/**
* @param array $matches A set of results of the `transform` function
* @return string
*/
protected function _email_callback($matches)
{
$address = $matches[1];
Kernel::addConfig('urls', $address);
$block = Kernel::get('OutputFormatBag')->buildTag('link', $address, array('email' => $address));
return parent::hashPart($block);
}
/**
* Process the fenced code blocks
*
* @param array $matches Results form the `transform()` function
* @return string
*/
protected function _callback($matches)
{
$language = $matches[2];
$codeblock = Helper::escapeCodeContent($matches[3]);
$codeblock = preg_replace_callback('/^\\n+/', array($this, '_newlines'), $codeblock);
$attributes = array();
if (!empty($language)) {
$attributes['language'] = $language;
}
$codeblock = Kernel::get('OutputFormatBag')->buildTag('preformatted', $codeblock, $attributes);
return "\n\n" . parent::hashBlock($codeblock) . "\n\n";
}
/**
* @param string $text
* @return string
*/
public function transform($text)
{
return preg_replace('{
^[ ]{0,3} # Leading space
([-*_]) # $1: First marker
(?> # Repeated marker group
[ ]{0,2} # Zero, one, or two spaces.
\\1 # Marker character
){2,} # Group repeated at least twice
[ ]* # Tailing spaces
$ # End of line.
}mx', "\n" . parent::hashBlock(Kernel::get('OutputFormatBag')->buildTag('horizontal_rule')) . "\n", $text);
}
/**
* Build each blockquote block
*
* @param array $matches A set of results of the `transform()` function
* @return string
*/
protected function _callback($matches)
{
$blockq = $matches[1];
$cite = isset($matches[2]) ? $matches[2] : null;
// trim one level of quoting - trim whitespace-only lines
$blockq = preg_replace('/^[ ]*>[ ]?(\\((.+?)\\))?|^[ ]+$/m', '', $blockq);
$blockq = Lexer::runGamut('html_block_gamut', $blockq);
# recurse
$blockq = preg_replace('/^/m', " ", $blockq);
// These leading spaces cause problem with <pre> content,
// so we need to fix that:
$blockq = preg_replace_callback('{(\\s*<pre>.+?</pre>)}sx', array($this, '_callback_spaces'), $blockq);
$attributes = array();
if (!empty($cite)) {
$attributes['cite'] = $cite;
}
$block = Kernel::get('OutputFormatBag')->buildTag('blockquote', $blockq, $attributes);
return "\n" . parent::hashBlock($block) . "\n\n";
}
/**
* Build each maths span
*
* @param string $texblock
* @return string
*/
public function span($texblock)
{
$texblock = trim($texblock);
$block = Kernel::get('OutputFormatBag')->buildTag('maths_span', $texblock, array());
return parent::hashPart($block);
}
/**
* Process ATX-style headers
*
* @param array $matches The results from the `transform()` function
* @return string
*/
protected function _atx_callback($matches)
{
$level = strlen($matches[1]) + $this->_getRebasedHeaderLevel();
$domid = !empty($matches[3]) ? $matches[3] : Helper::header2Label($matches[2]);
$domid = Kernel::get('DomId')->set($domid);
$title = Lexer::runGamut('span_gamut', $matches[2]);
Kernel::addConfig('menu', array('level' => $level, 'text' => parent::unhash($title)), $domid);
$block = Kernel::get('OutputFormatBag')->buildTag('title', $title, array('level' => $level, 'id' => $domid));
$this->_setContentTitle($title);
return "\n" . parent::hashBlock($block) . "\n\n";
}
/**
* Parse markdown text, calling _HashHTMLBlocks_InHTML for block tags.
*
* * $indent is the number of space to be ignored when checking for code
* blocks. This is important because if we don't take the indent into
* account, something like this (which looks right) won't work as expected:
*
* <div>
* <div markdown="1">
* Hello World. <-- Is this a Markdown code block or text?
* </div> <-- Is this a Markdown code block or a real tag?
* <div>
*
* If you don't like this, just don't indent the tag on which
* you apply the markdown="1" attribute.
*
* * If $enclosing_tag_re is not empty, stops at the first unmatched closing
* tag with that name. Nested tags supported.
*
* * If $span is true, text inside must treated as span. So any double
* newline will be replaced by a single newline so that it does not create
* paragraphs.
*
* Returns an array of that form: ( processed text , remaining text )
*
* @param string $text The text to be parsed
* @param int $indent The indentation to use
* @param string $enclosing_tag_re The closing tag to use
* @param bool $span Are we in a span element (false by default)
* @return array ( processed text , remaining text )
*/
protected function _hashBlocks_inMarkdown($text, $indent = 0, $enclosing_tag_re = '', $span = false)
{
if ($text === '') {
return array('', '');
}
// Regex to check for the presense of newlines around a block tag.
$newline_before_re = '/(?:^\\n?|\\n\\n)*$/';
$newline_after_re = '{
^ # Start of text following the tag.
(?>[ ]*<!--.*?-->)? # Optional comment.
[ ]*\\n # Must be followed by newline.
}xs';
// Regex to match any tag.
$block_tag_re = '{
( # $2: Capture hole tag.
</? # Any opening or closing tag.
(?> # Tag name.
' . $this->block_tags_re . ' |
' . $this->blocks_tags_re . ' |
' . $this->clean_tags_re . ' |
(?!\\s)' . $enclosing_tag_re . '
)
(?:
(?=[\\s"\'/a-zA-Z0-9]) # Allowed characters after tag name.
(?>
".*?" | # Double quotes (can contain `>`)
\'.*?\' | # Single quotes (can contain `>`)
.+? # Anything but quotes and `>`.
)*?
)?
> # End of tag.
|
<!-- .*? --> # HTML Comment
|
<\\?.*?\\?> | <%.*?%> # Processing instruction
|
<!\\[CDATA\\[.*?\\]\\]> # CData Block
|
# Code span marker
`+
' . (!$span ? ' # If not in span.
|
# Indented code block
(?: ^[ ]*\\n | ^ | \\n[ ]*\\n )
[ ]{' . ($indent + 4) . '}[^\\n]* \\n
(?>
(?: [ ]{' . ($indent + 4) . '}[^\\n]* | [ ]* ) \\n
)*
|
# Fenced code block marker
(?> ^ | \\n )
[ ]{0,' . $indent . '}~~~+[ ]*\\n
' : '') . ' # End (if not is span).
)
}xs';
$depth = 0;
// Current depth inside the tag tree.
$parsed = "";
// Parsed text that will be returned.
// Loop through every tag until we find the closing tag of the parent
// or loop until reaching the end of text if no parent tag specified.
do {
// Split the text using the first $tag_match pattern found.
// Text before pattern will be first in the array, text after
// pattern will be at the end, and between will be any catches made
// by the pattern.
$parts = preg_split($block_tag_re, $text, 2, PREG_SPLIT_DELIM_CAPTURE);
// If in Markdown span mode, add a empty-string span-level hash
// after each newline to prevent triggering any block element.
if ($span) {
$void = parent::hashPart("", ':');
$newline = "{$void}\n";
$parts[0] = $void . str_replace("\n", $newline, $parts[0]) . $void;
}
$parsed .= $parts[0];
// Text before current tag.
// If end of $text has been reached. Stop loop.
if (count($parts) < 3) {
$text = "";
break;
}
$tag = $parts[1];
// Tag to handle.
$text = $parts[2];
// Remaining text after current tag.
$tag_re = preg_quote($tag);
// For use in a regular expression.
// Check for: Code span marker
if ($tag[0] == "`") {
// Find corresponding end marker.
$tag_re = preg_quote($tag);
// End marker found: pass text unchanged until marker.
if (preg_match('{^(?>.+?|\\n(?!\\n))*?(?<!`)' . $tag_re . '(?!`)}', $text, $matches)) {
$parsed .= $tag . $matches[0];
$text = substr($text, strlen($matches[0]));
// Unmatched marker: just skip it.
} else {
$parsed .= $tag;
}
} elseif (preg_match('{^\\n?[ ]{0,' . ($indent + 3) . '}~}', $tag)) {
// Fenced code block marker: find matching end marker.
$tag_re = preg_quote(trim($tag));
// End marker found: pass text unchanged until marker.
if (preg_match('{^(?>.*\\n)+?[ ]{0,' . $indent . '}' . $tag_re . '[ ]*\\n}', $text, $matches)) {
$parsed .= $tag . $matches[0];
$text = substr($text, strlen($matches[0]));
// No end marker: just skip it.
} else {
$parsed .= $tag;
}
} elseif ($tag[0] == "\n" || $tag[0] == " ") {
// Indented code block: pass it unchanged, will be handled later.
$parsed .= $tag;
} elseif (preg_match('{^<(?:' . $this->block_tags_re . ')\\b}', $tag) || preg_match('{^<(?:' . $this->blocks_tags_re . ')\\b}', $tag) && preg_match($newline_before_re, $parsed) && preg_match($newline_after_re, $text)) {
// Need to parse tag and following text using the HTML parser.
list($block_text, $text) = self::_hashBlocks_inHTML($tag . $text, "hashBlock", true);
// Make sure it stays outside of any paragraph by adding newlines.
$parsed .= "\n\n{$block_text}\n\n";
} elseif (preg_match('{^<(?:' . $this->clean_tags_re . ')\\b}', $tag) || $tag[1] == '!' || $tag[1] == '?') {
// Need to parse tag and following text using the HTML parser.
// (don't check for markdown attribute)
list($block_text, $text) = $this->_hashBlocks_inHTML($tag . $text, "hashClean", false);
$parsed .= $block_text;
} elseif ($enclosing_tag_re !== '' && preg_match('{^</?(?:' . $enclosing_tag_re . ')\\b}', $tag)) {
// Increase/decrease nested tag count.
if ($tag[1] == '/') {
$depth--;
} elseif ($tag[strlen($tag) - 2] != '/') {
$depth++;
}
if ($depth < 0) {
// Going out of parent element. Clean up and break so we
// return to the calling function.
$text = $tag . $text;
break;
}
$parsed .= $tag;
} else {
$parsed .= $tag;
}
} while ($depth >= 0);
return array($parsed, $text);
}
/**
* @param array $matches A set of results of the `transform` function
* @return string
*/
protected function _inline_callback($matches)
{
$alt_text = $matches[2];
$url = $matches[3] == '' ? $matches[4] : $matches[3];
$title =& $matches[7];
$attributes = array();
$attributes['alt'] = Lexer::runGamut(GamutLoader::TOOL_ALIAS . ':EncodeAttribute', $alt_text);
$attributes['src'] = Lexer::runGamut(GamutLoader::TOOL_ALIAS . ':EncodeAttribute', $url);
if (!empty($title)) {
$attributes['title'] = Lexer::runGamut(GamutLoader::TOOL_ALIAS . ':EncodeAttribute', $title);
}
$block = Kernel::get('OutputFormatBag')->buildTag('image', null, $attributes);
return parent::hashPart($block);
}
/**
* Turn double returns into triple returns, so that we can make a
* paragraph for the last item in a list, if necessary
*
* @param array $matches The results form the `transform()` method
* @return string
*/
protected function _callback($matches)
{
// Re-usable patterns to match list item bullets and number markers:
$result = trim(self::transformItems($matches[1]));
$result = str_replace('<!--dt-->', '', $result);
$result = Kernel::get('OutputFormatBag')->buildTag('definition_list', $result);
return parent::hashBlock($result) . "\n\n";
}
/**
* @param array $matches A set of results of the `transform()` function
* @return string
*/
protected function _callback($matches)
{
return parent::hashPart(Kernel::get('OutputFormatBag')->buildTag('new_line') . "\n");
}
/**
* @param array $matches A set of results of the `transform` function
* @return string
*/
protected function _callback($matches)
{
$marker_any_re = '(?:' . self::$marker_ul_re . '|' . self::$marker_ol_re . ')';
$list = $matches[1] . "\n";
$list_type = preg_match('/' . self::$marker_ul_re . '/', $matches[4]) ? "unordered" : "ordered";
$marker_any_re = $list_type == "unordered" ? self::$marker_ul_re : self::$marker_ol_re;
$list = self::transformItems($list, $marker_any_re);
$block = Kernel::get('OutputFormatBag')->buildTag($list_type . '_list', $list);
return "\n" . parent::hashBlock($block) . "\n\n";
}
/**
* Form HTML tables: parses table contents
*
* @param array $matches
* @return string
*/
protected function _callback($matches)
{
$attributes = array();
//self::doDebug($matches);
// The head string may have a begin slash
$caption = count($matches) > 3 ? $matches[1] : null;
$head = count($matches) > 3 ? preg_replace('/^ *[|]/m', '', $matches[2]) : preg_replace('/^ *[|]/m', '', $matches[1]);
$underline = count($matches) > 3 ? $matches[3] : $matches[2];
$content = count($matches) > 3 ? preg_replace('/^ *[|]/m', '', $matches[4]) : preg_replace('/^ *[|]/m', '', $matches[3]);
// Remove any tailing pipes for each line.
$underline = preg_replace('/[|] *$/m', '', $underline);
$content = preg_replace('/[|] *$/m', '', $content);
// Reading alignement from header underline.
$separators = preg_split('/ *[|] */', $underline);
foreach ($separators as $n => $s) {
$attributes[$n] = array();
if (preg_match('/^ *-+: *$/', $s)) {
$attributes[$n]['style'] = 'text-align:right;';
} elseif (preg_match('/^ *:-+: *$/', $s)) {
$attributes[$n]['style'] = 'text-align:center;';
} elseif (preg_match('/^ *:-+ *$/', $s)) {
$attributes[$n]['style'] = 'text-align:left;';
}
}
// Split content by row.
$headers = explode("\n", trim($head, "\n"));
$text = '';
if (!empty($caption)) {
$this->table_id = Helper::header2Label($caption);
$text .= preg_replace_callback('/\\[(.*)\\]/', array($this, '_doCaption'), Lexer::runGamut('span_gamut', $caption));
}
$lines = '';
foreach ($headers as $_header) {
$line = '';
// Parsing span elements, including code spans, character escapes,
// and inline HTML tags, so that pipes inside those gets ignored.
$_header = Lexer::runGamut('filter:Span', $_header);
// Split row by cell.
$_header = preg_replace('/[|] *$/m', '', $_header);
$_headers = preg_split('/[|]/', $_header);
$col_count = count($_headers);
// Write column headers.
// we first loop for colspans
$headspans = array();
foreach ($_headers as $_i => $_cell) {
if ($_cell == '') {
if ($_i == 0) {
$headspans[1] = 2;
} else {
if (isset($headspans[$_i - 1])) {
$headspans[$_i - 1]++;
} else {
$headspans[$_i - 1] = 2;
}
}
}
}
foreach ($_headers as $n => $__header) {
if ($__header != '') {
$cell_attributes = $attributes[$n];
if (isset($headspans[$n])) {
$cell_attributes['colspan'] = $headspans[$n];
}
$line .= Kernel::get('OutputFormatBag')->buildTag('table_cell_head', Lexer::runGamut('span_gamut', trim($__header)), $cell_attributes) . "\n";
}
}
$lines .= Kernel::get('OutputFormatBag')->buildTag('table_line', $line) . "\n";
}
$text .= Kernel::get('OutputFormatBag')->buildTag('table_header', $lines);
// Split content by row.
$rows = explode("\n", trim($content, "\n"));
$lines = '';
foreach ($rows as $row) {
$line = '';
// Parsing span elements, including code spans, character escapes,
// and inline HTML tags, so that pipes inside those gets ignored.
$row = Lexer::runGamut('filter:Span', $row);
// Split row by cell.
$row_cells = preg_split('/ *[|] */', $row, $col_count);
$row_cells = array_pad($row_cells, $col_count, '');
// we first loop for colspans
$colspans = array();
foreach ($row_cells as $_i => $_cell) {
if ($_cell == '') {
if ($_i == 0) {
$colspans[1] = 2;
} else {
if (isset($colspans[$_i - 1])) {
$colspans[$_i - 1]++;
} else {
$colspans[$_i - 1] = 2;
}
}
}
}
foreach ($row_cells as $n => $cell) {
if ($cell != '') {
$cell_attributes = $attributes[$n];
if (isset($colspans[$n])) {
$cell_attributes['colspan'] = $colspans[$n];
}
$line .= Kernel::get('OutputFormatBag')->buildTag('table_cell', Lexer::runGamut('span_gamut', trim($cell)), $cell_attributes) . "\n";
}
}
$lines .= Kernel::get('OutputFormatBag')->buildTag('table_line', $line) . "\n";
}
$text .= Kernel::get('OutputFormatBag')->buildTag('table_body', $lines);
$table = Kernel::get('OutputFormatBag')->buildTag('table', $text);
return parent::hashBlock($table) . "\n";
}
/**
* Create a code span markup for $code. Called from handleSpanToken.
*
* @param string $code
* @return string
*/
public function span($code)
{
$codeblock = Kernel::get('OutputFormatBag')->buildTag('code', Helper::escapeCodeContent(trim($code)));
return parent::hashPart($codeblock);
}
/**
* Handle $token provided by parseSpan by determining its nature and
* returning the corresponding value that should replace it.
*
* @param string $token
* @param string $str
* @return string
*/
public function handleSpanToken($token, &$str)
{
switch ($token[0]) {
case "\\":
if ($token[1] == "(") {
$texend = strpos($str, '\\)');
if ($texend) {
$eqn = substr($str, 0, $texend);
$str = substr($str, $texend + 2);
$texspan = Lexer::runGamut('filter:Maths:span', $eqn);
return parent::hashPart($texspan);
} else {
return $str;
}
} else {
return parent::hashPart("&#" . ord($token[1]) . ";");
}
case "`":
// Search for end marker in remaining text.
if (preg_match('/^(.*?[^`])' . preg_quote($token) . '(?!`)(.*)$/sm', $str, $matches)) {
$str = $matches[2];
$codespan = Lexer::runGamut('filter:CodeBlock:span', $matches[1], true);
return parent::hashPart($codespan);
}
return $token;
// return as text since no ending marker found.
// return as text since no ending marker found.
default:
return parent::hashPart($token);
}
}
/**
* @param string $text
* @return string
*/
public function transform($text)
{
$token_stack = array('');
$text_stack = array('');
$italic = '';
$strong = '';
$tree_char_em = false;
while (1) {
// Get prepared regular expression for seraching emphasis tokens in current context.
$token_re = self::$em_strong_prepared["{$italic}{$strong}"];
// Each loop iteration search for the next emphasis token.
// Each token is then passed to handleSpanToken.
$parts = preg_split($token_re, $text, 2, PREG_SPLIT_DELIM_CAPTURE);
$text_stack[0] .= $parts[0];
$token =& $parts[1];
$text =& $parts[2];
if (empty($token)) {
// Reached end of text span: empty stack without emitting any more emphasis.
while ($token_stack[0]) {
$text_stack[1] .= array_shift($token_stack);
$text_stack[0] .= array_shift($text_stack);
}
break;
}
$token_len = strlen($token);
if ($tree_char_em) {
// Reached closing marker while inside a three-char emphasis.
if ($token_len == 3) {
// Three-char closing marker, close em and strong.
array_shift($token_stack);
$span = Lexer::runGamut('span_gamut', array_shift($text_stack));
$span = Kernel::get('OutputFormatBag')->buildTag('italic', $span);
$span = Kernel::get('OutputFormatBag')->buildTag('bold', $span);
$text_stack[0] .= parent::hashPart($span);
$italic = '';
$strong = '';
} else {
// Other closing marker: close one em or strong and
// change current token state to match the other
$token_stack[0] = str_repeat($token[0], 3 - $token_len);
$tag = $token_len == 2 ? "bold" : "italic";
$span = Lexer::runGamut('span_gamut', $text_stack[0]);
$span = Kernel::get('OutputFormatBag')->buildTag($tag, $span);
$text_stack[0] = parent::hashPart($span);
${$tag} = '';
// $$tag stands for $italic or $strong
}
$tree_char_em = false;
} elseif ($token_len == 3) {
if ($italic) {
// Reached closing marker for both em and strong.
// Closing strong marker:
for ($i = 0; $i < 2; ++$i) {
$shifted_token = array_shift($token_stack);
$tag = strlen($shifted_token) == 2 ? "bold" : "italic";
$span = Lexer::runGamut('span_gamut', array_shift($text_stack));
$span = Kernel::get('OutputFormatBag')->buildTag($tag, $span);
$text_stack[0] .= parent::hashPart($span);
${$tag} = '';
// $$tag stands for $italic or $strong
}
} else {
// Reached opening three-char emphasis marker. Push on token
// stack; will be handled by the special condition above.
$italic = $token[0];
$strong = "{$italic}{$italic}";
array_unshift($token_stack, $token);
array_unshift($text_stack, '');
$tree_char_em = true;
}
} elseif ($token_len == 2) {
if ($strong) {
// Unwind any dangling emphasis marker:
if (strlen($token_stack[0]) == 1) {
$text_stack[1] .= array_shift($token_stack);
$text_stack[0] .= array_shift($text_stack);
}
// Closing strong marker:
array_shift($token_stack);
$span = Lexer::runGamut('span_gamut', array_shift($text_stack));
$span = Kernel::get('OutputFormatBag')->buildTag('bold', $span);
$text_stack[0] .= parent::hashPart($span);
$strong = '';
} else {
array_unshift($token_stack, $token);
array_unshift($text_stack, '');
$strong = $token;
}
} else {
// Here $token_len == 1
if ($italic) {
if (strlen($token_stack[0]) == 1) {
// Closing emphasis marker:
array_shift($token_stack);
$span = Lexer::runGamut('span_gamut', array_shift($text_stack));
$span = Kernel::get('OutputFormatBag')->buildTag('italic', $span);
$text_stack[0] .= parent::hashPart($span);
$italic = '';
} else {
$text_stack[0] .= $token;
}
} else {
array_unshift($token_stack, $token);
array_unshift($text_stack, '');
$italic = $token;
}
}
}
return $text_stack[0];
}
/**
* Process each abbreviation
*
* @param array $matches One set of results form the `transform()` function
* @return string
*/
protected function _callback($matches)
{
$abbr = $matches[0];
$abbr_desciptions = Kernel::getConfig('abbr_desciptions');
if (isset($abbr_desciptions[$abbr])) {
$attributes = array();
$desc = trim($abbr_desciptions[$abbr]);
if (!empty($desc)) {
$attributes['title'] = Lexer::runGamut(GamutLoader::TOOL_ALIAS . ':EncodeAttribute', $desc);
}
$abbr = Kernel::get('OutputFormatBag')->buildTag('abbreviation', $abbr, $attributes);
return parent::hashBlock($abbr);
} else {
return $abbr;
}
}
