trenni 2.0.1 → 2.0.2

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA1:
-  metadata.gz: 5f6e48e9727b8f616cace8e7090227e13388ef58
-  data.tar.gz: 6e871441287d048b9dda16dd0dc428510b8055d2
+  metadata.gz: 7f978da77175a9071cf75459ecc67b76c46611a5
+  data.tar.gz: 69e8b067fdee4bcd19fcc9380c58c7e905571e9f
 SHA512:
-  metadata.gz: 9f45256745878e56128149e0a90a0f0e1cc171e8e17a1ce7feb02caff0f4d5a6a787be74e392e0d56991f40d1321eb2148ded8e1e23c30d14933ac160ffc91cf
-  data.tar.gz: 40f78cceea83878d176d71de6f0c29c45111e9b5a2dcb8fa1f83cd48f739192cacbb8508d44c293d97244e8679fd981fcf400963d32eed5bfa06bfa541ba6fe6
+  metadata.gz: 061165c578720fe3647d7290a1dfc293826a3366f22ad4d4ea52596ede57398535b29734cc5da009a540ae21059e7f226df8c93d0cb37d4ca8870194e64e0361
+  data.tar.gz: e7c2baa36ea1f88c9b16ccbccc26639c454ca3e5e812e20c7e6505b55a97043e592e46efb455a71b7b2156a7e837916f8948d307a0535a9b11ab794407d77ddf

data/Gemfile

@@ -10,6 +10,9 @@ group :development do
 	gem "benchmark-ips"
 	
 	gem "rake-compiler"
+	
+	# For comparisons:
+	gem "nokogiri"
 end
 
 group :test do

data/README.md

@@ -18,6 +18,36 @@ In addition, I wanted a simple markup parser and builder for HTML style markup.
 
 The 2nd release of Trenni in 2016 saw an overhaul of the internal parsers. I used [Ragel](http://www.colm.net/open-source/ragel/) to implement efficient event-based markup and template parsers, which can be compiled to both C and Ruby. This provides a native code path where possible giving speed-ups between 10x - 20x. In addition, the formal grammar is more robust.
 
+## Is it fast?
+
+It's faster than Nokogiri for parsing markup:
+
+	Trenni::Native
+	Warming up --------------------------------------
+	      Large (Trenni)    71.000  i/100ms
+	    Large (Nokogiri)    28.000  i/100ms
+	Calculating -------------------------------------
+	      Large (Trenni)    662.050  (± 3.9%) i/s -      3.337k in   5.048115s
+	    Large (Nokogiri)    266.878  (±10.9%) i/s -      1.316k in   5.008464s
+
+	Comparison:
+	      Large (Trenni):      662.1 i/s
+	    Large (Nokogiri):      266.9 i/s - 2.48x slower
+
+It's significantly faster than ERB:
+
+	Trenni::Template
+	Warming up --------------------------------------
+	     Trenni (object)    75.667k i/100ms
+	       ERB (binding)     6.940k i/100ms
+	Calculating -------------------------------------
+	     Trenni (object)      1.095M (± 7.9%) i/s -      5.448M in   5.007244s
+	       ERB (binding)     69.381k (± 7.1%) i/s -    347.000k in   5.027333s
+
+	Comparison:
+	     Trenni (object):  1094979.9 i/s
+	       ERB (binding):    69381.1 i/s - 15.78x  slower
+
 ## Installation
 
 Add this line to your application's Gemfile:

data/Rakefile

@@ -81,7 +81,7 @@ task :visualize_parsers do
 			pdf_path = parser_path + ".pdf"
 			sh("dot", "-Tpdf", "-o", pdf_path, dot_path)
 			
-			sh("open", pdf_path)
+			sh("open", pdf_path) rescue nil
 		end
 	end
 end

data/ext/trenni/markup.c

@@ -15,7 +15,7 @@ static const int Trenni_markup_parser_en_parse_entity = 42;
 static const int Trenni_markup_parser_en_main = 48;
 
 
-#line 192 "markup.rl"
+#line 190 "markup.rl"
 
 
 VALUE Trenni_Native_parse_markup(VALUE self, VALUE buffer, VALUE delegate, VALUE entities) {
@@ -127,150 +127,131 @@ case 48:
 		case 60: goto tr89;
 	}
 	goto tr87;
-tr90:
-#line 29 "markup.rl"
-	{
-		characters.end = p;
-		
-		Trenni_append_token(&pcdata, encoding, characters);
-	}
-#line 25 "markup.rl"
-	{
-		characters.begin = p;
-	}
-	goto st49;
 tr93:
-#line 25 "markup.rl"
+#line 32 "markup.rl"
 	{
 		characters.begin = p;
 	}
 	goto st49;
 tr87:
-#line 180 "markup.rl"
+#line 25 "markup.rl"
 	{
-	
 	}
 #line 18 "markup.rl"
 	{
 		pcdata = Qnil;
 	}
-#line 25 "markup.rl"
+#line 32 "markup.rl"
 	{
 		characters.begin = p;
 	}
 	goto st49;
 tr96:
-#line 151 "markup.rl"
+#line 158 "markup.rl"
 	{
 		rb_funcall(delegate, id_open_tag_end, 1, self_closing == 1 ? Qtrue : Qfalse);
 	}
-#line 180 "markup.rl"
+#line 25 "markup.rl"
 	{
-	
 	}
 #line 18 "markup.rl"
 	{
 		pcdata = Qnil;
 	}
-#line 25 "markup.rl"
+#line 32 "markup.rl"
 	{
 		characters.begin = p;
 	}
 	goto st49;
 tr99:
-#line 85 "markup.rl"
+#line 92 "markup.rl"
 	{
 		comment.end = p;
 		
 		rb_funcall(delegate, id_comment, 1, Trenni_token(comment, encoding));
 	}
-#line 180 "markup.rl"
+#line 25 "markup.rl"
 	{
-	
 	}
 #line 18 "markup.rl"
 	{
 		pcdata = Qnil;
 	}
-#line 25 "markup.rl"
+#line 32 "markup.rl"
 	{
 		characters.begin = p;
 	}
 	goto st49;
 tr102:
-#line 71 "markup.rl"
+#line 78 "markup.rl"
 	{
 		doctype.end = p;
 		
 		rb_funcall(delegate, id_doctype, 1, Trenni_token(doctype, encoding));
 	}
-#line 180 "markup.rl"
+#line 25 "markup.rl"
 	{
-	
 	}
 #line 18 "markup.rl"
 	{
 		pcdata = Qnil;
 	}
-#line 25 "markup.rl"
+#line 32 "markup.rl"
 	{
 		characters.begin = p;
 	}
 	goto st49;
 tr105:
-#line 170 "markup.rl"
+#line 177 "markup.rl"
 	{
 		cdata.end = p;
 		
 		rb_funcall(delegate, id_cdata, 1, Trenni_token(cdata, encoding));
 	}
-#line 180 "markup.rl"
+#line 25 "markup.rl"
 	{
-	
 	}
 #line 18 "markup.rl"
 	{
 		pcdata = Qnil;
 	}
-#line 25 "markup.rl"
+#line 32 "markup.rl"
 	{
 		characters.begin = p;
 	}
 	goto st49;
 tr108:
-#line 158 "markup.rl"
+#line 165 "markup.rl"
 	{
 		rb_funcall(delegate, id_close_tag, 2, Trenni_token(identifier, encoding), ULONG2NUM(identifier.begin-s));
 	}
-#line 180 "markup.rl"
+#line 25 "markup.rl"
 	{
-	
 	}
 #line 18 "markup.rl"
 	{
 		pcdata = Qnil;
 	}
-#line 25 "markup.rl"
+#line 32 "markup.rl"
 	{
 		characters.begin = p;
 	}
 	goto st49;
 tr111:
-#line 105 "markup.rl"
+#line 112 "markup.rl"
 	{
 		instruction.end = p;
 		
 		rb_funcall(delegate, id_instruction, 1, Trenni_token(instruction, encoding));
 	}
-#line 180 "markup.rl"
+#line 25 "markup.rl"
 	{
-	
 	}
 #line 18 "markup.rl"
 	{
 		pcdata = Qnil;
 	}
-#line 25 "markup.rl"
+#line 32 "markup.rl"
 	{
 		characters.begin = p;
 	}
@@ -279,14 +260,14 @@ st49:
 	if ( ++p == pe )
 		goto _test_eof49;
 case 49:
-#line 283 "markup.c"
+#line 264 "markup.c"
 	switch( (*p) ) {
 		case 38: goto tr91;
 		case 60: goto tr92;
 	}
-	goto tr90;
+	goto st49;
 tr91:
-#line 29 "markup.rl"
+#line 36 "markup.rl"
 	{
 		characters.end = p;
 		
@@ -300,9 +281,8 @@ tr94:
 	{{stack[top++] = 50; goto st42;}}
 	goto st50;
 tr88:
-#line 180 "markup.rl"
+#line 25 "markup.rl"
 	{
-	
 	}
 #line 18 "markup.rl"
 	{
@@ -312,13 +292,12 @@ tr88:
 	{{stack[top++] = 50; goto st42;}}
 	goto st50;
 tr97:
-#line 151 "markup.rl"
+#line 158 "markup.rl"
 	{
 		rb_funcall(delegate, id_open_tag_end, 1, self_closing == 1 ? Qtrue : Qfalse);
 	}
-#line 180 "markup.rl"
+#line 25 "markup.rl"
 	{
-	
 	}
 #line 18 "markup.rl"
 	{
@@ -328,15 +307,14 @@ tr97:
 	{{stack[top++] = 50; goto st42;}}
 	goto st50;
 tr100:
-#line 85 "markup.rl"
+#line 92 "markup.rl"
 	{
 		comment.end = p;
 		
 		rb_funcall(delegate, id_comment, 1, Trenni_token(comment, encoding));
 	}
-#line 180 "markup.rl"
+#line 25 "markup.rl"
 	{
-	
 	}
 #line 18 "markup.rl"
 	{
@@ -346,15 +324,14 @@ tr100:
 	{{stack[top++] = 50; goto st42;}}
 	goto st50;
 tr103:
-#line 71 "markup.rl"
+#line 78 "markup.rl"
 	{
 		doctype.end = p;
 		
 		rb_funcall(delegate, id_doctype, 1, Trenni_token(doctype, encoding));
 	}
-#line 180 "markup.rl"
+#line 25 "markup.rl"
 	{
-	
 	}
 #line 18 "markup.rl"
 	{
@@ -364,15 +341,14 @@ tr103:
 	{{stack[top++] = 50; goto st42;}}
 	goto st50;
 tr106:
-#line 170 "markup.rl"
+#line 177 "markup.rl"
 	{
 		cdata.end = p;
 		
 		rb_funcall(delegate, id_cdata, 1, Trenni_token(cdata, encoding));
 	}
-#line 180 "markup.rl"
+#line 25 "markup.rl"
 	{
-	
 	}
 #line 18 "markup.rl"
 	{
@@ -382,13 +358,12 @@ tr106:
 	{{stack[top++] = 50; goto st42;}}
 	goto st50;
 tr109:
-#line 158 "markup.rl"
+#line 165 "markup.rl"
 	{
 		rb_funcall(delegate, id_close_tag, 2, Trenni_token(identifier, encoding), ULONG2NUM(identifier.begin-s));
 	}
-#line 180 "markup.rl"
+#line 25 "markup.rl"
 	{
-	
 	}
 #line 18 "markup.rl"
 	{
@@ -398,15 +373,14 @@ tr109:
 	{{stack[top++] = 50; goto st42;}}
 	goto st50;
 tr112:
-#line 105 "markup.rl"
+#line 112 "markup.rl"
 	{
 		instruction.end = p;
 		
 		rb_funcall(delegate, id_instruction, 1, Trenni_token(instruction, encoding));
 	}
-#line 180 "markup.rl"
+#line 25 "markup.rl"
 	{
-	
 	}
 #line 18 "markup.rl"
 	{
@@ -419,38 +393,38 @@ st50:
 	if ( ++p == pe )
 		goto _test_eof50;
 case 50:
-#line 423 "markup.c"
+#line 397 "markup.c"
 	switch( (*p) ) {
 		case 38: goto tr94;
 		case 60: goto tr95;
 	}
 	goto tr93;
 tr89:
-#line 122 "markup.rl"
+#line 129 "markup.rl"
 	{
 	}
-#line 155 "markup.rl"
+#line 162 "markup.rl"
 	{
 	}
-#line 95 "markup.rl"
+#line 102 "markup.rl"
 	{
 		instruction.begin = p;
 	}
-#line 81 "markup.rl"
+#line 88 "markup.rl"
 	{
 		comment.begin = p;
 	}
-#line 67 "markup.rl"
+#line 74 "markup.rl"
 	{
 		doctype.begin = p;
 	}
-#line 166 "markup.rl"
+#line 173 "markup.rl"
 	{
 		cdata.begin = p;
 	}
 	goto st1;
 tr92:
-#line 29 "markup.rl"
+#line 36 "markup.rl"
 	{
 		characters.end = p;
 		
@@ -459,30 +433,29 @@ tr92:
 #line 22 "markup.rl"
 	{
 	}
-#line 184 "markup.rl"
+#line 28 "markup.rl"
 	{
-		// Entities are handled separately:
 		rb_funcall(delegate, id_text, 1, pcdata);
 	}
-#line 122 "markup.rl"
+#line 129 "markup.rl"
 	{
 	}
-#line 155 "markup.rl"
+#line 162 "markup.rl"
 	{
 	}
-#line 95 "markup.rl"
+#line 102 "markup.rl"
 	{
 		instruction.begin = p;
 	}
-#line 81 "markup.rl"
+#line 88 "markup.rl"
 	{
 		comment.begin = p;
 	}
-#line 67 "markup.rl"
+#line 74 "markup.rl"
 	{
 		doctype.begin = p;
 	}
-#line 166 "markup.rl"
+#line 173 "markup.rl"
 	{
 		cdata.begin = p;
 	}
@@ -491,206 +464,205 @@ tr95:
 #line 22 "markup.rl"
 	{
 	}
-#line 184 "markup.rl"
+#line 28 "markup.rl"
 	{
-		// Entities are handled separately:
 		rb_funcall(delegate, id_text, 1, pcdata);
 	}
-#line 122 "markup.rl"
+#line 129 "markup.rl"
 	{
 	}
-#line 155 "markup.rl"
+#line 162 "markup.rl"
 	{
 	}
-#line 95 "markup.rl"
+#line 102 "markup.rl"
 	{
 		instruction.begin = p;
 	}
-#line 81 "markup.rl"
+#line 88 "markup.rl"
 	{
 		comment.begin = p;
 	}
-#line 67 "markup.rl"
+#line 74 "markup.rl"
 	{
 		doctype.begin = p;
 	}
-#line 166 "markup.rl"
+#line 173 "markup.rl"
 	{
 		cdata.begin = p;
 	}
 	goto st1;
 tr98:
-#line 151 "markup.rl"
+#line 158 "markup.rl"
 	{
 		rb_funcall(delegate, id_open_tag_end, 1, self_closing == 1 ? Qtrue : Qfalse);
 	}
-#line 122 "markup.rl"
+#line 129 "markup.rl"
 	{
 	}
-#line 155 "markup.rl"
+#line 162 "markup.rl"
 	{
 	}
-#line 95 "markup.rl"
+#line 102 "markup.rl"
 	{
 		instruction.begin = p;
 	}
-#line 81 "markup.rl"
+#line 88 "markup.rl"
 	{
 		comment.begin = p;
 	}
-#line 67 "markup.rl"
+#line 74 "markup.rl"
 	{
 		doctype.begin = p;
 	}
-#line 166 "markup.rl"
+#line 173 "markup.rl"
 	{
 		cdata.begin = p;
 	}
 	goto st1;
 tr101:
-#line 85 "markup.rl"
+#line 92 "markup.rl"
 	{
 		comment.end = p;
 		
 		rb_funcall(delegate, id_comment, 1, Trenni_token(comment, encoding));
 	}
-#line 122 "markup.rl"
+#line 129 "markup.rl"
 	{
 	}
-#line 155 "markup.rl"
+#line 162 "markup.rl"
 	{
 	}
-#line 95 "markup.rl"
+#line 102 "markup.rl"
 	{
 		instruction.begin = p;
 	}
-#line 81 "markup.rl"
+#line 88 "markup.rl"
 	{
 		comment.begin = p;
 	}
-#line 67 "markup.rl"
+#line 74 "markup.rl"
 	{
 		doctype.begin = p;
 	}
-#line 166 "markup.rl"
+#line 173 "markup.rl"
 	{
 		cdata.begin = p;
 	}
 	goto st1;
 tr104:
-#line 71 "markup.rl"
+#line 78 "markup.rl"
 	{
 		doctype.end = p;
 		
 		rb_funcall(delegate, id_doctype, 1, Trenni_token(doctype, encoding));
 	}
-#line 122 "markup.rl"
+#line 129 "markup.rl"
 	{
 	}
-#line 155 "markup.rl"
+#line 162 "markup.rl"
 	{
 	}
-#line 95 "markup.rl"
+#line 102 "markup.rl"
 	{
 		instruction.begin = p;
 	}
-#line 81 "markup.rl"
+#line 88 "markup.rl"
 	{
 		comment.begin = p;
 	}
-#line 67 "markup.rl"
+#line 74 "markup.rl"
 	{
 		doctype.begin = p;
 	}
-#line 166 "markup.rl"
+#line 173 "markup.rl"
 	{
 		cdata.begin = p;
 	}
 	goto st1;
 tr107:
-#line 170 "markup.rl"
+#line 177 "markup.rl"
 	{
 		cdata.end = p;
 		
 		rb_funcall(delegate, id_cdata, 1, Trenni_token(cdata, encoding));
 	}
-#line 122 "markup.rl"
+#line 129 "markup.rl"
 	{
 	}
-#line 155 "markup.rl"
+#line 162 "markup.rl"
 	{
 	}
-#line 95 "markup.rl"
+#line 102 "markup.rl"
 	{
 		instruction.begin = p;
 	}
-#line 81 "markup.rl"
+#line 88 "markup.rl"
 	{
 		comment.begin = p;
 	}
-#line 67 "markup.rl"
+#line 74 "markup.rl"
 	{
 		doctype.begin = p;
 	}
-#line 166 "markup.rl"
+#line 173 "markup.rl"
 	{
 		cdata.begin = p;
 	}
 	goto st1;
 tr110:
-#line 158 "markup.rl"
+#line 165 "markup.rl"
 	{
 		rb_funcall(delegate, id_close_tag, 2, Trenni_token(identifier, encoding), ULONG2NUM(identifier.begin-s));
 	}
-#line 122 "markup.rl"
+#line 129 "markup.rl"
 	{
 	}
-#line 155 "markup.rl"
+#line 162 "markup.rl"
 	{
 	}
-#line 95 "markup.rl"
+#line 102 "markup.rl"
 	{
 		instruction.begin = p;
 	}
-#line 81 "markup.rl"
+#line 88 "markup.rl"
 	{
 		comment.begin = p;
 	}
-#line 67 "markup.rl"
+#line 74 "markup.rl"
 	{
 		doctype.begin = p;
 	}
-#line 166 "markup.rl"
+#line 173 "markup.rl"
 	{
 		cdata.begin = p;
 	}
 	goto st1;
 tr113:
-#line 105 "markup.rl"
+#line 112 "markup.rl"
 	{
 		instruction.end = p;
 		
 		rb_funcall(delegate, id_instruction, 1, Trenni_token(instruction, encoding));
 	}
-#line 122 "markup.rl"
+#line 129 "markup.rl"
 	{
 	}
-#line 155 "markup.rl"
+#line 162 "markup.rl"
 	{
 	}
-#line 95 "markup.rl"
+#line 102 "markup.rl"
 	{
 		instruction.begin = p;
 	}
-#line 81 "markup.rl"
+#line 88 "markup.rl"
 	{
 		comment.begin = p;
 	}
-#line 67 "markup.rl"
+#line 74 "markup.rl"
 	{
 		doctype.begin = p;
 	}
-#line 166 "markup.rl"
+#line 173 "markup.rl"
 	{
 		cdata.begin = p;
 	}
@@ -699,7 +671,7 @@ st1:
 	if ( ++p == pe )
 		goto _test_eof1;
 case 1:
-#line 703 "markup.c"
+#line 675 "markup.c"
 	switch( (*p) ) {
 		case 33: goto st15;
 		case 47: goto st36;
@@ -728,7 +700,7 @@ st2:
 	if ( ++p == pe )
 		goto _test_eof2;
 case 2:
-#line 732 "markup.c"
+#line 704 "markup.c"
 	switch( (*p) ) {
 		case 32: goto tr6;
 		case 47: goto tr7;
@@ -754,24 +726,24 @@ case 2:
 		goto tr1;
 	goto st2;
 tr1:
-#line 162 "markup.rl"
+#line 169 "markup.rl"
 	{
 		Trenni_raise_error("could not parse tag", buffer, p-s);
 	}
 	goto st0;
 tr69:
-#line 111 "markup.rl"
+#line 118 "markup.rl"
 	{
 		Trenni_raise_error("could not parse instruction", buffer, p-s);
 	}
 	goto st0;
 tr75:
-#line 35 "markup.rl"
+#line 42 "markup.rl"
 	{
 		Trenni_raise_error("could not parse entity", buffer, p-s);
 	}
 	goto st0;
-#line 775 "markup.c"
+#line 747 "markup.c"
 st0:
 cs = 0;
 	goto _out;
@@ -780,7 +752,7 @@ tr6:
 	{
 		identifier.end = p;
 	}
-#line 115 "markup.rl"
+#line 122 "markup.rl"
 	{
 		// Reset self-closing state - we don't know yet.
 		self_closing = 0;
@@ -793,7 +765,7 @@ tr14:
 	{
 		identifier.end = p;
 	}
-#line 141 "markup.rl"
+#line 148 "markup.rl"
 	{
 		if (has_value == 1) {
 			rb_funcall(delegate, id_attribute, 2, Trenni_token(identifier, encoding), pcdata);
@@ -805,11 +777,11 @@ tr14:
 	}
 	goto st3;
 tr26:
-#line 133 "markup.rl"
+#line 140 "markup.rl"
 	{
 		has_value = 1;
 	}
-#line 141 "markup.rl"
+#line 148 "markup.rl"
 	{
 		if (has_value == 1) {
 			rb_funcall(delegate, id_attribute, 2, Trenni_token(identifier, encoding), pcdata);
@@ -821,11 +793,11 @@ tr26:
 	}
 	goto st3;
 tr32:
-#line 137 "markup.rl"
+#line 144 "markup.rl"
 	{
 		has_value = 2;
 	}
-#line 141 "markup.rl"
+#line 148 "markup.rl"
 	{
 		if (has_value == 1) {
 			rb_funcall(delegate, id_attribute, 2, Trenni_token(identifier, encoding), pcdata);
@@ -840,7 +812,7 @@ st3:
 	if ( ++p == pe )
 		goto _test_eof3;
 case 3:
-#line 844 "markup.c"
+#line 816 "markup.c"
 	switch( (*p) ) {
 		case 32: goto st3;
 		case 47: goto tr11;
@@ -866,7 +838,7 @@ case 3:
 		goto tr1;
 	goto tr9;
 tr9:
-#line 129 "markup.rl"
+#line 136 "markup.rl"
 	{
 		has_value = 0;
 	}
@@ -879,7 +851,7 @@ st4:
 	if ( ++p == pe )
 		goto _test_eof4;
 case 4:
-#line 883 "markup.c"
+#line 855 "markup.c"
 	switch( (*p) ) {
 		case 32: goto tr14;
 		case 47: goto tr15;
@@ -910,20 +882,20 @@ tr7:
 	{
 		identifier.end = p;
 	}
-#line 115 "markup.rl"
+#line 122 "markup.rl"
 	{
 		// Reset self-closing state - we don't know yet.
 		self_closing = 0;
 		
 		rb_funcall(delegate, id_open_tag_begin, 2, Trenni_token(identifier, encoding), ULONG2NUM(identifier.begin-s));
 	}
-#line 125 "markup.rl"
+#line 132 "markup.rl"
 	{
 		self_closing = 1;
 	}
 	goto st5;
 tr11:
-#line 125 "markup.rl"
+#line 132 "markup.rl"
 	{
 		self_closing = 1;
 	}
@@ -933,7 +905,7 @@ tr15:
 	{
 		identifier.end = p;
 	}
-#line 141 "markup.rl"
+#line 148 "markup.rl"
 	{
 		if (has_value == 1) {
 			rb_funcall(delegate, id_attribute, 2, Trenni_token(identifier, encoding), pcdata);
@@ -943,17 +915,17 @@ tr15:
 			rb_funcall(delegate, id_attribute, 2, Trenni_token(identifier, encoding), Qtrue);
 		}
 	}
-#line 125 "markup.rl"
+#line 132 "markup.rl"
 	{
 		self_closing = 1;
 	}
 	goto st5;
 tr27:
-#line 133 "markup.rl"
+#line 140 "markup.rl"
 	{
 		has_value = 1;
 	}
-#line 141 "markup.rl"
+#line 148 "markup.rl"
 	{
 		if (has_value == 1) {
 			rb_funcall(delegate, id_attribute, 2, Trenni_token(identifier, encoding), pcdata);
@@ -963,17 +935,17 @@ tr27:
 			rb_funcall(delegate, id_attribute, 2, Trenni_token(identifier, encoding), Qtrue);
 		}
 	}
-#line 125 "markup.rl"
+#line 132 "markup.rl"
 	{
 		self_closing = 1;
 	}
 	goto st5;
 tr33:
-#line 137 "markup.rl"
+#line 144 "markup.rl"
 	{
 		has_value = 2;
 	}
-#line 141 "markup.rl"
+#line 148 "markup.rl"
 	{
 		if (has_value == 1) {
 			rb_funcall(delegate, id_attribute, 2, Trenni_token(identifier, encoding), pcdata);
@@ -983,7 +955,7 @@ tr33:
 			rb_funcall(delegate, id_attribute, 2, Trenni_token(identifier, encoding), Qtrue);
 		}
 	}
-#line 125 "markup.rl"
+#line 132 "markup.rl"
 	{
 		self_closing = 1;
 	}
@@ -992,7 +964,7 @@ st5:
 	if ( ++p == pe )
 		goto _test_eof5;
 case 5:
-#line 996 "markup.c"
+#line 968 "markup.c"
 	if ( (*p) == 62 )
 		goto st51;
 	goto tr1;
@@ -1001,7 +973,7 @@ tr8:
 	{
 		identifier.end = p;
 	}
-#line 115 "markup.rl"
+#line 122 "markup.rl"
 	{
 		// Reset self-closing state - we don't know yet.
 		self_closing = 0;
@@ -1014,7 +986,7 @@ tr17:
 	{
 		identifier.end = p;
 	}
-#line 141 "markup.rl"
+#line 148 "markup.rl"
 	{
 		if (has_value == 1) {
 			rb_funcall(delegate, id_attribute, 2, Trenni_token(identifier, encoding), pcdata);
@@ -1026,11 +998,11 @@ tr17:
 	}
 	goto st51;
 tr28:
-#line 133 "markup.rl"
+#line 140 "markup.rl"
 	{
 		has_value = 1;
 	}
-#line 141 "markup.rl"
+#line 148 "markup.rl"
 	{
 		if (has_value == 1) {
 			rb_funcall(delegate, id_attribute, 2, Trenni_token(identifier, encoding), pcdata);
@@ -1042,11 +1014,11 @@ tr28:
 	}
 	goto st51;
 tr34:
-#line 137 "markup.rl"
+#line 144 "markup.rl"
 	{
 		has_value = 2;
 	}
-#line 141 "markup.rl"
+#line 148 "markup.rl"
 	{
 		if (has_value == 1) {
 			rb_funcall(delegate, id_attribute, 2, Trenni_token(identifier, encoding), pcdata);
@@ -1061,7 +1033,7 @@ st51:
 	if ( ++p == pe )
 		goto _test_eof51;
 case 51:
-#line 1065 "markup.c"
+#line 1037 "markup.c"
 	switch( (*p) ) {
 		case 38: goto tr97;
 		case 60: goto tr98;
@@ -1077,7 +1049,7 @@ st6:
 	if ( ++p == pe )
 		goto _test_eof6;
 case 6:
-#line 1081 "markup.c"
+#line 1053 "markup.c"
 	switch( (*p) ) {
 		case 34: goto st7;
 		case 39: goto st12;
@@ -1098,25 +1070,13 @@ tr20:
 	{
 		pcdata = Qnil;
 	}
-#line 25 "markup.rl"
-	{
-		characters.begin = p;
-	}
-	goto st8;
-tr23:
-#line 29 "markup.rl"
-	{
-		characters.end = p;
-		
-		Trenni_append_token(&pcdata, encoding, characters);
-	}
-#line 25 "markup.rl"
+#line 32 "markup.rl"
 	{
 		characters.begin = p;
 	}
 	goto st8;
 tr29:
-#line 25 "markup.rl"
+#line 32 "markup.rl"
 	{
 		characters.begin = p;
 	}
@@ -1125,15 +1085,15 @@ st8:
 	if ( ++p == pe )
 		goto _test_eof8;
 case 8:
-#line 1129 "markup.c"
+#line 1089 "markup.c"
 	switch( (*p) ) {
 		case 34: goto tr24;
 		case 38: goto tr25;
 		case 60: goto tr1;
 	}
-	goto tr23;
+	goto st8;
 tr24:
-#line 29 "markup.rl"
+#line 36 "markup.rl"
 	{
 		characters.end = p;
 		
@@ -1152,7 +1112,7 @@ st9:
 	if ( ++p == pe )
 		goto _test_eof9;
 case 9:
-#line 1156 "markup.c"
+#line 1116 "markup.c"
 	switch( (*p) ) {
 		case 32: goto tr26;
 		case 47: goto tr27;
@@ -1170,7 +1130,7 @@ tr22:
 	{{stack[top++] = 10; goto st42;}}
 	goto st10;
 tr25:
-#line 29 "markup.rl"
+#line 36 "markup.rl"
 	{
 		characters.end = p;
 		
@@ -1187,7 +1147,7 @@ st10:
 	if ( ++p == pe )
 		goto _test_eof10;
 case 10:
-#line 1191 "markup.c"
+#line 1151 "markup.c"
 	switch( (*p) ) {
 		case 34: goto tr30;
 		case 38: goto tr31;
@@ -1221,25 +1181,13 @@ tr35:
 	{
 		pcdata = Qnil;
 	}
-#line 25 "markup.rl"
-	{
-		characters.begin = p;
-	}
-	goto st13;
-tr37:
-#line 29 "markup.rl"
-	{
-		characters.end = p;
-		
-		Trenni_append_token(&pcdata, encoding, characters);
-	}
-#line 25 "markup.rl"
+#line 32 "markup.rl"
 	{
 		characters.begin = p;
 	}
 	goto st13;
 tr39:
-#line 25 "markup.rl"
+#line 32 "markup.rl"
 	{
 		characters.begin = p;
 	}
@@ -1248,13 +1196,13 @@ st13:
 	if ( ++p == pe )
 		goto _test_eof13;
 case 13:
-#line 1252 "markup.c"
+#line 1200 "markup.c"
 	switch( (*p) ) {
 		case 38: goto tr38;
 		case 39: goto tr24;
 		case 60: goto tr1;
 	}
-	goto tr37;
+	goto st13;
 tr36:
 #line 18 "markup.rl"
 	{
@@ -1264,7 +1212,7 @@ tr36:
 	{{stack[top++] = 14; goto st42;}}
 	goto st14;
 tr38:
-#line 29 "markup.rl"
+#line 36 "markup.rl"
 	{
 		characters.end = p;
 		
@@ -1281,7 +1229,7 @@ st14:
 	if ( ++p == pe )
 		goto _test_eof14;
 case 14:
-#line 1285 "markup.c"
+#line 1233 "markup.c"
 	switch( (*p) ) {
 		case 38: goto tr40;
 		case 39: goto tr30;
@@ -1499,7 +1447,7 @@ st37:
 	if ( ++p == pe )
 		goto _test_eof37;
 case 37:
-#line 1503 "markup.c"
+#line 1451 "markup.c"
 	switch( (*p) ) {
 		case 47: goto tr1;
 		case 62: goto tr67;
@@ -1527,7 +1475,7 @@ st55:
 	if ( ++p == pe )
 		goto _test_eof55;
 case 55:
-#line 1531 "markup.c"
+#line 1479 "markup.c"
 	switch( (*p) ) {
 		case 38: goto tr109;
 		case 60: goto tr110;
@@ -1563,7 +1511,7 @@ st39:
 	if ( ++p == pe )
 		goto _test_eof39;
 case 39:
-#line 1567 "markup.c"
+#line 1515 "markup.c"
 	switch( (*p) ) {
 		case 32: goto tr71;
 		case 47: goto tr69;
@@ -1592,7 +1540,7 @@ tr71:
 	{
 		identifier.end = p;
 	}
-#line 99 "markup.rl"
+#line 106 "markup.rl"
 	{
 	}
 	goto st40;
@@ -1600,12 +1548,12 @@ st40:
 	if ( ++p == pe )
 		goto _test_eof40;
 case 40:
-#line 1604 "markup.c"
+#line 1552 "markup.c"
 	if ( (*p) == 63 )
 		goto tr73;
 	goto st40;
 tr73:
-#line 102 "markup.rl"
+#line 109 "markup.rl"
 	{
 	}
 	goto st41;
@@ -1613,7 +1561,7 @@ st41:
 	if ( ++p == pe )
 		goto _test_eof41;
 case 41:
-#line 1617 "markup.c"
+#line 1565 "markup.c"
 	switch( (*p) ) {
 		case 62: goto st56;
 		case 63: goto tr73;
@@ -1653,7 +1601,7 @@ case 43:
 		goto tr78;
 	goto tr75;
 tr78:
-#line 39 "markup.rl"
+#line 46 "markup.rl"
 	{
 		entity.begin = p;
 	}
@@ -1662,14 +1610,14 @@ st44:
 	if ( ++p == pe )
 		goto _test_eof44;
 case 44:
-#line 1666 "markup.c"
+#line 1614 "markup.c"
 	if ( (*p) == 59 )
 		goto tr81;
 	if ( 48 <= (*p) && (*p) <= 57 )
 		goto st44;
 	goto tr75;
 tr81:
-#line 59 "markup.rl"
+#line 66 "markup.rl"
 	{
 		entity.end = p;
 		
@@ -1681,7 +1629,7 @@ tr81:
 	{{cs = stack[--top];goto _again;}}
 	goto st57;
 tr84:
-#line 51 "markup.rl"
+#line 58 "markup.rl"
 	{
 		entity.end = p;
 		
@@ -1693,12 +1641,12 @@ tr84:
 	{{cs = stack[--top];goto _again;}}
 	goto st57;
 tr86:
-#line 43 "markup.rl"
+#line 50 "markup.rl"
 	{
 		entity.end = p;
 		
 		Trenni_append_string(&pcdata, encoding, 
-			rb_funcall(entities, rb_intern("[]"), 1, Trenni_token(entity, encoding))
+			rb_funcall(entities, id_key_get, 1, Trenni_token(entity, encoding))
 		);
 	}
 #line 8 "/Users/samuel/Documents/Programming/ioquatix/trenni/parsers/trenni/entities.rl"
@@ -1708,7 +1656,7 @@ st57:
 	if ( ++p == pe )
 		goto _test_eof57;
 case 57:
-#line 1712 "markup.c"
+#line 1660 "markup.c"
 	goto st0;
 st45:
 	if ( ++p == pe )
@@ -1724,7 +1672,7 @@ case 45:
 		goto tr82;
 	goto tr75;
 tr82:
-#line 39 "markup.rl"
+#line 46 "markup.rl"
 	{
 		entity.begin = p;
 	}
@@ -1733,7 +1681,7 @@ st46:
 	if ( ++p == pe )
 		goto _test_eof46;
 case 46:
-#line 1737 "markup.c"
+#line 1685 "markup.c"
 	if ( (*p) == 59 )
 		goto tr84;
 	if ( (*p) < 65 ) {
@@ -1746,7 +1694,7 @@ case 46:
 		goto st46;
 	goto tr75;
 tr77:
-#line 39 "markup.rl"
+#line 46 "markup.rl"
 	{
 		entity.begin = p;
 	}
@@ -1755,7 +1703,7 @@ st47:
 	if ( ++p == pe )
 		goto _test_eof47;
 case 47:
-#line 1759 "markup.c"
+#line 1707 "markup.c"
 	if ( (*p) == 59 )
 		goto tr86;
 	if ( (*p) < 65 ) {
@@ -1836,13 +1784,13 @@ case 47:
 	case 45: 
 	case 46: 
 	case 47: 
-#line 35 "markup.rl"
+#line 42 "markup.rl"
 	{
 		Trenni_raise_error("could not parse entity", buffer, p-s);
 	}
 	break;
 	case 53: 
-#line 71 "markup.rl"
+#line 78 "markup.rl"
 	{
 		doctype.end = p;
 		
@@ -1850,13 +1798,13 @@ case 47:
 	}
 	break;
 	case 26: 
-#line 77 "markup.rl"
+#line 84 "markup.rl"
 	{
 		Trenni_raise_error("could not parse doctype", buffer, p-s);
 	}
 	break;
 	case 52: 
-#line 85 "markup.rl"
+#line 92 "markup.rl"
 	{
 		comment.end = p;
 		
@@ -1866,13 +1814,13 @@ case 47:
 	case 17: 
 	case 18: 
 	case 19: 
-#line 91 "markup.rl"
+#line 98 "markup.rl"
 	{
 		Trenni_raise_error("could not parse comment", buffer, p-s);
 	}
 	break;
 	case 56: 
-#line 105 "markup.rl"
+#line 112 "markup.rl"
 	{
 		instruction.end = p;
 		
@@ -1883,19 +1831,19 @@ case 47:
 	case 39: 
 	case 40: 
 	case 41: 
-#line 111 "markup.rl"
+#line 118 "markup.rl"
 	{
 		Trenni_raise_error("could not parse instruction", buffer, p-s);
 	}
 	break;
 	case 51: 
-#line 151 "markup.rl"
+#line 158 "markup.rl"
 	{
 		rb_funcall(delegate, id_open_tag_end, 1, self_closing == 1 ? Qtrue : Qfalse);
 	}
 	break;
 	case 55: 
-#line 158 "markup.rl"
+#line 165 "markup.rl"
 	{
 		rb_funcall(delegate, id_close_tag, 2, Trenni_token(identifier, encoding), ULONG2NUM(identifier.begin-s));
 	}
@@ -1916,13 +1864,13 @@ case 47:
 	case 14: 
 	case 36: 
 	case 37: 
-#line 162 "markup.rl"
+#line 169 "markup.rl"
 	{
 		Trenni_raise_error("could not parse tag", buffer, p-s);
 	}
 	break;
 	case 54: 
-#line 170 "markup.rl"
+#line 177 "markup.rl"
 	{
 		cdata.end = p;
 		
@@ -1932,7 +1880,7 @@ case 47:
 	case 33: 
 	case 34: 
 	case 35: 
-#line 176 "markup.rl"
+#line 183 "markup.rl"
 	{
 		Trenni_raise_error("could not parse cdata", buffer, p-s);
 	}
@@ -1941,14 +1889,13 @@ case 47:
 #line 22 "markup.rl"
 	{
 	}
-#line 184 "markup.rl"
+#line 28 "markup.rl"
 	{
-		// Entities are handled separately:
 		rb_funcall(delegate, id_text, 1, pcdata);
 	}
 	break;
 	case 49: 
-#line 29 "markup.rl"
+#line 36 "markup.rl"
 	{
 		characters.end = p;
 		
@@ -1957,20 +1904,19 @@ case 47:
 #line 22 "markup.rl"
 	{
 	}
-#line 184 "markup.rl"
+#line 28 "markup.rl"
 	{
-		// Entities are handled separately:
 		rb_funcall(delegate, id_text, 1, pcdata);
 	}
 	break;
-#line 1967 "markup.c"
+#line 1913 "markup.c"
 	}
 	}
 
 	_out: {}
 	}
 
-#line 216 "markup.rl"
+#line 214 "markup.rl"
 
 	
 	if (p != eof) {