From cb9fc852cf39d91dfa1dc915073914fd153e774a Mon Sep 17 00:00:00 2001
From: alecpl <alec@alec.pl>
Date: Wed, 23 Sep 2009 13:25:05 -0400
Subject: [PATCH] - Fix regular expression for e-mail address (#1486152)
---
program/include/rcube_vcard.php | 107 ++++++++++++++++++++++++++++++++++++++++-------------
1 files changed, 81 insertions(+), 26 deletions(-)
diff --git a/program/include/rcube_vcard.php b/program/include/rcube_vcard.php
index 3ad47a5..7dbbb3f 100644
--- a/program/include/rcube_vcard.php
+++ b/program/include/rcube_vcard.php
@@ -5,7 +5,7 @@
| program/include/rcube_vcard.php |
| |
| This file is part of the RoundCube Webmail client |
- | Copyright (C) 2008, RoundCube Dev. - Switzerland |
+ | Copyright (C) 2008-2009, RoundCube Dev. - Switzerland |
| Licensed under the GNU GPL |
| |
| PURPOSE: |
@@ -73,7 +73,7 @@
$this->business = ($this->raw['X-ABShowAs'][0] == 'COMPANY') || (join('', (array)$this->raw['N'][0]) == '' && !empty($this->organization));
foreach ((array)$this->raw['EMAIL'] as $i => $raw_email)
- $this->email[$i] = $raw_email[0];
+ $this->email[$i] = is_array($raw_email) ? $raw_email[0] : $raw_email;
// make the pref e-mail address the first entry in $this->email
$pref_index = $this->get_type_index('EMAIL', 'pref');
@@ -172,6 +172,7 @@
$encoding = self::detect_encoding($data);
if ($encoding && $encoding != RCMAIL_CHARSET) {
$data = rcube_charset_convert($data, $encoding);
+ $data = preg_replace(array('/^[\xFE\xFF]{2}/', '/^\xEF\xBB\xBF/', '/^\x00+/'), '', $data); // also remove BOM
}
$vcard_block = '';
@@ -217,15 +218,22 @@
$vcard = preg_replace(array('/^item\d*\.X-AB.*$/m', '/^item\d*\./m', "/\n+/"), array('', '', "\n"), $vcard);
// remove vcard 2.1 charset definitions
- $vcard = preg_replace('/;CHARSET=[^:]+/', '', $vcard);
+ $vcard = preg_replace('/;CHARSET=[^:;]+/', '', $vcard);
+
+ // if N doesn't have any semicolons, add some
+ $vcard = preg_replace('/^(N:[^;\R]*)$/m', '\1;;;;', $vcard);
return $vcard;
}
+ private static function rfc2425_fold_callback($matches)
+ {
+ return ":\n ".rtrim(chunk_split($matches[1], 72, "\n "));
+ }
private static function rfc2425_fold($val)
{
- return preg_replace('/:([^\n]{72,})/e', '":\n ".rtrim(chunk_split("\\1", 72, "\n "))', $val) . "\n";
+ return preg_replace_callback('/:([^\n]{72,})/', array('self', 'rfc2425_fold_callback'), $val) . "\n";
}
@@ -241,31 +249,47 @@
// Perform RFC2425 line unfolding
$vcard = preg_replace(array("/\r/", "/\n\s+/"), '', $vcard);
+ $lines = preg_split('/\r?\n/', $vcard);
$data = array();
- if (preg_match_all('/^([^\\:]*):(.+)$/m', $vcard, $regs, PREG_SET_ORDER)) {
- foreach($regs as $line) {
- // convert 2.1-style "EMAIL;internet;home:" to 3.0-style "EMAIL;TYPE=internet,home:"
- if(($data['VERSION'][0] == "2.1") && preg_match('/^([^;]+);([^:]+)/', $line[1], $regs2) && !preg_match('/^TYPE=/i', $regs2[2])) {
- $line[1] = $regs2[1] . ";TYPE=" . strtr($regs2[2], array(";" => ","));
- }
+
+ for ($i=0; $i < count($lines); $i++) {
+ if (!preg_match('/^([^\\:]*):(.+)$/', $lines[$i], $line))
+ continue;
- if (!preg_match('/^(BEGIN|END)$/', $line[1]) && preg_match_all('/([^\\;]+);?/', $line[1], $regs2)) {
- $entry = array(self::vcard_unquote($line[2]));
-
- foreach($regs2[1] as $attrid => $attr) {
- if ((list($key, $value) = explode('=', $attr)) && $value)
- $entry[strtolower($key)] = array_merge((array)$entry[strtolower($key)], (array)self::vcard_unquote($value, ','));
- elseif ($attrid > 0)
- $entry[$key] = true; # true means attr without =value
- }
-
- $data[$regs2[1][0]][] = count($entry) > 1 ? $entry : $entry[0];
- }
+ // convert 2.1-style "EMAIL;internet;home:" to 3.0-style "EMAIL;TYPE=internet;TYPE=home:"
+ if (($data['VERSION'][0] == "2.1") && preg_match('/^([^;]+);([^:]+)/', $line[1], $regs2) && !preg_match('/^TYPE=/i', $regs2[2])) {
+ $line[1] = $regs2[1];
+ foreach (explode(';', $regs2[2]) as $prop)
+ $line[1] .= ';' . (strpos($prop, '=') ? $prop : 'TYPE='.$prop);
}
- unset($data['VERSION']);
+ if (!preg_match('/^(BEGIN|END)$/i', $line[1]) && preg_match_all('/([^\\;]+);?/', $line[1], $regs2)) {
+ $entry = array();
+ $field = strtoupper($regs2[1][0]);
+
+ foreach($regs2[1] as $attrid => $attr) {
+ if ((list($key, $value) = explode('=', $attr)) && $value) {
+ if ($key == 'ENCODING') {
+ // add next line(s) to value string if QP line end detected
+ while ($value == 'QUOTED-PRINTABLE' && preg_match('/=$/', $lines[$i]))
+ $line[2] .= "\n" . $lines[++$i];
+
+ $line[2] = self::decode_value($line[2], $value);
+ }
+ else
+ $entry[strtolower($key)] = array_merge((array)$entry[strtolower($key)], (array)self::vcard_unquote($value, ','));
+ }
+ else if ($attrid > 0) {
+ $entry[$key] = true; // true means attr without =value
+ }
+ }
+
+ $entry = array_merge($entry, (array)self::vcard_unquote($line[2]));
+ $data[$field][] = count($entry) > 1 ? $entry : $entry[0];
+ }
}
+ unset($data['VERSION']);
return $data;
}
@@ -293,6 +317,28 @@
/**
+ * Decode a given string with the encoding rule from ENCODING attributes
+ *
+ * @param string String to decode
+ * @param string Encoding type (quoted-printable and base64 supported)
+ * @return string Decoded 8bit value
+ */
+ private static function decode_value($value, $encoding)
+ {
+ switch (strtolower($encoding)) {
+ case 'quoted-printable':
+ return quoted_printable_decode($value);
+
+ case 'base64':
+ return base64_decode($value);
+
+ default:
+ return $value;
+ }
+ }
+
+
+ /**
* Encodes an entry for storage in our database (vcard 3.0 format, unfolded)
*
* @param array Raw data structure to encode
@@ -302,7 +348,7 @@
{
foreach((array)$data as $type => $entries) {
/* valid N has 5 properties */
- while ($type == "N" && count($entries[0]) < 5)
+ while ($type == "N" && is_array($entries[0]) && count($entries[0]) < 5)
$entries[0][] = "";
foreach((array)$entries as $entry) {
@@ -313,7 +359,7 @@
if (is_int($attrname))
$value[] = $attrvalues;
elseif ($attrvalues === true)
- $attr .= ";$attrname"; # true means just tag, not tag=value, as in PHOTO;BASE64:...
+ $attr .= ";$attrname"; // true means just tag, not tag=value, as in PHOTO;BASE64:...
else {
foreach((array)$attrvalues as $attrvalue)
$attr .= ";$attrname=" . self::vcard_quote($attrvalue, ',');
@@ -367,6 +413,15 @@
if (substr($string, 0, 2) == "\xFF\xFE") return 'UTF-16LE'; // Little Endian
if (substr($string, 0, 3) == "\xEF\xBB\xBF") return 'UTF-8';
+ // use mb_detect_encoding()
+ $encodings = array('UTF-8', 'ISO-8859-1', 'ISO-8859-2', 'ISO-8859-3',
+ 'ISO-8859-4', 'ISO-8859-5', 'ISO-8859-6', 'ISO-8859-7', 'ISO-8859-8', 'ISO-8859-9',
+ 'ISO-8859-10', 'ISO-8859-13', 'ISO-8859-14', 'ISO-8859-15', 'ISO-8859-16',
+ 'WINDOWS-1252', 'WINDOWS-1251', 'BIG5', 'GB2312');
+
+ if (function_exists('mb_detect_encoding') && ($enc = mb_detect_encoding($string, $encodings)))
+ return $enc;
+
// No match, check for UTF-8
// from http://w3.org/International/questions/qa-forms-utf-8.html
if (preg_match('/\A(
@@ -381,7 +436,7 @@
)*\z/xs', substr($string, 0, 2048)))
return 'UTF-8';
- return null;
+ return rcmail::get_instance()->config->get('default_charset', 'ISO-8859-1'); # fallback to Latin-1
}
}
--
Gitblit v1.9.1