From ba2177e2cfda958ea12c5b04dbf60907aaa8816d Mon Sep 17 00:00:00 2001
From: Yorick Peterse <yorickpeterse@gmail.com>
Date: Mon, 2 Mar 2015 16:20:54 +0100
Subject: [PATCH] Lex contents of <script> tags as plain text.

When lexing input in HTML mode the lexer has to treat _all_ content of a
<script> tag as plain text. This ensures that the lexer can process input such
as "x <y" and "// <foo>" correctly.

Fixes #70.
---
 ext/c/lexer.rl                   | 11 +++++++++-
 ext/java/org/liboga/xml/Lexer.rl | 11 ++++++++++
 ext/ragel/base_lexer.rl          | 36 +++++++++++++++++++++++++++++++-
 lib/oga/xml/lexer.rb             | 17 +++++++++++++++
 4 files changed, 73 insertions(+), 2 deletions(-)

diff --git a/ext/c/lexer.rl b/ext/c/lexer.rl
index 0defe64..73039dd 100644
--- a/ext/c/lexer.rl
+++ b/ext/c/lexer.rl
@@ -17,7 +17,13 @@ on `ts` and `te`) so the macro ignores this argument.
     liboga_xml_lexer_callback_simple(self, name);
 
 #define advance_line(amount) \
-    rb_funcall(self, rb_intern("advance_line"), 1, INT2NUM(amount));
+    rb_funcall(self, id_advance_line, 1, INT2NUM(amount));
+
+#define inside_html_script_p() \
+    rb_funcall(self, id_inside_html_script_p, 0) == Qtrue
+
+ID id_advance_line;
+ID id_inside_html_script_p;
 
 %%machine c_lexer;
 
@@ -167,6 +173,9 @@ void Init_liboga_xml_lexer()
     VALUE mXML   = rb_const_get(mOga, rb_intern("XML"));
     VALUE cLexer = rb_define_class_under(mXML, "Lexer", rb_cObject);
 
+    id_advance_line         = rb_intern("advance_line");
+    id_inside_html_script_p = rb_intern("inside_html_script?");
+
     rb_define_method(cLexer, "advance_native", oga_xml_lexer_advance, 1);
     rb_define_method(cLexer, "reset_native", oga_xml_lexer_reset, 0);
 
diff --git a/ext/java/org/liboga/xml/Lexer.rl b/ext/java/org/liboga/xml/Lexer.rl
index d746dd2..a359234 100644
--- a/ext/java/org/liboga/xml/Lexer.rl
+++ b/ext/java/org/liboga/xml/Lexer.rl
@@ -185,6 +185,17 @@ public class Lexer extends RubyObject
 
         this.callMethod(context, "advance_line", lines);
     }
+
+    /**
+     * Returns true if we're in an HTML script tag. See
+     * Oga::XML::Lexer#inside_html_script? for more information.
+     */
+    public Boolean inside_html_script_p()
+    {
+        ThreadContext context = this.runtime.getCurrentContext();
+
+        return this.callMethod(context, "inside_html_script?").isTrue();
+    }
 }
 
 %%{
diff --git a/ext/ragel/base_lexer.rl b/ext/ragel/base_lexer.rl
index 305f9e2..9b107d9 100644
--- a/ext/ragel/base_lexer.rl
+++ b/ext/ragel/base_lexer.rl
@@ -327,7 +327,17 @@
         # We're done with the open tag of the element.
         '>' => {
             callback_simple(id_on_element_open_end);
-            fnext main;
+
+            if ( inside_html_script_p() )
+            {
+                mark = ts + 1;
+
+                fnext script_text;
+            }
+            else
+            {
+                fnext main;
+            }
         };
 
         # Self closing tags.
@@ -391,6 +401,30 @@
         };
     *|;
 
+    # <script> tags in HTML can contain basically anything except for the
+    # literal "</script>". As a result of this we can't use the regular text
+    # machine.
+    script_text := |*
+        '</script>' => {
+            callback(id_on_text, data, encoding, mark, ts);
+
+            mark = 0;
+
+            if ( lines > 0 )
+            {
+                advance_line(lines);
+
+                lines = 0;
+            }
+
+            callback_simple(id_on_element_end);
+
+            fnext main;
+        };
+
+        any $count_newlines;
+    *|;
+
     # The main machine aka the entry point of Ragel.
     main := |*
         doctype_start  => start_doctype;
diff --git a/lib/oga/xml/lexer.rb b/lib/oga/xml/lexer.rb
index fa1757c..14f0784 100644
--- a/lib/oga/xml/lexer.rb
+++ b/lib/oga/xml/lexer.rb
@@ -40,6 +40,14 @@ module Oga
     class Lexer
       attr_reader :html
 
+      ##
+      # Element name used to determine if a tag being processed is a Javascript
+      # tag.
+      #
+      # @return [String]
+      #
+      SCRIPT_TAG = 'script'.freeze
+
       ##
       # @param [String|IO] data The data to lex. This can either be a String or
       #  an IO instance.
@@ -181,6 +189,15 @@ module Oga
         return @elements.last
       end
 
+      ##
+      # Returns true if the current element is the HTML `<script>` element.
+      #
+      # @return [TrueClass|FalseClass]
+      #
+      def inside_html_script?
+        return html? && current_element == SCRIPT_TAG
+      end
+
       ##
       # Called when processing a single quote.
       #