json_mend 0.3.1 → 0.3.2

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: cef80671ecf0296d0b0742bb94342c93a9cdf52e5def2ee4592a11f473ed6d6b
-  data.tar.gz: 831e6b821e39dff4adcf605e4f0beb53500ca846ac02a96e8c49e9822c7b1475
+  metadata.gz: 923008e3c63e24de16c3ee6b26097cf4064a32a16bac749c5501e313996238f1
+  data.tar.gz: c6c6040f9d54fe7604ae7126402886af159aae5e001dca93cbceebf254839a55
 SHA512:
-  metadata.gz: a43f2089605d476414103b4333fe7d248856f27d0a8589041d7826905c611619a18fe45841dcf8e97bdfae90cfb6718d39323d0839c8fe97dbb584a54db4712f
-  data.tar.gz: 65c72349a8a6b721ef0f1a077bb69f82fe36fb499aa30da57f36fbc0dddfd4a7129082ffc32f940569af876fa8f43d4f626507652f5480b06fc7717b648cf418
+  metadata.gz: c869c17b06f5ed0e46e3f74ccf59c8c374af1431b43ee7d9d59be2170e5ba88e10f83753db36af1e4ebba2cf519273b5f5c32c42fa386978f6a96c22095b063d
+  data.tar.gz: 770eb1238b3f73261a2130b2a900dfff0bd2770cf0f6b3b38ecee30cafdbfb09e59b87e4df4dd8aba722c5dbbd14bb1b35df509fdc92efbe081d8342ed7e435b

data/lib/json_mend/parser.rb

@@ -70,7 +70,7 @@ module JsonMend
             break
           else
             # Ignore strings that look like closing braces garbage (e.g. "}", " ] ")
-            next if new_json.is_a?(String) && new_json.strip.match?(/^[}\]]+$/)
+            next if new_json.is_a?(String) && new_json.match?(/\A\s*[}\]]+\s*\z/)
 
             if both_hash?(json.last, new_json)
               json[-1] = deep_merge_hashes(json.last, new_json)
@@ -265,20 +265,14 @@ module JsonMend
 
       if value == :inferred_true
         if %w[true false null].include?(key.downcase)
-          # Look back: If it's concatenated to the previous value (like falsetrue), keep it.
-          # If it's separated by space/delimiters, it's trailing garbage, so drop it.
-          if pos_before_key.positive?
-            prev_byte = @scanner.string.getbyte(pos_before_key - 1)
-            # Check ASCII byte ranges for a-z, A-Z, 0-9, $, -, and _
-            is_concatenated = prev_byte && (
-              prev_byte.between?(48, 57)  || # 0-9
-              prev_byte.between?(65, 90)  || # A-Z
-              prev_byte.between?(97, 122) || # a-z
-              [36, 45, 95].include?(prev_byte) # $, -, _
-            )
-          else
-            is_concatenated = false
-          end
+          prev_byte = @scanner.string.getbyte(pos_before_key - 1)
+          # Check ASCII byte ranges for a-z, A-Z, 0-9, $, -, and _
+          is_concatenated = prev_byte && (
+            prev_byte.between?(48, 57)  || # 0-9
+            prev_byte.between?(65, 90)  || # A-Z
+            prev_byte.between?(97, 122) || # a-z
+            [36, 45, 95].include?(prev_byte) # $, -, _
+          )
 
           return [nil, nil, false] unless is_concatenated
         end
@@ -335,7 +329,7 @@ module JsonMend
       @context.pop
 
       # If parse_json returned JSON_STOP_TOKEN (nothing found due to garbage->terminator),
-      # treat it as nil (null) for object values to be safe.
+      # treat it as empty string for object values to be safe.
       value == JSON_STOP_TOKEN ? '' : value
     end
 
@@ -426,10 +420,10 @@ module JsonMend
     # many common errors found in LLM-generated JSON, such as missing quotes,
     # incorrect escape sequences, and ambiguous string terminators
     def parse_string
-      char = prepare_string_parsing
+      char = peek_char
 
       # A valid string can only start with a valid quote or, in our case, with a literal
-      while !@scanner.eos? && !STRING_DELIMITERS.include?(char) && !char.match?(/[\p{L}0-9$_-]/)
+      while !@scanner.eos? && !STRING_DELIMITERS.include?(char) && !char&.match?(/[\p{L}0-9$_-]/)
         return '' if TERMINATORS_STRING_GUESSED.include?(char)
 
         @scanner.getch
@@ -485,18 +479,6 @@ module JsonMend
 
     # string helper methods
 
-    def prepare_string_parsing
-      char = peek_char
-
-      # Consume comments that appear before the string starts
-      while COMMENT_DELIMETERS.include?(char)
-        parse_comment
-        char = peek_char
-      end
-
-      char
-    end
-
     def determine_delimiters(char:)
       missing_quotes = false
       lstring_delimiter = rstring_delimiter = '"'
@@ -1251,10 +1233,12 @@ module JsonMend
     # returns the index (offset) from the scanner
     def skip_to_character(characters, start_idx: 0)
       pattern = SKIP_CHARS_REGEX_CACHE.fetch(characters, nil)
+      # :nocov:
       if pattern.nil?
         chars = Array(characters).map { |c| Regexp.escape(c.to_s) }
         pattern = Regexp.new(chars.join('|'))
       end
+      # :nocov:
 
       saved_pos = @scanner.pos
       # Skip start_idx
@@ -1315,7 +1299,7 @@ module JsonMend
               (matched.length - 1) + start_idx
             else
               # No non-space found.
-              @scanner.rest.length + start_idx
+              (@scanner.string.length - @scanner.charpos) + start_idx
             end
 
       @scanner.pos = saved_pos
@@ -1346,12 +1330,12 @@ module JsonMend
       # Handle the common 0-offset case
       if offset.zero?
         # peek(1) returns the next BYTE, not character
-        byte_str = @scanner.peek(1)
-        return nil if byte_str.empty?
+        byte = @scanner.string.getbyte(@scanner.pos)
+        return nil unless byte
 
         # Fast path: If it's a standard ASCII char (0-127), return it directly.
-        # This avoids the regex overhead for standard JSON characters ({, [, ", etc).
-        return byte_str if byte_str.getbyte(0) < 128
+        # Enforcing UTF-8 ensures we don't mix US-ASCII and UTF-8 strings later.
+        return byte.chr(Encoding::UTF_8) if byte < 128
 
         # Slow path: If it's a multibyte char (e.g. “), use regex to match the full character.
         return @scanner.check(/./m)

data/lib/json_mend/version.rb

@@ -1,5 +1,5 @@
 # frozen_string_literal: true
 
 module JsonMend
-  VERSION = '0.3.1'
+  VERSION = '0.3.2'
 end

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: json_mend
 version: !ruby/object:Gem::Version
-  version: 0.3.1
+  version: 0.3.2
 platform: ruby
 authors:
 - Oleksii Vasyliev