parser.c: include line and column in error messages

Author: byroot

CHANGES.md

@@ -2,6 +2,7 @@
 
 ### Unreleased
 
+* Include line and column in parser errors. Both in the message and as exception attributes.
 * Handle non-string hash keys with broken `to_s` implementations.
 * `JSON.generate` now uses SSE2 (x86) or NEON (arm64) instructions when available to escape strings.
 

ext/json/ext/parser/parser.c

@@ -395,6 +395,23 @@ static void raise_parse_error(const char *format, JSON_ParserState *state)
 {
     unsigned char buffer[PARSE_ERROR_FRAGMENT_LEN + 1];
 
+    const char *cursor = state->cursor;
+    long column = 0;
+    long line = 1;
+
+    while (cursor >= state->start) {
+        if (*cursor-- == '\n') {
+            break;
+        }
+        column++;
+    }
+
+    while (cursor >= state->start) {
+        if (*cursor-- == '\n') {
+            line++;
+        }
+    }
+
     const char *ptr = state->cursor;
     size_t len = ptr ? strnlen(ptr, PARSE_ERROR_FRAGMENT_LEN) : 0;
 
@@ -413,7 +430,14 @@ static void raise_parse_error(const char *format, JSON_ParserState *state)
         ptr = (const char *)buffer;
     }
 
-    rb_enc_raise(enc_utf8, rb_path2class("JSON::ParserError"), format, ptr);
+    VALUE msg = rb_sprintf(format, ptr);
+    VALUE message = rb_enc_sprintf(enc_utf8, "%s at line %ld column %ld", RSTRING_PTR(msg), line, column);
+    RB_GC_GUARD(msg);
+
+    VALUE exc = rb_exc_new_str(rb_path2class("JSON::ParserError"), message);
+    rb_ivar_set(exc, rb_intern("@line"), LONG2NUM(line));
+    rb_ivar_set(exc, rb_intern("@column"), LONG2NUM(column));
+    rb_exc_raise(exc);
 }
 
 #ifdef RBIMPL_ATTR_NORETURN
@@ -508,11 +532,11 @@ json_eat_comments(JSON_ParserState *state)
                 break;
             }
             default:
-                raise_parse_error("unexpected token at '%s'", state);
+                raise_parse_error("unexpected token '%s'", state);
                 break;
         }
     } else {
-        raise_parse_error("unexpected token at '%s'", state);
+        raise_parse_error("unexpected token '%s'", state);
     }
 }
 
@@ -870,15 +894,15 @@ static VALUE json_parse_any(JSON_ParserState *state, JSON_ParserConfig *config)
                 return json_push_value(state, config, Qnil);
             }
 
-            raise_parse_error("unexpected token at '%s'", state);
+            raise_parse_error("unexpected token '%s'", state);
             break;
         case 't':
             if ((state->end - state->cursor >= 4) && (memcmp(state->cursor, "true", 4) == 0)) {
                 state->cursor += 4;
                 return json_push_value(state, config, Qtrue);
             }
 
-            raise_parse_error("unexpected token at '%s'", state);
+            raise_parse_error("unexpected token '%s'", state);
             break;
         case 'f':
             // Note: memcmp with a small power of two compile to an integer comparison
@@ -887,7 +911,7 @@ static VALUE json_parse_any(JSON_ParserState *state, JSON_ParserConfig *config)
                 return json_push_value(state, config, Qfalse);
             }
 
-            raise_parse_error("unexpected token at '%s'", state);
+            raise_parse_error("unexpected token '%s'", state);
             break;
         case 'N':
             // Note: memcmp with a small power of two compile to an integer comparison
@@ -896,15 +920,15 @@ static VALUE json_parse_any(JSON_ParserState *state, JSON_ParserConfig *config)
                 return json_push_value(state, config, CNaN);
             }
 
-            raise_parse_error("unexpected token at '%s'", state);
+            raise_parse_error("unexpected token '%s'", state);
             break;
         case 'I':
             if (config->allow_nan && (state->end - state->cursor >= 8) && (memcmp(state->cursor, "Infinity", 8) == 0)) {
                 state->cursor += 8;
                 return json_push_value(state, config, CInfinity);
             }
 
-            raise_parse_error("unexpected token at '%s'", state);
+            raise_parse_error("unexpected token '%s'", state);
             break;
         case '-':
             // Note: memcmp with a small power of two compile to an integer comparison
@@ -913,7 +937,7 @@ static VALUE json_parse_any(JSON_ParserState *state, JSON_ParserConfig *config)
                     state->cursor += 9;
                     return json_push_value(state, config, CMinusInfinity);
                 } else {
-                    raise_parse_error("unexpected token at '%s'", state);
+                    raise_parse_error("unexpected token '%s'", state);
                 }
             }
             // Fallthrough

lib/json/common.rb

@@ -230,7 +230,9 @@ def self.create_id
   class JSONError < StandardError; end
 
   # This exception is raised if a parser error occurs.
-  class ParserError < JSONError; end
+  class ParserError < JSONError
+    attr_reader :line, :column
+  end
 
   # This exception is raised if the nesting of parsed data structures is too
   # deep.

test/json/json_ext_parser_test.rb

@@ -15,15 +15,19 @@ def test_allocate
 
   def test_error_messages
     ex = assert_raise(ParserError) { parse('Infinity') }
-    assert_equal "unexpected token at 'Infinity'", ex.message
+    unless RUBY_PLATFORM =~ /java/
+      assert_equal "unexpected token 'Infinity' at line 1 column 1", ex.message
+    end
 
+    ex = assert_raise(ParserError) { parse('-Infinity') }
     unless RUBY_PLATFORM =~ /java/
-      ex = assert_raise(ParserError) { parse('-Infinity') }
-      assert_equal "unexpected token at '-Infinity'", ex.message
+      assert_equal "unexpected token '-Infinity' at line 1 column 1", ex.message
     end
 
     ex = assert_raise(ParserError) { parse('NaN') }
-    assert_equal "unexpected token at 'NaN'", ex.message
+    unless RUBY_PLATFORM =~ /java/
+      assert_equal "unexpected token 'NaN' at line 1 column 1", ex.message
+    end
   end
 
   if GC.respond_to?(:stress=)

test/json/json_parser_test.rb

@@ -638,32 +638,32 @@ def test_parse_error_message_length
     error = assert_raise(JSON::ParserError) do
       JSON.parse('{"foo": ' + ('A' * 500) + '}')
     end
-    assert_operator 60, :>, error.message.bytesize
+    assert_operator 80, :>, error.message.bytesize
   end
 
   def test_parse_error_incomplete_hash
     error = assert_raise(JSON::ParserError) do
       JSON.parse('{"input":{"firstName":"Bob","lastName":"Mob","email":"[email protected]"}')
     end
     if RUBY_ENGINE == "ruby"
-      assert_equal %(expected ',' or '}' after object value, got: ''), error.message
+      assert_equal %(expected ',' or '}' after object value, got: '' at line 1 column 72), error.message
     end
   end
 
   def test_parse_error_snippet
     omit "C ext only test" unless RUBY_ENGINE == "ruby"
 
     error = assert_raise(JSON::ParserError) { JSON.parse("あああああああああああああああああああああああ") }
-    assert_equal "unexpected character: 'ああああああああああ'", error.message
+    assert_equal "unexpected character: 'ああああああああああ' at line 1 column 1", error.message
 
     error = assert_raise(JSON::ParserError) { JSON.parse("aあああああああああああああああああああああああ") }
-    assert_equal "unexpected character: 'aああああああああああ'", error.message
+    assert_equal "unexpected character: 'aああああああああああ' at line 1 column 1", error.message
 
     error = assert_raise(JSON::ParserError) { JSON.parse("abあああああああああああああああああああああああ") }
-    assert_equal "unexpected character: 'abあああああああああ'", error.message
+    assert_equal "unexpected character: 'abあああああああああ' at line 1 column 1", error.message
 
     error = assert_raise(JSON::ParserError) { JSON.parse("abcあああああああああああああああああああああああ") }
-    assert_equal "unexpected character: 'abcあああああああああ'", error.message
+    assert_equal "unexpected character: 'abcあああああああああ' at line 1 column 1", error.message
   end
 
   def test_parse_leading_slash