mb_decode_mimeheader obeys RFC 2047 regarding underscores and QPrint encoding

Author: alexdowad

NEWS

@@ -52,7 +52,8 @@ PHP                                                                        NEWS
   . Added json_validate(). (Juan Morales)
 
 - MBString:
-  . mb_detect_encoding is better able to identify the correct encoding for Turkish text. (Alex Dowad)
+  . mb_detect_encoding is better able to identify the correct encoding for
+    Turkish text. (Alex Dowad)
   . mb_detect_encoding's "non-strict" mode now behaves as described in the
     documentation. Previously, it would return false if the very first byte
     of the input string was invalid in all candidate encodings. (Alex Dowad)
@@ -62,6 +63,10 @@ PHP                                                                        NEWS
     MB_CASE_LOWER_SIMPLE and MB_CASE_TITLE_SIMPLE. (Alex Dowad)
   . mb_detect_encoding is better able to identify UTF-8 and UTF-16 strings
     with a byte-order mark. (Alex Dowad)
+  . mb_decode_mimeheader interprets underscores in QPrint-encoded MIME
+    encoded words as required by RFC 2047; they are converted to spaces.
+    Underscores must be encoded as "=5F" in such MIME encoded words.
+    (Alex Dowad)
 
 - Opcache:
   . Added start, restart and force restart time to opcache's

UPGRADING

@@ -78,6 +78,10 @@ PHP 8.3 UPGRADE NOTES
     casing rules for the Greek letter sigma. For mb_convert_case, conditional
     casing only applies to MB_CASE_LOWER and MB_CASE_TITLE modes, not to
     MB_CASE_LOWER_SIMPLE and MB_CASE_TITLE_SIMPLE. (Alex Dowad)
+  . mb_decode_mimeheader interprets underscores in QPrint-encoded MIME
+    encoded words as required by RFC 2047; they are converted to spaces.
+    Underscores must be encoded as "=5F" in such MIME encoded words.
+    (Alex Dowad)
 
 - Standard:
   . E_NOTICEs emitted by unserialized() have been promoted to E_WARNING.

ext/mbstring/mbstring.c

@@ -5705,7 +5705,10 @@ static unsigned char* mime_header_decode_encoded_word(unsigned char *p, unsigned
 		/* Fill `buf` with bytes from decoding QPrint */
 		while (p < e) {
 			unsigned char c = *p++;
-			if (c == '=' && (e - p) >= 2) {
+			if (c == '_') {
+				*bufp++ = ' ';
+				continue;
+			} else if (c == '=' && (e - p) >= 2) {
 				unsigned char c2 = *p++;
 				unsigned char c3 = *p++;
 				if (qprint_map[c2] >= 0 && qprint_map[c3] >= 0) {

ext/mbstring/tests/mb_decode_mimeheader_variation5.phpt

@@ -0,0 +1,23 @@
+--TEST--
+Test mb_decode_mimeheader() function: use of underscores in QPrint-encoded data
+--EXTENSIONS--
+mbstring
+--FILE--
+<?php
+
+// RFC 2047 says that in a QPrint-encoded MIME encoded word, underscores should be converted to spaces
+var_dump(mb_decode_mimeheader("=?UTF-8?Q?abc?="));
+var_dump(mb_decode_mimeheader("=?UTF-8?Q?abc_def?="));
+var_dump(mb_decode_mimeheader("_=?UTF-8?Q?abc_def?=_"));
+var_dump(mb_decode_mimeheader("=?UTF-8?Q?__=E6=B1=89=E5=AD=97__?="));
+
+// This is how underscores should be encoded in MIME encoded words with QPrint
+var_dump(mb_decode_mimeheader("=?UTF-8?Q?=5F?="));
+
+?>
+--EXPECT--
+string(3) "abc"
+string(7) "abc def"
+string(9) "_abc def_"
+string(10) "  汉字  "
+string(1) "_"