[pull] master from php:master

NEWS

@@ -16,6 +16,8 @@ PHP                                                                        NEWS
 
 - Standard:
   . Add HEIF/HEIC support to getimagesize. (Benstone Zhang)
+  . Implement #71517 (Implement SVG support for getimagesize() and friends).
+    (nielsdos)
 
 - URI:
   . Empty host handling is fixed. (Máté Kocsis)

UPGRADING

@@ -246,6 +246,15 @@ PHP 8.5 UPGRADE NOTES
     ignored. This change affects only the sendmail transport.
   . getimagesize() now supports HEIF/HEIC images.
 
+- Standard:
+  . getimagesize() now supports SVG images when ext-libxml is also loaded.
+    Similarly, image_type_to_extension() and image_type_to_extension()
+    now also handle IMAGETYPE_SVG.
+  . The array returned by getimagesize() now has two additional entries:
+    "width_unit" and "height_unit" to indicate in which units the dimensions
+    are expressed. These units are px by default. They are not necessarily
+    the same (just to give one example: one may be cm and the other may be px).
+
 - XSL:
   . The $namespace argument of XSLTProcessor::getParameter(),
     XSLTProcessor::setParameter() and XSLTProcessor::removeParameter()
@@ -566,6 +575,9 @@ PHP 8.5 UPGRADE NOTES
   . T_VOID_CAST.
   . T_PIPE.
 
+- Standard:
+  . IMAGETYPE_SVG when libxml is loaded.
+
 ========================================
 11. Changes to INI File Handling
 ========================================

UPGRADING.INTERNALS

@@ -134,6 +134,10 @@ PHP 8.5 INTERNALS UPGRADE NOTES
   . The php_std_date() function has been removed. Use php_format_date() with
     the "D, d M Y H:i:s \\G\\M\\T" format instead.
   . Added php_url_encode_to_smart_str() to encode a URL to a smart_str buffer.
+  . The functionality of getimagesize(), image_type_to_mime_type(),
+    and image_type_to_extension() is now extensible using the internal APIs
+    php_image_register_handler() and php_image_unregister_handler() in
+    php_image.h.
 
 ========================
 4. OpCode changes

ext/intl/calendar/gregoriancalendar_methods.cpp

@@ -134,7 +134,7 @@ static void _php_intlgregcal_constructor_body(
 
 	// instantion of ICU object
 	Calendar_object *co = Z_INTL_CALENDAR_P(return_value);
-	GregorianCalendar *gcal = NULL;
+	std::unique_ptr<GregorianCalendar> gcal;
 
 	if (co->ucal) {
 		zend_throw_error(NULL, "IntlGregorianCalendar object is already constructed");
@@ -159,15 +159,12 @@ static void _php_intlgregcal_constructor_body(
 			locale = const_cast<char*>(intl_locale_get_default());
 		}
 
-		gcal = new GregorianCalendar(tz, Locale::createFromName(locale),
-			status);
+		gcal = std::unique_ptr<GregorianCalendar>(new GregorianCalendar(tz, Locale::createFromName(locale),
+			status));
 			// Should this throw?
 		if (U_FAILURE(status)) {
 			intl_error_set(NULL, status, "intlgregcal_create_instance: error "
 				"creating ICU GregorianCalendar from time zone and locale", 0);
-			if (gcal) {
-				delete gcal;
-			}
 			delete tz;
 			if (!is_constructor) {
 				zval_ptr_dtor(return_value);
@@ -177,26 +174,28 @@ static void _php_intlgregcal_constructor_body(
 		}
 	} else {
 		// From date/time (3, 5 or 6 arguments)
+		GregorianCalendar *tmp;
 		for (int i = 0; i < variant; i++) {
 			ZEND_VALUE_ERROR_OUT_OF_BOUND_VALUE(largs[i], hasThis() ? (i-1) : i);
 		}
 
 		if (variant == 3) {
-			gcal = new GregorianCalendar((int32_t)largs[0], (int32_t)largs[1],
+			tmp = new GregorianCalendar((int32_t)largs[0], (int32_t)largs[1],
 				(int32_t)largs[2], status);
 		} else if (variant == 5) {
-			gcal = new GregorianCalendar((int32_t)largs[0], (int32_t)largs[1],
+			tmp = new GregorianCalendar((int32_t)largs[0], (int32_t)largs[1],
 				(int32_t)largs[2], (int32_t)largs[3], (int32_t)largs[4], status);
 		} else if (variant == 6) {
-			gcal = new GregorianCalendar((int32_t)largs[0], (int32_t)largs[1],
+			tmp = new GregorianCalendar((int32_t)largs[0], (int32_t)largs[1],
 				(int32_t)largs[2], (int32_t)largs[3], (int32_t)largs[4], (int32_t)largs[5],
 				status);
 		} else {
 			ZEND_UNREACHABLE();
 		}
 
-		if (!set_gregorian_calendar_time_zone(gcal, status)) {
-			delete gcal;
+		gcal = std::unique_ptr<GregorianCalendar>(tmp);
+
+		if (!set_gregorian_calendar_time_zone(gcal.get(), status)) {
 			if (!is_constructor) {
 				zval_ptr_dtor(return_value);
 				RETVAL_NULL();
@@ -205,7 +204,7 @@ static void _php_intlgregcal_constructor_body(
 		}
 	}
 
-	co->ucal = gcal;
+	co->ucal = gcal.release();
 }
 
 U_CFUNC PHP_FUNCTION(intlgregcal_create_instance)
@@ -234,7 +233,7 @@ U_CFUNC PHP_METHOD(IntlGregorianCalendar, createFromDate)
 	UErrorCode status = U_ZERO_ERROR;
 	zend_error_handling error_handling;
 	Calendar_object *co;
-	GregorianCalendar *gcal;
+	std::unique_ptr<GregorianCalendar> gcal;
 
 	intl_error_reset(NULL);
 
@@ -250,15 +249,14 @@ U_CFUNC PHP_METHOD(IntlGregorianCalendar, createFromDate)
 
 	zend_replace_error_handling(EH_THROW, IntlException_ce_ptr, &error_handling);
 
-	gcal = new GregorianCalendar((int32_t) year, (int32_t) month, (int32_t) day, status);
-	if (!set_gregorian_calendar_time_zone(gcal, status)) {
-		delete gcal;
+	gcal = std::unique_ptr<GregorianCalendar>(new GregorianCalendar((int32_t) year, (int32_t) month, (int32_t) day, status));
+	if (!set_gregorian_calendar_time_zone(gcal.get(), status)) {
 		goto cleanup;
 	}
 
 	object_init_ex(return_value, GregorianCalendar_ce_ptr);
 	co = Z_INTL_CALENDAR_P(return_value);
-	co->ucal = gcal;
+	co->ucal = gcal.release();
 
 cleanup:
 	zend_restore_error_handling(&error_handling);
@@ -271,7 +269,7 @@ U_CFUNC PHP_METHOD(IntlGregorianCalendar, createFromDateTime)
 	UErrorCode status = U_ZERO_ERROR;
 	zend_error_handling error_handling;
 	Calendar_object *co;
-	GregorianCalendar *gcal;
+	GregorianCalendar *tmp;
 
 	intl_error_reset(NULL);
 
@@ -294,19 +292,20 @@ U_CFUNC PHP_METHOD(IntlGregorianCalendar, createFromDateTime)
 	zend_replace_error_handling(EH_THROW, IntlException_ce_ptr, &error_handling);
 
 	if (second_is_null) {
-		gcal = new GregorianCalendar((int32_t) year, (int32_t) month, (int32_t) day, (int32_t) hour, (int32_t) minute, status);
+		tmp = new GregorianCalendar((int32_t) year, (int32_t) month, (int32_t) day, (int32_t) hour, (int32_t) minute, status);
 	} else {
 		ZEND_VALUE_ERROR_OUT_OF_BOUND_VALUE(second, 6);
-		gcal = new GregorianCalendar((int32_t) year, (int32_t) month, (int32_t) day, (int32_t) hour, (int32_t) minute, (int32_t) second, status);
+		tmp = new GregorianCalendar((int32_t) year, (int32_t) month, (int32_t) day, (int32_t) hour, (int32_t) minute, (int32_t) second, status);
 	}
-	if (!set_gregorian_calendar_time_zone(gcal, status)) {
-		delete gcal;
+	auto gcal = std::unique_ptr<GregorianCalendar>(tmp);
+	if (!set_gregorian_calendar_time_zone(gcal.get(), status)) {
 		goto cleanup;
 	}
 
 	object_init_ex(return_value, GregorianCalendar_ce_ptr);
 	co = Z_INTL_CALENDAR_P(return_value);
-	co->ucal = gcal;
+	// TODO: trying to get passed the ownership change step
+	co->ucal = gcal.release();
 
 cleanup:
 	zend_restore_error_handling(&error_handling);

ext/libxml/config.w32

@@ -11,7 +11,7 @@ if (PHP_LIBXML == "yes") {
 		if (GREP_HEADER("libxml/xmlversion.h", "#define\\s+LIBXML_VERSION\\s+(\\d+)", PHP_PHP_BUILD + "\\include\\libxml2") &&
 				+RegExp.$1 >= 20904) {
 
-			EXTENSION("libxml", "libxml.c mime_sniff.c", false /* never shared */, "/DZEND_ENABLE_STATIC_TSRMLS_CACHE=1");
+			EXTENSION("libxml", "libxml.c mime_sniff.c image_svg.c", false /* never shared */, "/DZEND_ENABLE_STATIC_TSRMLS_CACHE=1");
 			AC_DEFINE("HAVE_LIBXML", 1, "Define to 1 if the PHP extension 'libxml' is available.");
 			ADD_FLAG("CFLAGS_LIBXML", "/D LIBXML_STATIC /D LIBXML_STATIC_FOR_DLL /D HAVE_WIN32_THREADS ");
 			if (!PHP_LIBXML_SHARED) {

ext/libxml/config0.m4

@@ -12,7 +12,7 @@ if test "$PHP_LIBXML" != "no"; then
     AC_DEFINE([HAVE_LIBXML], [1],
       [Define to 1 if the PHP extension 'libxml' is available.])
     PHP_NEW_EXTENSION([libxml],
-      [libxml.c mime_sniff.c],
+      [libxml.c mime_sniff.c image_svg.c],
       [$ext_shared],,
       [-DZEND_ENABLE_STATIC_TSRMLS_CACHE=1])
     PHP_INSTALL_HEADERS([ext/libxml], [php_libxml.h])

ext/libxml/image_svg.c

@@ -0,0 +1,178 @@
+/*
+   +----------------------------------------------------------------------+
+   | Copyright (c) The PHP Group                                          |
+   +----------------------------------------------------------------------+
+   | This source file is subject to version 3.01 of the PHP license,      |
+   | that is bundled with this package in the file LICENSE, and is        |
+   | available through the world-wide-web at the following url:           |
+   | https://www.php.net/license/3_01.txt                                 |
+   | If you did not receive a copy of the PHP license and are unable to   |
+   | obtain it through the world-wide-web, please send a note to          |
+   | license@php.net so we can mail you a copy immediately.               |
+   +----------------------------------------------------------------------+
+   | Authors: Niels Dossche <nielsdos@php.net>                            |
+   +----------------------------------------------------------------------+
+ */
+
+#ifdef HAVE_CONFIG_H
+#include "config.h"
+#endif
+
+#include "php.h"
+#include "image_svg.h"
+#include "php_libxml.h"
+
+#include "ext/standard/php_image.h"
+
+#include <libxml/xmlreader.h>
+
+#ifdef HAVE_LIBXML
+
+static int svg_image_type_id;
+
+static int php_libxml_svg_stream_read(void *context, char *buffer, int len)
+{
+	return php_stream_read(context, buffer, len);
+}
+
+/* Sanity check that the input only contains characters valid for a dimension (numbers with units, e.g. 5cm).
+ * This also protects the user against injecting XSS.
+ * Only accept [0-9]+[a-zA-Z]* */
+static bool php_libxml_parse_dimension(const xmlChar *input, const xmlChar **unit_position)
+{
+	if (!(*input >= '0' && *input <= '9')) {
+		return false;
+	}
+
+	input++;
+
+	while (*input) {
+		if (!(*input >= '0' && *input <= '9')) {
+			if ((*input >= 'a' && *input <= 'z') || (*input >= 'A' && *input <= 'Z')) {
+				break;
+			}
+			return false;
+		}
+		input++;
+	}
+
+	*unit_position = input;
+
+	while (*input) {
+		if (!((*input >= 'a' && *input <= 'z') || (*input >= 'A' && *input <= 'Z'))) {
+			return false;
+		}
+		input++;
+	}
+
+	return true;
+}
+
+zend_result php_libxml_svg_image_handle(php_stream *stream, struct php_gfxinfo **result)
+{
+	if (php_stream_rewind(stream)) {
+		return FAILURE;
+	}
+
+	/* Early check before doing more expensive work */
+	if (php_stream_getc(stream) != '<') {
+		return FAILURE;
+	}
+
+	if (php_stream_rewind(stream)) {
+		return FAILURE;
+	}
+
+	PHP_LIBXML_SANITIZE_GLOBALS(reader_for_stream);
+	xmlTextReaderPtr reader = xmlReaderForIO(
+		php_libxml_svg_stream_read,
+		NULL,
+		stream,
+		NULL,
+		NULL,
+		XML_PARSE_NOWARNING | XML_PARSE_NOERROR | XML_PARSE_NONET
+	);
+	PHP_LIBXML_RESTORE_GLOBALS(reader_for_stream);
+
+	if (!reader) {
+		return FAILURE;
+	}
+
+	bool is_svg = false;
+	while (xmlTextReaderRead(reader) == 1) {
+		if (xmlTextReaderNodeType(reader) == XML_READER_TYPE_ELEMENT) {
+			/* Root must be an svg element */
+			const xmlChar *name = xmlTextReaderConstLocalName(reader);
+			if (!name || strcasecmp((const char *) name, "svg") != 0) {
+				break;
+			}
+
+			xmlChar *width = xmlTextReaderGetAttribute(reader, BAD_CAST "width");
+			xmlChar *height = xmlTextReaderGetAttribute(reader, BAD_CAST "height");
+			const xmlChar *width_unit_position, *height_unit_position;
+			if (!width || !height
+				|| !php_libxml_parse_dimension(width, &width_unit_position)
+				|| !php_libxml_parse_dimension(height, &height_unit_position)) {
+				xmlFree(width);
+				xmlFree(height);
+				break;
+			}
+
+			is_svg = true;
+			if (result) {
+				*result = ecalloc(1, sizeof(**result));
+				(*result)->width = ZEND_STRTOL((const char *) width, NULL, 10);
+				(*result)->height = ZEND_STRTOL((const char *) height, NULL, 10);
+				if (*width_unit_position) {
+					(*result)->width_unit = zend_string_init((const char*) width_unit_position,
+															 xmlStrlen(width_unit_position), false);
+				}
+				if (*height_unit_position) {
+					(*result)->height_unit = zend_string_init((const char*) height_unit_position,
+															  xmlStrlen(height_unit_position), false);
+				}
+			}
+
+			xmlFree(width);
+			xmlFree(height);
+			break;
+		}
+	}
+
+	xmlFreeTextReader(reader);
+
+	return is_svg ? SUCCESS : FAILURE;
+}
+
+zend_result php_libxml_svg_image_identify(php_stream *stream)
+{
+	return php_libxml_svg_image_handle(stream, NULL);
+}
+
+struct php_gfxinfo *php_libxml_svg_image_get_info(php_stream *stream)
+{
+	struct php_gfxinfo *result = NULL;
+	zend_result status = php_libxml_svg_image_handle(stream, &result);
+	ZEND_ASSERT((status == SUCCESS) == (result != NULL));
+	return result;
+}
+
+static const struct php_image_handler svg_image_handler = {
+	"image/svg+xml",
+	".svg",
+	PHP_IMAGE_CONST_NAME("SVG"),
+	php_libxml_svg_image_identify,
+	php_libxml_svg_image_get_info,
+};
+
+void php_libxml_register_image_svg_handler(void)
+{
+	svg_image_type_id = php_image_register_handler(&svg_image_handler);
+}
+
+zend_result php_libxml_unregister_image_svg_handler(void)
+{
+	return php_image_unregister_handler(svg_image_type_id);
+}
+
+#endif