[RFC] Add grapheme_limit_codepoints function

Unicode's grapheme cluster is no limit codepoints.
But 1 grapheme cluster and many codepoints can maybe crash.
So I would like set limit for codepoints per grapheme cluster.

Author: youkidearitai

ext/intl/grapheme/grapheme.h

@@ -31,5 +31,6 @@ void grapheme_close_global_iterator( void );
 #define GRAPHEME_EXTRACT_TYPE_MAXCHARS	2
 #define GRAPHEME_EXTRACT_TYPE_MIN	GRAPHEME_EXTRACT_TYPE_COUNT
 #define GRAPHEME_EXTRACT_TYPE_MAX	GRAPHEME_EXTRACT_TYPE_MAXCHARS
+#define GRAPHEME_LIMIT_CODEPOINTS	32
 
 #endif // GRAPHEME_GRAPHEME_H

ext/intl/grapheme/grapheme_string.cpp

@@ -1135,4 +1135,68 @@ U_CFUNC PHP_FUNCTION(grapheme_levenshtein)
 	efree(ustring1);
 }
 
+U_CFUNC PHP_FUNCTION(grapheme_limit_codepoints)
+{
+	char *string;
+	size_t string_len = 0;
+	zend_long limit_codepoint = GRAPHEME_LIMIT_CODEPOINTS;
+	int ustring_len = 0;
+	UErrorCode status;
+	unsigned char u_break_iterator_buffer[U_BRK_SAFECLONE_BUFFERSIZE];
+
+	ZEND_PARSE_PARAMETERS_START(1, 2)
+		Z_PARAM_STRING(string, string_len)
+		Z_PARAM_OPTIONAL
+		Z_PARAM_LONG(limit_codepoint)
+	ZEND_PARSE_PARAMETERS_END();
+
+	status = U_ZERO_ERROR;
+	UBreakIterator *bi;
+	UText ut = UTEXT_INITIALIZER;
+	bi = grapheme_get_break_iterator((void*)u_break_iterator_buffer, &status );
+
+	if( U_FAILURE(status) ) {
+		intl_error_set_code( nullptr, status );
+
+		/* Set error messages. */
+		intl_error_set_custom_msg( nullptr, "Error in grapheme_get_break_iterator" );
+		RETURN_FALSE;
+	}
+
+	utext_openUTF8(&ut, string, string_len, &status);
+	ubrk_setUText(bi, &ut, &status);
+
+	if ( U_FAILURE( status ) ) {
+		/* Set global error code. */
+		intl_error_set_code( nullptr, status );
+
+		/* Set error messages. */
+		intl_error_set_custom_msg( nullptr, "Error opening UTF-8 text");
+
+		RETURN_FALSE;
+	}
+
+	zend_ulong pos, before;
+	zend_bool ret = true;
+	for (before = pos = 0; pos != UBRK_DONE; ) {
+		pos = ubrk_next(bi);
+		if (pos != UBRK_DONE) {
+			for (zend_ulong i = before; i < (pos - before); i++) {
+				U8_FWD_1(string, before, (pos - before) - i);
+				if (i >= limit_codepoint) {
+					ret = false;
+					goto bi_close;
+				}
+			}
+		}
+		before = pos;
+	}
+
+bi_close:
+	ubrk_close(bi);
+	utext_close(&ut);
+
+	RETURN_BOOL(ret);
+
+}
 /* }}} */

ext/intl/php_intl.stub.php

@@ -166,6 +166,11 @@
  * @cvalue UIDNA_ERROR_CONTEXTJ
  */
 const IDNA_ERROR_CONTEXTJ = UNKNOWN;
+/**
+ * @var int
+ * @cvalue GRAPHEME_LIMIT_CODEPOINTS
+ */
+const GRAPHEME_LIMIT_CODEPOINTS = UNKNOWN;
 
 class IntlException extends Exception
 {
@@ -445,6 +450,8 @@ function grapheme_str_split(string $string, int $length = 1): array|false {}
 
 function grapheme_levenshtein(string $string1, string $string2, int $insertion_cost = 1, int $replacement_cost = 1, int $deletion_cost = 1, string $locale = ""): int|false {}
 
+function grapheme_limit_codepoints(string $string, int $limit = GRAPHEME_LIMIT_CODEPOINTS): bool {}
+
 /** @param int $next */
 function grapheme_extract(string $haystack, int $size, int $type = GRAPHEME_EXTR_COUNT, int $offset = 0, &$next = null): string|false {}
 

ext/intl/php_intl_arginfo.h

@@ -0,0 +1,20 @@
+--TEST--
+grapheme_limit_codepoints() function test
+--EXTENSIONS--
+intl
+--FILE--
+<?php
+$f = "あい👨‍👨‍👦‍👦‍👦‍👦‍👦‍👦‍👦‍👦‍👦‍👦‍👦‍👦‍👦‍👦‍👦‍👦‍👦‍👦‍👦‍👦‍👦‍👦‍👦‍👦‍👦‍👦‍👦‍👦‍👦‍👦うえお";
+var_dump(grapheme_limit_codepoints($f));
+$f = "あいうえお👨‍👨‍👦";
+var_dump(grapheme_limit_codepoints($f));
+$f = "あいうえおH̵̛͕̞̦̰̜͍̰̥̟͆̏͂̌͑ͅä̷͔̟͓̬̯̟͍̭͉͈̮͙̣̯̬͚̞̭̍̀̾͠m̴̡̧̛̝̯̹̗̹̤̲̺̟̥̈̏͊̔̑̍͆̌̀̚͝͝b̴̢̢̫̝̠̗̼̬̻̮̺̭͔̘͑̆̎̚ư̵̧̡̥̙̭̿̈̀̒̐̊͒͑r̷̡̡̲̼̖͎̫̮̜͇̬͌͘g̷̹͍͎̬͕͓͕̐̃̈́̓̆̚͝ẻ̵̡̼̬̥̹͇̭͔̯̉͛̈́̕r̸̮̖̻̮̣̗͚͖̝̂͌̾̓̀̿̔̀͋̈́͌̈́̋͜👨‍👨‍👦";
+var_dump(grapheme_limit_codepoints($f));
+$f = "ཧྐྵྨླྺྼྻྂ";
+var_dump(grapheme_limit_codepoints($f));
+?>
+--EXPECT--
+bool(false)
+bool(true)
+bool(true)
+bool(true)