mirror of
https://github.com/OpenXE-org/OpenXE.git
synced 2024-12-27 07:00:29 +01:00
672 lines
22 KiB
PHP
672 lines
22 KiB
PHP
<?php
|
|
|
|
namespace Sabre\VObject\Parser;
|
|
|
|
use Sabre\VObject\Component;
|
|
use Sabre\VObject\Component\VCalendar;
|
|
use Sabre\VObject\Component\VCard;
|
|
use Sabre\VObject\Document;
|
|
use Sabre\VObject\EofException;
|
|
use Sabre\VObject\Node;
|
|
use Sabre\VObject\ParseException;
|
|
|
|
/**
|
|
* MimeDir parser.
|
|
*
|
|
* This class parses iCalendar 2.0 and vCard 2.1, 3.0 and 4.0 files. This
|
|
* parser will return one of the following two objects from the parse method:
|
|
*
|
|
* Sabre\VObject\Component\VCalendar
|
|
* Sabre\VObject\Component\VCard
|
|
*
|
|
* @copyright Copyright (C) fruux GmbH (https://fruux.com/)
|
|
* @author Evert Pot (http://evertpot.com/)
|
|
* @license http://sabre.io/license/ Modified BSD License
|
|
*/
|
|
class MimeDir extends Parser
|
|
{
|
|
/**
|
|
* The input stream.
|
|
*
|
|
* @var resource
|
|
*/
|
|
protected $input;
|
|
|
|
/**
|
|
* Root component.
|
|
*
|
|
* @var Component
|
|
*/
|
|
protected $root;
|
|
|
|
/**
|
|
* By default all input will be assumed to be UTF-8.
|
|
*
|
|
* However, both iCalendar and vCard might be encoded using different
|
|
* character sets. The character set is usually set in the mime-type.
|
|
*
|
|
* If this is the case, use setEncoding to specify that a different
|
|
* encoding will be used. If this is set, the parser will automatically
|
|
* convert all incoming data to UTF-8.
|
|
*
|
|
* @var string
|
|
*/
|
|
protected $charset = 'UTF-8';
|
|
|
|
/**
|
|
* The list of character sets we support when decoding.
|
|
*
|
|
* This would be a const expression but for now we need to support PHP 5.5
|
|
*/
|
|
protected static $SUPPORTED_CHARSETS = [
|
|
'UTF-8',
|
|
'ISO-8859-1',
|
|
'Windows-1252',
|
|
];
|
|
|
|
/**
|
|
* Parses an iCalendar or vCard file.
|
|
*
|
|
* Pass a stream or a string. If null is parsed, the existing buffer is
|
|
* used.
|
|
*
|
|
* @param string|resource|null $input
|
|
* @param int $options
|
|
*
|
|
* @return \Sabre\VObject\Document
|
|
*/
|
|
public function parse($input = null, $options = 0)
|
|
{
|
|
$this->root = null;
|
|
|
|
if (!is_null($input)) {
|
|
$this->setInput($input);
|
|
}
|
|
|
|
if (0 !== $options) {
|
|
$this->options = $options;
|
|
}
|
|
|
|
$this->parseDocument();
|
|
|
|
return $this->root;
|
|
}
|
|
|
|
/**
|
|
* By default all input will be assumed to be UTF-8.
|
|
*
|
|
* However, both iCalendar and vCard might be encoded using different
|
|
* character sets. The character set is usually set in the mime-type.
|
|
*
|
|
* If this is the case, use setEncoding to specify that a different
|
|
* encoding will be used. If this is set, the parser will automatically
|
|
* convert all incoming data to UTF-8.
|
|
*
|
|
* @param string $charset
|
|
*/
|
|
public function setCharset($charset)
|
|
{
|
|
if (!in_array($charset, self::$SUPPORTED_CHARSETS)) {
|
|
throw new \InvalidArgumentException('Unsupported encoding. (Supported encodings: '.implode(', ', self::$SUPPORTED_CHARSETS).')');
|
|
}
|
|
$this->charset = $charset;
|
|
}
|
|
|
|
/**
|
|
* Sets the input buffer. Must be a string or stream.
|
|
*
|
|
* @param resource|string $input
|
|
*/
|
|
public function setInput($input)
|
|
{
|
|
// Resetting the parser
|
|
$this->lineIndex = 0;
|
|
$this->startLine = 0;
|
|
|
|
if (is_string($input)) {
|
|
// Converting to a stream.
|
|
$stream = fopen('php://temp', 'r+');
|
|
fwrite($stream, $input);
|
|
rewind($stream);
|
|
$this->input = $stream;
|
|
} elseif (is_resource($input)) {
|
|
$this->input = $input;
|
|
} else {
|
|
throw new \InvalidArgumentException('This parser can only read from strings or streams.');
|
|
}
|
|
}
|
|
|
|
/**
|
|
* Parses an entire document.
|
|
*/
|
|
protected function parseDocument()
|
|
{
|
|
$line = $this->readLine();
|
|
|
|
// BOM is ZERO WIDTH NO-BREAK SPACE (U+FEFF).
|
|
// It's 0xEF 0xBB 0xBF in UTF-8 hex.
|
|
if (3 <= strlen($line)
|
|
&& 0xef === ord($line[0])
|
|
&& 0xbb === ord($line[1])
|
|
&& 0xbf === ord($line[2])) {
|
|
$line = substr($line, 3);
|
|
}
|
|
|
|
switch (strtoupper($line)) {
|
|
case 'BEGIN:VCALENDAR':
|
|
$class = VCalendar::$componentMap['VCALENDAR'];
|
|
break;
|
|
case 'BEGIN:VCARD':
|
|
$class = VCard::$componentMap['VCARD'];
|
|
break;
|
|
default:
|
|
throw new ParseException('This parser only supports VCARD and VCALENDAR files');
|
|
}
|
|
|
|
$this->root = new $class([], false);
|
|
|
|
while (true) {
|
|
// Reading until we hit END:
|
|
$line = $this->readLine();
|
|
if ('END:' === strtoupper(substr($line, 0, 4))) {
|
|
break;
|
|
}
|
|
$result = $this->parseLine($line);
|
|
if ($result) {
|
|
$this->root->add($result);
|
|
}
|
|
}
|
|
|
|
$name = strtoupper(substr($line, 4));
|
|
if ($name !== $this->root->name) {
|
|
throw new ParseException('Invalid MimeDir file. expected: "END:'.$this->root->name.'" got: "END:'.$name.'"');
|
|
}
|
|
}
|
|
|
|
/**
|
|
* Parses a line, and if it hits a component, it will also attempt to parse
|
|
* the entire component.
|
|
*
|
|
* @param string $line Unfolded line
|
|
*
|
|
* @return Node
|
|
*/
|
|
protected function parseLine($line)
|
|
{
|
|
// Start of a new component
|
|
if ('BEGIN:' === strtoupper(substr($line, 0, 6))) {
|
|
if (substr($line, 6) === $this->root->name) {
|
|
throw new ParseException('Invalid MimeDir file. Unexpected component: "'.$line.'" in document type '.$this->root->name);
|
|
}
|
|
$component = $this->root->createComponent(substr($line, 6), [], false);
|
|
|
|
while (true) {
|
|
// Reading until we hit END:
|
|
$line = $this->readLine();
|
|
if ('END:' === strtoupper(substr($line, 0, 4))) {
|
|
break;
|
|
}
|
|
$result = $this->parseLine($line);
|
|
if ($result) {
|
|
$component->add($result);
|
|
}
|
|
}
|
|
|
|
$name = strtoupper(substr($line, 4));
|
|
if ($name !== $component->name) {
|
|
throw new ParseException('Invalid MimeDir file. expected: "END:'.$component->name.'" got: "END:'.$name.'"');
|
|
}
|
|
|
|
return $component;
|
|
} else {
|
|
// Property reader
|
|
$property = $this->readProperty($line);
|
|
if (!$property) {
|
|
// Ignored line
|
|
return false;
|
|
}
|
|
|
|
return $property;
|
|
}
|
|
}
|
|
|
|
/**
|
|
* We need to look ahead 1 line every time to see if we need to 'unfold'
|
|
* the next line.
|
|
*
|
|
* If that was not the case, we store it here.
|
|
*
|
|
* @var string|null
|
|
*/
|
|
protected $lineBuffer;
|
|
|
|
/**
|
|
* The real current line number.
|
|
*/
|
|
protected $lineIndex = 0;
|
|
|
|
/**
|
|
* In the case of unfolded lines, this property holds the line number for
|
|
* the start of the line.
|
|
*
|
|
* @var int
|
|
*/
|
|
protected $startLine = 0;
|
|
|
|
/**
|
|
* Contains a 'raw' representation of the current line.
|
|
*
|
|
* @var string
|
|
*/
|
|
protected $rawLine;
|
|
|
|
/**
|
|
* Reads a single line from the buffer.
|
|
*
|
|
* This method strips any newlines and also takes care of unfolding.
|
|
*
|
|
* @throws \Sabre\VObject\EofException
|
|
*
|
|
* @return string
|
|
*/
|
|
protected function readLine()
|
|
{
|
|
if (!\is_null($this->lineBuffer)) {
|
|
$rawLine = $this->lineBuffer;
|
|
$this->lineBuffer = null;
|
|
} else {
|
|
do {
|
|
$eof = \feof($this->input);
|
|
|
|
$rawLine = \fgets($this->input);
|
|
|
|
if ($eof || (\feof($this->input) && false === $rawLine)) {
|
|
throw new EofException('End of document reached prematurely');
|
|
}
|
|
if (false === $rawLine) {
|
|
throw new ParseException('Error reading from input stream');
|
|
}
|
|
$rawLine = \rtrim($rawLine, "\r\n");
|
|
} while ('' === $rawLine); // Skipping empty lines
|
|
++$this->lineIndex;
|
|
}
|
|
$line = $rawLine;
|
|
|
|
$this->startLine = $this->lineIndex;
|
|
|
|
// Looking ahead for folded lines.
|
|
while (true) {
|
|
$nextLine = \rtrim(\fgets($this->input), "\r\n");
|
|
++$this->lineIndex;
|
|
if (!$nextLine) {
|
|
break;
|
|
}
|
|
if ("\t" === $nextLine[0] || ' ' === $nextLine[0]) {
|
|
$curLine = \substr($nextLine, 1);
|
|
$line .= $curLine;
|
|
$rawLine .= "\n ".$curLine;
|
|
} else {
|
|
$this->lineBuffer = $nextLine;
|
|
break;
|
|
}
|
|
}
|
|
$this->rawLine = $rawLine;
|
|
|
|
return $line;
|
|
}
|
|
|
|
/**
|
|
* Reads a property or component from a line.
|
|
*/
|
|
protected function readProperty($line)
|
|
{
|
|
if ($this->options & self::OPTION_FORGIVING) {
|
|
$propNameToken = 'A-Z0-9\-\._\\/';
|
|
} else {
|
|
$propNameToken = 'A-Z0-9\-\.';
|
|
}
|
|
|
|
$paramNameToken = 'A-Z0-9\-';
|
|
$safeChar = '^";:,';
|
|
$qSafeChar = '^"';
|
|
|
|
$regex = "/
|
|
^(?P<name> [$propNameToken]+ ) (?=[;:]) # property name
|
|
|
|
|
(?<=:)(?P<propValue> .+)$ # property value
|
|
|
|
|
;(?P<paramName> [$paramNameToken]+) (?=[=;:]) # parameter name
|
|
|
|
|
(=|,)(?P<paramValue> # parameter value
|
|
(?: [$safeChar]*) |
|
|
\"(?: [$qSafeChar]+)\"
|
|
) (?=[;:,])
|
|
/xi";
|
|
|
|
//echo $regex, "\n"; die();
|
|
preg_match_all($regex, $line, $matches, PREG_SET_ORDER);
|
|
|
|
$property = [
|
|
'name' => null,
|
|
'parameters' => [],
|
|
'value' => null,
|
|
];
|
|
|
|
$lastParam = null;
|
|
|
|
/*
|
|
* Looping through all the tokens.
|
|
*
|
|
* Note that we are looping through them in reverse order, because if a
|
|
* sub-pattern matched, the subsequent named patterns will not show up
|
|
* in the result.
|
|
*/
|
|
foreach ($matches as $match) {
|
|
if (isset($match['paramValue'])) {
|
|
if ($match['paramValue'] && '"' === $match['paramValue'][0]) {
|
|
$value = substr($match['paramValue'], 1, -1);
|
|
} else {
|
|
$value = $match['paramValue'];
|
|
}
|
|
|
|
$value = $this->unescapeParam($value);
|
|
|
|
if (is_null($lastParam)) {
|
|
throw new ParseException('Invalid Mimedir file. Line starting at '.$this->startLine.' did not follow iCalendar/vCard conventions');
|
|
}
|
|
if (is_null($property['parameters'][$lastParam])) {
|
|
$property['parameters'][$lastParam] = $value;
|
|
} elseif (is_array($property['parameters'][$lastParam])) {
|
|
$property['parameters'][$lastParam][] = $value;
|
|
} else {
|
|
$property['parameters'][$lastParam] = [
|
|
$property['parameters'][$lastParam],
|
|
$value,
|
|
];
|
|
}
|
|
continue;
|
|
}
|
|
if (isset($match['paramName'])) {
|
|
$lastParam = strtoupper($match['paramName']);
|
|
if (!isset($property['parameters'][$lastParam])) {
|
|
$property['parameters'][$lastParam] = null;
|
|
}
|
|
continue;
|
|
}
|
|
if (isset($match['propValue'])) {
|
|
$property['value'] = $match['propValue'];
|
|
continue;
|
|
}
|
|
if (isset($match['name']) && $match['name']) {
|
|
$property['name'] = strtoupper($match['name']);
|
|
continue;
|
|
}
|
|
|
|
// @codeCoverageIgnoreStart
|
|
throw new \LogicException('This code should not be reachable');
|
|
// @codeCoverageIgnoreEnd
|
|
}
|
|
|
|
if (is_null($property['value'])) {
|
|
$property['value'] = '';
|
|
}
|
|
if (!$property['name']) {
|
|
if ($this->options & self::OPTION_IGNORE_INVALID_LINES) {
|
|
return false;
|
|
}
|
|
throw new ParseException('Invalid Mimedir file. Line starting at '.$this->startLine.' did not follow iCalendar/vCard conventions');
|
|
}
|
|
|
|
// vCard 2.1 states that parameters may appear without a name, and only
|
|
// a value. We can deduce the value based on its name.
|
|
//
|
|
// Our parser will get those as parameters without a value instead, so
|
|
// we're filtering these parameters out first.
|
|
$namedParameters = [];
|
|
$namelessParameters = [];
|
|
|
|
foreach ($property['parameters'] as $name => $value) {
|
|
if (!is_null($value)) {
|
|
$namedParameters[$name] = $value;
|
|
} else {
|
|
$namelessParameters[] = $name;
|
|
}
|
|
}
|
|
|
|
$propObj = $this->root->createProperty($property['name'], null, $namedParameters);
|
|
|
|
foreach ($namelessParameters as $namelessParameter) {
|
|
$propObj->add(null, $namelessParameter);
|
|
}
|
|
|
|
if ('QUOTED-PRINTABLE' === strtoupper($propObj['ENCODING'])) {
|
|
$propObj->setQuotedPrintableValue($this->extractQuotedPrintableValue());
|
|
} else {
|
|
$charset = $this->charset;
|
|
if (Document::VCARD21 === $this->root->getDocumentType() && isset($propObj['CHARSET'])) {
|
|
// vCard 2.1 allows the character set to be specified per property.
|
|
$charset = (string) $propObj['CHARSET'];
|
|
}
|
|
switch (strtolower($charset)) {
|
|
case 'utf-8':
|
|
break;
|
|
case 'iso-8859-1':
|
|
$property['value'] = utf8_encode($property['value']);
|
|
break;
|
|
case 'windows-1252':
|
|
$property['value'] = mb_convert_encoding($property['value'], 'UTF-8', $charset);
|
|
break;
|
|
default:
|
|
throw new ParseException('Unsupported CHARSET: '.$propObj['CHARSET']);
|
|
}
|
|
$propObj->setRawMimeDirValue($property['value']);
|
|
}
|
|
|
|
return $propObj;
|
|
}
|
|
|
|
/**
|
|
* Unescapes a property value.
|
|
*
|
|
* vCard 2.1 says:
|
|
* * Semi-colons must be escaped in some property values, specifically
|
|
* ADR, ORG and N.
|
|
* * Semi-colons must be escaped in parameter values, because semi-colons
|
|
* are also use to separate values.
|
|
* * No mention of escaping backslashes with another backslash.
|
|
* * newlines are not escaped either, instead QUOTED-PRINTABLE is used to
|
|
* span values over more than 1 line.
|
|
*
|
|
* vCard 3.0 says:
|
|
* * (rfc2425) Backslashes, newlines (\n or \N) and comma's must be
|
|
* escaped, all time time.
|
|
* * Comma's are used for delimiters in multiple values
|
|
* * (rfc2426) Adds to to this that the semi-colon MUST also be escaped,
|
|
* as in some properties semi-colon is used for separators.
|
|
* * Properties using semi-colons: N, ADR, GEO, ORG
|
|
* * Both ADR and N's individual parts may be broken up further with a
|
|
* comma.
|
|
* * Properties using commas: NICKNAME, CATEGORIES
|
|
*
|
|
* vCard 4.0 (rfc6350) says:
|
|
* * Commas must be escaped.
|
|
* * Semi-colons may be escaped, an unescaped semi-colon _may_ be a
|
|
* delimiter, depending on the property.
|
|
* * Backslashes must be escaped
|
|
* * Newlines must be escaped as either \N or \n.
|
|
* * Some compound properties may contain multiple parts themselves, so a
|
|
* comma within a semi-colon delimited property may also be unescaped
|
|
* to denote multiple parts _within_ the compound property.
|
|
* * Text-properties using semi-colons: N, ADR, ORG, CLIENTPIDMAP.
|
|
* * Text-properties using commas: NICKNAME, RELATED, CATEGORIES, PID.
|
|
*
|
|
* Even though the spec says that commas must always be escaped, the
|
|
* example for GEO in Section 6.5.2 seems to violate this.
|
|
*
|
|
* iCalendar 2.0 (rfc5545) says:
|
|
* * Commas or semi-colons may be used as delimiters, depending on the
|
|
* property.
|
|
* * Commas, semi-colons, backslashes, newline (\N or \n) are always
|
|
* escaped, unless they are delimiters.
|
|
* * Colons shall not be escaped.
|
|
* * Commas can be considered the 'default delimiter' and is described as
|
|
* the delimiter in cases where the order of the multiple values is
|
|
* insignificant.
|
|
* * Semi-colons are described as the delimiter for 'structured values'.
|
|
* They are specifically used in Semi-colons are used as a delimiter in
|
|
* REQUEST-STATUS, RRULE, GEO and EXRULE. EXRULE is deprecated however.
|
|
*
|
|
* Now for the parameters
|
|
*
|
|
* If delimiter is not set (null) this method will just return a string.
|
|
* If it's a comma or a semi-colon the string will be split on those
|
|
* characters, and always return an array.
|
|
*
|
|
* @param string $input
|
|
* @param string $delimiter
|
|
*
|
|
* @return string|string[]
|
|
*/
|
|
public static function unescapeValue($input, $delimiter = ';')
|
|
{
|
|
$regex = '# (?: (\\\\ (?: \\\\ | N | n | ; | , ) )';
|
|
if ($delimiter) {
|
|
$regex .= ' | ('.$delimiter.')';
|
|
}
|
|
$regex .= ') #x';
|
|
|
|
$matches = preg_split($regex, $input, -1, PREG_SPLIT_DELIM_CAPTURE | PREG_SPLIT_NO_EMPTY);
|
|
|
|
$resultArray = [];
|
|
$result = '';
|
|
|
|
foreach ($matches as $match) {
|
|
switch ($match) {
|
|
case '\\\\':
|
|
$result .= '\\';
|
|
break;
|
|
case '\N':
|
|
case '\n':
|
|
$result .= "\n";
|
|
break;
|
|
case '\;':
|
|
$result .= ';';
|
|
break;
|
|
case '\,':
|
|
$result .= ',';
|
|
break;
|
|
case $delimiter:
|
|
$resultArray[] = $result;
|
|
$result = '';
|
|
break;
|
|
default:
|
|
$result .= $match;
|
|
break;
|
|
}
|
|
}
|
|
|
|
$resultArray[] = $result;
|
|
|
|
return $delimiter ? $resultArray : $result;
|
|
}
|
|
|
|
/**
|
|
* Unescapes a parameter value.
|
|
*
|
|
* vCard 2.1:
|
|
* * Does not mention a mechanism for this. In addition, double quotes
|
|
* are never used to wrap values.
|
|
* * This means that parameters can simply not contain colons or
|
|
* semi-colons.
|
|
*
|
|
* vCard 3.0 (rfc2425, rfc2426):
|
|
* * Parameters _may_ be surrounded by double quotes.
|
|
* * If this is not the case, semi-colon, colon and comma may simply not
|
|
* occur (the comma used for multiple parameter values though).
|
|
* * If it is surrounded by double-quotes, it may simply not contain
|
|
* double-quotes.
|
|
* * This means that a parameter can in no case encode double-quotes, or
|
|
* newlines.
|
|
*
|
|
* vCard 4.0 (rfc6350)
|
|
* * Behavior seems to be identical to vCard 3.0
|
|
*
|
|
* iCalendar 2.0 (rfc5545)
|
|
* * Behavior seems to be identical to vCard 3.0
|
|
*
|
|
* Parameter escaping mechanism (rfc6868) :
|
|
* * This rfc describes a new way to escape parameter values.
|
|
* * New-line is encoded as ^n
|
|
* * ^ is encoded as ^^.
|
|
* * " is encoded as ^'
|
|
*
|
|
* @param string $input
|
|
*/
|
|
private function unescapeParam($input)
|
|
{
|
|
return
|
|
preg_replace_callback(
|
|
'#(\^(\^|n|\'))#',
|
|
function ($matches) {
|
|
switch ($matches[2]) {
|
|
case 'n':
|
|
return "\n";
|
|
case '^':
|
|
return '^';
|
|
case '\'':
|
|
return '"';
|
|
|
|
// @codeCoverageIgnoreStart
|
|
}
|
|
// @codeCoverageIgnoreEnd
|
|
},
|
|
$input
|
|
);
|
|
}
|
|
|
|
/**
|
|
* Gets the full quoted printable value.
|
|
*
|
|
* We need a special method for this, because newlines have both a meaning
|
|
* in vCards, and in QuotedPrintable.
|
|
*
|
|
* This method does not do any decoding.
|
|
*
|
|
* @return string
|
|
*/
|
|
private function extractQuotedPrintableValue()
|
|
{
|
|
// We need to parse the raw line again to get the start of the value.
|
|
//
|
|
// We are basically looking for the first colon (:), but we need to
|
|
// skip over the parameters first, as they may contain one.
|
|
$regex = '/^
|
|
(?: [^:])+ # Anything but a colon
|
|
(?: "[^"]")* # A parameter in double quotes
|
|
: # start of the value we really care about
|
|
(.*)$
|
|
/xs';
|
|
|
|
preg_match($regex, $this->rawLine, $matches);
|
|
|
|
$value = $matches[1];
|
|
// Removing the first whitespace character from every line. Kind of
|
|
// like unfolding, but we keep the newline.
|
|
$value = str_replace("\n ", "\n", $value);
|
|
|
|
// Microsoft products don't always correctly fold lines, they may be
|
|
// missing a whitespace. So if 'forgiving' is turned on, we will take
|
|
// those as well.
|
|
if ($this->options & self::OPTION_FORGIVING) {
|
|
while ('=' === substr($value, -1) && $this->lineBuffer) {
|
|
// Reading the line
|
|
$this->readLine();
|
|
// Grabbing the raw form
|
|
$value .= "\n".$this->rawLine;
|
|
}
|
|
}
|
|
|
|
return $value;
|
|
}
|
|
}
|