From b5b0a53b09776f74ab4ddf6b9468ab1b41267ab5 Mon Sep 17 00:00:00 2001
From: Michael M Slusarz <slusarz@curecanti.org>
Date: Sun, 24 Feb 2013 18:34:35 -0700
Subject: [PATCH] Fix #64166: quoted-printable-encode stream filter incorrectly
 discarding whitespace

Second attempt: need to use lookaheadto determine whether to encode ws
---
 ext/standard/filters.c                   | 58 +++++++++++++++++-------
 ext/standard/tests/streams/bug64166.phpt | 52 +++++++++++++--------
 2 files changed, 74 insertions(+), 36 deletions(-)

diff --git a/ext/standard/filters.c b/ext/standard/filters.c
index 21e165b6c2..99a39be9a1 100644
--- a/ext/standard/filters.c
+++ b/ext/standard/filters.c
@@ -791,7 +791,7 @@ static php_conv_err_t php_conv_qprint_encode_convert(php_conv_qprint_encode *ins
 	unsigned int line_ccnt;
 	unsigned int lb_ptr;
 	unsigned int lb_cnt;
-	unsigned int prev_ws;
+	unsigned int trail_ws;
 	int opts;
 	static char qp_digits[] = "0123456789ABCDEF";
 
@@ -808,7 +808,7 @@ static php_conv_err_t php_conv_qprint_encode_convert(php_conv_qprint_encode *ins
 	icnt = *in_left_p;
 	pd = (unsigned char *)(*out_pp);
 	ocnt = *out_left_p;
-	prev_ws = 0;
+	trail_ws = 0;
 
 	for (;;) {
 		if (!(opts & PHP_CONV_QPRINT_OPT_BINARY) && inst->lbchars != NULL && inst->lbchars_len > 0) {
@@ -827,14 +827,6 @@ static php_conv_err_t php_conv_qprint_encode_convert(php_conv_qprint_encode *ins
 						break;
 					}
 
-					/* If the character(s) immediately before the line break
-					 * is whitespace, need to convert to soft linebreak to
-					 * preserve that data. */
-					if (prev_ws > 0) {
-						*(pd++) = '=';
-						ocnt--;
-					}
-
 					for (i = 0; i < lb_cnt; i++) {
 						*(pd++) = inst->lbchars[i];
 						ocnt--;
@@ -852,9 +844,10 @@ static php_conv_err_t php_conv_qprint_encode_convert(php_conv_qprint_encode *ins
 		} 
 
 		c = NEXT_CHAR(ps, icnt, lb_ptr, lb_cnt, inst->lbchars);
-		prev_ws = 0;
 
-		if (!(opts & PHP_CONV_QPRINT_OPT_BINARY) && (c == '\t' || c == ' ')) {
+		if (!(opts & PHP_CONV_QPRINT_OPT_BINARY) &&
+			(trail_ws == 0) &&
+			(c == '\t' || c == ' ')) {
 			if (line_ccnt < 2 && inst->lbchars != NULL) {
 				if (ocnt < inst->lbchars_len + 1) {
 					err = PHP_CONV_ERR_TOO_BIG;
@@ -874,11 +867,41 @@ static php_conv_err_t php_conv_qprint_encode_convert(php_conv_qprint_encode *ins
 					err = PHP_CONV_ERR_TOO_BIG;
 					break;
 				}
-				*(pd++) = c;
-				ocnt--;
-				line_ccnt--;
-				prev_ws = 1;
-				CONSUME_CHAR(ps, icnt, lb_ptr, lb_cnt);
+
+				/* Check to see if this is EOL whitespace. */
+				if (inst->lbchars != NULL) {
+					unsigned int j, lb_cnt2;
+					lb_cnt2 = 0;
+					unsigned char *ps2;
+					ps2 = ps;
+					trail_ws = 1;
+
+					for (j = icnt - 1; j > 0; j--, ps2++) {
+						if (*ps2 == inst->lbchars[lb_cnt2]) {
+							lb_cnt2++;
+							if (lb_cnt2 >= inst->lbchars_len) {
+								/* Found trailing ws. Reset to top of main
+								 * for loop to allow for code to do necessary
+								 * wrapping/encoding. */
+								break;
+							}
+						} else if (lb_cnt2 != 0 || (*ps2 != '\t' && *ps2 != ' ')) {
+							/* At least one non-EOL character following, so
+							 * don't need to encode ws. */
+							trail_ws = 0;
+							break;
+						} else {
+							trail_ws++;
+						}
+					}
+				}
+
+				if (trail_ws == 0) {
+					*(pd++) = c;
+					ocnt--;
+					line_ccnt--;
+					CONSUME_CHAR(ps, icnt, lb_ptr, lb_cnt);
+				}
 			}
 		} else if ((!(opts & PHP_CONV_QPRINT_OPT_FORCE_ENCODE_FIRST) || line_ccnt < inst->line_len) && ((c >= 33 && c <= 60) || (c >= 62 && c <= 126))) { 
 			if (line_ccnt < 2 && inst->lbchars != NULL) {
@@ -927,6 +950,7 @@ static php_conv_err_t php_conv_qprint_encode_convert(php_conv_qprint_encode *ins
 			*(pd++) = qp_digits[(c & 0x0f)]; 
 			ocnt -= 3;
 			line_ccnt -= 3;
+			trail_ws--;
 			CONSUME_CHAR(ps, icnt, lb_ptr, lb_cnt);
 		}
 	}
diff --git a/ext/standard/tests/streams/bug64166.phpt b/ext/standard/tests/streams/bug64166.phpt
index b9a7a052a4..e725a4b698 100644
--- a/ext/standard/tests/streams/bug64166.phpt
+++ b/ext/standard/tests/streams/bug64166.phpt
@@ -2,31 +2,45 @@
 Bug #64166: quoted-printable-encode stream filter incorrectly discarding whitespace
 --FILE--
 <?php
-$data = "FIRST \nSECOND";
 
-$fd = fopen('php://temp', 'w+');
-fwrite($fd, $data);
-rewind($fd);
+function test_64166($data) {
+    $fd = fopen('php://temp', 'w+');
+    fwrite($fd, $data);
+    rewind($fd);
 
-$res = stream_filter_append($fd, 'convert.quoted-printable-encode', STREAM_FILTER_READ, array(
-	'line-break-chars' => "\n",
-	'line-length' => 7
-));
-var_dump(stream_get_contents($fd, -1, 0));
+    $res = stream_filter_append($fd, 'convert.quoted-printable-encode', STREAM_FILTER_READ, array(
+		'line-break-chars' => "\n",
+		'line-length' => 74
+    ));
+    var_dump(stream_get_contents($fd, -1, 0));
 
-stream_filter_remove($res);
+    stream_filter_remove($res);
+
+    rewind($fd);
+    stream_filter_append($fd, 'convert.quoted-printable-encode', STREAM_FILTER_READ, array(
+		'line-break-chars' => "\n",
+		'line-length' => 6
+    ));
+    var_dump(stream_get_contents($fd, -1, 0));
+
+    fclose($fd);
+}
+
+test_64166("FIRST \nSECOND");
+test_64166("FIRST  \nSECOND");
 
-rewind($fd);
-stream_filter_append($fd, 'convert.quoted-printable-encode', STREAM_FILTER_READ, array(
-	'line-break-chars' => "\n",
-	'line-length' => 6
-));
-var_dump(stream_get_contents($fd, -1, 0));
 ?>
 --EXPECT--
-string(14) "FIRST =
+string(15) "FIRST=20
+SECOND"
+string(19) "FIRST=
+=20
+SECON=
+D"
+string(18) "FIRST=20=20
 SECOND"
-string(18) "FIRST=
- =
+string(24) "FIRST=
+=20=
+=20
 SECON=
 D"
-- 
2.40.0