s around # "paragraphs" that are wrapped in non-block-level tags, such as anchors, # phrase emphasis, and spans. The list of tags we're looking for is # hard-coded: # # * List "a" is made of tags which can be both inline or block-level. # These will be treated block-level when the start tag is alone on # its line, otherwise they're not matched here and will be taken as # inline later. # * List "b" is made of tags which are always block-level; # $block_tags_a = 'ins|del'; $block_tags_b = 'p|div|h[1-6]|blockquote|pre|table|dl|ol|ul|address|'. 'script|noscript|form|fieldset|iframe|math'; # Regular expression for the content of a block tag. $nested_tags_level = 4; $attr = ' (?> # optional tag attributes \s # starts with whitespace (?> [^>"/]+ # text outside quotes | /+(?!>) # slash not followed by ">" | "[^"]*" # text inside double quotes (tolerate ">") | \'[^\']*\' # text inside single quotes (tolerate ">") )* )? '; $content = str_repeat(' (?> [^<]+ # content without tag | <\2 # nested opening tag '.$attr.' # attributes (?> /> | >', $nested_tags_level). # end of opening tag '.*?'. # last level nested tag content str_repeat(' # closing nested tag ) | <(?!/\2\s*> # other tags with a different name ) )*', $nested_tags_level); $content2 = str_replace('\2', '\3', $content); # First, look for nested blocks, e.g.: #
#
#
# The outermost tags must start at the left margin for this to match, and
# the inner nested divs must be indented.
# We need to do this before the next, more liberal match, because the next
# match will start at the first `
# tags for inner block must be indented.
#
# ` and stop at the first `
`.
$text = preg_replace_callback('{(?>
(?>
(?<=\n\n) # Starting after a blank line
| # or
\A\n? # the beginning of the doc
)
( # save in $1
# Match from `\n. It was easier to make a special # case than to make the other regex more complicated. [ ]{0,'.$less_than_tab.'} <(hr) # start tag = $2 \b # word break ([^<>])*? # /?> # the matching end tag [ ]* (?=\n{2,}|\Z) # followed by a blank line or end of document | # Special case for standalone HTML comments: [ ]{0,'.$less_than_tab.'} (?s: ) [ ]* (?=\n{2,}|\Z) # followed by a blank line or end of document | # PHP and ASP-style processor instructions ( ) [ ]* (?=\n{2,}|\Z) # followed by a blank line or end of document ) )}Sxmi', array(&$this, '_hashHTMLBlocks_callback'), $text); return $text; } function _hashHTMLBlocks_callback($matches) { $text = $matches[1]; $key = $this->hashBlock($text); return "\n\n$key\n\n"; } function hashPart($text, $boundary = 'X') { # # Called whenever a tag must be hashed when a function insert an atomic # element in the text stream. Passing $text to through this function gives # a unique text-token which will be reverted back when calling unhash. # # The $boundary argument specify what character should be used to surround # the token. By convension, "B" is used for block elements that needs not # to be wrapped into paragraph tags at the end, ":" is used for elements # that are word separators and "S" is used for general span-level elements. # # Swap back any tag hash found in $text so we do not have to `unhash` # multiple times at the end. $text = $this->unhash($text); # Then hash the block. static $i = 0; $key = "$boundary\x1A" . ++$i . $boundary; $this->html_hashes[$key] = $text; return $key; # String that will replace the tag. } function hashBlock($text) { # # Shortcut function for hashPart with block-level boundaries. # return $this->hashPart($text, 'B'); } var $block_gamut = array( # # These are all the transformations that form block-level # tags like paragraphs, headers, and list items. # "doHeaders" => 10, "doHorizontalRules" => 20, "doLists" => 40, "doCodeBlocks" => 50, "doBlockQuotes" => 60, ); function runBlockGamut($text) { # # Run block gamut tranformations. # # We need to escape raw HTML in Markdown source before doing anything # else. This need to be done for each block, and not only at the # begining in the Markdown function since hashed blocks can be part of # list items and could have been indented. Indented blocks would have # been seen as a code block in a previous pass of hashHTMLBlocks. $text = $this->hashHTMLBlocks($text); return $this->runBasicBlockGamut($text); } function runBasicBlockGamut($text) { # # Run block gamut tranformations, without hashing HTML blocks. This is # useful when HTML blocks are known to be already hashed, like in the first # whole-document pass. # foreach ($this->block_gamut as $method => $priority) { $text = $this->$method($text); } # Finally form paragraph and restore hashed blocks. $text = $this->formParagraphs($text); return $text; } function doHorizontalRules($text) { # Do Horizontal Rules: 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".$this->hashBlock("
` blocks.
#
$text = preg_replace_callback('{
(?:\n\n|\A)
( # $1 = the code block -- one or more lines, starting with a space/tab
(?>
[ ]{'.$this->tab_width.'} # Lines must start with a tab or a tab-width of spaces
.*\n+
)+
)
((?=^[ ]{0,'.$this->tab_width.'}\S)|\Z) # Lookahead for non-space at line-start, or end of doc
}xm',
array(&$this, '_doCodeBlocks_callback'), $text);
return $text;
}
function _doCodeBlocks_callback($matches) {
$codeblock = $matches[1];
$codeblock = $this->outdent($codeblock);
$codeblock = htmlspecialchars($codeblock, ENT_NOQUOTES);
# trim leading newlines and trailing newlines
$codeblock = preg_replace('/\A\n+|\n+\z/', '', $codeblock);
$codeblock = "$codeblock\n
";
return "\n\n".$this->hashBlock($codeblock)."\n\n";
}
function makeCodeSpan($code) {
#
# Create a code span markup for $code. Called from handleSpanToken.
#
$code = htmlspecialchars(trim($code), ENT_NOQUOTES);
return $this->hashPart("$code");
}
function doItalicsAndBold($text) {
# must go first:
$text = preg_replace_callback('{
( # $1: Marker
(?
[^*_]+? # Anthing not em markers.
|
# Balence any regular emphasis inside.
\1 (?=\S) .+? (?<=\S) \1
|
. # Allow unbalenced * and _.
)+?
)
(?<=\S) \1\1 # End mark not preceded by whitespace.
}sx',
array(&$this, '_doItalicAndBold_strong_callback'), $text);
# Then :
$text = preg_replace_callback(
'{ ( (?runSpanGamut($text);
return $this->hashPart("$text");
}
function _doItalicAndBold_strong_callback($matches) {
$text = $matches[2];
$text = $this->runSpanGamut($text);
return $this->hashPart("$text");
}
function doBlockQuotes($text) {
$text = preg_replace_callback('/
( # Wrap whole match in $1
(?>
^[ ]*>[ ]? # ">" at the start of a line
.+\n # rest of the first line
(.+\n)* # subsequent consecutive lines
\n* # blanks
)+
)
/xm',
array(&$this, '_doBlockQuotes_callback'), $text);
return $text;
}
function _doBlockQuotes_callback($matches) {
$bq = $matches[1];
# trim one level of quoting - trim whitespace-only lines
$bq = preg_replace('/^[ ]*>[ ]?|^[ ]+$/m', '', $bq);
$bq = $this->runBlockGamut($bq); # recurse
$bq = preg_replace('/^/m', " ", $bq);
# These leading spaces cause problem with content,
# so we need to fix that:
$bq = preg_replace_callback('{(\s*.+?
)}sx',
array(&$this, '_DoBlockQuotes_callback2'), $bq);
return "\n". $this->hashBlock("\n$bq\n
")."\n\n";
}
function _doBlockQuotes_callback2($matches) {
$pre = $matches[1];
$pre = preg_replace('/^ /m', '', $pre);
return $pre;
}
function formParagraphs($text) {
#
# Params:
# $text - string to process with html tags
#
# Strip leading and trailing lines:
$text = preg_replace('/\A\n+|\n+\z/', '', $text);
$grafs = preg_split('/\n{2,}/', $text, -1, PREG_SPLIT_NO_EMPTY);
#
# Wrap
tags and unhashify HTML blocks
#
foreach ($grafs as $key => $value) {
if (!preg_match('/^B\x1A[0-9]+B$/', $value)) {
# Is a paragraph.
$value = $this->runSpanGamut($value);
$value = preg_replace('/^([ ]*)/', "
", $value);
$value .= "
";
$grafs[$key] = $this->unhash($value);
}
else {
# Is a block.
# Modify elements of @grafs in-place...
$graf = $value;
$block = $this->html_hashes[$graf];
$graf = $block;
// if (preg_match('{
// \A
// ( # $1 = tag
// ]*
// \b
// markdown\s*=\s* ([\'"]) # $2 = attr quote char
// 1
// \2
// [^>]*
// >
// )
// ( # $3 = contents
// .*
// )
// () # $4 = closing tag
// \z
// }xs', $block, $matches))
// {
// list(, $div_open, , $div_content, $div_close) = $matches;
//
// # We can't call Markdown(), because that resets the hash;
// # that initialization code should be pulled into its own sub, though.
// $div_content = $this->hashHTMLBlocks($div_content);
//
// # Run document gamut methods on the content.
// foreach ($this->document_gamut as $method => $priority) {
// $div_content = $this->$method($div_content);
// }
//
// $div_open = preg_replace(
// '{\smarkdown\s*=\s*([\'"]).+?\1}', '', $div_open);
//
// $graf = $div_open . "\n" . $div_content . "\n" . $div_close;
// }
$grafs[$key] = $graf;
}
}
return implode("\n\n", $grafs);
}
function encodeAmpsAndAngles($text) {
# Smart processing for ampersands and angle brackets that need to be encoded.
if ($this->no_entities) {
$text = str_replace('&', '&', $text);
$text = str_replace('<', '<', $text);
return $text;
}
# Ampersand-encoding based entirely on Nat Irons's Amputator MT plugin:
# http://bumppo.net/projects/amputator/
$text = preg_replace('/&(?!#?[xX]?(?:[0-9a-fA-F]+|\w+);)/',
'&', $text);;
# Encode naked <'s
$text = preg_replace('{<(?![a-z/?\$!%])}i', '<', $text);
return $text;
}
function doAutoLinks($text) {
$text = preg_replace_callback('{<((https?|ftp|dict):[^\'">\s]+)>}',
array(&$this, '_doAutoLinks_url_callback'), $text);
# Email addresses:
$text = preg_replace_callback('{
<
(?:mailto:)?
(
[-.\w\x80-\xFF]+
\@
[-a-z0-9\x80-\xFF]+(\.[-a-z0-9\x80-\xFF]+)*\.[a-z]+
)
>
}xi',
array(&$this, '_doAutoLinks_email_callback'), $text);
return $text;
}
function _doAutoLinks_url_callback($matches) {
$url = $this->encodeAmpsAndAngles($matches[1]);
$link = "$url";
return $this->hashPart($link);
}
function _doAutoLinks_email_callback($matches) {
$address = $matches[1];
$link = $this->encodeEmailAddress($address);
return $this->hashPart($link);
}
function encodeEmailAddress($addr) {
#
# Input: an email address, e.g. "foo@example.com"
#
# Output: the email address as a mailto link, with each character
# of the address encoded as either a decimal or hex entity, in
# the hopes of foiling most address harvesting spam bots. E.g.:
#
#
#
# Based by a filter by Matthew Wickline, posted to BBEdit-Talk.
# With some optimizations by Milian Wolff.
#
$addr = "mailto:" . $addr;
$chars = preg_split('/(? $char) {
$ord = ord($char);
# Ignore non-ascii chars.
if ($ord < 128) {
$r = ($seed * (1 + $key)) % 100; # Pseudo-random function.
# roughly 10% raw, 45% hex, 45% dec
# '@' *must* be encoded. I insist.
if ($r > 90 && $char != '@') /* do nothing */;
else if ($r < 45) $chars[$key] = '&#x'.dechex($ord).';';
else $chars[$key] = '&#'.$ord.';';
}
}
$addr = implode('', $chars);
$text = implode('', array_slice($chars, 7)); # text without `mailto:`
$addr = "$text";
return $addr;
}
function parseSpan($str) {
#
# Take the string $str and parse it into tokens, hashing embeded HTML,
# escaped characters and handling code spans.
#
$output = '';
$regex = '{
(
\\\\['.preg_quote($this->escape_chars).']
|
(?no_markup ? '' : '
|
# comment
|
<\?.*?\?> | <%.*?%> # processing instruction
|
<[/!$]?[-a-zA-Z0-9:]+ # regular tags
(?>
\s
(?>[^"\'>]+|"[^"]*"|\'[^\']*\')*
)?
>
').'
)
}xs';
while (1) {
#
# Each loop iteration seach for either the next tag, the next
# openning code span marker, or the next escaped character.
# Each token is then passed to handleSpanToken.
#
$parts = preg_split($regex, $str, 2, PREG_SPLIT_DELIM_CAPTURE);
# Create token from text preceding tag.
if ($parts[0] != "") {
$output .= $parts[0];
}
# Check if we reach the end.
if (isset($parts[1])) {
$output .= $this->handleSpanToken($parts[1], $parts[2]);
$str = $parts[2];
}
else {
break;
}
}
return $output;
}
function handleSpanToken($token, &$str) {
#
# Handle $token provided by parseSpan by determining its nature and
# returning the corresponding value that should replace it.
#
switch ($token{0}) {
case "\\":
return $this->hashPart("&#". ord($token{1}). ";");
case "`":
# Search for end marker in remaining text.
if (preg_match('/^(.*?[^`])'.$token.'(?!`)(.*)$/sm',
$str, $matches))
{
$str = $matches[2];
$codespan = $this->makeCodeSpan($matches[1]);
return $this->hashPart($codespan);
}
return $token; // return as text since no ending marker found.
default:
return $this->hashPart($token);
}
}
function outdent($text) {
#
# Remove one level of line-leading tabs or spaces
#
return preg_replace('/^(\t|[ ]{1,'.$this->tab_width.'})/m', '', $text);
}
# String length function for detab. `_initDetab` will create a function to
# hanlde UTF-8 if the default function does not exist.
var $utf8_strlen = 'mb_strlen';
function detab($text) {
#
# Replace tabs with the appropriate amount of space.
#
# For each line we separate the line in blocks delemited by
# tab characters. Then we reconstruct every line by adding the
# appropriate number of space between each blocks.
$text = preg_replace_callback('/^.*\t.*$/m',
array(&$this, '_detab_callback'), $text);
return $text;
}
function _detab_callback($matches) {
$line = $matches[0];
$strlen = $this->utf8_strlen; # strlen function for UTF-8.
# Split in blocks.
$blocks = explode("\t", $line);
# Add each blocks to the line.
$line = $blocks[0];
unset($blocks[0]); # Do not add first block twice.
foreach ($blocks as $block) {
# Calculate amount of space, insert spaces, insert block.
$amount = $this->tab_width -
$strlen($line, 'UTF-8') % $this->tab_width;
$line .= str_repeat(" ", $amount) . $block;
}
return $line;
}
function _initDetab() {
#
# Check for the availability of the function in the `utf8_strlen` property
# (initially `mb_strlen`). If the function is not available, create a
# function that will loosely count the number of UTF-8 characters with a
# regular expression.
#
if (function_exists($this->utf8_strlen)) return;
$this->utf8_strlen = create_function('$text', 'return preg_match_all(
"/[\\\\x00-\\\\xBF]|[\\\\xC0-\\\\xFF][\\\\x80-\\\\xBF]*/",
$text, $m);');
}
function unhash($text) {
#
# Swap back in all the tags hashed by _HashHTMLBlocks.
#
return preg_replace_callback('/(.)\x1A[0-9]+\1/',
array(&$this, '_unhash_callback'), $text);
}
function _unhash_callback($matches) {
return $this->html_hashes[$matches[0]];
}
}
#
# Markdown Extra Parser Class
#
class MarkdownExtra_Parser extends Markdown_Parser {
# Prefix for footnote ids.
var $fn_id_prefix = "";
# Optional title attribute for footnote links and backlinks.
var $fn_link_title = MARKDOWN_FN_LINK_TITLE;
var $fn_backlink_title = MARKDOWN_FN_BACKLINK_TITLE;
# Optional class attribute for footnote links and backlinks.
var $fn_link_class = MARKDOWN_FN_LINK_CLASS;
var $fn_backlink_class = MARKDOWN_FN_BACKLINK_CLASS;
function MarkdownExtra_Parser() {
#
# Constructor function. Initialize the parser object.
#
# Add extra escapable characters before parent constructor
# initialize the table.
$this->escape_chars .= ':|';
# Insert extra document, block, and span transformations.
# Parent constructor will do the sorting.
$this->document_gamut += array(
"stripFootnotes" => 15,
"stripAbbreviations" => 25,
"appendFootnotes" => 50,
);
$this->block_gamut += array(
"doTables" => 15,
"doDefLists" => 45,
);
$this->span_gamut += array(
"doFootnotes" => 5,
"doAbbreviations" => 70,
);
parent::Markdown_Parser();
}
# Extra hashes used during extra transformations.
var $footnotes = array();
var $footnotes_ordered = array();
var $abbr_desciptions = array();
var $abbr_matches = array();
# Status flag to avoid invalid nesting.
var $in_footnote = false;
function transform($text) {
#
# Added clear to the new $html_hashes, reordered `hashHTMLBlocks` before
# blank line stripping and added extra parameter to `runBlockGamut`.
#
# Clear the global hashes. If we don't clear these, you get conflicts
# from other articles when generating a page which contains more than
# one article (e.g. an index page that shows the N most recent
# articles):
$this->footnotes = array();
$this->footnotes_ordered = array();
$this->abbr_desciptions = array();
$this->abbr_matches = array();
return parent::transform($text);
}
### HTML Block Parser ###
# Tags that are always treated as block tags:
var $block_tags = 'p|div|h[1-6]|blockquote|pre|table|dl|ol|ul|address|form|fieldset|iframe|hr|legend';
# Tags treated as block tags only if the opening tag is alone on it's line:
var $context_block_tags = 'script|noscript|math|ins|del';
# Tags where markdown="1" default to span mode:
var $contain_span_tags = 'p|h[1-6]|li|dd|dt|td|th|legend|address';
# Tags which must not have their contents modified, no matter where
# they appear:
var $clean_tags = 'script|math';
# Tags that do not need to be closed.
var $auto_close_tags = 'hr|img';
function hashHTMLBlocks($text) {
#
# Hashify HTML Blocks and "clean tags".
#
# We only want to do this for block-level HTML tags, such as headers,
# lists, and tables. That's because we still want to wrap s around
# "paragraphs" that are wrapped in non-block-level tags, such as anchors,
# phrase emphasis, and spans. The list of tags we're looking for is
# hard-coded.
#
# This works by calling _HashHTMLBlocks_InMarkdown, which then calls
# _HashHTMLBlocks_InHTML when it encounter block tags. When the markdown="1"
# attribute is found whitin a tag, _HashHTMLBlocks_InHTML calls back
# _HashHTMLBlocks_InMarkdown to handle the Markdown syntax within the tag.
# These two functions are calling each other. It's recursive!
#
#
# Call the HTML-in-Markdown hasher.
#
list($text, ) = $this->_hashHTMLBlocks_inMarkdown($text);
return $text;
}
function _hashHTMLBlocks_inMarkdown($text, $indent = 0,
$enclosing_tag = '', $span = false)
{
#
# 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:
#
#
#
# Hello World. <-- Is this a Markdown code block or text?
# <-- Is this a Markdown code block or a real tag?
#
#
# If you don't like this, just don't indent the tag on which
# you apply the markdown="1" attribute.
#
# * If $enclosing_tag is not empty, stops at the first unmatched closing
# tag with that name. Nested tags supported.
#
# * If $span is true, text inside must treated as span. So any double
# newline will be replaced by a single newline so that it does not create
# paragraphs.
#
# Returns an array of that form: ( processed text , remaining text )
#
if ($text === '') return array('', '');
# Regex to check for the presense of newlines around a block tag.
$newline_match_before = '/(?:^\n?|\n\n)*$/';
$newline_match_after =
'{
^ # Start of text following the tag.
(?:[ ]*)? # Optional comment.
[ ]*\n # Must be followed by newline.
}xs';
# Regex to match any tag.
$block_tag_match =
'{
( # $2: Capture hole tag.
block_tags.' |
'.$this->context_block_tags.' |
'.$this->clean_tags.' |
(?!\s)'.$enclosing_tag.'
)
\s* # Whitespace.
(?>
".*?" | # Double quotes (can contain `>`)
\'.*?\' | # Single quotes (can contain `>`)
.+? # Anything but quotes and `>`.
)*?
> # End of tag.
|
# HTML Comment
|
<\?.*?\?> | <%.*?%> # Processing instruction
|
# CData Block
)
}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_match, $text, 2,
PREG_SPLIT_DELIM_CAPTURE);
# If in Markdown span mode, add a empty-string span-level hash
# after each newline to prevent triggering any block element.
if ($span) {
$void = $this->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 = $parts[2]; # Remaining text after current tag.
#
# Check for: Tag inside code block or span
#
if (# Find current paragraph
preg_match('/(?>^\n?|\n\n)((?>.+\n?)*?)$/', $parsed, $matches) &&
(
# Then match in it either a code block...
preg_match('/^ {'.($indent+4).'}.*(?>\n {'.($indent+4).'}.*)*'.
'(?!\n)$/', $matches[1], $x) ||
# ...or unbalenced code span markers. (the regex matches balenced)
!preg_match('/^(?>[^`]+|(`+)(?>[^`]+|(?!\1[^`])`)*?\1(?!`))*$/s',
$matches[1])
))
{
# Tag is in code block or span and may not be a tag at all. So we
# simply skip the first char (should be a `<`).
$parsed .= $tag{0};
$text = substr($tag, 1) . $text; # Put back $tag minus first char.
}
#
# Check for: Opening Block level tag or
# Opening Content Block tag (like ins and del)
# used as a block tag (tag is alone on it's line).
#
else if (preg_match("{^<(?:$this->block_tags)\b}", $tag) ||
( preg_match("{^<(?:$this->context_block_tags)\b}", $tag) &&
preg_match($newline_match_before, $parsed) &&
preg_match($newline_match_after, $text) )
)
{
# Need to parse tag and following text using the HTML parser.
list($block_text, $text) =
$this->_hashHTMLBlocks_inHTML($tag . $text, "hashBlock", true);
# Make sure it stays outside of any paragraph by adding newlines.
$parsed .= "\n\n$block_text\n\n";
}
#
# Check for: Clean tag (like script, math)
# HTML Comments, processing instructions.
#
else if (preg_match("{^<(?:$this->clean_tags)\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) =
$this->_hashHTMLBlocks_inHTML($tag . $text, "hashClean", false);
$parsed .= $block_text;
}
#
# Check for: Tag with same name as enclosing tag.
#
else if ($enclosing_tag !== '' &&
# Same name as enclosing tag.
preg_match("{^= 0);
return array($parsed, $text);
}
function _hashHTMLBlocks_inHTML($text, $hash_method, $md_attr) {
#
# Parse HTML, calling _HashHTMLBlocks_InMarkdown for block tags.
#
# * Calls $hash_method to convert any blocks.
# * Stops when the first opening tag closes.
# * $md_attr indicate if the use of the `markdown="1"` attribute is allowed.
# (it is not inside clean tags)
#
# Returns an array of that form: ( processed text , remaining text )
#
if ($text === '') return array('', '');
# Regex to match `markdown` attribute inside of a tag.
$markdown_attr_match = '
{
\s* # Eat whitespace before the `markdown` attribute
markdown
\s*=\s*
(?:
(["\']) # $1: quote delimiter
(.*?) # $2: attribute value
\1 # matching delimiter
|
([^\s>]*) # $3: unquoted attribute value
)
() # $4: make $3 always defined (avoid warnings)
}xs';
# Regex to match any tag.
$tag_match = '{
( # $2: Capture hole tag.
".*?" | # Double quotes (can contain `>`)
\'.*?\' | # Single quotes (can contain `>`)
.+? # Anything but quotes and `>`.
)*?
> # End of tag.
|
# HTML Comment
|
<\?.*?\?> | <%.*?%> # Processing instruction
|
# CData Block
)
}xs';
$original_text = $text; # Save original text in case of faliure.
$depth = 0; # Current depth inside the tag tree.
$block_text = ""; # Temporary text holder for current text.
$parsed = ""; # Parsed text that will be returned.
#
# Get the name of the starting tag.
#
if (preg_match("/^<([\w:$]*)\b/", $text, $matches))
$base_tag_name = $matches[1];
#
# Loop through every tag until we find the corresponding closing tag.
#
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.
#
$parts = preg_split($tag_match, $text, 2, PREG_SPLIT_DELIM_CAPTURE);
if (count($parts) < 3) {
#
# End of $text reached with unbalenced tag(s).
# In that case, we return original text unchanged and pass the
# first character as filtered to prevent an infinite loop in the
# parent function.
#
return array($original_text{0}, substr($original_text, 1));
}
$block_text .= $parts[0]; # Text before current tag.
$tag = $parts[1]; # Tag to handle.
$text = $parts[2]; # Remaining text after current tag.
#
# Check for: Auto-close tag (like
)
# Comments and Processing Instructions.
#
if (preg_match("{^auto_close_tags)\b}", $tag) ||
$tag{1} == '!' || $tag{1} == '?')
{
# Just add the tag to the block as if it was text.
$block_text .= $tag;
}
else {
#
# Increase/decrease nested tag count. Only do so if
# the tag's name match base tag's.
#
if (preg_match("{^mode = $attr_m[2] . $attr_m[3];
$span_mode = $this->mode == 'span' || $this->mode != 'block' &&
preg_match("{^<(?:$this->contain_span_tags)\b}", $tag);
# Calculate indent before tag.
preg_match('/(?:^|\n)( *?)(?! ).*?$/', $block_text, $matches);
$indent = strlen($matches[1]);
# End preceding block with this tag.
$block_text .= $tag;
$parsed .= $this->$hash_method($block_text);
# Get enclosing tag name for the ParseMarkdown function.
preg_match('/^<([\w:$]*)\b/', $tag, $matches);
$tag_name = $matches[1];
# Parse the content using the HTML-in-Markdown parser.
list ($block_text, $text)
= $this->_hashHTMLBlocks_inMarkdown($text, $indent,
$tag_name, $span_mode);
# Outdent markdown text.
if ($indent > 0) {
$block_text = preg_replace("/^[ ]{1,$indent}/m", "",
$block_text);
}
# Append tag content to parsed text.
if (!$span_mode) $parsed .= "\n\n$block_text\n\n";
else $parsed .= "$block_text";
# Start over a new block.
$block_text = "";
}
else $block_text .= $tag;
}
} while ($depth > 0);
#
# Hash last block text that wasn't processed inside the loop.
#
$parsed .= $this->$hash_method($block_text);
return array($parsed, $text);
}
function hashClean($text) {
#
# Called whenever a tag must be hashed when a function insert a "clean" tag
# in $text, it pass through this function and is automaticaly escaped,
# blocking invalid nested overlap.
#
return $this->hashPart($text, 'C');
}
function doHeaders($text) {
#
# Redefined to add id attribute support.
#
# Setext-style headers:
# Header 1 {#header1}
# ========
#
# Header 2 {#header2}
# --------
#
$text = preg_replace_callback(
'{
(^.+?) # $1: Header text
(?:[ ]+\{\#([-_:a-zA-Z0-9]+)\})? # $2: Id attribute
[ ]*\n(=+|-+)[ ]*\n+ # $3: Header footer
}mx',
array(&$this, '_doHeaders_callback_setext'), $text);
# atx-style headers:
# # Header 1 {#header1}
# ## Header 2 {#header2}
# ## Header 2 with closing hashes ## {#header3}
# ...
# ###### Header 6 {#header2}
#
$text = preg_replace_callback('{
^(\#{1,6}) # $1 = string of #\'s
[ ]*
(.+?) # $2 = Header text
[ ]*
\#* # optional closing #\'s (not counted)
(?:[ ]+\{\#([-_:a-zA-Z0-9]+)\})? # id attribute
[ ]*
\n+
}xm',
array(&$this, '_doHeaders_callback_atx'), $text);
return $text;
}
function _doHeaders_attr($attr) {
if (empty($attr)) return "";
return " id=\"$attr\"";
}
function _doHeaders_callback_setext($matches) {
$level = $matches[3]{0} == '=' ? 1 : 2;
$attr = $this->_doHeaders_attr($id =& $matches[2]);
$block = "".$this->runSpanGamut($matches[1])."";
return "\n" . $this->hashBlock($block) . "\n\n";
}
function _doHeaders_callback_atx($matches) {
$level = strlen($matches[1]);
$attr = $this->_doHeaders_attr($id =& $matches[3]);
$block = "".$this->runSpanGamut($matches[2])."";
return "\n" . $this->hashBlock($block) . "\n\n";
}
function doTables($text) {
#
# Form HTML tables.
#
$less_than_tab = $this->tab_width - 1;
#
# Find tables with leading pipe.
#
# | Header 1 | Header 2
# | -------- | --------
# | Cell 1 | Cell 2
# | Cell 3 | Cell 4
#
$text = preg_replace_callback('
{
^ # Start of a line
[ ]{0,'.$less_than_tab.'} # Allowed whitespace.
[|] # Optional leading pipe (present)
(.+) \n # $1: Header row (at least one pipe)
[ ]{0,'.$less_than_tab.'} # Allowed whitespace.
[|] ([ ]*[-:]+[-| :]*) \n # $2: Header underline
( # $3: Cells
(?>
[ ]* # Allowed whitespace.
[|] .* \n # Row content.
)*
)
(?=\n|\Z) # Stop at final double newline.
}xm',
array(&$this, '_doTable_leadingPipe_callback'), $text);
#
# Find tables without leading pipe.
#
# Header 1 | Header 2
# -------- | --------
# Cell 1 | Cell 2
# Cell 3 | Cell 4
#
$text = preg_replace_callback('
{
^ # Start of a line
[ ]{0,'.$less_than_tab.'} # Allowed whitespace.
(\S.*[|].*) \n # $1: Header row (at least one pipe)
[ ]{0,'.$less_than_tab.'} # Allowed whitespace.
([-:]+[ ]*[|][-| :]*) \n # $2: Header underline
( # $3: Cells
(?>
.* [|] .* \n # Row content
)*
)
(?=\n|\Z) # Stop at final double newline.
}xm',
array(&$this, '_DoTable_callback'), $text);
return $text;
}
function _doTable_leadingPipe_callback($matches) {
$head = $matches[1];
$underline = $matches[2];
$content = $matches[3];
# Remove leading pipe for each row.
$content = preg_replace('/^ *[|]/m', '', $content);
return $this->_doTable_callback(array($matches[0], $head, $underline, $content));
}
function _doTable_callback($matches) {
$head = $matches[1];
$underline = $matches[2];
$content = $matches[3];
# Remove any tailing pipes for each line.
$head = preg_replace('/[|] *$/m', '', $head);
$underline = preg_replace('/[|] *$/m', '', $underline);
$content = preg_replace('/[|] *$/m', '', $content);
# Reading alignement from header underline.
$separators = preg_split('/ *[|] */', $underline);
foreach ($separators as $n => $s) {
if (preg_match('/^ *-+: *$/', $s)) $attr[$n] = ' align="right"';
else if (preg_match('/^ *:-+: *$/', $s))$attr[$n] = ' align="center"';
else if (preg_match('/^ *:-+ *$/', $s)) $attr[$n] = ' align="left"';
else $attr[$n] = '';
}
# Parsing span elements, including code spans, character escapes,
# and inline HTML tags, so that pipes inside those gets ignored.
$head = $this->parseSpan($head);
$headers = preg_split('/ *[|] */', $head);
$col_count = count($headers);
# Write column headers.
$text = "\n";
$text .= "\n";
$text .= "\n";
foreach ($headers as $n => $header)
$text .= " ".$this->runSpanGamut(trim($header))."\n";
$text .= " \n";
$text .= "\n";
# Split content by row.
$rows = explode("\n", trim($content, "\n"));
$text .= "\n";
foreach ($rows as $row) {
# Parsing span elements, including code spans, character escapes,
# and inline HTML tags, so that pipes inside those gets ignored.
$row = $this->parseSpan($row);
# Split row by cell.
$row_cells = preg_split('/ *[|] */', $row, $col_count);
$row_cells = array_pad($row_cells, $col_count, '');
$text .= "\n";
foreach ($row_cells as $n => $cell)
$text .= " ".$this->runSpanGamut(trim($cell))."\n";
$text .= " \n";
}
$text .= "\n";
$text .= "
";
return $this->hashBlock($text) . "\n";
}
function doDefLists($text) {
#
# Form HTML definition lists.
#
$less_than_tab = $this->tab_width - 1;
# Re-usable pattern to match any entire dl list:
$whole_list = '(?>
( # $1 = whole list
( # $2
[ ]{0,'.$less_than_tab.'}
((?>.*\S.*\n)+) # $3 = defined term
\n?
[ ]{0,'.$less_than_tab.'}:[ ]+ # colon starting definition
)
(?s:.+?)
( # $4
\z
|
\n{2,}
(?=\S)
(?! # Negative lookahead for another term
[ ]{0,'.$less_than_tab.'}
(?: \S.*\n )+? # defined term
\n?
[ ]{0,'.$less_than_tab.'}:[ ]+ # colon starting definition
)
(?! # Negative lookahead for another definition
[ ]{0,'.$less_than_tab.'}:[ ]+ # colon starting definition
)
)
)
)'; // mx
$text = preg_replace_callback('{
(?:(?<=\n\n)|\A\n?)
'.$whole_list.'
}mx',
array(&$this, '_doDefLists_callback'), $text);
return $text;
}
function _doDefLists_callback($matches) {
# Re-usable patterns to match list item bullets and number markers:
$list = $matches[1];
# Turn double returns into triple returns, so that we can make a
# paragraph for the last item in a list, if necessary:
$result = trim($this->processDefListItems($list));
$result = "\n" . $result . "\n
";
return $this->hashBlock($result) . "\n\n";
}
function processDefListItems($list_str) {
#
# Process the contents of a single definition list, splitting it
# into individual term and definition list items.
#
$less_than_tab = $this->tab_width - 1;
# trim trailing blank lines:
$list_str = preg_replace("/\n{2,}\\z/", "\n", $list_str);
# Process definition terms.
$list_str = preg_replace_callback('{
(?:\n\n+|\A\n?) # leading line
( # definition terms = $1
[ ]{0,'.$less_than_tab.'} # leading whitespace
(?![:][ ]|[ ]) # negative lookahead for a definition
# mark (colon) or more whitespace.
(?: \S.* \n)+? # actual term (not whitespace).
)
(?=\n?[ ]{0,3}:[ ]) # lookahead for following line feed
# with a definition mark.
}xm',
array(&$this, '_processDefListItems_callback_dt'), $list_str);
# Process actual definitions.
$list_str = preg_replace_callback('{
\n(\n+)? # leading line = $1
[ ]{0,'.$less_than_tab.'} # whitespace before colon
[:][ ]+ # definition mark (colon)
((?s:.+?)) # definition text = $2
(?= \n+ # stop at next definition mark,
(?: # next term or end of text
[ ]{0,'.$less_than_tab.'} [:][ ] |
| \z
)
)
}xm',
array(&$this, '_processDefListItems_callback_dd'), $list_str);
return $list_str;
}
function _processDefListItems_callback_dt($matches) {
$terms = explode("\n", trim($matches[1]));
$text = '';
foreach ($terms as $term) {
$term = $this->runSpanGamut(trim($term));
$text .= "\n" . $term . " ";
}
return $text . "\n";
}
function _processDefListItems_callback_dd($matches) {
$leading_line = $matches[1];
$def = $matches[2];
if ($leading_line || preg_match('/\n{2,}/', $def)) {
$def = $this->runBlockGamut($this->outdent($def . "\n\n"));
$def = "\n". $def ."\n";
}
else {
$def = rtrim($def);
$def = $this->runSpanGamut($this->outdent($def));
}
return "\n " . $def . " \n";
}
function doItalicsAndBold($text) {
#
# Redefined to change emphasis by underscore behaviour so that it does not
# work in the middle of a word.
#
# must go first:
$text = preg_replace_callback(array(
'{
( # $1: Marker
(?
[^_]+? # Anthing not em markers.
|
# Balence any regular _ emphasis inside.
(?
[^*]+? # Anthing not em markers.
|
# Balence any regular * emphasis inside.
\* (?=\S) (.+?) (?<=\S) \*
|
\* # Allow unbalenced as last resort.
)+?
)
(?<=\S) \*\* # End mark not preceded by whitespace.
}sx',
),
array(&$this, '_doItalicAndBold_strong_callback'), $text);
# Then :
$text = preg_replace_callback(array(
'{ ( (? tags
#
# Strip leading and trailing lines:
$text = preg_replace('/\A\n+|\n+\z/', '', $text);
$grafs = preg_split('/\n{2,}/', $text, -1, PREG_SPLIT_NO_EMPTY);
#
# Wrap tags and unhashify HTML blocks
#
foreach ($grafs as $key => $value) {
$value = trim($this->runSpanGamut($value));
# Check if this should be enclosed in a paragraph.
# Clean tag hashes & block tag hashes are left alone.
$is_p = !preg_match('/^B\x1A[0-9]+B|^C\x1A[0-9]+C$/', $value);
if ($is_p) {
$value = "
$value
";
}
$grafs[$key] = $value;
}
# Join grafs in one text, then unhash HTML tags.
$text = implode("\n\n", $grafs);
# Finish by removing any tag hashes still present in $text.
$text = $this->unhash($text);
return $text;
}
### Footnotes
function stripFootnotes($text) {
#
# Strips link definitions from text, stores the URLs and titles in
# hash references.
#
$less_than_tab = $this->tab_width - 1;
# Link defs are in the form: [^id]: url "optional title"
$text = preg_replace_callback('{
^[ ]{0,'.$less_than_tab.'}\[\^(.+?)\][ ]?: # note_id = $1
[ ]*
\n? # maybe *one* newline
( # text = $2 (no blank lines allowed)
(?:
.+ # actual text
|
\n # newlines but
(?!\[\^.+?\]:\s)# negative lookahead for footnote marker.
(?!\n+[ ]{0,3}\S)# ensure line is not blank and followed
# by non-indented content
)*
)
}xm',
array(&$this, '_stripFootnotes_callback'),
$text);
return $text;
}
function _stripFootnotes_callback($matches) {
$note_id = $this->fn_id_prefix . $matches[1];
$this->footnotes[$note_id] = $this->outdent($matches[2]);
return ''; # String that will replace the block
}
function doFootnotes($text) {
#
# Replace footnote references in $text [^id] with a special text-token
# which will be can be
#
if (!$this->in_footnote && !$this->in_anchor) {
$text = preg_replace('{\[\^(.+?)\]}', "F\x1Afn:\\1\x1A:", $text);
}
return $text;
}
function appendFootnotes($text) {
#
# Append footnote list to text.
#
$text = preg_replace_callback('{F\x1Afn:(.*?)\x1A:}',
array(&$this, '_appendFootnotes_callback'), $text);
if (!empty($this->footnotes_ordered)) {
$text .= "\n\n";
$text .= "\n";
$text .= "fn_backlink_class != "") {
$class = $this->fn_backlink_class;
$class = $this->encodeAmpsAndAngles($class);
$class = str_replace('"', '"', $class);
$attr .= " class=\"$class\"";
}
if ($this->fn_backlink_title != "") {
$title = $this->fn_backlink_title;
$title = $this->encodeAmpsAndAngles($title);
$title = str_replace('"', '"', $title);
$attr .= " title=\"$title\"";
}
$num = 0;
$this->in_footnote = true;
foreach ($this->footnotes_ordered as $note_id => $footnote) {
$footnote .= "\n"; # Need to append newline before parsing.
$footnote = $this->runBlockGamut("$footnote\n");
$attr2 = str_replace("%%", ++$num, $attr);
# Add backlink to last paragraph; create new paragraph if needed.
$backlink = "↩";
if (preg_match('{$}', $footnote)) {
$footnote = substr($footnote, 0, -4) . " $backlink";
} else {
$footnote .= "\n\n$backlink
";
}
$text .= "\n";
$text .= $footnote . "\n";
$text .= " \n\n";
}
$this->in_footnote = false;
$text .= "\n";
$text .= " ";
}
return $text;
}
function _appendFootnotes_callback($matches) {
$node_id = $this->fn_id_prefix . $matches[1];
# Create footnote marker only if it has a corresponding footnote *and*
# the footnote hasn't been used by another marker.
if (isset($this->footnotes[$node_id])) {
# Transfert footnote content to the ordered list.
$this->footnotes_ordered[$node_id] = $this->footnotes[$node_id];
unset($this->footnotes[$node_id]);
$num = count($this->footnotes_ordered);
$attr = " rel=\"footnote\"";
if ($this->fn_link_class != "") {
$class = $this->fn_link_class;
$class = $this->encodeAmpsAndAngles($class);
$class = str_replace('"', '"', $class);
$attr .= " class=\"$class\"";
}
if ($this->fn_link_title != "") {
$title = $this->fn_link_title;
$title = $this->encodeAmpsAndAngles($title);
$title = str_replace('"', '"', $title);
$attr .= " title=\"$title\"";
}
$attr = str_replace("%%", $num, $attr);
return
"".
"$num".
"";
}
return "[^".$matches[1]."]";
}
### Abbreviations ###
function stripAbbreviations($text) {
#
# Strips abbreviations from text, stores titles in hash references.
#
$less_than_tab = $this->tab_width - 1;
# Link defs are in the form: [id]*: url "optional title"
$text = preg_replace_callback('{
^[ ]{0,'.$less_than_tab.'}\*\[(.+?)\][ ]?: # abbr_id = $1
(.*) # text = $2 (no blank lines allowed)
}xm',
array(&$this, '_stripAbbreviations_callback'),
$text);
return $text;
}
function _stripAbbreviations_callback($matches) {
$abbr_word = $matches[1];
$abbr_desc = $matches[2];
$this->abbr_matches[] = preg_quote($abbr_word);
$this->abbr_desciptions[$abbr_word] = trim($abbr_desc);
return ''; # String that will replace the block
}
function doAbbreviations($text) {
#
# Find defined abbreviations in text and wrap them in elements.
#
if ($this->abbr_matches) {
// cannot use the /x modifier because abbr_matches may
// contain spaces:
$text = preg_replace_callback('{'.
'(?abbr_matches) .')'.
'(?![\w\x1A])'.
'}',
array(&$this, '_doAbbreviations_callback'), $text);
}
return $text;
}
function _doAbbreviations_callback($matches) {
$abbr = $matches[0];
if (isset($this->abbr_desciptions[$abbr])) {
$desc = $this->abbr_desciptions[$abbr];
if (empty($desc)) {
return $this->hashPart("$abbr");
} else {
$desc = htmlspecialchars($desc, ENT_NOQUOTES);
return $this->hashPart("$abbr");
}
} else {
return $matches[0];
}
}
}
/*
PHP Markdown Extra
==================
Description
-----------
This is a PHP port of the original Markdown formatter written in Perl
by John Gruber. This special "Extra" version of PHP Markdown features
further enhancements to the syntax for making additional constructs
such as tables and definition list.
Markdown is a text-to-HTML filter; it translates an easy-to-read /
easy-to-write structured text format into HTML. Markdown's text format
is most similar to that of plain text email, and supports features such
as headers, *emphasis*, code blocks, blockquotes, and links.
Markdown's syntax is designed not as a generic markup language, but
specifically to serve as a front-end to (X)HTML. You can use span-level
HTML tags anywhere in a Markdown document, and you can use block level
HTML tags (like and as well).
For more information about Markdown's syntax, see:
Please include with your report: (1) the example input; (2) the output you
expected; (3) the output Markdown actually produced.
Version History
---------------
See the readme file for detailed release notes for this version.
Copyright and License
---------------------
PHP Markdown & Extra
Copyright (c) 2004-2007 Michel Fortin