HTML escaping, switchable HTML processing

[cgilite] / markdown.awk

diff --git a/markdown.awk b/markdown.awk
index a5c2eb1f587d79f203df3858819d33da565236bf..785ece87fbad731d7d4de1343752a2a63ccc9632 100755 (executable)
--- a/markdown.awk
+++ b/markdown.awk
@@ -23,7 +23,8 @@
 # - [x] Blockquotes
 # - [x] Lists (ordered, unordered)
 # - [x] Code blocks (using indention)
-# - [ ] Horizontal rules
+# - [x] Horizontal rules
+# - [x] Verbatim HTML block (disabled by default)
 #
 # Basic Markdown - Inline elements:
 # ---------------------------------
@@ -31,9 +32,13 @@
 # - [x] Reference style links
 # - [x] Emphasis *em*/**strong** (*Asterisk*, _Underscore_)
 # - [x] `code`, also ``code containing `backticks` ``
-# - [ ] Images / reference style images
+# - [x] Images / reference style images
 # - [x] <automatic links>
 # - [x] backslash escapes
+# - [x] Verbatim HTML inline (disabled by default)
+# - [x] HTML escaping
+#
+# NOTE: Set the environment variable MD_HTML=true to enable verbatim HTML
 #
 # Extensions - Block elements:
 # ----------------------------
@@ -146,6 +151,32 @@ function inline( line, LOCAL, len, code, href, guard ) {
       return "" substr(line, 1, len) inline( substr(line, len + 1) );
     }
 
+  # inline images
+  } else if ( match(line, /^!\[([^]]+)\]\(([^"\)]+)([ \t]+"([^"]+)")?\)/) ) {
+    len = RLENGTH;
+    text  = gensub(/^!\[([^]]+)\]\(([^"\)]+)([ \t]+"([^"]+)")?\)/, "\\1", "g", line);
+    href  = gensub(/^!\[([^]]+)\]\(([^"\)]+)([ \t]+"([^"]+)")?\)/, "\\2", "g", line);
+    title = gensub(/^!\[([^]]+)\]\(([^"\)]+)([ \t]+"([^"]+)")?\)/, "\\4", "g", line);
+    if ( title ) {
+      return "<img src=\"" HTML(href) "\" alt=\"" HTML(text) "\" title=\"" HTML(title) "\" />" inline( substr( line, len + 1) );
+    } else {
+      return "<img src=\"" HTML(href) "\" alt=\"" HTML(text) "\" />" inline( substr( line, len + 1) );
+    }
+
+  # reference style images
+  } else if ( match(line, /^!\[([^]]+)\] ?\[([^]]*)\]/ ) ) {
+    len = RLENGTH;
+    text = gensub(/^!\[([^\n]+)\] ?\[([^\n]*)\].*/, "\\1", 1, line);
+      id = gensub(/^!\[([^\n]+)\] ?\[([^\n]*)\].*/, "\\2", 1, line);
+    if ( ! id ) id = text;
+    if ( rl_href[id] && rl_title[id] ) {
+      return "<img src=\"" rl_href[id] "\" alt=\"" HTML(text) "\" title=\"" rl_title[id] "\" />" inline( substr( line, len + 1) );
+    } else if ( rl_href[id] ) {
+      return "<img src=\"" rl_href[id] "\" alt=\"" HTML(text) "\" />" inline( substr( line, len + 1) );
+    } else {
+      return "" substr(line, 1, len) inline( substr(line, len + 1) );
+    }
+
   #  ~~strikeout~~ (pandoc)
   } else if ( match(line, /^~~([[:graph:]]|[[:graph:]]([^~]|~[^~])*[[:graph:]])~~/) ) {
     len = RLENGTH;
@@ -171,7 +202,7 @@ function inline( line, LOCAL, len, code, href, guard ) {
     return "<strong>" inline( substr( line, 3, len - 4 ) ) "</strong>" inline( substr( line, len + 1 ) );
 
   #  __strong__
-  } else if ( match(line, "^__(([^_[:space:]]|" ieu ")|([^_[:space:]]|" ieu ")(" nu "|" ieu ")*([^_[:space:]]|" ieu "))__[[:space:]]") ) {
+  } else if ( match(line, "^__(([^_[:space:]]|" ieu ")|([^_[:space:]]|" ieu ")(" nu "|" ieu ")*([^_[:space:]]|" ieu "))__[[:space:][:punct:]]") ) {
     len = RLENGTH;
     return "<strong>" inline( substr( line, 3, len - 5 ) ) "</strong>" inline( substr( line, len) );
 
@@ -186,7 +217,7 @@ function inline( line, LOCAL, len, code, href, guard ) {
     return "<em>" inline( substr( line, 2, len - 2 ) ) "</em>" inline( substr( line, len + 1 ) );
 
   #  _em_
-  } else if ( match(line, "^_(([^_[:space:]]|" isu ")|([^_[:space:]]|" isu ")(" nu "|" isu ")*([^_[:space:]]|" isu "))_[[:space:]]") ) {
+  } else if ( match(line, "^_(([^_[:space:]]|" isu ")|([^_[:space:]]|" isu ")(" nu "|" isu ")*([^_[:space:]]|" isu "))_[[:space:][:punct:]]") ) {
     len = RLENGTH;
     return "<em>" inline( substr( line, 2, len - 3 ) ) "</em>" inline( substr( line, len ) );
 
@@ -195,6 +226,20 @@ function inline( line, LOCAL, len, code, href, guard ) {
     len = RLENGTH;
     return "<em>" inline( substr( line, 2, len - 2 ) ) "</em>" inline( substr( line, len + 1 ) );
 
+  # Verbatim inline HTML
+  } else if ( AllowHTML && match( line, /^(<!--([^-]|-[^-]|--[^>])*-->|<\?([^\?]|\?[^>])*\?>|<![A-Z][^>]*>|<!\[CDATA\[([^\]]|\][^\]]|\]\][^>])*\]\]>|<\/[A-Za-z][A-Za-z0-9-]*[[:space:]]*>|<[A-Za-z][A-Za-z0-9-]*([[:space:]]+[A-Za-z_:][A-Za-z0-9_\.:-]*([[:space:]]*=[[:space:]]*([[:space:]"'=<>`]+|"[^"]*"|'[^']*'))?)*[[:space:]]*\/?>)/) ) {
+    len = RLENGTH;
+    return substr( line, 1, len) inline(substr(line, len + 1));
+
+  # Literal HTML entities
+  } else if ( match( line, /^&([a-zA-Z]{2,32}|#[0-9]{1,7}|#[xX][0-9a-fA-F]{1,6});/) ) {
+    len = RLENGTH;
+    return substr( line, 1, len ) inline(substr(line, len + 1));
+
+  # Escape lone HTML character
+  } else if ( match( line, /^[&<>"']/) ) {
+    return HTML(substr(line, 1, 1)) inline(substr(line, 2));
+
   #  continue walk over string
   } else {
     return substr(line, 1, 1) inline( substr(line, 2) );
@@ -206,6 +251,33 @@ function _block( block, LOCAL, st, len, hlvl, htxt, guard, code, indent ) {
 
   if ( block == "" ) {
     return "";
+
+  # HTML #2 #3 #4 $5
+  } else if ( AllowHTML && match( block, /(^|\n) ? ? ?(<!--([^-]|-[^-]|--[^>])*(-->|$)|<\?([^\?]|\?[^>])*(\?>|$)|<![A-Z][^>]*(>|$)|<!\[CDATA\[([^\]]|\][^\]]|\]\][^>])*(\]\]>|$))/) ) {
+    len = RLENGTH; st = RSTART;
+    return _block(substr(block, 1, st - 1)) substr(block, st, len) _block(substr(block, st + len));
+
+  # HTML #6
+  } else if ( AllowHTML && match( tolower(block), /(^|\n) ? ? ?<\/?(address|article|aside|base|basefont|blockquote|body|caption|center|col|colgroup|dd|details|dialog|dir|div|dl|dt|fieldset|figcaption|figure|footer|form|frame|frameset|h[123456]|head|header|hr|html|iframe|legend|li|link|main|menu|menuitem|nav|noframes|ol|optgroup|option|p|param|section|source|summary|table|tbody|td|tfoot|th|thead|title|tr|track|ul)([[:space:]\n>]|\/>)([^\n]|\n[ \t]*[^\n])*(\n[[:space:]]*\n|$)/) ) {
+    len = RLENGTH; st = RSTART;
+    return _block(substr(block, 1, st - 1)) substr(block, st, len) _block(substr(block, st + len));
+
+  # HTML #1
+  } else if ( AllowHTML && match( tolower(block), /(^|\n) ? ? ?<(script|pre|style)([[:space:]\n>]).*(<\/script>|<\/pre>|<\/style>|$)/) ) {
+    len = RLENGTH; st = RSTART;
+    match( tolower(substr(block, st, len)), /(<\/script>|<\/pre>|<\/style>)/);
+    len = RSTART + RLENGTH;
+    return _block(substr(block, 1, st - 1)) substr(block, st, len) _block(substr(block, st + len));
+
+  # HTML #7
+  } else if ( AllowHTML && match( block, /^ ? ? ?(<\/[A-Za-z][A-Za-z0-9-]*[[:space:]]*>|<[A-Za-z][A-Za-z0-9-]*([[:space:]]+[A-Za-z_:][A-Za-z0-9_\.:-]*([[:space:]]*=[[:space:]]*([[:space:]"'=<>`]+|"[^"]*"|'[^']*'))?)*[[:space:]]*\/?>)([[:space:]]*\n)([^\n]|\n[ \t]*[^\n])*(\n[[:space:]]*\n|$)/) ) {
+    len = RLENGTH; st = RSTART;
+    return substr(block, st, len) _block(substr(block, st + len));
+
+  # Horizontal rule
+  } else if ( match( block, /(^|\n) ? ? ?((\* *){3,}|(- *){3,}|(_ *){3,})($|\n)/) ) {
+    len = RLENGTH; st = RSTART;
+    return _block(substr(block, 1, st - 1)) "<hr />\n" _block(substr(block, st + len));
  
   # Blockquote (leading >)
   } else if ( match( block, /^> /) ) {
@@ -330,6 +402,7 @@ function _list( block, last, LOCAL, p) {
 BEGIN {
   # Global Vars
   file = ""; rl_href[""] = ""; rl_title[""] = "";
+  if (ENVIRON["MD_HTML"] == "true") { AllowHTML = "true"; }
 
   # Buffering of full file ist necessary, e.g. to find reference links
   while (getline) { file = file $0 "\n"; }