psych 3.1.0 → 4.0.0

data/ext/psych/yaml/scanner.c

@@ -38,8 +38,8 @@
  *      BLOCK-END                       # Indentation decrease.
  *      FLOW-SEQUENCE-START             # '['
  *      FLOW-SEQUENCE-END               # ']'
- *      BLOCK-SEQUENCE-START            # '{'
- *      BLOCK-SEQUENCE-END              # '}'
+ *      FLOW-MAPPING-START              # '{'
+ *      FLOW-MAPPING-END                # '}'
  *      BLOCK-ENTRY                     # '-'
  *      FLOW-ENTRY                      # ','
  *      KEY                             # '?' or nothing (simple keys).
@@ -273,7 +273,7 @@
  * The tokens BLOCK-SEQUENCE-START and BLOCK-MAPPING-START denote indentation
  * increase that precedes a block collection (cf. the INDENT token in Python).
  * The token BLOCK-END denote indentation decrease that ends a block collection
- * (cf. the DEDENT token in Python).  However YAML has some syntax pecularities
+ * (cf. the DEDENT token in Python).  However YAML has some syntax peculiarities
  * that makes detections of these tokens more complex.
  *
  * The tokens BLOCK-ENTRY, KEY, and VALUE are used to represent the indicators
@@ -348,6 +348,7 @@
  *          SCALAR("another value",plain)
  *          KEY
  *          SCALAR("a mapping",plain)
+ *          VALUE
  *          BLOCK-MAPPING-START
  *          KEY
  *          SCALAR("key 1",plain)
@@ -711,7 +712,7 @@ yaml_parser_scan_tag_handle(yaml_parser_t *parser, int directive,
         yaml_mark_t start_mark, yaml_char_t **handle);
 
 static int
-yaml_parser_scan_tag_uri(yaml_parser_t *parser, int directive,
+yaml_parser_scan_tag_uri(yaml_parser_t *parser, int uri_char, int directive,
         yaml_char_t *head, yaml_mark_t start_mark, yaml_char_t **uri);
 
 static int
@@ -1227,7 +1228,7 @@ yaml_parser_roll_indent(yaml_parser_t *parser, ptrdiff_t column,
             return 0;
         }
 
-        parser->indent = column;
+        parser->indent = (int)column;
 
         /* Create a token and insert it into the queue. */
 
@@ -2292,7 +2293,7 @@ yaml_parser_scan_tag_directive_value(yaml_parser_t *parser,
 
     /* Scan a prefix. */
 
-    if (!yaml_parser_scan_tag_uri(parser, 1, NULL, start_mark, &prefix_value))
+    if (!yaml_parser_scan_tag_uri(parser, 1, 1, NULL, start_mark, &prefix_value))
         goto error;
 
     /* Expect a whitespace or line break. */
@@ -2410,7 +2411,7 @@ yaml_parser_scan_tag(yaml_parser_t *parser, yaml_token_t *token)
 
         /* Consume the tag value. */
 
-        if (!yaml_parser_scan_tag_uri(parser, 0, NULL, start_mark, &suffix))
+        if (!yaml_parser_scan_tag_uri(parser, 1, 0, NULL, start_mark, &suffix))
             goto error;
 
         /* Check for '>' and eat it. */
@@ -2438,14 +2439,14 @@ yaml_parser_scan_tag(yaml_parser_t *parser, yaml_token_t *token)
         {
             /* Scan the suffix now. */
 
-            if (!yaml_parser_scan_tag_uri(parser, 0, NULL, start_mark, &suffix))
+            if (!yaml_parser_scan_tag_uri(parser, 0, 0, NULL, start_mark, &suffix))
                 goto error;
         }
         else
         {
             /* It wasn't a handle after all.  Scan the rest of the tag. */
 
-            if (!yaml_parser_scan_tag_uri(parser, 0, handle, start_mark, &suffix))
+            if (!yaml_parser_scan_tag_uri(parser, 0, 0, handle, start_mark, &suffix))
                 goto error;
 
             /* Set the handle to '!'. */
@@ -2474,9 +2475,11 @@ yaml_parser_scan_tag(yaml_parser_t *parser, yaml_token_t *token)
     if (!CACHE(parser, 1)) goto error;
 
     if (!IS_BLANKZ(parser->buffer)) {
-        yaml_parser_set_scanner_error(parser, "while scanning a tag",
-                start_mark, "did not find expected whitespace or line break");
-        goto error;
+        if (!parser->flow_level || !CHECK(parser->buffer, ',') ) {
+            yaml_parser_set_scanner_error(parser, "while scanning a tag",
+                    start_mark, "did not find expected whitespace or line break");
+            goto error;
+        }
     }
 
     end_mark = parser->mark;
@@ -2565,7 +2568,7 @@ error:
  */
 
 static int
-yaml_parser_scan_tag_uri(yaml_parser_t *parser, int directive,
+yaml_parser_scan_tag_uri(yaml_parser_t *parser, int uri_char, int directive,
         yaml_char_t *head, yaml_mark_t start_mark, yaml_char_t **uri)
 {
     size_t length = head ? strlen((char *)head) : 0;
@@ -2601,8 +2604,11 @@ yaml_parser_scan_tag_uri(yaml_parser_t *parser, int directive,
      * The set of characters that may appear in URI is as follows:
      *
      *      '0'-'9', 'A'-'Z', 'a'-'z', '_', '-', ';', '/', '?', ':', '@', '&',
-     *      '=', '+', '$', ',', '.', '!', '~', '*', '\'', '(', ')', '[', ']',
-     *      '%'.
+     *      '=', '+', '$', '.', '!', '~', '*', '\'', '(', ')', '%'.
+     *
+     * If we are inside a verbatim tag <...> (parameter uri_char is true)
+     * then also the following flow indicators are allowed:
+     *      ',', '[', ']'
      */
 
     while (IS_ALPHA(parser->buffer) || CHECK(parser->buffer, ';')
@@ -2610,12 +2616,15 @@ yaml_parser_scan_tag_uri(yaml_parser_t *parser, int directive,
             || CHECK(parser->buffer, ':') || CHECK(parser->buffer, '@')
             || CHECK(parser->buffer, '&') || CHECK(parser->buffer, '=')
             || CHECK(parser->buffer, '+') || CHECK(parser->buffer, '$')
-            || CHECK(parser->buffer, ',') || CHECK(parser->buffer, '.')
+            || CHECK(parser->buffer, '.') || CHECK(parser->buffer, '%')
             || CHECK(parser->buffer, '!') || CHECK(parser->buffer, '~')
             || CHECK(parser->buffer, '*') || CHECK(parser->buffer, '\'')
             || CHECK(parser->buffer, '(') || CHECK(parser->buffer, ')')
-            || CHECK(parser->buffer, '[') || CHECK(parser->buffer, ']')
-            || CHECK(parser->buffer, '%'))
+            || (uri_char && (
+                CHECK(parser->buffer, ',')
+                || CHECK(parser->buffer, '[') || CHECK(parser->buffer, ']')
+                )
+            ))
     {
         /* Check if it is a URI-escape sequence. */
 
@@ -2860,7 +2869,7 @@ yaml_parser_scan_block_scalar(yaml_parser_t *parser, yaml_token_t *token,
 
     if (!CACHE(parser, 1)) goto error;
 
-    while ((int)parser->mark.column == indent && !IS_Z(parser->buffer))
+    while ((int)parser->mark.column == indent && !(IS_Z(parser->buffer)))
     {
         /*
          * We are at the beginning of a non-empty line.
@@ -3278,7 +3287,7 @@ yaml_parser_scan_flow_scalar(yaml_parser_t *parser, yaml_token_t *token,
 
         /* Check if we are at the end of the scalar. */
 
-        /* Fix for crash unitialized value crash
+        /* Fix for crash uninitialized value crash
          * Credit for the bug and input is to OSS Fuzz
          * Credit for the fix to Alex Gaynor
          */
@@ -3430,11 +3439,22 @@ yaml_parser_scan_plain_scalar(yaml_parser_t *parser, yaml_token_t *token)
 
         while (!IS_BLANKZ(parser->buffer))
         {
-            /* Check for 'x:x' in the flow context. TODO: Fix the test "spec-08-13". */
+            /* Check for "x:" + one of ',?[]{}' in the flow context. TODO: Fix the test "spec-08-13".
+             * This is not completely according to the spec
+             * See http://yaml.org/spec/1.1/#id907281 9.1.3. Plain
+             */
 
             if (parser->flow_level
                     && CHECK(parser->buffer, ':')
-                    && !IS_BLANKZ_AT(parser->buffer, 1)) {
+                    && (
+                        CHECK_AT(parser->buffer, ',', 1)
+                        || CHECK_AT(parser->buffer, '?', 1)
+                        || CHECK_AT(parser->buffer, '[', 1)
+                        || CHECK_AT(parser->buffer, ']', 1)
+                        || CHECK_AT(parser->buffer, '{', 1)
+                        || CHECK_AT(parser->buffer, '}', 1)
+                    )
+                    ) {
                 yaml_parser_set_scanner_error(parser, "while scanning a plain scalar",
                         start_mark, "found unexpected ':'");
                 goto error;
@@ -3444,8 +3464,8 @@ yaml_parser_scan_plain_scalar(yaml_parser_t *parser, yaml_token_t *token)
 
             if ((CHECK(parser->buffer, ':') && IS_BLANKZ_AT(parser->buffer, 1))
                     || (parser->flow_level &&
-                        (CHECK(parser->buffer, ',') || CHECK(parser->buffer, ':')
-                         || CHECK(parser->buffer, '?') || CHECK(parser->buffer, '[')
+                        (CHECK(parser->buffer, ',')
+                         || CHECK(parser->buffer, '[')
                          || CHECK(parser->buffer, ']') || CHECK(parser->buffer, '{')
                          || CHECK(parser->buffer, '}'))))
                 break;
@@ -3512,7 +3532,7 @@ yaml_parser_scan_plain_scalar(yaml_parser_t *parser, yaml_token_t *token)
                 if (leading_blanks && (int)parser->mark.column < indent
                         && IS_TAB(parser->buffer)) {
                     yaml_parser_set_scanner_error(parser, "while scanning a plain scalar",
-                            start_mark, "found a tab character that violate indentation");
+                            start_mark, "found a tab character that violates indentation");
                     goto error;
                 }
 

data/ext/psych/yaml/yaml.h

@@ -26,7 +26,9 @@ extern "C" {
 
 /** The public API declaration. */
 
-#ifdef _WIN32
+#if defined(__MINGW32__)
+#   define  YAML_DECLARE(type)  type
+#elif defined(_WIN32)
 #   if defined(YAML_DECLARE_STATIC)
 #       define  YAML_DECLARE(type)  type
 #   elif defined(YAML_DECLARE_EXPORT)
@@ -230,7 +232,7 @@ typedef enum yaml_token_type_e {
 
     /** A BLOCK-SEQUENCE-START token. */
     YAML_BLOCK_SEQUENCE_START_TOKEN,
-    /** A BLOCK-SEQUENCE-END token. */
+    /** A BLOCK-MAPPING-START token. */
     YAML_BLOCK_MAPPING_START_TOKEN,
     /** A BLOCK-END token. */
     YAML_BLOCK_END_TOKEN,
@@ -550,7 +552,7 @@ yaml_document_end_event_initialize(yaml_event_t *event, int implicit);
  */
 
 YAML_DECLARE(int)
-yaml_alias_event_initialize(yaml_event_t *event, yaml_char_t *anchor);
+yaml_alias_event_initialize(yaml_event_t *event, const yaml_char_t *anchor);
 
 /**
  * Create a SCALAR event.
@@ -576,8 +578,8 @@ yaml_alias_event_initialize(yaml_event_t *event, yaml_char_t *anchor);
 
 YAML_DECLARE(int)
 yaml_scalar_event_initialize(yaml_event_t *event,
-        yaml_char_t *anchor, yaml_char_t *tag,
-        yaml_char_t *value, int length,
+        const yaml_char_t *anchor, const yaml_char_t *tag,
+        const yaml_char_t *value, int length,
         int plain_implicit, int quoted_implicit,
         yaml_scalar_style_t style);
 
@@ -599,7 +601,7 @@ yaml_scalar_event_initialize(yaml_event_t *event,
 
 YAML_DECLARE(int)
 yaml_sequence_start_event_initialize(yaml_event_t *event,
-        yaml_char_t *anchor, yaml_char_t *tag, int implicit,
+        const yaml_char_t *anchor, const yaml_char_t *tag, int implicit,
         yaml_sequence_style_t style);
 
 /**
@@ -631,7 +633,7 @@ yaml_sequence_end_event_initialize(yaml_event_t *event);
 
 YAML_DECLARE(int)
 yaml_mapping_start_event_initialize(yaml_event_t *event,
-        yaml_char_t *anchor, yaml_char_t *tag, int implicit,
+        const yaml_char_t *anchor, const yaml_char_t *tag, int implicit,
         yaml_mapping_style_t style);
 
 /**
@@ -663,7 +665,7 @@ yaml_event_delete(yaml_event_t *event);
 
 /** The tag @c !!null with the only possible value: @c null. */
 #define YAML_NULL_TAG       "tag:yaml.org,2002:null"
-/** The tag @c !!bool with the values: @c true and @c falce. */
+/** The tag @c !!bool with the values: @c true and @c false. */
 #define YAML_BOOL_TAG       "tag:yaml.org,2002:bool"
 /** The tag @c !!str for string values. */
 #define YAML_STR_TAG        "tag:yaml.org,2002:str"
@@ -894,7 +896,7 @@ yaml_document_get_root_node(yaml_document_t *document);
 
 YAML_DECLARE(int)
 yaml_document_add_scalar(yaml_document_t *document,
-        yaml_char_t *tag, yaml_char_t *value, int length,
+        const yaml_char_t *tag, const yaml_char_t *value, int length,
         yaml_scalar_style_t style);
 
 /**
@@ -911,7 +913,7 @@ yaml_document_add_scalar(yaml_document_t *document,
 
 YAML_DECLARE(int)
 yaml_document_add_sequence(yaml_document_t *document,
-        yaml_char_t *tag, yaml_sequence_style_t style);
+        const yaml_char_t *tag, yaml_sequence_style_t style);
 
 /**
  * Create a MAPPING node and attach it to the document.
@@ -927,7 +929,7 @@ yaml_document_add_sequence(yaml_document_t *document,
 
 YAML_DECLARE(int)
 yaml_document_add_mapping(yaml_document_t *document,
-        yaml_char_t *tag, yaml_mapping_style_t style);
+        const yaml_char_t *tag, yaml_mapping_style_t style);
 
 /**
  * Add an item to a SEQUENCE node.
@@ -935,7 +937,7 @@ yaml_document_add_mapping(yaml_document_t *document,
  * @param[in,out]   document    A document object.
  * @param[in]       sequence    The sequence node id.
  * @param[in]       item        The item node id.
-*
+ *
  * @returns @c 1 if the function succeeded, @c 0 on error.
  */
 
@@ -950,7 +952,7 @@ yaml_document_append_sequence_item(yaml_document_t *document,
  * @param[in]       mapping     The mapping node id.
  * @param[in]       key         The key node id.
  * @param[in]       value       The value node id.
-*
+ *
  * @returns @c 1 if the function succeeded, @c 0 on error.
  */
 
@@ -1018,6 +1020,7 @@ typedef enum yaml_parser_state_e {
     YAML_PARSE_DOCUMENT_CONTENT_STATE,
     /** Expect DOCUMENT-END. */
     YAML_PARSE_DOCUMENT_END_STATE,
+
     /** Expect a block node. */
     YAML_PARSE_BLOCK_NODE_STATE,
     /** Expect a block node or indentless sequence. */
@@ -1028,6 +1031,7 @@ typedef enum yaml_parser_state_e {
     YAML_PARSE_BLOCK_SEQUENCE_FIRST_ENTRY_STATE,
     /** Expect an entry of a block sequence. */
     YAML_PARSE_BLOCK_SEQUENCE_ENTRY_STATE,
+
     /** Expect an entry of an indentless sequence. */
     YAML_PARSE_INDENTLESS_SEQUENCE_ENTRY_STATE,
     /** Expect the first key of a block mapping. */
@@ -1038,6 +1042,7 @@ typedef enum yaml_parser_state_e {
     YAML_PARSE_BLOCK_MAPPING_VALUE_STATE,
     /** Expect the first entry of a flow sequence. */
     YAML_PARSE_FLOW_SEQUENCE_FIRST_ENTRY_STATE,
+
     /** Expect an entry of a flow sequence. */
     YAML_PARSE_FLOW_SEQUENCE_ENTRY_STATE,
     /** Expect a key of an ordered mapping. */
@@ -1049,6 +1054,7 @@ typedef enum yaml_parser_state_e {
     /** Expect the first key of a flow mapping. */
     YAML_PARSE_FLOW_MAPPING_FIRST_KEY_STATE,
     /** Expect a key of a flow mapping. */
+
     YAML_PARSE_FLOW_MAPPING_KEY_STATE,
     /** Expect a value of a flow mapping. */
     YAML_PARSE_FLOW_MAPPING_VALUE_STATE,
@@ -1203,7 +1209,7 @@ typedef struct yaml_parser_s {
     /** The number of tokens fetched from the queue. */
     size_t tokens_parsed;
 
-    /* Does the tokens queue contain a token ready for dequeueing. */
+    /** Does the tokens queue contain a token ready for dequeueing. */
     int token_available;
 
     /** The indentation levels stack. */
@@ -1329,7 +1335,7 @@ yaml_parser_delete(yaml_parser_t *parser);
  * Set a string input.
  *
  * Note that the @a input pointer must be valid while the @a parser object
- * exists.  The application is responsible for destroing @a input after
+ * exists.  The application is responsible for destroying @a input after
  * destroying the @a parser.
  *
  * @param[in,out]   parser  A parser object.
@@ -1444,7 +1450,7 @@ yaml_parser_parse(yaml_parser_t *parser, yaml_event_t *event);
  * @param[in,out]   parser      A parser object.
  * @param[out]      document    An empty document object.
  *
- * @return @c 1 if the function succeeded, @c 0 on error.
+ * @returns @c 1 if the function succeeded, @c 0 on error.
  */
 
 YAML_DECLARE(int)
@@ -1487,6 +1493,7 @@ typedef enum yaml_emitter_state_e {
     YAML_EMIT_DOCUMENT_CONTENT_STATE,
     /** Expect DOCUMENT-END. */
     YAML_EMIT_DOCUMENT_END_STATE,
+
     /** Expect the first item of a flow sequence. */
     YAML_EMIT_FLOW_SEQUENCE_FIRST_ITEM_STATE,
     /** Expect an item of a flow sequence. */
@@ -1497,6 +1504,7 @@ typedef enum yaml_emitter_state_e {
     YAML_EMIT_FLOW_MAPPING_KEY_STATE,
     /** Expect a value for a simple key of a flow mapping. */
     YAML_EMIT_FLOW_MAPPING_SIMPLE_VALUE_STATE,
+
     /** Expect a value of a flow mapping. */
     YAML_EMIT_FLOW_MAPPING_VALUE_STATE,
     /** Expect the first item of a block sequence. */
@@ -1507,6 +1515,7 @@ typedef enum yaml_emitter_state_e {
     YAML_EMIT_BLOCK_MAPPING_FIRST_KEY_STATE,
     /** Expect the key of a block mapping. */
     YAML_EMIT_BLOCK_MAPPING_KEY_STATE,
+
     /** Expect a value for a simple key of a block mapping. */
     YAML_EMIT_BLOCK_MAPPING_SIMPLE_VALUE_STATE,
     /** Expect a value of a block mapping. */
@@ -1515,6 +1524,18 @@ typedef enum yaml_emitter_state_e {
     YAML_EMIT_END_STATE
 } yaml_emitter_state_t;
 
+
+/* This is needed for C++ */
+
+typedef struct yaml_anchors_s {
+    /** The number of references. */
+    int references;
+    /** The anchor id. */
+    int anchor;
+    /** If the node has been emitted? */
+    int serialized;
+} yaml_anchors_t;
+
 /**
  * The emitter structure.
  *
@@ -1546,7 +1567,7 @@ typedef struct yaml_emitter_s {
     /** Write handler. */
     yaml_write_handler_t *write_handler;
 
-    /** A pointer for passing to the white handler. */
+    /** A pointer for passing to the write handler. */
     void *write_handler_data;
 
     /** Standard (string or file) output data. */
@@ -1713,7 +1734,7 @@ typedef struct yaml_emitter_s {
         size_t length;
         /** Does the scalar contain line breaks? */
         int multiline;
-        /** Can the scalar be expessed in the flow plain style? */
+        /** Can the scalar be expressed in the flow plain style? */
         int flow_plain_allowed;
         /** Can the scalar be expressed in the block plain style? */
         int block_plain_allowed;
@@ -1740,14 +1761,7 @@ typedef struct yaml_emitter_s {
     int closed;
 
     /** The information associated with the document nodes. */
-    struct {
-        /** The number of references. */
-        int references;
-        /** The anchor id. */
-        int anchor;
-        /** If the node has been emitted? */
-        int serialized;
-    } *anchors;
+    yaml_anchors_t *anchors;
 
     /** The last assigned anchor id. */
     int last_anchor_id;
@@ -1936,10 +1950,10 @@ yaml_emitter_close(yaml_emitter_t *emitter);
 /**
  * Emit a YAML document.
  *
- * The documen object may be generated using the yaml_parser_load() function
+ * The document object may be generated using the yaml_parser_load() function
  * or the yaml_document_initialize() function.  The emitter takes the
- * responsibility for the document object and destoys its content after
- * it is emitted. The document object is destroyedeven if the function fails.
+ * responsibility for the document object and destroys its content after
+ * it is emitted. The document object is destroyed even if the function fails.
  *
  * @param[in,out]   emitter     An emitter object.
  * @param[in,out]   document    A document object.

data/ext/psych/yaml/yaml_private.h

@@ -2,8 +2,8 @@
 #include RUBY_EXTCONF_H
 #endif
 
-#if HAVE_CONFIG_H
-#include <config.h>
+#ifdef HAVE_CONFIG_H
+#include "config.h"
 #endif
 
 #include <yaml.h>
@@ -175,14 +175,14 @@ yaml_string_join(
  * Check the octet at the specified position.
  */
 
-#define CHECK_AT(string,octet,offset)                                           \
+#define CHECK_AT(string,octet,offset)                   \
     ((string).pointer[offset] == (yaml_char_t)(octet))
 
 /*
  * Check the current octet in the buffer.
  */
 
-#define CHECK(string,octet) CHECK_AT((string),(octet),0)
+#define CHECK(string,octet) (CHECK_AT((string),(octet),0))
 
 /*
  * Check if the character at the specified position is an alphabetical

data/lib/psych/class_loader.rb

@@ -35,9 +35,11 @@ module Psych
 
     constants.each do |const|
       konst = const_get const
-      define_method(const.to_s.downcase) do
-        load konst
-      end
+      class_eval <<~RUBY
+        def #{const.to_s.downcase}
+          load #{konst.inspect}
+        end
+      RUBY
     end
 
     private
@@ -69,7 +71,7 @@ module Psych
       rescue
         nil
       end
-    }.compact]
+    }.compact].freeze
 
     class Restricted < ClassLoader
       def initialize classes, symbols

data/lib/psych/handler.rb

@@ -119,7 +119,7 @@ module Psych
     # +tag+ is an associated tag or nil
     # +plain+ is a boolean value
     # +quoted+ is a boolean value
-    # +style+ is an integer idicating the string style
+    # +style+ is an integer indicating the string style
     #
     # See the constants in Psych::Nodes::Scalar for the possible values of
     # +style+

data/lib/psych/nodes/node.rb

@@ -46,8 +46,8 @@ module Psych
       # Convert this node to Ruby.
       #
       # See also Psych::Visitors::ToRuby
-      def to_ruby
-        Visitors::ToRuby.create.accept(self)
+      def to_ruby(symbolize_names: false, freeze: false)
+        Visitors::ToRuby.create(symbolize_names: symbolize_names, freeze: freeze).accept(self)
       end
       alias :transform :to_ruby
 

data/lib/psych/nodes/scalar.rb

@@ -50,7 +50,7 @@ module Psych
       # +tag+ is an associated tag or nil
       # +plain+ is a boolean value
       # +quoted+ is a boolean value
-      # +style+ is an integer idicating the string style
+      # +style+ is an integer indicating the string style
       #
       # == See Also
       #

data/lib/psych/scalar_scanner.rb

@@ -14,16 +14,15 @@ module Psych
               |\.(nan|NaN|NAN)(?# not a number))$/x
 
     # Taken from http://yaml.org/type/int.html
-    INTEGER = /^(?:[-+]?0b[0-1_]+          (?# base 2)
-                  |[-+]?0[0-7_]+           (?# base 8)
-                  |[-+]?(?:0|[1-9][0-9_]*) (?# base 10)
-                  |[-+]?0x[0-9a-fA-F_]+    (?# base 16))$/x
+    INTEGER = /^(?:[-+]?0b[0-1_,]+          (?# base 2)
+                  |[-+]?0[0-7_,]+           (?# base 8)
+                  |[-+]?(?:0|[1-9][0-9_,]*) (?# base 10)
+                  |[-+]?0x[0-9a-fA-F_,]+    (?# base 16))$/x
 
     attr_reader :class_loader
 
     # Create a new scanner
     def initialize class_loader
-      @string_cache = {}
       @symbol_cache = {}
       @class_loader = class_loader
     end
@@ -31,81 +30,70 @@ module Psych
     # Tokenize +string+ returning the Ruby object
     def tokenize string
       return nil if string.empty?
-      return string if @string_cache.key?(string)
       return @symbol_cache[string] if @symbol_cache.key?(string)
 
-      case string
       # Check for a String type, being careful not to get caught by hash keys, hex values, and
       # special floats (e.g., -.inf).
-      when /^[^\d\.:-]?[A-Za-z_\s!@#\$%\^&\*\(\)\{\}\<\>\|\/\\~;=]+/, /\n/
-        if string.length > 5
-          @string_cache[string] = true
-          return string
-        end
+      if string.match?(/^[^\d\.:-]?[A-Za-z_\s!@#\$%\^&\*\(\)\{\}\<\>\|\/\\~;=]+/) || string.match?(/\n/)
+        return string if string.length > 5
 
-        case string
-        when /^[^ytonf~]/i
-          @string_cache[string] = true
+        if string.match?(/^[^ytonf~]/i)
           string
-        when '~', /^null$/i
+        elsif string == '~' || string.match?(/^null$/i)
           nil
-        when /^(yes|true|on)$/i
+        elsif string.match?(/^(yes|true|on)$/i)
           true
-        when /^(no|false|off)$/i
+        elsif string.match?(/^(no|false|off)$/i)
           false
         else
-          @string_cache[string] = true
           string
         end
-      when TIME
+      elsif string.match?(TIME)
         begin
           parse_time string
         rescue ArgumentError
           string
         end
-      when /^\d{4}-(?:1[012]|0\d|\d)-(?:[12]\d|3[01]|0\d|\d)$/
+      elsif string.match?(/^\d{4}-(?:1[012]|0\d|\d)-(?:[12]\d|3[01]|0\d|\d)$/)
         require 'date'
         begin
           class_loader.date.strptime(string, '%Y-%m-%d')
         rescue ArgumentError
           string
         end
-      when /^\.inf$/i
+      elsif string.match?(/^\.inf$/i)
         Float::INFINITY
-      when /^-\.inf$/i
+      elsif string.match?(/^-\.inf$/i)
         -Float::INFINITY
-      when /^\.nan$/i
+      elsif string.match?(/^\.nan$/i)
         Float::NAN
-      when /^:./
+      elsif string.match?(/^:./)
         if string =~ /^:(["'])(.*)\1/
           @symbol_cache[string] = class_loader.symbolize($2.sub(/^:/, ''))
         else
           @symbol_cache[string] = class_loader.symbolize(string.sub(/^:/, ''))
         end
-      when /^[-+]?[0-9][0-9_]*(:[0-5]?[0-9]){1,2}$/
+      elsif string.match?(/^[-+]?[0-9][0-9_]*(:[0-5]?[0-9]){1,2}$/)
         i = 0
         string.split(':').each_with_index do |n,e|
           i += (n.to_i * 60 ** (e - 2).abs)
         end
         i
-      when /^[-+]?[0-9][0-9_]*(:[0-5]?[0-9]){1,2}\.[0-9_]*$/
+      elsif string.match?(/^[-+]?[0-9][0-9_]*(:[0-5]?[0-9]){1,2}\.[0-9_]*$/)
         i = 0
         string.split(':').each_with_index do |n,e|
           i += (n.to_f * 60 ** (e - 2).abs)
         end
         i
-      when FLOAT
-        if string =~ /\A[-+]?\.\Z/
-          @string_cache[string] = true
+      elsif string.match?(FLOAT)
+        if string.match?(/\A[-+]?\.\Z/)
           string
         else
           Float(string.gsub(/[,_]|\.([Ee]|$)/, '\1'))
         end
+      elsif string.match?(INTEGER)
+        parse_int string
       else
-        int = parse_int string.gsub(/[,_]/, '')
-        return int if int
-
-        @string_cache[string] = true
         string
       end
     end
@@ -113,8 +101,7 @@ module Psych
     ###
     # Parse and return an int from +string+
     def parse_int string
-      return unless INTEGER === string
-      Integer(string)
+      Integer(string.gsub(/[,_]/, ''))
     end
 
     ###