X-Git-Url: https://git.plutz.net/?p=cgilite;a=blobdiff_plain;f=markdown.awk;h=361e6005c3fc42fa5e1f81859d69ec59d0c0a138;hp=e652be1d7a992cd4c129251c41ad1d917102069a;hb=142f5b0b60216dddb94ef30684c300b633f269ed;hpb=987f4ef7db8df774e437d36f199be7c2e1396584

diff --git a/markdown.awk b/markdown.awk
index e652be1..361e600 100755
--- a/markdown.awk
+++ b/markdown.awk
@@ -24,7 +24,7 @@
 # - [x] Lists (ordered, unordered)
 # - [x] Code blocks (using indention)
 # - [x] Horizontal rules
-# - [x] Verbatim HTML block (gfm)
+# - [x] Verbatim HTML block (disabled by default)
 #
 # Basic Markdown - Inline elements:
 # ---------------------------------
@@ -35,11 +35,15 @@
 # - [x] Images / reference style images
 # - [x] <automatic links>
 # - [x] backslash escapes
-# - [ ] Verbatim HTML inline
+# - [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:
 # ----------------------------
 # -  ?  Heading identifiers (php md, pandoc)
+# - [x] Automatic heading identifiers (custom)
 # - [x] Fenced code blocks (php md, pandoc)
 #   - [-] Fenced code attributes
 # - [ ] Tables
@@ -48,7 +52,7 @@
 #   -  ?  Grid table (pandoc)
 #   -  ?  Pipe table (php md pandoc)
 # - [x] Line blocks (pandoc)
-# - [ ] Task lists (pandoc)
+# - [x] Task lists (pandoc)
 # - [ ] Definition lists (php md, pandoc)
 # - [-] Numbered example lists (pandoc)
 # - [-] Metadata blocks (pandoc)
@@ -223,6 +227,20 @@ 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) );
@@ -235,45 +253,27 @@ function _block( block, LOCAL, st, len, hlvl, htxt, guard, code, indent ) {
   if ( block == "" ) {
     return "";
 
-  # HTML #2
-  } else if ( match( block, /(^|\n) ? ? ?<!--([^-]|-[^-]|--[^>])*(-->|$)/) ) {
-    len = RLENGTH; st = RSTART;
-    return _block(substr(block, 1, st - 1)) substr(block, st, len) _block(substr(block, st + len));
-
-  # HTML #3
-  } else if ( match( block, /(^|\n) ? ? ?<\?([^\?]|\?[^>])*(\?>|$)/) ) {
-    len = RLENGTH; st = RSTART;
-    return _block(substr(block, 1, st - 1)) substr(block, st, len) _block(substr(block, st + len));
-
-  # HTML #4
-  } else if ( match( block, /(^|\n) ? ? ?<![A-Z][^>]*(>|$)/) ) {
-    len = RLENGTH; st = RSTART;
-    return _block(substr(block, 1, st - 1)) substr(block, st, len) _block(substr(block, st + len));
-
-  # HTML #5
-  } else if ( match( block, /(^|\n) ? ? ?<!\[CDATA\[([^\]]|\][^\]]|\]\][^>])*(\]\]>|$)/) ) {
+  # 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 ( 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|$)/) ) {
+  } else if ( AllowHTML && match( tolower(block), /(^|\n) ? ? ?<\/?(address|article|aside|base|basefont|blockquote|body|caption|center|col|colgroup|dd|details|dialog|dir|div|dl|dt|fieldset|figcaption|figure|footer|form|frame|frameset|h[123456]|head|header|hr|html|iframe|legend|li|link|main|menu|menuitem|nav|noframes|ol|optgroup|option|p|param|section|source|summary|table|tbody|td|tfoot|th|thead|title|tr|track|ul)([[:space:]\n>]|\/>)([^\n]|\n[ \t]*[^\n])*(\n[[:space:]]*\n|$)/) ) {
     len = RLENGTH; st = RSTART;
     return _block(substr(block, 1, st - 1)) substr(block, st, len) _block(substr(block, st + len));
 
   # HTML #1
-  } else if ( match( tolower(block), /(^|\n) ? ? ?<(script|pre|style)([[:space:]\n>]).*(<\/script>|<\/pre>|<\/style>|$)/) ) {
+  } else if ( AllowHTML && match( tolower(block), /(^|\n) ? ? ?<(script|pre|style)([[:space:]\n>]).*(<\/script>|<\/pre>|<\/style>|$)/) ) {
     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 ( match( block, /^ ? ? ?(<\/[A-Za-z][A-Za-z0-9-]* \t]*>|<[A-Za-z][A-Za-z0-9-]*([ \t]+[A-Za-z_:][A-Za-z0-9_\.:-]*[ \t]*=[ \t]*([[:space:]"'=<>`]+|"[^"]*"|'[^']*'))*[ \t]*\/?>)([[:space:]]*\n).*(\n[[:space:]]*\n|$)/) ) {
+  } else if ( AllowHTML && match( block, /^ ? ? ?(<\/[A-Za-z][A-Za-z0-9-]*[[:space:]]*>|<[A-Za-z][A-Za-z0-9-]*([[:space:]]+[A-Za-z_:][A-Za-z0-9_\.:-]*([[:space:]]*=[[:space:]]*([[:space:]"'=<>`]+|"[^"]*"|'[^']*'))?)*[[:space:]]*\/?>)([[:space:]]*\n)([^\n]|\n[ \t]*[^\n])*(\n[[:space:]]*\n|$)/) ) {
     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, /^> /) ) {
@@ -343,23 +343,32 @@ function _block( block, LOCAL, st, len, hlvl, htxt, guard, code, indent ) {
   # First Order Heading
   } else if ( match( block, /^[^\n]+\n===+(\n|$)/ ) ) {
     len = RLENGTH;
-    return "<h1>" inline( gensub( /\n.*$/, "", "g", block ) ) "</h1>\n\n" \
+    HL[1]++; HL[2] = 0; HL[3] = 0; HL[4] = 0; HL[5] = 0; HL[6] = 0;
+    return "<h1 id=\"" HL[1] " - " HTML(gensub( /\n.*$/, "", "g", block )) "\">" inline( gensub( /\n.*$/, "", "g", block ) ) "</h1>\n\n" \
            _block( substr( block, len + 1 ) );
 
   # Second Order Heading
   } else if ( match( block, /^[^\n]+\n---+(\n|$)/ ) ) {
     len = RLENGTH;
-    return "<h2>" inline( gensub( /\n.*$/, "", "g", block ) ) "</h2>\n\n" \
+    HL[2]++; HL[3] = 0; HL[4] = 0; HL[5] = 0; HL[6] = 0;
+    return "<h2 id=\"" HL[1] "." HL[2] " - " HTML(gensub( /\n.*$/, "", "g", block )) "\">" inline( gensub( /\n.*$/, "", "g", block ) ) "</h2>\n\n" \
            _block( substr( block, len + 1) );
 
   # Nth Order Heading
-  } else if ( match( block, /^#{1,6}[[:space:]]*[^\n]+([[:space:]]*#*)(\n|$)/ ) ) {
+  } else if ( match( block, /^#{1,6}[ \t]*[^\n]+([ \t]*#*)(\n|$)/ ) ) {
     len = RLENGTH;
     hlvl = length( gensub( /^(#{1,6}).*$/, "\\1", "g", block ) );
-    htxt = gensub( /[[:space:]]*#*$/, "", "1", gensub( /^#{1,6}[[:space:]]*([^\n]+)([[:space:]]*#*)\n.*$/, "\\1", "g", block ) )
-    return "<h" hlvl ">" inline( htxt ) "</h" hlvl ">\n\n" \
+    htxt = gensub(/^#{1,6}[ \t]*(([^ \t\n]+|[ \t]+[^ \t\n#]|[ \t]+#+[^\n#])+)([ \t]*#*)(\n.*)?$/, "\\1", 1, block);
+    HL[hlvl]++; for ( n = hlvl + 1; n < 7; n++) { HL[n] = 0;}
+    hid = HL[1]; for ( n = 2; n <= hlvl; n++) { hid = hid "." HL[n] ; }
+    return "<h" hlvl " id=\"" hid " - " HTML(htxt) "\">" inline( htxt ) "</h" hlvl ">\n\n" \
            _block( substr( block, len + 1) );
 
+  # Horizontal rule
+  } else if ( match( block, /(^|\n) ? ? ?((\* *){3,}|(- *){3,}|(_ *){3,})($|\n)/) ) {
+    len = RLENGTH; st = RSTART;
+    return _block(substr(block, 1, st - 1)) "<hr />\n" _block(substr(block, st + len));
+
   # Plain paragraph
   } else {
     match( block, /(^|\n)[[:space:]]*(\n|$)/ ) || match( block, /$/ );
@@ -392,15 +401,25 @@ function _list( block, last, LOCAL, p) {
     sub( /<\/p>\n/, "", p );
   }
   sub( /\n$/, "", p );
+
+  # Task List (pandoc)
+       if ( p ~ /^\[ \].*/ )       { p = "<input type=checkbox disabled />" substr(p, 4); }
+  else if ( p ~ /^\[[xX]\].*/ )    { p = "<input type=checkbox disabled checked />" substr(p, 4); }
+  else if ( p ~ /^<p>\[ \].*/ )    { p = "<p><input type=checkbox disabled />" substr(p, 7); }
+  else if ( p ~ /^<p>\[[xX]\].*/ ) { p = "<p><input type=checkbox disabled checked />" substr(p, 7); }
   return "<li>" p "</li>\n" _list( block, last );
 }
 
 BEGIN {
   # Global Vars
   file = ""; rl_href[""] = ""; rl_title[""] = "";
+  if (ENVIRON["MD_HTML"] == "true") { AllowHTML = "true"; }
+  HL[1] = 0; HL[2] = 0; HL[3] = 0; HL[4] = 0; HL[5] = 0; HL[6] = 0;
 
   # Buffering of full file ist necessary, e.g. to find reference links
   while (getline) { file = file $0 "\n"; }
+  # Clean up MS-DOS line breaks
+  gsub(/\r\n/, "\n", file);
 
   # Fill array of reference links
   f = file; rl_id;