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: var block_tags_a = "p|div|h[1-6]|blockquote|pre|table|dl|ol|ul|script|noscript|form|fieldset|iframe|math|ins|del" var block_tags_b = "p|div|h[1-6]|blockquote|pre|table|dl|ol|ul|script|noscript|form|fieldset|iframe|math" // First, look for nested blocks, e.g.: //
. It was easier to make a special case than // to make the other regex more complicated. /* text = text.replace(/ ( // save in $1 \n\n // Starting after a blank line [ ]{0,3} (<(hr) // start tag = $2 \b // word break ([^<>])*? // \/?>) // the matching end tag [ \t]* (?=\n{2,}) // followed by a blank line ) /g,hashElement); */ text = text.replace(/(\n[ ]{0,3}(<(hr)\b([^<>])*?\/?>)[ \t]*(?=\n{2,}))/g,hashElement); // Special case for standalone HTML comments: /* text = text.replace(/ ( // save in $1 \n\n // Starting after a blank line [ ]{0,3} // attacklab: g_tab_width - 1 [ \t]* (?=\n{2,}) // followed by a blank line ) /g,hashElement); */ text = text.replace(/(\n\n[ ]{0,3}[ \t]*(?=\n{2,}))/g,hashElement); // PHP and ASP-style processor instructions ( and <%...%>) /* text = text.replace(/ (?: \n\n // Starting after a blank line ) ( // save in $1 [ ]{0,3} // attacklab: g_tab_width - 1 (?: <([?%]) // $2 [^\r]*? \2> ) [ \t]* (?=\n{2,}) // followed by a blank line ) /g,hashElement); */ text = text.replace(/(?:\n\n)([ ]{0,3}(?:<([?%])[^\r]*?\2>)[ \t]*(?=\n{2,}))/g,hashElement); // attacklab: Undo double lines (see comment at top of this function) text = text.replace(/\n\n/g,"\n"); return text; } var hashElement = function(wholeMatch,m1) { var blockText = m1; // Undo double lines blockText = blockText.replace(/\n\n/g,"\n"); blockText = blockText.replace(/^\n/,""); // strip trailing blank lines blockText = blockText.replace(/\n+$/g,""); // Replace the element text with a marker ("~KxK" where x is its key) blockText = "\n\n~K" + (g_html_blocks.push(blockText)-1) + "K\n\n"; return blockText; }; var _RunBlockGamut = function(text) { // // These are all the transformations that form block-level // tags like paragraphs, headers, and list items. // text = _DoHeaders(text); // Do Horizontal Rules: var key = hashBlock("
"); text = text.replace(/^[ ]{0,2}([ ]?\*[ ]?){3,}[ \t]*$/gm,key); text = text.replace(/^[ ]{0,2}([ ]?\-[ ]?){3,}[ \t]*$/gm,key); text = text.replace(/^[ ]{0,2}([ ]?\_[ ]?){3,}[ \t]*$/gm,key); text = _DoLists(text); text = _DoCodeBlocks(text); text = _DoBlockQuotes(text); // We already ran _HashHTMLBlocks() before, in Markdown(), but that // was to escape raw HTML in the original Markdown source. This time, // we're escaping the markup we've just created, so that we don't wrap //
tags around block-level tags.
text = _HashHTMLBlocks(text);
text = _FormParagraphs(text);
return text;
}
var _RunSpanGamut = function(text) {
//
// These are all the transformations that occur *within* block-level
// tags like paragraphs, headers, and list items.
//
text = _DoCodeSpans(text);
text = _EscapeSpecialCharsWithinTagAttributes(text);
text = _EncodeBackslashEscapes(text);
// Process anchor and image tags. Images must come first,
// because ![foo][f] looks like an anchor.
text = _DoImages(text);
text = _DoAnchors(text);
// Make links out of things like ` Just type tags
//
// Strip leading and trailing lines:
text = text.replace(/^\n+/g,"");
text = text.replace(/\n+$/g,"");
var grafs = text.split(/\n{2,}/g);
var grafsOut = new Array();
//
// Wrap tags.
//
var end = grafs.length;
for (var i=0; i ");
str += " 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:
my $block_tags_a = qr/p|div|h[1-6]|blockquote|pre|table|dl|ol|ul|script|noscript|form|fieldset|iframe|math|ins|del/;
my $block_tags_b = qr/p|div|h[1-6]|blockquote|pre|table|dl|ol|ul|script|noscript|form|fieldset|iframe|math/;
# First, look for nested blocks, e.g.:
# Just type tags
#
my $text = shift;
# Strip leading and trailing lines:
$text =~ s/\A\n+//;
$text =~ s/\n+\z//;
my @grafs = split(/\n{2,}/, $text);
#
# Wrap tags.
#
foreach (@grafs) {
unless (defined( $g_html_blocks{$_} )) {
$_ = _RunSpanGamut($_);
s/^([ \t]*)/ /;
$_ .= "
\n");
return text;
}
var _EscapeSpecialCharsWithinTagAttributes = function(text) {
//
// Within tags -- meaning between < and > -- encode [\ ` * _] so they
// don't conflict with their use in Markdown for code, italics and strong.
//
// Build a regex to find HTML tags and comments. See Friedl's
// "Mastering Regular Expressions", 2nd Ed., pp. 200-201.
var regex = /(<[a-z\/!$]("[^"]*"|'[^']*'|[^'">])*>|)/gi;
text = text.replace(regex, function(wholeMatch) {
var tag = wholeMatch.replace(/(.)<\/?code>(?=.)/g,"$1`");
tag = escapeCharacters(tag,"\\`*_");
return tag;
});
return text;
}
var _DoAnchors = function(text) {
//
// Turn Markdown link shortcuts into XHTML tags.
//
//
// First, handle reference-style links: [link text] [id]
//
/*
text = text.replace(/
( // wrap whole match in $1
\[
(
(?:
\[[^\]]*\] // allow brackets nested one level
|
[^\[] // or anything else
)*
)
\]
[ ]? // one optional space
(?:\n[ ]*)? // one optional newline followed by spaces
\[
(.*?) // id = $3
\]
)()()()() // pad remaining backreferences
/g,_DoAnchors_callback);
*/
text = text.replace(/(\[((?:\[[^\]]*\]|[^\[\]])*)\][ ]?(?:\n[ ]*)?\[(.*?)\])()()()()/g,writeAnchorTag);
//
// Next, inline-style links: [link text](url "optional title")
//
/*
text = text.replace(/
( // wrap whole match in $1
\[
(
(?:
\[[^\]]*\] // allow brackets nested one level
|
[^\[\]] // or anything else
)
)
\]
\( // literal paren
[ \t]*
() // no id, so leave $3 empty
? // href = $4
[ \t]*
( // $5
(['"]) // quote char = $6
(.*?) // Title = $7
\6 // matching quote
[ \t]* // ignore any spaces/tabs between closing quote and )
)? // title is optional
\)
)
/g,writeAnchorTag);
*/
text = text.replace(/(\[((?:\[[^\]]*\]|[^\[\]])*)\]\([ \t]*()?[ \t]*((['"])(.*?)\6[ \t]*)?\))/g,writeAnchorTag);
//
// Last, handle reference-style shortcuts: [link text]
// These must come last in case you've also got [link test][1]
// or [link test](/foo)
//
/*
text = text.replace(/
( // wrap whole match in $1
\[
([^\[\]]+) // link text = $2; can't contain '[' or ']'
\]
)()()()()() // pad rest of backreferences
/g, writeAnchorTag);
*/
text = text.replace(/(\[([^\[\]]+)\])()()()()()/g, writeAnchorTag);
return text;
}
var writeAnchorTag = function(wholeMatch,m1,m2,m3,m4,m5,m6,m7) {
if (m7 == undefined) m7 = "";
var whole_match = m1;
var link_text = m2;
var link_id = m3.toLowerCase();
var url = m4;
var title = m7;
if (url == "") {
if (link_id == "") {
// lower-case and turn embedded newlines into spaces
link_id = link_text.toLowerCase().replace(/ ?\n/g," ");
}
url = "#"+link_id;
if (g_urls[link_id] != undefined) {
url = g_urls[link_id];
if (g_titles[link_id] != undefined) {
title = g_titles[link_id];
}
}
else {
if (whole_match.search(/\(\s*\)$/m)>-1) {
// Special case for explicit empty url
url = "";
} else {
return whole_match;
}
}
}
url = escapeCharacters(url,"*_");
var result = "";
return result;
}
var _DoImages = function(text) {
//
// Turn Markdown image shortcuts into tags.
//
//
// First, handle reference-style labeled images: ![alt text][id]
//
/*
text = text.replace(/
( // wrap whole match in $1
!\[
(.*?) // alt text = $2
\]
[ ]? // one optional space
(?:\n[ ]*)? // one optional newline followed by spaces
\[
(.*?) // id = $3
\]
)()()()() // pad rest of backreferences
/g,writeImageTag);
*/
text = text.replace(/(!\[(.*?)\][ ]?(?:\n[ ]*)?\[(.*?)\])()()()()/g,writeImageTag);
//
// Next, handle inline images: 
// Don't forget: encode * and _
/*
text = text.replace(/
( // wrap whole match in $1
!\[
(.*?) // alt text = $2
\]
\s? // One optional whitespace character
\( // literal paren
[ \t]*
() // no id, so leave $3 empty
? // src url = $4
[ \t]*
( // $5
(['"]) // quote char = $6
(.*?) // title = $7
\6 // matching quote
[ \t]*
)? // title is optional
\)
)
/g,writeImageTag);
*/
text = text.replace(/(!\[(.*?)\]\s?\([ \t]*()?[ \t]*((['"])(.*?)\6[ \t]*)?\))/g,writeImageTag);
return text;
}
var writeImageTag = function(wholeMatch,m1,m2,m3,m4,m5,m6,m7) {
var whole_match = m1;
var alt_text = m2;
var link_id = m3.toLowerCase();
var url = m4;
var title = m7;
if (!title) title = "";
if (url == "") {
if (link_id == "") {
// lower-case and turn embedded newlines into spaces
link_id = alt_text.toLowerCase().replace(/ ?\n/g," ");
}
url = "#"+link_id;
if (g_urls[link_id] != undefined) {
url = g_urls[link_id];
if (g_titles[link_id] != undefined) {
title = g_titles[link_id];
}
}
else {
return whole_match;
}
}
alt_text = alt_text.replace(/"/g,""");
url = escapeCharacters(url,"*_");
var result = "
" + _RunSpanGamut(m1) + "");});
text = text.replace(/^(.+)[ \t]*\n-+[ \t]*\n+/gm,
function(matchFound,m1){return hashBlock("
" + _RunSpanGamut(m1) + "
");});
// atx-style headers:
// # Header 1
// ## Header 2
// ## Header 2 with closing hashes ##
// ...
// ###### Header 6
//
/*
text = text.replace(/
^(\#{1,6}) // $1 = string of #'s
[ \t]*
(.+?) // $2 = Header text
[ \t]*
\#* // optional closing #'s (not counted)
\n+
/gm, function() {...});
*/
text = text.replace(/^(\#{1,6})[ \t]*(.+?)[ \t]*\#*\n+/gm,
function(wholeMatch,m1,m2) {
var h_level = m1.length;
return hashBlock("` blocks.
//
/*
text = text.replace(text,
/(?:\n\n|^)
( // $1 = the code block -- one or more lines, starting with a space/tab
(?:
(?:[ ]{4}|\t) // Lines must start with a tab or a tab-width of spaces - attacklab: g_tab_width
.*\n+
)+
)
(\n*[ ]{0,3}[^ \t\n]|(?=~0)) // attacklab: g_tab_width
/g,function(){...});
*/
// attacklab: sentinel workarounds for lack of \A and \Z, safari\khtml bug
text += "~0";
text = text.replace(/(?:\n\n|^)((?:(?:[ ]{4}|\t).*\n+)+)(\n*[ ]{0,3}[^ \t\n]|(?=~0))/g,
function(wholeMatch,m1,m2) {
var codeblock = m1;
var nextChar = m2;
codeblock = _EncodeCode( _Outdent(codeblock));
codeblock = _Detab(codeblock);
codeblock = codeblock.replace(/^\n+/g,""); // trim leading newlines
codeblock = codeblock.replace(/\n+$/g,""); // trim trailing whitespace
codeblock = "
";
return hashBlock(codeblock) + nextChar;
}
);
// attacklab: strip sentinel
text = text.replace(/~0/,"");
return text;
}
var hashBlock = function(text) {
text = text.replace(/(^\n+|\n+$)/g,"");
return "\n\n~K" + (g_html_blocks.push(text)-1) + "K\n\n";
}
var _DoCodeSpans = function(text) {
//
// * Backtick quotes are used for " + codeblock + "\n spans.
//
// * You can use multiple backticks as the delimiters if you want to
// include literal backticks in the code span. So, this input:
//
// Just type ``foo `bar` baz`` at the prompt.
//
// Will translate to:
//
// foo `bar` baz at the prompt.`bar` ...
//
/*
text = text.replace(/
(^|[^\\]) // Character before opening ` can't be a backslash
(`+) // $2 = Opening run of `
( // $3 = The code block
[^\r]*?
[^`] // attacklab: work around lack of lookbehind
)
\2 // Matching closer
(?!`)
/gm, function(){...});
*/
text = text.replace(/(^|[^\\])(`+)([^\r]*?[^`])\2(?!`)/gm,
function(wholeMatch,m1,m2,m3,m4) {
var c = m3;
c = c.replace(/^([ \t]*)/g,""); // leading whitespace
c = c.replace(/[ \t]*$/g,""); // trailing whitespace
c = _EncodeCode(c);
return m1+""+c+"";
});
return text;
}
var _EncodeCode = function(text) {
//
// Encode/escape certain characters inside Markdown code runs.
// The point is that in code, these characters are literals,
// and lose their special Markdown meanings.
//
// Encode all ampersands; HTML entities are not
// entities within a Markdown code span.
text = text.replace(/&/g,"&");
// Do the angle bracket song and dance:
text = text.replace(//g,">");
// Now, escape characters that are magic in Markdown:
text = escapeCharacters(text,"\*_{}[]\\",false);
// jj the line above breaks this:
//---
//* Item
// 1. Subitem
// special char: *
//---
return text;
}
var _DoItalicsAndBold = function(text) {
// must go first:
text = text.replace(/(\*\*|__)(?=\S)([^\r]*?\S[*_]*)\1/g,
"$2");
text = text.replace(/(\*|_)(?=\S)([^\r]*?\S)\1/g,
"$2");
return text;
}
var _DoBlockQuotes = function(text) {
/*
text = text.replace(/
( // Wrap whole match in $1
(
^[ \t]*>[ \t]? // '>' at the start of a line
.+\n // rest of the first line
(.+\n)* // subsequent consecutive lines
\n* // blanks
)+
)
/gm, function(){...});
*/
text = text.replace(/((^[ \t]*>[ \t]?.+\n(.+\n)*\n*)+)/gm,
function(wholeMatch,m1) {
var bq = m1;
// attacklab: hack around Konqueror 3.5.4 bug:
// "----------bug".replace(/^-/g,"") == "bug"
bq = bq.replace(/^[ \t]*>[ \t]?/gm,"~0"); // trim one level of quoting
// attacklab: clean up hack
bq = bq.replace(/~0/g,"");
bq = bq.replace(/^[ \t]+$/gm,""); // trim whitespace-only lines
bq = _RunBlockGamut(bq); // recurse
bq = bq.replace(/(^|\n)/g,"$1 ");
// These leading spaces screw with content, so we need to fix that:
bq = bq.replace(
/(\s*
[^\r]+?<\/pre>)/gm,
function(wholeMatch,m1) {
var pre = m1;
// attacklab: hack around Konqueror 3.5.4 bug:
pre = pre.replace(/^ /mg,"~0");
pre = pre.replace(/~0/g,"");
return pre;
});
return hashBlock("\n" + bq + "\n
");
});
return text;
}
var _FormParagraphs = function(text) {
//
// Params:
// $text - string to process with html tags get encoded.
#
my $text = shift;
# 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):
%g_urls = ();
%g_titles = ();
%g_html_blocks = ();
# Standardize line endings:
$text =~ s{\r\n}{\n}g; # DOS to Unix
$text =~ s{\r}{\n}g; # Mac to Unix
# Make sure $text ends with a couple of newlines:
$text .= "\n\n";
# Convert all tabs to spaces.
$text = _Detab($text);
# Strip any lines consisting only of spaces and tabs.
# This makes subsequent regexen easier to write, because we can
# match consecutive blank lines with /\n+/ instead of something
# contorted like /[ \t]*\n+/ .
$text =~ s/^[ \t]+$//mg;
# Turn block-level HTML blocks into hash entries
$text = _HashHTMLBlocks($text);
# Strip link definitions, store in hashes.
$text = _StripLinkDefinitions($text);
$text = _RunBlockGamut($text);
$text = _UnescapeSpecialChars($text);
return $text . "\n";
}
sub _StripLinkDefinitions {
#
# Strips link definitions from text, stores the URLs and titles in
# hash references.
#
my $text = shift;
my $less_than_tab = $g_tab_width - 1;
# Link defs are in the form: ^[id]: url "optional title"
while ($text =~ s{
^[ ]{0,$less_than_tab}\[(.+)\]: # id = $1
[ \t]*
\n? # maybe *one* newline
[ \t]*
? # url = $2
[ \t]*
\n? # maybe one newline
[ \t]*
(?:
(?<=\s) # lookbehind for whitespace
["(]
(.+?) # title = $3
[")]
[ \t]*
)? # title is optional
(?:\n+|\Z)
}
{}mx) {
$g_urls{lc $1} = _EncodeAmpsAndAngles( $2 ); # Link IDs are case-insensitive
if ($3) {
$g_titles{lc $1} = $3;
$g_titles{lc $1} =~ s/"/"/g;
}
}
return $text;
}
sub _HashHTMLBlocks {
my $text = shift;
my $less_than_tab = $g_tab_width - 1;
# Hashify HTML blocks:
# 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
. It was easier to make a special case than
# to make the other regex more complicated.
$text =~ s{
(?:
(?<=\n\n) # Starting after a blank line
| # or
\A\n? # the beginning of the doc
)
( # save in $1
[ ]{0,$less_than_tab}
<(hr) # start tag = $2
\b # word break
([^<>])*? #
/?> # the matching end tag
[ \t]*
(?=\n{2,}|\Z) # followed by a blank line or end of document
)
}{
my $key = md5_hex($1);
$g_html_blocks{$key} = $1;
"\n\n" . $key . "\n\n";
}egx;
# Special case for standalone HTML comments:
$text =~ s{
(?:
(?<=\n\n) # Starting after a blank line
| # or
\A\n? # the beginning of the doc
)
( # save in $1
[ ]{0,$less_than_tab}
(?s:
)
[ \t]*
(?=\n{2,}|\Z) # followed by a blank line or end of document
)
}{
my $key = md5_hex($1);
$g_html_blocks{$key} = $1;
"\n\n" . $key . "\n\n";
}egx;
return $text;
}
sub _RunBlockGamut {
#
# These are all the transformations that form block-level
# tags like paragraphs, headers, and list items.
#
my $text = shift;
$text = _DoHeaders($text);
# Do Horizontal Rules:
$text =~ s{^[ ]{0,2}([ ]?\*[ ]?){3,}[ \t]*$}{\n? # src url = $3
[ \t]*
( # $4
(['"]) # quote char = $5
(.*?) # title = $6
\5 # matching quote
[ \t]*
)? # title is optional
\)
)
}{
my $result;
my $whole_match = $1;
my $alt_text = $2;
my $url = $3;
my $title = '';
if (defined($6)) {
$title = $6;
}
$alt_text =~ s/"/"/g;
$title =~ s/"/"/g;
$url =~ s! \* !$g_escape_table{'*'}!gx; # We've got to encode these to avoid
$url =~ s! _ !$g_escape_table{'_'}!gx; # conflicting with italics/bold.
$result = "
" . _RunSpanGamut($1) . "\n\n";
}egmx;
$text =~ s{ ^(.+)[ \t]*\n-+[ \t]*\n+ }{
"
" . _RunSpanGamut($1) . "
\n\n";
}egmx;
# atx-style headers:
# # Header 1
# ## Header 2
# ## Header 2 with closing hashes ##
# ...
# ###### Header 6
#
$text =~ s{
^(\#{1,6}) # $1 = string of #'s
[ \t]*
(.+?) # $2 = Header text
[ \t]*
\#* # optional closing #'s (not counted)
\n+
}{
my $h_level = length($1);
"` blocks.
#
my $text = shift;
$text =~ s{
(?:\n\n|\A)
( # $1 = the code block -- one or more lines, starting with a space/tab
(?:
(?:[ ]{$g_tab_width} | \t) # Lines must start with a tab or a tab-width of spaces
.*\n+
)+
)
((?=^[ ]{0,$g_tab_width}\S)|\Z) # Lookahead for non-space at line-start, or end of doc
}{
my $codeblock = $1;
my $result; # return value
$codeblock = _EncodeCode(_Outdent($codeblock));
$codeblock = _Detab($codeblock);
$codeblock =~ s/\A\n+//; # trim leading newlines
$codeblock =~ s/\s+\z//; # trim trailing whitespace
$result = "\n\n";
@egsx;
return $text;
}
sub _EncodeCode {
#
# Encode/escape certain characters inside Markdown code runs.
# The point is that in code, these characters are literals,
# and lose their special Markdown meanings.
#
local $_ = shift;
# Encode all ampersands; HTML entities are not
# entities within a Markdown code span.
s/&/&/g;
# Encode $'s, but only if we're running under Blosxom.
# (Blosxom interpolates Perl variables in article bodies.)
{
no warnings 'once';
if (defined($blosxom::version)) {
s/\$/$/g;
}
}
# Do the angle bracket song and dance:
s! < !<!gx;
s! > !>!gx;
# Now, escape characters that are magic in Markdown:
s! \* !$g_escape_table{'*'}!gx;
s! _ !$g_escape_table{'_'}!gx;
s! { !$g_escape_table{'{'}!gx;
s! } !$g_escape_table{'}'}!gx;
s! \[ !$g_escape_table{'['}!gx;
s! \] !$g_escape_table{']'}!gx;
s! \\ !$g_escape_table{'\\'}!gx;
return $_;
}
sub _DoItalicsAndBold {
my $text = shift;
# must go first:
$text =~ s{ (\*\*|__) (?=\S) (.+?[*_]*) (?<=\S) \1 }
{$2}gsx;
$text =~ s{ (\*|_) (?=\S) (.+?) (?<=\S) \1 }
{$2}gsx;
return $text;
}
sub _DoBlockQuotes {
my $text = shift;
$text =~ s{
( # Wrap whole match in $1
(
^[ \t]*>[ \t]? # '>' at the start of a line
.+\n # rest of the first line
(.+\n)* # subsequent consecutive lines
\n* # blanks
)+
)
}{
my $bq = $1;
$bq =~ s/^[ \t]*>[ \t]?//gm; # trim one level of quoting
$bq =~ s/^[ \t]+$//mg; # trim whitespace-only lines
$bq = _RunBlockGamut($bq); # recurse
$bq =~ s/^/ /g;
# These leading spaces screw with
\n\n";
$result;
}egmx;
return $text;
}
sub _DoCodeSpans {
#
# * Backtick quotes are used for " . $codeblock . "\n spans.
#
# * You can use multiple backticks as the delimiters if you want to
# include literal backticks in the code span. So, this input:
#
# Just type ``foo `bar` baz`` at the prompt.
#
# Will translate to:
#
# foo `bar` baz at the prompt.`bar` ...
#
my $text = shift;
$text =~ s@
(?$c content, so we need to fix that:
$bq =~ s{
(\s*.+?
)
}{
my $pre = $1;
$pre =~ s/^ //mg;
$pre;
}egsx;
"\n$bq\n
\n\n";
}egmx;
return $text;
}
sub _FormParagraphs {
#
# Params:
# $text - string to process with html