Add support for single-quoted strings and fix stringifying whitespace

Fixes hjson#18

src/HJSON/HJSONParser.php

@@ -15,6 +15,7 @@ public function __construct()
     {
         $this->escapee = [
             '"'  => '"',
+            '\'' => '\'',
             "\\" => "\\",
             '/'  => '/',
             'b'  => chr(8),
@@ -99,23 +100,27 @@ private function value()
             case '[':
                 return $this->_array();
             case '"':
-                return $this->string();
+                return $this->string('"');
+            case '\'':
+                if ($this->peek(0) !== '\'' || $this->peek(1) !== '\'') {
+                    return $this->string('\'');
+                }
             default:
                 return $this->tfnns();
         }
     }
 
-    private function string()
+    private function string($quote)
     {
         // Parse a string value.
         $hex;
         $string = '';
         $uffff;
 
         // When parsing for string values, we must look for " and \ characters.
-        if ($this->ch === '"') {
+        if ($this->ch === $quote) {
             while ($this->next() !== null) {
-                if ($this->ch === '"') {
+                if ($this->ch === $quote) {
                     $this->next();
                     return $string;
                 }
@@ -410,7 +415,9 @@ private function keyname()
         // unless they include {}[],: or whitespace.
 
         if ($this->ch === '"') {
-            return $this->string();
+            return $this->string('"');
+        } else if ($this->ch === '\'') {
+            return $this->string('\'');
         }
 
         $name = "";

src/HJSON/HJSONStringifier.php

@@ -2,9 +2,13 @@
 
 namespace HJSON;
 
+/**
+ * NOTE: this may return an empty string at the end of the array when the input
+ * string ends with a newline character
+ */
 function mb_str_split($string)
 {
-    return preg_split('/(?<!^)(?!$)/u', $string);
+    return preg_split('/(?<!^)/u', $string);
 }
 
 class HJSONStringifier
@@ -13,11 +17,12 @@ class HJSONStringifier
     // needsEscape tests if the string can be written without escapes
     private $needsEscape = '/[\\\"\x00-\x1f\x7f-\x9f\x{00ad}\x{0600}-\x{0604}\x{070f}\x{17b4}\x{17b5}\x{200c}-\x{200f}\x{2028}-\x{202f}\x{2060}-\x{206f}\x{feff}\x{fff0}-\x{ffff}\x]/u';
     // needsQuotes tests if the string can be written as a quoteless string (includes needsEscape but without \\ and \")
-    private $needsQuotes = '/^\\s|^"|^\'\'\'|^#|^\\/\\*|^\\/\\/|^\\{|^\\}|^\\[|^\\]|^:|^,|\\s$|[\x00-\x1f\x7f-\x9f\x{00ad}\x{0600}-\x{0604}\x{070f}\x{17b4}\x{17b5}\x{200c}-\x{200f}\x{2028}-\x{202f}\x{2060}-\x{206f}\x{feff}\x{fff0}-\x{ffff}\x]/u';
+    private $needsQuotes = '/^\\s|^"|^\'|^\'\'\'|^#|^\\/\\*|^\\/\\/|^\\{|^\\}|^\\[|^\\]|^:|^,|\\s$|[\x00-\x1f\x7f-\x9f\x{00ad}\x{0600}-\x{0604}\x{070f}\x{17b4}\x{17b5}\x{200c}-\x{200f}\x{2028}-\x{202f}\x{2060}-\x{206f}\x{feff}\x{fff0}-\x{ffff}\x]/u';
     // needsEscapeML tests if the string can be written as a multiline string (includes needsEscape but without \n, \r, \\ and \")
-    private $needsEscapeML = '/\'\'\'|[\x00-\x09\x0b\x0c\x0e-\x1f\x7f-\x9f\x{00ad}\x{0600}-\x{0604}\x{070f}\x{17b4}\x{17b5}\x{200c}-\x{200f}\x{2028}-\x{202f}\x{2060}-\x{206f}\x{feff}\x{fff0}-\x{ffff}\x]/u';
+    private $needsEscapeML = '/\'\'\'|[\x00-\x08\x0b\x0c\x0e-\x1f\x7f-\x9f\x{00ad}\x{0600}-\x{0604}\x{070f}\x{17b4}\x{17b5}\x{200c}-\x{200f}\x{2028}-\x{202f}\x{2060}-\x{206f}\x{feff}\x{fff0}-\x{ffff}\x]/u';
+    private $onlyWhitespace = '/^\\s+$/';
     private $startsWithKeyword = '/^(true|false|null)\s*((,|\]|\}|#|\/\/|\/\*).*)?$/';
-    private $needsEscapeName = '/[,\{\[\}\]\s:#"]|\/\/|\/\*|\'\'\'/';
+    private $needsEscapeName = '/[,\{\[\}\]\s:#"\']|\/\/|\/\*|\'\'\'/';
     private $gap = '';
     private $indent = '  ';
 
@@ -38,6 +43,7 @@ public function __construct()
             "\n" => "\\n",
             "\r" => "\\r",
             '"'  => '\\"',
+            '\''  => '\\\'',
             '\\' => "\\\\"
         ];
         $this->meta[chr(8)] = '\\b';
@@ -136,7 +142,11 @@ private function quote($string = null, $gap = null, $hasComment = null, $isRootO
 
             if (!preg_match($this->needsEscape, $string)) {
                 return '"' . $string . '"';
-            } elseif (!preg_match($this->needsEscapeML, $string) && !$isRootObject) {
+            } elseif (
+                !preg_match($this->needsEscapeML, $string) &&
+                !preg_match($this->onlyWhitespace, $string) &&
+                !$isRootObject
+            ) {
                 return $this->mlString($string, $gap);
             } else {
                 return '"' . $this->quoteReplace($string) . '"';

tests/assets/failKey5_test.hjson

@@ -0,0 +1,4 @@
+{
+  # invalid name
+  '''foo''': 0
+}

tests/assets/failStr8a_test.hjson

@@ -0,0 +1,4 @@
+{
+  # invalid ml-string
+  foo : ""'text'''
+}

tests/assets/keys_result.hjson

@@ -18,6 +18,9 @@
   "foo\"bar": test
   "'''": test
   "foo'''bar": test
+  "'": test
+  "'foo": test
+  "foo'bar": test
   ":": test
   "foo:bar": test
   "{": test

tests/assets/keys_result.json

@@ -18,6 +18,9 @@
   "foo\"bar": "test",
   "'''": "test",
   "foo'''bar": "test",
+  "'": "test",
+  "'foo": "test",
+  "foo'bar": "test",
   ":": "test",
   "foo:bar": "test",
   "{": "test",

tests/assets/keys_test.hjson

@@ -22,6 +22,9 @@
   "foo\"bar": test
   "'''": test
   "foo'''bar": test
+  "'": test
+  "'foo": test
+  "foo'bar": test
   # control char in key name
   ":": test
   "foo:bar": test

tests/assets/strings2_result.hjson

@@ -0,0 +1,29 @@
+{
+  key1: a key in single quotes
+  "key 2": a key in single quotes
+  "key \"": a key in single quotes
+  text:
+  [
+    single quoted string
+    '''You need quotes	for escapes'''
+    " untrimmed "
+    "untrimmed "
+    containing " double quotes
+    containing " double quotes
+    containing " double quotes
+    '''"containing more " double quotes"'''
+    containing ' single quotes
+    containing ' single quotes
+    containing ' single quotes
+    "'containing more ' single quotes'"
+    "'containing more ' single quotes'"
+    "\n"
+    "  \n"
+    "\n  \n  \n  \n"
+    "\t\n"
+  ]
+  foo3a: asdf'''
+  foo3b: "'''asdf"
+  foo4a: "asdf'''\nasdf"
+  foo4b: "asdf\n'''asdf"
+}

tests/assets/strings2_result.json

@@ -0,0 +1,28 @@
+{
+  "key1": "a key in single quotes",
+  "key 2": "a key in single quotes",
+  "key \"": "a key in single quotes",
+  "text": [
+    "single quoted string",
+    "You need quotes\tfor escapes",
+    " untrimmed ",
+    "untrimmed ",
+    "containing \" double quotes",
+    "containing \" double quotes",
+    "containing \" double quotes",
+    "\"containing more \" double quotes\"",
+    "containing ' single quotes",
+    "containing ' single quotes",
+    "containing ' single quotes",
+    "'containing more ' single quotes'",
+    "'containing more ' single quotes'",
+    "\n",
+    "  \n",
+    "\n  \n  \n  \n",
+    "\t\n"
+  ],
+  "foo3a": "asdf'''",
+  "foo3b": "'''asdf",
+  "foo4a": "asdf'''\nasdf",
+  "foo4b": "asdf\n'''asdf"
+}

tests/assets/strings2_test.hjson

@@ -0,0 +1,36 @@
+{
+  # Hjson 3 allows the use of single quotes
+
+  'key1': a key in single quotes
+  'key 2': a key in single quotes
+  'key "': a key in single quotes
+
+  text: [
+    'single quoted string'
+    'You need quotes\tfor escapes'
+    ' untrimmed '
+    'untrimmed '
+    'containing " double quotes'
+    'containing \" double quotes'
+    "containing \" double quotes"
+    '"containing more " double quotes"'
+    'containing \' single quotes'
+    "containing ' single quotes"
+    "containing \' single quotes"
+    "'containing more ' single quotes'"
+    "\'containing more \' single quotes\'"
+
+    '\n'
+    '  \n'
+    '\n  \n  \n  \n'
+    '\t\n'
+  ]
+
+  # escapes/no escape
+
+  foo3a: 'asdf\'\'\''
+  foo3b: '\'\'\'asdf'
+
+  foo4a: 'asdf\'\'\'\nasdf'
+  foo4b: 'asdf\n\'\'\'asdf'
+}

tests/assets/strings_result.hjson

@@ -1,10 +1,14 @@
 {
   text1: This is a valid string value.
   text2: a \ is just a \
-  text3: "You need quotes\tfor escapes"
+  text3: '''You need quotes	for escapes'''
   text4a: " untrimmed "
   text4b: " untrimmed"
   text4c: "untrimmed "
+  notml1: "\n"
+  notml2: "  \n"
+  notml3: "\n  \n  \n  \n"
+  notml4: "\t\n"
   multiline1:
     '''
     first line

tests/assets/strings_result.json

@@ -5,6 +5,10 @@
   "text4a": " untrimmed ",
   "text4b": " untrimmed",
   "text4c": "untrimmed ",
+  "notml1": "\n",
+  "notml2": "  \n",
+  "notml3": "\n  \n  \n  \n",
+  "notml4": "\t\n",
   "multiline1": "first line\n  indented line\nlast line",
   "multiline2": "first line\n  indented line\nlast line",
   "multiline3": "first line\n  indented line\nlast line\n",

tests/assets/strings_test.hjson

@@ -10,6 +10,11 @@
   text4b: " untrimmed"
   text4c: "untrimmed "
 
+  notml1: "\n"
+  notml2: "  \n"
+  notml3: "\n  \n  \n  \n"
+  notml4: "\t\n"
+
   # multiline string
 
   multiline1:

tests/assets/testlist.txt

@@ -20,7 +20,6 @@ failJSON20_test.json
 failJSON21_test.json
 failJSON22_test.json
 failJSON23_test.json
-failJSON24_test.json
 failJSON26_test.json
 failJSON28_test.json
 failJSON29_test.json
@@ -33,6 +32,7 @@ failKey1_test.hjson
 failKey2_test.hjson
 failKey3_test.hjson
 failKey4_test.hjson
+failKey5_test.hjson
 failMLStr1_test.hjson
 failObj1_test.hjson
 failObj2_test.hjson
@@ -61,8 +61,10 @@ failStr6a_test.hjson
 failStr6b_test.hjson
 failStr6c_test.hjson
 failStr6d_test.hjson
+failStr8a_test.hjson
 kan_test.hjson
 keys_test.hjson
+mltabs_test.json
 oa_test.hjson
 pass1_test.json
 pass2_test.json
@@ -71,5 +73,6 @@ pass4_test.json
 passSingle_test.hjson
 root_test.hjson
 stringify1_test.hjson
+strings2_test.hjson
 strings_test.hjson
 trail_test.hjson