encoder: Add support for base64url encoding

Author: JakubOnderka

README.md

@@ -136,6 +136,8 @@ $fileContentEncoded = new SimdJsonBase64Encode($fileContent);
 simdjson_encode(['image' => $fileContentEncoded]); // returns {"image":"TWFueSBoYW5kcyBtYWtlIGxpZ2h0IHdvcmsu..."}
 ```
 
+You can also use base64url encoding (RFC 4648 §5) by setting second argument to true: `new SimdJsonBase64Encode($fileContent, true);`
+
 ### Encode to stream
 
 For large data sets, simdjson_php provides the `simdjson_encode_to_stream()` function to save data directly to a file or output buffer.

php_simdjson.cpp

@@ -591,23 +591,29 @@ PHP_FUNCTION(simdjson_encode_to_stream) {
 
 PHP_METHOD(SimdJsonBase64Encode, __construct) {
     zend_string *binary_string = NULL;
+    bool base64_url = false;
 
-    ZEND_PARSE_PARAMETERS_START(1, 1)
+    ZEND_PARSE_PARAMETERS_START(1, 2)
         Z_PARAM_STR(binary_string)
+        Z_PARAM_OPTIONAL
+        Z_PARAM_BOOL(base64_url)
     ZEND_PARSE_PARAMETERS_END();
 
     return_value = ZEND_THIS;
     ZVAL_STR_COPY(OBJ_PROP_NUM(Z_OBJ_P(ZEND_THIS), 0), binary_string);
+    ZVAL_BOOL(OBJ_PROP_NUM(Z_OBJ_P(ZEND_THIS), 1), base64_url);
 }
 
 PHP_METHOD(SimdJsonBase64Encode, jsonSerialize) {
     ZEND_PARSE_PARAMETERS_NONE();
 
-    zend_string *binary_string = Z_STR(Z_OBJ_P(ZEND_THIS)->properties_table[0]);
+    zend_string *binary_string = Z_STR_P(OBJ_PROP_NUM(Z_OBJ_P(ZEND_THIS), 0));
+    bool base64_url = Z_TYPE_INFO_P(OBJ_PROP_NUM(Z_OBJ_P(ZEND_THIS), 1)) == IS_TRUE;
+    auto options = base64_url ? simdutf::base64_url : simdutf::base64_default;
 
-    size_t encoded_length = simdutf::base64_length_from_binary(ZSTR_LEN(binary_string));
+    size_t encoded_length = simdutf::base64_length_from_binary(ZSTR_LEN(binary_string), options);
     zend_string *result = zend_string_alloc(encoded_length, 0);
-    simdutf::binary_to_base64(ZSTR_VAL(binary_string), ZSTR_LEN(binary_string), ZSTR_VAL(result));
+    simdutf::binary_to_base64(ZSTR_VAL(binary_string), ZSTR_LEN(binary_string), ZSTR_VAL(result), options);
     ZSTR_VAL(result)[encoded_length] = '\0';
     GC_ADD_FLAGS(result, IS_STR_VALID_UTF8); // base64 encoded string must be always valid UTF-8 string
 

simdjson.stub.php

@@ -364,11 +364,13 @@ class SimdJsonEncoderException extends SimdJsonException {}
  */
 final class SimdJsonBase64Encode implements JsonSerializable {
     private string $string;
+    private bool $base64url = false;
 
     /**
      * @param string $string The data to encode.
+     * @param bool $base64url When true, use base64url encoding (RFC 4648 §5).
      */
-    public function __construct(string $string) {}
+    public function __construct(string $string, bool $base64url = false) {}
 
     /**
      * @return mixed Base64 encoded string

simdjson_arginfo.h

@@ -1,5 +1,5 @@
 /* This is a generated file, edit the .stub.php file instead.
- * Stub hash: 5b2147fe5e6865e7ab51e22a21cb235aeba0a12e */
+ * Stub hash: 21c3b21c513b1ce2b8c72912856cd6d7326dbfe5 */
 
 ZEND_BEGIN_ARG_WITH_RETURN_TYPE_INFO_EX(arginfo_simdjson_validate, 0, 1, _IS_BOOL, 0)
 	ZEND_ARG_TYPE_INFO(0, json, IS_STRING, 0)
@@ -62,6 +62,7 @@ ZEND_END_ARG_INFO()
 
 ZEND_BEGIN_ARG_INFO_EX(arginfo_class_SimdJsonBase64Encode___construct, 0, 0, 1)
 	ZEND_ARG_TYPE_INFO(0, string, IS_STRING, 0)
+	ZEND_ARG_TYPE_INFO_WITH_DEFAULT_VALUE(0, base64url, _IS_BOOL, 0, "false")
 ZEND_END_ARG_INFO()
 
 ZEND_BEGIN_ARG_WITH_RETURN_TYPE_INFO_EX(arginfo_class_SimdJsonBase64Encode_jsonSerialize, 0, 0, IS_MIXED, 0)
@@ -216,5 +217,11 @@ static zend_class_entry *register_class_SimdJsonBase64Encode(zend_class_entry *c
 	zend_declare_typed_property(class_entry, property_string_name, &property_string_default_value, ZEND_ACC_PRIVATE, NULL, (zend_type) ZEND_TYPE_INIT_MASK(MAY_BE_STRING));
 	zend_string_release(property_string_name);
 
+	zval property_base64url_default_value;
+	ZVAL_FALSE(&property_base64url_default_value);
+	zend_string *property_base64url_name = zend_string_init("base64url", sizeof("base64url") - 1, 1);
+	zend_declare_typed_property(class_entry, property_base64url_name, &property_base64url_default_value, ZEND_ACC_PRIVATE, NULL, (zend_type) ZEND_TYPE_INIT_MASK(MAY_BE_BOOL));
+	zend_string_release(property_base64url_name);
+
 	return class_entry;
 }

src/simdjson_encoder.cpp

@@ -709,14 +709,16 @@ static zend_result simdjson_escape_string(smart_str *buf, zend_string *str, simd
 }
 
 static void simdjson_encode_base64_object(smart_str *buf, const zval *val) {
-    zend_string *binary_string = Z_STR(Z_OBJ_P(val)->properties_table[0]);
+    zend_string *binary_string = Z_STR_P(OBJ_PROP_NUM(Z_OBJ_P(val), 0));
+    bool base64_url = Z_TYPE_INFO_P(OBJ_PROP_NUM(Z_OBJ_P(val), 1)) == IS_TRUE;
+    auto options = base64_url ? simdutf::base64_url : simdutf::base64_default;
 
     // As we are sure that base64 encoded string is always valid UTF-8 and do not contain any char that need to be
     // escaped, so we can skip all checks and just directly copy encoded string to output buffer
-    size_t encoded_length = simdutf::base64_length_from_binary(ZSTR_LEN(binary_string));
+    size_t encoded_length = simdutf::base64_length_from_binary(ZSTR_LEN(binary_string), options);
     char* output = simdjson_smart_str_extend(buf, encoded_length + 2);
     *output++ = '"';
-    simdutf::binary_to_base64(ZSTR_VAL(binary_string), ZSTR_LEN(binary_string), output);
+    simdutf::binary_to_base64(ZSTR_VAL(binary_string), ZSTR_LEN(binary_string), output, options);
     output += encoded_length;
     *output = '"';
 }

tests/encode/simdjson_base64_encode.phpt

@@ -2,16 +2,29 @@
 Test SimdJsonBase64Encode
 --FILE--
 <?php
+echo "Base64 encode\n";
 $base64Encode = new SimdJsonBase64Encode("ahoj");
 var_dump($base64Encode->jsonSerialize());
 var_dump((string)$base64Encode);
 var_dump(json_encode(base64_encode("ahoj")));
 var_dump(json_encode($base64Encode));
 var_dump(simdjson_encode($base64Encode));
+echo "Base64 URL encode\n";
+$base64Encode = new SimdJsonBase64Encode("ahoj", true);
+var_dump($base64Encode->jsonSerialize());
+var_dump((string)$base64Encode);
+var_dump(json_encode($base64Encode));
+var_dump(simdjson_encode($base64Encode));
 ?>
 --EXPECT--
+Base64 encode
 string(8) "YWhvag=="
 string(8) "YWhvag=="
 string(10) ""YWhvag==""
 string(10) ""YWhvag==""
 string(10) ""YWhvag==""
+Base64 URL encode
+string(6) "YWhvag"
+string(6) "YWhvag"
+string(8) ""YWhvag""
+string(8) ""YWhvag""