_ARROW FUNCTION UNARY CREMENT IF TRY CLASS SWITCH ]>
diff --git a/test/literals.co b/test/literals.co
@@ -2,7 +2,7 @@ a = [((x) -> x), ((x) -> x * x)]
 eq a.length, 2
 
 
-eq (3 -4), -1
+eq 3-4, -1
 
 
 # Decimal number literals.