/**
* Recieves parse data from the context tree. Sends the
* data on to the code parser, then to the renderer for
* building the result string
*/
function addParseData($code, $context_name, array $data = array(), $complex = false)
{
// @todo [blocking 1.1.5] test this, esp. not passing back anything and passing back multiple
// can use PHP code parser for this
// @todo [blocking 1.1.9] since we are only looking for whitespace at start and end this can
// be optimised
if (GESHI_COMPLEX_NO == $complex) {
$this->_addToParsedCode(array($code, $context_name, $data));
} elseif (GESHI_COMPLEX_PASSALL == $complex) {
// Parse all at once
$this->_addToParsedCode($this->_codeParser->parseToken($code, $context_name, $data));
} elseif (GESHI_COMPLEX_TOKENISE == $complex) {
$matches = array();
preg_match_all('/^(\\s*)(.*?)(\\s*)$/si', $code, $matches);
//echo 'START<br />';
//print_r($matches);
if ($matches[1][0]) {
$this->_addToParsedCode($this->_codeParser->parseToken($matches[1][0], $context_name, $data));
}
if ('' != $matches[2][0]) {
while ('' != $matches[2][0]) {
$pos = geshi_get_position($matches[2][0], 'REGEX#(\\s+)#');
if (false !== $pos['pos']) {
// Parse the token up to the whitespace
//echo 'code: |' . substr($matches[2][0], 0, $pos['pos']) . '|<br />';
$this->_addToParsedCode($this->_codeParser->parseToken(substr($matches[2][0], 0, $pos['pos']), $context_name, $data));
// Parse the whitespace
//echo 'ws: |' . substr($matches[2][0], $pos['pos'], $pos['len']) . '|<br />';
$this->_addToParsedCode($this->_codeParser->parseToken(substr($matches[2][0], $pos['pos'], $pos['len']), $context_name, $data));
// Trim what we just parsed
$matches[2][0] = substr($matches[2][0], $pos['pos'] + $pos['len']);
} else {
// No more whitespace
//echo 'no more whitespace: |' . $matches[2][0] . '<br />';
$this->_addToParsedCode($this->_codeParser->parseToken($matches[2][0], $context_name, $data));
break;
}
}
}
if ($matches[3][0]) {
$this->_addToParsedCode($this->_codeParser->parseToken($matches[3][0], $context_name, $data));
}
}
// else wtf???
}
/**
* Adds code detected as being in this context to the parse data
*/
function _addParseData($code, $first_char_of_next_context = '')
{
$parent_name = $this->_parentName;
geshi_dbg('GeSHiPHPDoubleStringContext::_addParseData(' . substr($code, 0, 15) . '...)');
while (true) {
$earliest_data = array('pos' => false, 'len' => 0);
foreach ($this->_regexes as $regex) {
$data = geshi_get_position($code, $regex, 0, false, true);
// request table
if (false != $data['pos'] && false === $earliest_data['pos'] || false !== $data['pos'] && ($data['pos'] < $earliest_data['pos'] || $data['pos'] == $earliest_data['pos'] && $data['len'] > $earliest_data['len'])) {
$earliest_data = $data;
}
}
if (false === $earliest_data['pos']) {
// No more variables in this string
break;
}
// bugfix: because we match a var, it might have been escaped.
// so only do to -1 so we can catch slash if it has been
$pos = $earliest_data['pos'] ? $earliest_data['pos'] - 1 : 0;
$len = $earliest_data['pos'] ? $earliest_data['len'] + 1 : $earliest_data['len'];
parent::_addParseData(substr($code, 0, $pos));
// Now the entire possible var is in:
$possible_var = substr($code, $pos, $len);
geshi_dbg('Found variable at position ' . $earliest_data['pos'] . '(' . $possible_var . ')');
// Check that the dollar sign that started this variable was not escaped
//$first_part = str_replace('\\\\', '', substr($code, 0, $pos));
//if ('\\' == substr($first_part, -1)) {
// If \\ before var and { is not next character after that...
if ('\\' == substr($possible_var, 0, 1) && '{' != substr($possible_var, 1, 1)) {
// This variable has been escaped, so add the escaped dollar sign
// as the correct context, and the rest of the variable (recurse to catch
// other variables inside this possible variable)
geshi_dbg('Variable was escaped');
$this->_styler->addParseData(substr($possible_var, 0, 2), $parent_name . '/esc', $this->_getExtraParseData(), $this->_complexFlag);
$this->_addParseData(substr($possible_var, 2));
} else {
// Add first character that might have been a \\ but in fact isn't to the parent
// but only do it if we had to modify the position
if ('$' != substr($possible_var, 0, 1)) {
parent::_addParseData(substr($possible_var, 0, 1));
$possible_var = substr($possible_var, 1);
}
// Many checks could go in here...
// @todo [blocking 1.1.5] check for ${foo} variables: start { matched by end }
// because at the moment ${foo is matched for example.
if ('{' == substr($possible_var, 0, 1)) {
if ('}' == substr($possible_var, -1)) {
$start_brace = '{';
} else {
$start_brace = '';
parent::_addParseData('{');
// remove brace from $possible_var. This will only be used
// if the variable isn't an OO variable anyway...
$possible_var = substr($possible_var, 1);
}
} else {
$start_brace = '';
}
if (isset($earliest_data['tab'][5])) {
// Then we matched off the third regex - the one that does objects
// The first { if there is one, and $this (which is in index 2
$this->_styler->addParseData($start_brace . $earliest_data['tab'][2], $parent_name . '/var', $this->_getExtraParseData(), $this->_complexFlag);
// The -> with any whitespace around it
$this->_styler->addParseData($earliest_data['tab'][3], $parent_name . '/symbol', $this->_getExtraParseData(), $this->_complexFlag);
// The method name
$this->_styler->addParseData($earliest_data['tab'][4], $parent_name . '/oodynamic', $this->_getExtraParseData(), $this->_complexFlag);
// The closing }, if any
if ($earliest_data['tab'][5]) {
if ($start_brace) {
$this->_styler->addParseData($earliest_data['tab'][5], $parent_name . '/var', $this->_getExtraParseData(), $this->_complexFlag);
} else {
parent::_addParseData('}');
}
}
} else {
$this->_styler->addParseData($possible_var, $parent_name . '/var', $this->_getExtraParseData(), $this->_complexFlag);
}
}
// Chop off what we have done
$code = substr($code, $earliest_data['pos'] + $earliest_data['len']);
}
// Add the rest
parent::_addParseData($code, $first_char_of_next_context);
}
/**
* Checks whether the character(s) at the start of the parameter string are
* characters that should be escaped.
*
* @param string The string to check the beginning of for escape characters
* @return int|false The length of the escape character sequence, else false
*/
function _shouldBeEscaped($code, $chars_to_escape)
{
geshi_dbg('Checking: ' . substr($code, 0, 15));
foreach ($chars_to_escape as $match) {
if ('REGEX' != substr($match, 0, 5)) {
geshi_dbg('Test: ' . $match);
if (substr($code, 0, 1) == $match) {
return 1;
}
} else {
geshi_dbg(' Testing via regex: ' . $match . '... ', false);
$data = geshi_get_position($code, $match, 0);
if (0 === $data['pos']) {
geshi_dbg('match, data = ' . print_r($data, true));
return $data['len'];
}
geshi_dbg('no match');
}
}
// No matches...
return false;
}
/**
* GetContextEndData
*/
function _getContextEndData($code, $context_open_key, $context_opener, $beginning_of_context)
{
geshi_dbg('GeSHiContext::_getContextEndData(' . $this->_contextName . ', ' . $context_open_key . ', ' . $context_opener . ', ' . $beginning_of_context . ')');
$context_end_pos = false;
$context_end_len = -1;
$context_end_dlm = '';
$offset = 0;
// Bail out if context open key tells us that there is no ender for this context
if (-1 == $context_open_key) {
geshi_dbg(' no opener so no ender');
return false;
}
// Balanced endings is handled here
if (isset($this->_contextDelimiters[$context_open_key][3])) {
$balance_opener = $this->_contextDelimiters[$context_open_key][3][0];
$balance_closer = $this->_contextDelimiters[$context_open_key][3][1];
// We get the first push for free
// @todo [blocking 1.1.4] if what we are balancing against is not related
// to the starter of the context then we have a problem... check $context_opener
// for starter stuff instead of assuming
$balance_count = 1;
geshi_dbg('@w Begun balancing');
while ($balance_count > 0) {
// Look for opener/closers.
$opener_pos = geshi_get_position($code, $balance_opener, $offset);
$closer_pos = geshi_get_position($code, $balance_closer, $offset);
geshi_dbg(' opener pos = ' . print_r($opener_pos, true) . ', closer pos = ' . print_r($closer_pos, true));
// Check what we found
if (false !== $opener_pos['pos']) {
if (false !== $closer_pos['pos']) {
// Opener and closer available
if ($opener_pos['pos'] < $closer_pos['pos']) {
// Opener is closer so inc. counter
++$balance_count;
geshi_dbg(' opener is closer so inc. to ' . $balance_count);
// Start searching from new pos just past where we found the opener
$offset = $opener_pos['pos'] + 1;
// @todo [blocking 1.1.4] could cache closer pos at this point?
} else {
// closer is closer (bad english heh)
--$balance_count;
$offset = $closer_pos['pos'] + 1;
geshi_dbg(' closer is closer so dec. to ' . $balance_count);
}
} else {
// No closer will ever be available yet we are still in this context...
// use end of code as end pos
// I've yet to test this case
geshi_dbg('@w No closer but still in this context!');
return array('pos' => strlen($code), 'len' => 0, 'dlm' => '');
}
} elseif (false !== $closer_pos['pos']) {
// No opener but closer. Nothing wrong with this
--$balance_count;
$offset = $closer_pos['pos'] + 1;
geshi_dbg(' only closer left, dec. to ' . $balance_count);
} else {
// No opener or closer
// Assume that we end this context at the end of the code, with
// no delimiter
geshi_dbg('@w No opener or closer but still in this context!');
return array('pos' => strlen($code), 'len' => 0, 'dlm' => '');
}
}
// start looking for real end from the position where balancing ends
// because we've found where balancing ends, but the end of the balancing
// is likely to be the same as the end of the context
--$offset;
}
foreach ($this->_contextDelimiters[$context_open_key][1] as $ender) {
geshi_dbg(' Checking ender: ' . str_replace("\n", '\\n', $ender), false);
$ender = $this->_substitutePlaceholders($ender);
geshi_dbg(' converted to ' . $ender);
// Use the offset we may have found when handling balancing of contexts (will
// be zero if balancing not done).
$position = geshi_get_position($code, $ender, $offset);
geshi_dbg(' Ender ' . $ender . ': ' . print_r($position, true));
$length = $position['len'];
$position = $position['pos'];
// BUGFIX:skip around crap starters
if (false === $position) {
continue;
}
if (false === $context_end_pos || $position < $context_end_pos || $position == $context_end_pos && strlen($ender) > $context_end_len) {
$context_end_pos = $position;
$context_end_len = $length;
$context_end_dlm = $ender;
}
}
geshi_dbg('Context ' . $this->_contextName . ' can finish at position ' . $context_end_pos);
if (false !== $context_end_pos) {
return array('pos' => $context_end_pos, 'len' => $context_end_len, 'dlm' => $context_end_dlm);
} else {
return false;
}
}
