feature: add option to allow comments in decode.

Author: arkriny

README.md

@@ -262,3 +262,16 @@ cjson.encode(t) -- {"my_array":[]} properly re-encoded as an array
 ```
 
 [Back to TOC](#table-of-contents)
+
+decode_allow_comments
+--------------------------
+**syntax:** `cjson.decode_allow_comments(enabled)`
+
+**default:** false
+
+If enabled, allows JavaScript-style comments in `cjson.decode` input. Comments
+are treated as whitespace and may appear anywhere whitespace is valid in JSON.
+Supports single-line comments beginning with '//' and block comments enclosed
+with '/* ... */'.
+
+[Back to TOC](#table-of-contents)

lua_cjson.c

@@ -89,6 +89,7 @@
 #define DEFAULT_ENCODE_NUMBER_PRECISION 14
 #define DEFAULT_ENCODE_EMPTY_TABLE_AS_OBJECT 1
 #define DEFAULT_DECODE_ARRAY_WITH_ARRAY_MT 0
+#define DEFAULT_DECODE_ALLOW_COMMENTS 0
 #define DEFAULT_ENCODE_ESCAPE_FORWARD_SLASH 1
 #define DEFAULT_ENCODE_SKIP_UNSUPPORTED_VALUE_TYPES 0
 #define DEFAULT_ENCODE_INDENT NULL
@@ -178,6 +179,7 @@ typedef struct {
     int decode_invalid_numbers;
     int decode_max_depth;
     int decode_array_with_array_mt;
+    int decode_allow_comments;
     int encode_skip_unsupported_value_types;
 } json_config_t;
 
@@ -383,6 +385,16 @@ static int json_cfg_decode_array_with_array_mt(lua_State *l)
     return 1;
 }
 
+/* Configures whether decoder allows comments */
+static int json_cfg_decode_allow_comments(lua_State *l)
+{
+    json_config_t *cfg = json_arg_init(l, 1);
+
+    json_enum_option(l, 1, &cfg->decode_allow_comments, NULL, 1);
+
+    return 1;
+}
+
 /* Configure how to treat invalid types */
 static int json_cfg_encode_skip_unsupported_value_types(lua_State *l)
 {
@@ -515,6 +527,7 @@ static void json_create_config(lua_State *l)
     cfg->encode_number_precision = DEFAULT_ENCODE_NUMBER_PRECISION;
     cfg->encode_empty_table_as_object = DEFAULT_ENCODE_EMPTY_TABLE_AS_OBJECT;
     cfg->decode_array_with_array_mt = DEFAULT_DECODE_ARRAY_WITH_ARRAY_MT;
+    cfg->decode_allow_comments = DEFAULT_DECODE_ALLOW_COMMENTS;
     cfg->encode_escape_forward_slash = DEFAULT_ENCODE_ESCAPE_FORWARD_SLASH;
     cfg->encode_skip_unsupported_value_types = DEFAULT_ENCODE_SKIP_UNSUPPORTED_VALUE_TYPES;
     cfg->encode_indent = DEFAULT_ENCODE_INDENT;
@@ -1279,13 +1292,46 @@ static void json_next_token(json_parse_t *json, json_token_t *token)
     const json_token_type_t *ch2token = json->cfg->ch2token;
     int ch;
 
-    /* Eat whitespace. */
     while (1) {
-        ch = (unsigned char)*(json->ptr);
-        token->type = ch2token[ch];
-        if (token->type != T_WHITESPACE)
+        /* Eat whitespace. */
+        while (1) {
+            ch = (unsigned char)*(json->ptr);
+            token->type = ch2token[ch];
+            if (token->type != T_WHITESPACE)
+                break;
+            json->ptr++;
+        }
+
+        if (!json->cfg->decode_allow_comments)
             break;
-        json->ptr++;
+
+        /* Eat comments. */
+        if ((unsigned char)json->ptr[0] != '/' ||
+            ((unsigned char)json->ptr[1] != '/' &&
+            (unsigned char)json->ptr[1] != '*')) {
+            break;
+        }
+
+        if (json->ptr[1] == '/') {
+            /* Handle single-line comment */
+            json->ptr += 2;
+            while (*json->ptr != '\0' && *json->ptr != '\n')
+                json->ptr++;
+        } else {
+            /* Handle multi-line comment */
+            json->ptr += 2;
+            while (1) {
+                if (*json->ptr == '\0') {
+                    json_set_token_error(token, json, "unclosed multi-line comment");
+                    return;
+                }
+                if (json->ptr[0] == '*' && json->ptr[1] == '/') {
+                    json->ptr += 2;
+                    break;
+                }
+                json->ptr++;
+            }
+        }
     }
 
     /* Store location of new token. Required when throwing errors
@@ -1625,6 +1671,7 @@ static int lua_cjson_new(lua_State *l)
         { "decode", json_decode },
         { "encode_empty_table_as_object", json_cfg_encode_empty_table_as_object },
         { "decode_array_with_array_mt", json_cfg_decode_array_with_array_mt },
+        { "decode_allow_comments", json_cfg_decode_allow_comments },
         { "encode_sparse_array", json_cfg_encode_sparse_array },
         { "encode_max_depth", json_cfg_encode_max_depth },
         { "decode_max_depth", json_cfg_decode_max_depth },

tests/test.lua

@@ -333,6 +333,32 @@ local cjson_tests = {
       json.decode, { [["\uDB00\uD"]] },
       false, { "Expected value but found invalid unicode escape code at character 2" } },
 
+    -- Test comments
+    { 'Set decode_allow_comments(true)',
+      json.decode_allow_comments, { true }, true, { true } },
+    { "Decode single-line comment",
+      json.decode, { '{//comment\n}' }, true, { {} } },
+    { "Decode single-line comment with windows newline",
+      json.decode, { '{//comment\r\n}' }, true, { {} } },
+    { "Decode single-line comment after string value",
+      json.decode, { '"test // /* */ string"//comment' }, true, { "test // /* */ string" } },
+    { "Decode multi-line comment",
+      json.decode, { '{/* A multi-line\ncomment*/}' }, true, { {} } },
+    { "Decode multi-line comment before colon",
+      json.decode, { '{"a" /* Comment */: 1}' }, true, { { a = 1 } } },
+    { "Decode multi-line comment after colon",
+      json.decode, { '{"a": /* Comment */ 1}' }, true, { { a = 1 } } },
+    { "Decode multiple comments in a row",
+      json.decode, { '/*first*//*second*/{}' }, true, { {} } },
+    { "Decode unclosed multi-line comment [throw error]",
+      json.decode, { '{}/*Unclosed' },
+      false, { "Expected the end but found unclosed multi-line comment at character 13" } },
+    { "Decode comment inside number [throw error]",
+      json.decode, { '{"a":1/*x*/0}' },
+      false, { "Expected comma or object end but found T_INTEGER at character 12" } },
+    { 'Set decode_allow_comments(false)',
+      json.decode_allow_comments, { false }, true, { false } },
+
     -- Test indenting
     { 'Set encode_indent("  ")',
       json.encode_indent, { "  " }, true, { "  " } },