[RFC] Implement mb_str_pad()

NEWS

@@ -2,6 +2,8 @@ PHP                                                                        NEWS
 |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
 ?? ??? ????, PHP 8.3.0alpha3
 
+- MBString:
+  . Implement mb_str_pad() RFC. (nielsdos)
 
 22 Jun 2023, PHP 8.3.0alpha2
 

UPGRADING

@@ -208,6 +208,10 @@ PHP 8.3 UPGRADE NOTES
     the given $depth and $options.
     RFC: https://wiki.php.net/rfc/json_validate
 
+- MBString:
+  . Added mb_str_pad(), which is the mbstring equivalent of str_pad().
+    RFC: https://wiki.php.net/rfc/mb_str_pad
+
 - Posix:
   . Added posix_sysconf call to get runtime informations.
   . Added posix_pathconf call to get configuration value from a directory/file.

ext/mbstring/mbstring.c

@@ -5522,6 +5522,132 @@ PHP_FUNCTION(mb_chr)
 }
 /* }}} */
 
+PHP_FUNCTION(mb_str_pad)
+{
+	zend_string *input, *encoding_str = NULL, *pad = NULL;
+	zend_long pad_to_length;
+	zend_long pad_type_val = PHP_STR_PAD_RIGHT;
+
+	ZEND_PARSE_PARAMETERS_START(2, 5)
+		Z_PARAM_STR(input)
+		Z_PARAM_LONG(pad_to_length)
+		Z_PARAM_OPTIONAL
+		Z_PARAM_STR(pad)
+		Z_PARAM_LONG(pad_type_val)
+		Z_PARAM_STR_OR_NULL(encoding_str)
+	ZEND_PARSE_PARAMETERS_END();
+
+	const mbfl_encoding *encoding = php_mb_get_encoding(encoding_str, 5);
+	if (!encoding) {
+		RETURN_THROWS();
+	}
+
+	size_t input_length = mb_get_strlen(input, encoding);
+
+	/* If resulting string turns out to be shorter than input string,
+	   we simply copy the input and return. */
+	if (pad_to_length < 0 || (size_t)pad_to_length <= input_length) {
+		RETURN_STR_COPY(input);
+	}
+
+	if (ZSTR_LEN(pad) == 0) {
+		zend_argument_value_error(3, "must be a non-empty string");
+		RETURN_THROWS();
+	}
+
+	if (pad_type_val < PHP_STR_PAD_LEFT || pad_type_val > PHP_STR_PAD_BOTH) {
+		zend_argument_value_error(4, "must be STR_PAD_LEFT, STR_PAD_RIGHT, or STR_PAD_BOTH");
+		RETURN_THROWS();
+	}
+
+	size_t pad_length = mb_get_strlen(pad, encoding);
+
+	size_t num_mb_pad_chars = pad_to_length - input_length;
+
+	/* We need to figure out the left/right padding lengths. */
+	size_t left_pad = 0, right_pad = 0; /* Initialize here to silence compiler warnings. */
+	switch (pad_type_val) {
+		case PHP_STR_PAD_RIGHT:
+			right_pad = num_mb_pad_chars;
+			break;
+
+		case PHP_STR_PAD_LEFT:
+			left_pad = num_mb_pad_chars;
+			break;
+
+		case PHP_STR_PAD_BOTH:
+			left_pad = num_mb_pad_chars / 2;
+			right_pad = num_mb_pad_chars - left_pad;
+			break;
+	}
+
+	/* How many full block copies need to happen, and how many characters are then left over? */
+	size_t full_left_pad_copies = left_pad / pad_length;
+	size_t full_right_pad_copies = right_pad / pad_length;
+	size_t remaining_left_pad_chars = left_pad % pad_length;
+	size_t remaining_right_pad_chars = right_pad % pad_length;
+
+	if (UNEXPECTED(full_left_pad_copies > SIZE_MAX / ZSTR_LEN(pad) || full_right_pad_copies > SIZE_MAX / ZSTR_LEN(pad))) {
+		goto overflow_no_release;
+	}
+
+	/* Compute the number of bytes required for the padding */
+	size_t full_left_pad_bytes = full_left_pad_copies * ZSTR_LEN(pad);
+	size_t full_right_pad_bytes = full_right_pad_copies * ZSTR_LEN(pad);
+
+	/* No special fast-path handling necessary for zero-length pads because these functions will not
+	 * allocate memory in case a zero-length pad is required. */
+	zend_string *remaining_left_pad_str = mb_get_substr(pad, 0, remaining_left_pad_chars, encoding);
+	zend_string *remaining_right_pad_str = mb_get_substr(pad, 0, remaining_right_pad_chars, encoding);
+
+	if (UNEXPECTED(full_left_pad_bytes > ZSTR_MAX_LEN - ZSTR_LEN(remaining_left_pad_str)
+		|| full_right_pad_bytes > ZSTR_MAX_LEN - ZSTR_LEN(remaining_right_pad_str))) {
+		goto overflow;
+	}
+
+	size_t left_pad_bytes = full_left_pad_bytes + ZSTR_LEN(remaining_left_pad_str);
+	size_t right_pad_bytes = full_right_pad_bytes + ZSTR_LEN(remaining_right_pad_str);
+
+	if (UNEXPECTED(left_pad_bytes > ZSTR_MAX_LEN - right_pad_bytes
+		|| ZSTR_LEN(input) > ZSTR_MAX_LEN - left_pad_bytes - right_pad_bytes)) {
+		goto overflow;
+	}
+
+	zend_string *result = zend_string_alloc(ZSTR_LEN(input) + left_pad_bytes + right_pad_bytes, false);
+	char *buffer = ZSTR_VAL(result);
+
+	/* First we pad the left. */
+	for (size_t i = 0; i < full_left_pad_copies; i++, buffer += ZSTR_LEN(pad)) {
+		memcpy(buffer, ZSTR_VAL(pad), ZSTR_LEN(pad));
+	}
+	memcpy(buffer, ZSTR_VAL(remaining_left_pad_str), ZSTR_LEN(remaining_left_pad_str));
+	buffer += ZSTR_LEN(remaining_left_pad_str);
+
+	/* Then we copy the input string. */
+	memcpy(buffer, ZSTR_VAL(input), ZSTR_LEN(input));
+	buffer += ZSTR_LEN(input);
+
+	/* Finally, we pad on the right. */
+	for (size_t i = 0; i < full_right_pad_copies; i++, buffer += ZSTR_LEN(pad)) {
+		memcpy(buffer, ZSTR_VAL(pad), ZSTR_LEN(pad));
+	}
+	memcpy(buffer, ZSTR_VAL(remaining_right_pad_str), ZSTR_LEN(remaining_right_pad_str));
+
+	ZSTR_VAL(result)[ZSTR_LEN(result)] = '\0';
+
+	zend_string_release_ex(remaining_left_pad_str, false);
+	zend_string_release_ex(remaining_right_pad_str, false);
+
+	RETURN_NEW_STR(result);
+
+overflow:
+	zend_string_release_ex(remaining_left_pad_str, false);
+	zend_string_release_ex(remaining_right_pad_str, false);
+overflow_no_release:
+	zend_throw_error(NULL, "String size overflow");
+	RETURN_THROWS();
+}
+
 /* {{{ */
 PHP_FUNCTION(mb_scrub)
 {

ext/mbstring/mbstring.stub.php

@@ -183,6 +183,8 @@ function mb_ord(string $string, ?string $encoding = null): int|false {}
 
 function mb_chr(int $codepoint, ?string $encoding = null): string|false {}
 
+function mb_str_pad(string $string, int $length, string $pad_string = " ", int $pad_type = STR_PAD_RIGHT, ?string $encoding = null): string {}
+
 #ifdef HAVE_MBREGEX
 /** @refcount 1 */
 function mb_regex_encoding(?string $encoding = null): string|bool {}

ext/mbstring/tests/mb_str_pad.phpt

@@ -0,0 +1,138 @@
+--TEST--
+mb_str_pad()
+--EXTENSIONS--
+mbstring
+--FILE--
+<?php
+
+echo "--- Error conditions ---\n";
+try {
+    var_dump(mb_str_pad('▶▶', 6, '', STR_PAD_RIGHT));
+} catch (ValueError $e) {
+    var_dump($e->getMessage());
+}
+try {
+    var_dump(mb_str_pad('▶▶', 6, '', STR_PAD_LEFT));
+} catch (ValueError $e) {
+    var_dump($e->getMessage());
+}
+try {
+    var_dump(mb_str_pad('▶▶', 6, '', STR_PAD_BOTH));
+} catch (ValueError $e) {
+    var_dump($e->getMessage());
+}
+try {
+    var_dump(mb_str_pad('▶▶', 6, ' ', 123456));
+} catch (ValueError $e) {
+    var_dump($e->getMessage());
+}
+try {
+    var_dump(mb_str_pad('▶▶', 6, ' ', STR_PAD_BOTH, 'unexisting'));
+} catch (ValueError $e) {
+    var_dump($e->getMessage());
+}
+
+echo "--- Simple ASCII strings ---\n";
+var_dump(mb_str_pad('Hello', 7, '+-', STR_PAD_BOTH));
+var_dump(mb_str_pad('World', 10, '+-', STR_PAD_BOTH));
+var_dump(mb_str_pad('Hello', 7, '+-', STR_PAD_LEFT));
+var_dump(mb_str_pad('World', 10, '+-', STR_PAD_LEFT));
+var_dump(mb_str_pad('Hello', 7, '+-', STR_PAD_RIGHT));
+var_dump(mb_str_pad('World', 10, '+-', STR_PAD_RIGHT));
+
+echo "--- Edge cases pad length ---\n";
+var_dump(mb_str_pad('▶▶', 2, ' ', STR_PAD_BOTH));
+var_dump(mb_str_pad('▶▶', 1, ' ', STR_PAD_BOTH));
+var_dump(mb_str_pad('▶▶', 0, ' ', STR_PAD_BOTH));
+var_dump(mb_str_pad('▶▶', -1, ' ', STR_PAD_BOTH));
+
+echo "--- Empty input string ---\n";
+var_dump(mb_str_pad('', 2, ' ', STR_PAD_BOTH));
+var_dump(mb_str_pad('', 1, ' ', STR_PAD_BOTH));
+var_dump(mb_str_pad('', 0, ' ', STR_PAD_BOTH));
+var_dump(mb_str_pad('', -1, ' ', STR_PAD_BOTH));
+
+echo "--- No default argument ---\n";
+var_dump(mb_str_pad('▶▶', 6, pad_type: STR_PAD_RIGHT));
+var_dump(mb_str_pad('▶▶', 6, pad_type: STR_PAD_LEFT));
+var_dump(mb_str_pad('▶▶', 6, pad_type: STR_PAD_BOTH));
+
+echo "--- UTF-8 emojis ---\n";
+for ($i = 6; $i > 0; $i--) {
+    var_dump(mb_str_pad('▶▶', $i, '❤❓❇', STR_PAD_RIGHT));
+    var_dump(mb_str_pad('▶▶', $i, '❤❓❇', STR_PAD_LEFT));
+    var_dump(mb_str_pad('▶▶', $i, '❤❓❇', STR_PAD_BOTH));
+}
+
+echo "--- UTF-8, 32, 7 test ---\n";
+
+// Taken from mb_substr.phpt
+$utf8 = "Σὲ γνωρίζω ἀπὸ τὴν κόψη Зарегистрируйтесь";
+$utf32 = mb_convert_encoding($utf8, 'UTF-32', 'UTF-8');
+$utf7 = mb_convert_encoding($utf8, 'UTF-7', 'UTF-8');
+$tests = ["UTF-8" => $utf8, "UTF-32" => $utf32, "UTF-7" => $utf7];
+
+foreach ($tests as $encoding => $test) {
+    $pad_str = mb_convert_encoding('▶▶', $encoding, 'UTF-8');
+    var_dump(mb_convert_encoding(mb_str_pad($test, 44, $pad_str, STR_PAD_RIGHT, $encoding), 'UTF-8', $encoding));
+    var_dump(mb_convert_encoding(mb_str_pad($test, 44, $pad_str, STR_PAD_LEFT, $encoding), 'UTF-8', $encoding));
+    var_dump(mb_convert_encoding(mb_str_pad($test, 44, $pad_str, STR_PAD_BOTH, $encoding), 'UTF-8', $encoding));
+}
+?>
+--EXPECT--
+--- Error conditions ---
+string(66) "mb_str_pad(): Argument #3 ($pad_string) must be a non-empty string"
+string(66) "mb_str_pad(): Argument #3 ($pad_string) must be a non-empty string"
+string(66) "mb_str_pad(): Argument #3 ($pad_string) must be a non-empty string"
+string(90) "mb_str_pad(): Argument #4 ($pad_type) must be STR_PAD_LEFT, STR_PAD_RIGHT, or STR_PAD_BOTH"
+string(82) "mb_str_pad(): Argument #5 ($encoding) must be a valid encoding, "unexisting" given"
+--- Simple ASCII strings ---
+string(7) "+Hello+"
+string(10) "+-World+-+"
+string(7) "+-Hello"
+string(10) "+-+-+World"
+string(7) "Hello+-"
+string(10) "World+-+-+"
+--- Edge cases pad length ---
+string(6) "▶▶"
+string(6) "▶▶"
+string(6) "▶▶"
+string(6) "▶▶"
+--- Empty input string ---
+string(2) "  "
+string(1) " "
+string(0) ""
+string(0) ""
+--- No default argument ---
+string(10) "▶▶    "
+string(10) "    ▶▶"
+string(10) "  ▶▶  "
+--- UTF-8 emojis ---
+string(18) "▶▶❤❓❇❤"
+string(18) "❤❓❇❤▶▶"
+string(18) "❤❓▶▶❤❓"
+string(15) "▶▶❤❓❇"
+string(15) "❤❓❇▶▶"
+string(15) "❤▶▶❤❓"
+string(12) "▶▶❤❓"
+string(12) "❤❓▶▶"
+string(12) "❤▶▶❤"
+string(9) "▶▶❤"
+string(9) "❤▶▶"
+string(9) "▶▶❤"
+string(6) "▶▶"
+string(6) "▶▶"
+string(6) "▶▶"
+string(6) "▶▶"
+string(6) "▶▶"
+string(6) "▶▶"
+--- UTF-8, 32, 7 test ---
+string(92) "Σὲ γνωρίζω ἀπὸ τὴν κόψη Зарегистрируйтесь▶▶▶"
+string(92) "▶▶▶Σὲ γνωρίζω ἀπὸ τὴν κόψη Зарегистрируйтесь"
+string(92) "▶Σὲ γνωρίζω ἀπὸ τὴν κόψη Зарегистрируйтесь▶▶"
+string(92) "Σὲ γνωρίζω ἀπὸ τὴν κόψη Зарегистрируйтесь▶▶▶"
+string(92) "▶▶▶Σὲ γνωρίζω ἀπὸ τὴν κόψη Зарегистрируйтесь"
+string(92) "▶Σὲ γνωρίζω ἀπὸ τὴν κόψη Зарегистрируйтесь▶▶"
+string(92) "Σὲ γνωρίζω ἀπὸ τὴν κόψη Зарегистрируйтесь▶▶▶"
+string(92) "▶▶▶Σὲ γνωρίζω ἀπὸ τὴν κόψη Зарегистрируйтесь"
+string(92) "▶Σὲ γνωρίζω ἀπὸ τὴν κόψη Зарегистрируйтесь▶▶"