Bug #60494 [Csd]: iconv_mime_decode does ignore special characters

Edit report at https://bugs.php.net/bug.php?id=60494&edit=1 ID: 60494 Updated by: cmb@php.net Reported by: bin at webcandy dot at Summary: iconv_mime_decode does ignore special characters Status: Closed Type: Bug Package: ICONV related Operating System: Windows 7 PHP Version: Irrelevant Assigned To: cmb Block user comment: N Private report: N New Comment: > This fix has now broken all headers even with the > ICONV_MIME_DECODE_CONTINUE_ON_ERROR flag set. When this flag is > set it should continue anyway. Thanks for reporting! Tentatively, I agree with your statement. However, since this fix has already been shipped with PHP 7.1.22 and 7.2.10, please open a new ticket (and assign to me, if possible) to better be able to track the bug(fixe)s. Previous Comments: ------------------------------------------------------------------------ [2018-11-13 22:17:19] shafer_w2002 at yahoo dot com This fix has now broken all headers even with the ICONV_MIME_DECODE_CONTINUE_ON_ERROR flag set. When this flag is set it should continue anyway. ------------------------------------------------------------------------ [2018-08-25 15:54:07] cmb@php.net Automatic comment on behalf of cmbecker69@gmx.de Revision: http://git.php.net/?p=php-src.git;a=commit;h=e29c946c29afdb0bf89c5329fcf3038448d50e17 Log: Fix #60494: iconv_mime_decode does ignore special characters ------------------------------------------------------------------------ [2018-03-11 23:12:47] cmb@php.net iconv_mime_decode() is supposed to decode a MIME header according to RFC 2047. Neither 'ä', 'ö' nor 'ß' are valid characters in headers, so returning these characters unmodified would be wrong. However, it is not correct to simply skip invalid characters. Instead the function should fail (i.e. return FALSE). It seems the problem is that inside of _php_iconv_mime_decode() all calls to _php_iconv_appendc() ignore its return value (although all calls to _php_iconv_appendl() cater to it). In this case the culprit is: <https://github.com/php/php-src/blob/php-7.2.3/ext/iconv/iconv.c#L1538> ------------------------------------------------------------------------ [2011-12-11 21:16:22] bin at webcandy dot at Expected result should probably have 2 bytes as I expect it to be UTF-8 when using iconv.output_encoding = UTF-8 Exptected result: ---------------- string(2) "ä" string(2) "ö" string(2) "ß" ------------------------------------------------------------------------ [2011-12-11 21:12:35] bin at webcandy dot at Description: ------------ If the string $text passed to iconv_mime_decode($text) does contain special characters (tested with ä, ü and ß), it will just return an empty character for these characters. In my opinion it iconv_mime_decode should leave special characters as they are as long as they're not encoded via Q-encoding. Tested on Windows 7 and Debian. iconv.output_encoding is set to "UTF-8" Test script: --------------- <?php var_dump(iconv_mime_decode('ä')); var_dump(iconv_mime_decode('ö')); var_dump(iconv_mime_decode('ß')); ?> Expected result: ---------------- string(1) "ä" string(1) "ö" string(1) "ß" Actual result: -------------- string(0) "" string(0) "" string(0) "" ------------------------------------------------------------------------ -- Edit this bug report at https://bugs.php.net/bug.php?id=60494&edit=1