Markdown.pl: improve ```-delimited code blocks handling

Handle ```-delimited code blocks earlier so that tabs within them
can be correctly expanded to 8-character tab stop positions and
also to avoid the result being incorrectly interpreted any further.

Signed-off-by: Kyle J. McKay <mackyle@gmail.com>

Markdown.pl

@@ -35,7 +35,7 @@ $VERSION = '1.0.3';
 # Global default settings:
 #
 my $g_empty_element_suffix = " />";	# Change to ">" for HTML output
-my $g_tab_width = 4;
+my $g_tab_width = 4;			# Legacy even though it's wrong
 
 
 #
@@ -290,6 +290,9 @@ sub Markdown {
     # Make sure $text ends with a couple of newlines:
     $text .= "\n\n";
 
+    # Handle backticks-delimited code blocks
+    $text = _HashBTCodeBlocks($text);
+
     # Convert all tabs to spaces.
     $text = _Detab($text);
 
@@ -313,6 +316,40 @@ sub Markdown {
 }
 
 
+sub _HashBTCodeBlocks {
+#
+#   Process Markdown backticks (```) delimited code blocks
+#
+    my $text = shift;
+
+    $text =~ s{
+	    (?:\n|\A)
+		``(`+)[ \t]*(?:([\w.+-]+)[ \t]*)?\n
+	    (		    # $3 = the code block -- one or more lines, starting with ```
+	      (?:
+		.*\n+
+	      )+?
+	    )
+	    (?:(?:``\1[ \t]*(?:\n|\Z))|\Z) # and ending with ``` or end of document
+	}{
+	    # $2 contains syntax highlighting to use if defined
+	    my $codeblock = $3;
+	    $codeblock =~ s/[ \t]+$//mg; # trim trailing spaces on lines
+	    $codeblock = _Detab($codeblock, 8); # physical tab stops are always 8
+	    $codeblock =~ s/\A\n+//; # trim leading newlines
+	    $codeblock =~ s/\s+\z//; # trim trailing whitespace
+	    $codeblock = _EncodeCode($codeblock); # or run highlighter here
+	    $codeblock = "<pre><code>" . $codeblock . "\n</code></pre>";
+
+	    my $key = md5_hex($codeblock);
+	    $g_html_blocks{$key} = $codeblock;
+	    "\n\n" . $key . "\n\n";
+	}egmx;
+
+    return $text;
+}
+
+
 sub _StripLinkDefinitions {
 #
 # Strips link definitions from text, stores the URLs and titles in
@@ -1003,29 +1040,6 @@ sub _DoCodeBlocks {
 	    $result;
 	}egmx;
 
-    $text =~ s{
-	    (?:\n|\A)
-		``(`+)[ \t]*(?:[\w.-]+[ \t]*)?\n
-	    (		    # $1 = the code block -- one or more lines, starting with ```
-	      (?:
-		.*\n+
-	      )+?
-	    )
-	    (?:(?:``\1[ \t]*(?:\n|\Z))|\Z) # and ending with ``` or end of document
-	}{
-	    my $codeblock = $2;
-	    my $result; # return value
-
-	    $codeblock = _EncodeCode($codeblock);
-	    $codeblock = _Detab($codeblock);
-	    $codeblock =~ s/\A\n+//; # trim leading newlines
-	    $codeblock =~ s/\s+\z//; # trim trailing whitespace
-
-	    $result = "\n\n<pre><code>" . $codeblock . "\n</code></pre>\n\n";
-
-	    $result;
-	}egmx;
-
     return $text;
 }
 

basics.text

@@ -293,11 +293,40 @@ Output:
 
 To specify an entire block of pre-formatted code, indent every line of
 the block by 4 spaces or 1 tab. Just like with code spans, `&`, `<`,
-and `>` characters will be escaped automatically. Alternatively an
-entire block of pre-formatted code may be preceded with a line consisting
-of 3 backtick quotes (or more) and followed by a line consisting of the
-same number of backtick quotes -- in which case the code itself does not
-need to be additionally indented.
+and `>` characters will be escaped automatically.
+
+Alternatively an entire block of pre-formatted code may be preceded with a
+line consisting of 3 backtick quotes (or more) and followed by a line
+consisting of the same number of backtick quotes -- in which case the code
+itself does not need to be additionally indented.  The first line may
+optionally have a syntax specifier (e.g. sh, c, perl, etc.) appended.
+Note also that any physical tab characters within a 3-backtick-quotes,
+non-indented code block are always expanded correctly to 8-character tab-stop
+positions (which differs from the 4-character positions used in the rest of a
+markdown file).  This is to facilitate simple copy-and-paste to include code
+snippets.
+
+Markdown:
+
+    ```
+    # This is a simple code block with unspecified syntax
+    ```
+
+Output:
+
+    <pre><code># This is a simple code block with unspecified syntax
+    </code></pre>
+
+Markdown:
+
+    ``` perl
+    my $var = "value"; # this should be highlighted as Perl code
+    ```
+
+Output:
+
+    <pre><code>my $var = "value"; # this should be highlighted as Perl code
+    </code></pre>
 
 Markdown:
 

syntax.text

@@ -535,6 +535,9 @@ Note that the 3 backtick quotes (or more) must appear at the beginning
 of the line.  To include a code block within a list (or other indented
 element), the indentation technique must be used.
 
+Also note that within a backticks-delimited code block, tab characters
+are always expanded with the tab stop locations 8 characters apart.
+
 Within a code block, ampersands (`&`) and angle brackets (`<` and `>`)
 are automatically converted into HTML entities. This makes it very
 easy to include example HTML source code using Markdown -- just paste