From: steven at acko dot net Operating system: Windows 2000 PHP version: 4CVS-2003-12-14 (stable) PHP Bug Type: *XML functions Bug description: XML Parser ignores UTF-8 input encoding?
Description: ------------ PHP seems to ignore the encoding when parsing an UTF-8 encoded XML file, and assumes it is ISO-8859-1 instead. The code below contains a very short XML file (inline) to parse, with the character a-with-tilde as value (this is just to show what happens, there is nothing special about this character). The a-with-tilde takes 2 bytes in UTF-8 encoding, and is represented as such in the XML file string. The buggy behaviour is illustrated with the 3 possible PHP output encodings. Comment/uncomment the correct xml_parser_set_option() call to see the behaviour in all output encodings. Note that nothing changes in behaviour if you change the encoding="utf-8" in the source XML into 'iso-8859-1', or remove it altogether, which shows that it is being ignored. Reproduce code: --------------- <?php $xmlfile = "<?xml version=\"1.0\" encoding=\"utf-8\" ?><tag>\xC3\xA3</tag>"; function handler_data($parser, $data) { print "Data: $data\n"; print "Length: ". strlen($data) ."\n"; } $xml_parser = xml_parser_create(); xml_set_character_data_handler($xml_parser, "handler_data"); xml_parser_set_option($xml_parser, XML_OPTION_TARGET_ENCODING, "utf-8"); // xml_parser_set_option($xml_parser, XML_OPTION_TARGET_ENCODING, "iso-8859-1"); // xml_parser_set_option($xml_parser, XML_OPTION_TARGET_ENCODING, "us-ascii"); xml_parse($xml_parser, $xmlfile, 1); xml_parser_free($xml_parser); ?> Expected result: ---------------- With output encoding set to UTF-8, PHP should leave the data alone and should print 2 bytes (\xC3 and \xA3) for the a-with-tilde. With output encoding set to ISO-8859-1, PHP should encode the a-with-tile as a single byte (\xE3). With output encoding set to US-ASCII, PHP should output a single '?' to indicate the a-with-tilde is not available in the output encoding. Actual result: -------------- With output encoding set to UTF-8, PHP re-encodes the input into UTF-8, resulting in 4 bytes for a-with-tilde. With output encoding set to ISO-8859-1, PHP leaves the input untouched. With output encoding set to US-ASCII, PHP outputs two question marks. -- Edit bug report at http://bugs.php.net/?id=26623&edit=1 -- Try a CVS snapshot (php4): http://bugs.php.net/fix.php?id=26623&r=trysnapshot4 Try a CVS snapshot (php5): http://bugs.php.net/fix.php?id=26623&r=trysnapshot5 Fixed in CVS: http://bugs.php.net/fix.php?id=26623&r=fixedcvs Fixed in release: http://bugs.php.net/fix.php?id=26623&r=alreadyfixed Need backtrace: http://bugs.php.net/fix.php?id=26623&r=needtrace Need Reproduce Script: http://bugs.php.net/fix.php?id=26623&r=needscript Try newer version: http://bugs.php.net/fix.php?id=26623&r=oldversion Not developer issue: http://bugs.php.net/fix.php?id=26623&r=support Expected behavior: http://bugs.php.net/fix.php?id=26623&r=notwrong Not enough info: http://bugs.php.net/fix.php?id=26623&r=notenoughinfo Submitted twice: http://bugs.php.net/fix.php?id=26623&r=submittedtwice register_globals: http://bugs.php.net/fix.php?id=26623&r=globals PHP 3 support discontinued: http://bugs.php.net/fix.php?id=26623&r=php3 Daylight Savings: http://bugs.php.net/fix.php?id=26623&r=dst IIS Stability: http://bugs.php.net/fix.php?id=26623&r=isapi Install GNU Sed: http://bugs.php.net/fix.php?id=26623&r=gnused Floating point limitations: http://bugs.php.net/fix.php?id=26623&r=float