X-Git-Url: https://jasonwoof.com/gitweb/?a=blobdiff_plain;f=parse-html.coffee;h=85a91af5f83259a3507a98cb586c696ae0439f23;hb=6c4bd699f811de0097387b167097665b57c4b011;hp=d35dd885b413fedaac30c0c850d1371d3805a003;hpb=1dd403e5fec58bd28837e8d12c9c1455f1258062;p=peach-html5-editor.git

diff --git a/parse-html.coffee b/parse-html.coffee
index d35dd88..85a91af 100644
--- a/parse-html.coffee
+++ b/parse-html.coffee
@@ -47,6 +47,12 @@
 #   0: a "end of the list", "current node", "bottommost", "last"
 
 
+# browser
+# note: to get this to run outside a browser, you'll have to write a native
+# implementation of decode_named_char_ref()
+unless module?.exports?
+	window.wheic = {}
+	module = exports: window.wheic
 
 # Each node is an obect of the Node class. Here are the Node types:
 TYPE_TAG = 0 # name, {attributes}, [children]
@@ -85,6 +91,7 @@ class Node
 		@children = args.children ? []
 		@namespace = args.namespace ? NS_HTML
 		@parent = args.parent ? null
+		@token = args.token ? null
 		if args.id?
 			@id = "#{args.id}+"
 		else
@@ -93,9 +100,12 @@ class Node
 		# WARNING this doesn't work right on open tags that are still being parsed
 		attrs = {}
 		attrs[k] = v for k, v of @attrs
-		return new Node @type, name: @name, text: @text, attrs: attrs, namespace: @namespace, id: @id
+		return new Node @type, name: @name, text: @text, attrs: attrs, namespace: @namespace, id: @id, token: @token
 	acknowledge_self_closing: ->
-		@flag 'did_self_close', true
+		if @token?
+			@token.flag 'did_self_close'
+		else
+			@flag 'did_self_close', true
 	flag: ->
 		# fixfull
 	serialize: (shallow = false, show_ids = false) -> # for unit tests
@@ -154,8 +164,10 @@ new_element = (name) ->
 new_text_node = (txt) ->
 	return new Node TYPE_TEXT, text: txt
 new_character_token = new_text_node
-new_comment_node = (txt) ->
+new_comment_token = (txt) ->
 	return new Node TYPE_COMMENT, text: txt
+new_doctype_token = (name) ->
+	return new Node TYPE_DOCTYPE, name: name
 new_eof_token = ->
 	return new Node TYPE_EOF
 new_afe_marker = ->
@@ -362,6 +374,11 @@ parse_html = (txt, parse_error_cb = null) ->
 	temporary_buffer = null
 	pending_table_character_tokens = null
 	head_element_pointer = null
+	flag_fragment_parsing = null
+	context_element = null
+
+	stop_parsing = ->
+		flag_parsing = false
 
 	parse_error = ->
 		if parse_error_cb?
@@ -609,6 +626,14 @@ parse_html = (txt, parse_error_cb = null) ->
 			node = open_els[node_i]
 			# 19. Return to the step labeled loop.
 
+	# 8.2.3.2
+
+	# http://www.w3.org/TR/html5/syntax.html#adjusted-current-node
+	adjusted_current_node = ->
+		if open_els.length is 1 and flag_fragment_parsing
+			return context_element
+		return open_els[0]
+
 	# http://www.w3.org/TR/html5/syntax.html#reconstruct-the-active-formatting-elements
 	# this implementation is structured (mostly) as described at the link above.
 	# capitalized comments are the "labels" described at the link above.
@@ -1051,7 +1076,7 @@ parse_html = (txt, parse_error_cb = null) ->
 		while t.attrs_a.length
 			a = t.attrs_a.pop()
 			attrs[a[0]] = a[1] # TODO check what to do with dupilcate attrs
-		el = new Node TYPE_TAG, name: t.name, namespace: namespace, attrs: attrs
+		el = new Node TYPE_TAG, name: t.name, namespace: namespace, attrs: attrs, token: t
 
 		# TODO 2. If the newly created element has an xmlns attribute in the
 		# XMLNS namespace whose value is not exactly the same as the element's
@@ -1136,8 +1161,8 @@ parse_html = (txt, parse_error_cb = null) ->
 			doc.children.push t
 			return
 		if t.type is TYPE_DOCTYPE
+			# FIXME check identifiers, set quirks, etc
 			# fixfull
-			t.name = 'html'
 			doc.children.push t
 			insertion_mode = ins_mode_before_html
 			return
@@ -1230,16 +1255,16 @@ parse_html = (txt, parse_error_cb = null) ->
 		if t.type is TYPE_START_TAG and (t.name is 'base' or t.name is 'basefont' or t.name is 'bgsound' or t.name is 'link')
 			el = insert_html_element t
 			open_els.shift()
-			el.acknowledge_self_closing()
+			t.acknowledge_self_closing()
 			return
 		if t.type is TYPE_START_TAG and t.name is 'meta'
 			el = insert_html_element t
 			open_els.shift()
-			el.acknowledge_self_closing()
+			t.acknowledge_self_closing()
 			# fixfull encoding stuff
 			return
 		if t.type is TYPE_START_TAG and t.name is 'title'
-			parse_generic_rcdata_element t
+			parse_generic_rcdata_text t
 			return
 		if t.type is TYPE_START_TAG and ((t.name is 'noscript' and flag_scripting) or (t.name is 'noframes' or t.name is 'style'))
 			parse_generic_raw_text t
@@ -1251,7 +1276,7 @@ parse_html = (txt, parse_error_cb = null) ->
 		if t.type is TYPE_START_TAG and t.name is 'script'
 			ail = adjusted_insertion_location()
 			el = token_to_element t, NS_HTML, ail
-			el.flag_parser_inserted true # FIXME implement
+			el.flag 'parser-inserted', true # FIXME implement
 			# fixfull frament case
 			ail[0].children.splice ail[1], 0, el
 			open_els.unshift el
@@ -1292,12 +1317,12 @@ parse_html = (txt, parse_error_cb = null) ->
 			parse_error()
 			return
 		ins_mode_in_head_else t
-	
+
 	# 8.2.5.4.5 http://www.w3.org/TR/html5/syntax.html#parsing-main-inheadnoscript
 	ins_mode_in_head_noscript = (t) ->
 		# FIXME ?fixfull
 		console.log "ins_mode_in_head_noscript unimplemented"
-	
+
 	# 8.2.5.4.6 http://www.w3.org/TR/html5/syntax.html#the-after-head-insertion-mode
 	ins_mode_after_head_else = (t) ->
 		body_tok = new_open_tag 'body'
@@ -1458,7 +1483,7 @@ parse_html = (txt, parse_error_cb = null) ->
 						parse_error()
 						break
 				# TODO stack of template insertion modes thing
-				flag_parsing = false # stop parsing
+				stop_parsing()
 			when TYPE_END_TAG
 				switch t.name
 					when 'body'
@@ -1594,7 +1619,7 @@ parse_html = (txt, parse_error_cb = null) ->
 							parse_error()
 							el = insert_html_element t
 							open_els.shift()
-							el.acknowledge_self_closing()
+							t.acknowledge_self_closing()
 					when 'form'
 						parse_error()
 						if form_element_pointer?
@@ -1705,7 +1730,7 @@ parse_html = (txt, parse_error_cb = null) ->
 		if t.type is TYPE_START_TAG and t.name is 'col'
 			el = insert_html_element t
 			open_els.shift()
-			el.acknowledge_self_closing()
+			t.acknowledge_self_closing()
 			return
 		if t.type is TYPE_END_TAG and t.name is 'colgroup'
 			if open_els[0].name is 'colgroup'
@@ -1982,14 +2007,191 @@ parse_html = (txt, parse_error_cb = null) ->
 		ins_mode_in_select t
 		return
 
-	# CONTINUE more insertion modes!
-
-
+	# 8.2.5.4.18 http://www.w3.org/TR/html5/syntax.html#parsing-main-intemplate
+	ins_mode_in_template = (t) ->
+		if t.type is TYPE_TEXT or t.type is TYPE_COMMENT or t.type is TYPE_DOCTYPE
+			ins_mode_in_body t
+			return
+		if (t.type is TYPE_START_TAG and (t.name is 'base' or t.name is 'basefont' or t.name is 'bgsound' or t.name is 'link' or t.name is 'meta' or t.name is 'noframes' or t.name is 'script' or t.name is 'style' or t.name is 'template' or t.name is 'title')) or (t.type is TYPE_END_TAG and t.name is 'template')
+			ins_mode_in_head t
+			return
+		if t.type is TYPE_START_TAG and (t.name is 'caption' or t.name is 'colgroup' or t.name is 'tbody' or t.name is 'tfoot' or t.name is 'thead')
+			template_insertion_modes.shift()
+			template_insertion_modes.unshift ins_mode_in_table
+			insertion_mode = ins_mode_in_table
+			insertion_mode t
+			return
+		if t.type is TYPE_START_TAG and t.name is 'col'
+			template_insertion_modes.shift()
+			template_insertion_modes.unshift ins_mode_in_column_group
+			insertion_mode = ins_mode_in_column_group
+			insertion_mode t
+			return
+		if t.type is TYPE_START_TAG and t.name is 'tr'
+			template_insertion_modes.shift()
+			template_insertion_modes.unshift ins_mode_in_table_body
+			insertion_mode = ins_mode_in_table_body
+			insertion_mode t
+			return
+		if t.type is TYPE_START_TAG and (t.name is 'td' or t.name is 'th')
+			template_insertion_modes.shift()
+			template_insertion_modes.unshift ins_mode_in_row
+			insertion_mode = ins_mode_in_row
+			insertion_mode t
+			return
+		if t.type is TYPE_START_TAG
+			template_insertion_modes.shift()
+			template_insertion_modes.unshift ins_mode_in_body
+			insertion_mode = ins_mode_in_body
+			insertion_mode t
+			return
+		if t.type is TYPE_END_TAG
+			parse_error()
+			return
+		if t.type is EOF
+			unless template_tag_is_open()
+				stop_parsing()
+				return
+			parse_error()
+			loop
+				el = open_els.shift()
+				if el.name is 'template' # fixfull check namespace
+					break
+			clear_afe_to_marker()
+			template_insertion_modes.shift()
+			reset_insertion_mode()
+			insertion_mode t
 
+	# 8.2.5.4.19 http://www.w3.org/TR/html5/syntax.html#parsing-main-afterbody
+	ins_mode_after_body = (t) ->
+		if is_space_tok t
+			ins_mode_in_body t
+			return
+		if t.type is TYPE_COMMENT
+			insert_comment t, [open_els[0], open_els[0].children.length]
+			return
+		if t.type is TYPE_DOCTYPE
+			parse_error()
+			return
+		if t.type is TYPE_START_TAG and t.name is 'html'
+			ins_mode_in_body t
+			return
+		if t.type is TYPE_END_TAG and t.name is 'html'
+			# fixfull fragment case
+			insertion_mode = ins_mode_after_after_body
+			return
+		if t.type is TYPE_EOF
+			stop_parsing()
+			return
+		# Anything ELse
+		parse_error()
+		insertion_mode = ins_mode_in_body
+		insertion_mode t
 
+	# 8.2.5.4.20 http://www.w3.org/TR/html5/syntax.html#parsing-main-inframeset
+	ins_mode_in_frameset = (t) ->
+		if is_space_tok t
+			insert_character t
+			return
+		if t.type is TYPE_COMMENT
+			insert_comment t
+			return
+		if t.type is TYPE_DOCTYPE
+			parse_error()
+			return
+		if t.type is TYPE_START_TAG and t.name is 'html'
+			ins_mode_in_body t
+			return
+		if t.type is TYPE_START_TAG and t.name is 'frameset'
+			insert_html_element t
+			return
+		if t.type is TYPE_END_TAG and t.name is 'frameset'
+			# TODO ?correct for: "if the current node is the root html element"
+			if open_els.length is 1
+				parse_error()
+				return # fragment case
+			open_els.shift()
+			if flag_fragment_parsing is false and open_els[0].name isnt 'frameset'
+				insertion_mode = ins_mode_after_frameset
+			return
+		if t.type is TYPE_START_TAG and t.name is 'frame'
+			insert_html_element t
+			open_els.shift()
+			t.acknowledge_self_closing()
+			return
+		if t.type is TYPE_START TAG and t.name is 'noframes'
+			ins_mode_in_head t
+			return
+		if t.type is TYPE_EOF
+			# TODO ?correct for: "if the current node is not the root html element"
+			if open_els.length isnt 1
+				parse_error()
+			stop_parsing()
+			return
+		# Anything else
+		parse_error()
+		return
 
+	# 8.2.5.4.21 http://www.w3.org/TR/html5/syntax.html#parsing-main-afterframeset
+	ins_mode_after_frameset = (t) ->
+		if is_space_tok t
+			insert_character t
+			return
+		if t.type is TYPE_COMMENT
+			insert_comment t
+			return
+		if t.type is TYPE_DOCTYPE
+			parse_error()
+			return
+		if t.type is TYPE_START_TAG and t.name is 'html'
+			ins_mode_in_body t
+			return
+		if t.type is TYPE_END_TAG and t.name is 'html'
+			insert_mode = ins_mode_after_after_frameset
+			return
+		if t.type is TYPE_START_TAG and t.name is 'noframes'
+			ins_mode_in_head t
+			return
+		if t.type is TYPE_EOF
+			stop_parsing()
+			return
+		# Anything else
+		parse_error()
+		return
 
+	# 8.2.5.4.22 http://www.w3.org/TR/html5/syntax.html#the-after-after-body-insertion-mode
+	ins_mode_after_after_body = (t) ->
+		if t.type is TYPE_COMMENT
+			insert_comment t, [doc, doc.children.length]
+			return
+		if t.type is TYPE_DOCTYPE or is_space_tok(t) or (t.type is TYPE_START_TAG and t.name is 'html')
+			ins_mode_in_body t
+			return
+		if t.type is TYPE_EOF
+			stop_parsing()
+			return
+		# Anything else
+		parse_error()
+		insertion_mode = ins_mode_in_body
+		return
 
+	# 8.2.5.4.23 http://www.w3.org/TR/html5/syntax.html#the-after-after-frameset-insertion-mode
+	ins_mode_after_after_frameset = (t) ->
+		if t.type is TYPE_COMMENT
+			insert_comment t, [doc, doc.children.length]
+			return
+		if t.type is TYPE_DOCTYPE or is_space_tok(t) or (t.type is TYPE_START_TAG and t.name is 'html')
+			ins_mode_in_body t
+			return
+		if t.type is TYPE_EOF
+			stop_parsing()
+			return
+		if t.type is TYPE_START_TAG and t.name is 'noframes'
+			ins_mode_in_head t
+			return
+		# Anything else
+		parse_error()
+		return
 
 
 
@@ -2085,6 +2287,7 @@ parse_html = (txt, parse_error_cb = null) ->
 				tok_state = tok_state_end_tag_open
 			when '?'
 				parse_error()
+				tok_cur_tag = new_comment_token '?'
 				tok_state = tok_state_bogus_comment
 			else
 				if lc_alpha.indexOf(c) > -1
@@ -2119,6 +2322,7 @@ parse_html = (txt, parse_error_cb = null) ->
 					tok_state = tok_state_tag_name
 				else
 					parse_error()
+					tok_cur_tag = new_comment_token '/'
 					tok_state = tok_state_bogus_comment
 		return null
 
@@ -2486,6 +2690,601 @@ parse_html = (txt, parse_error_cb = null) ->
 				cur -= 1 # we didn't handle that char
 		return null
 
+	# 8.2.4.44 http://www.w3.org/TR/html5/syntax.html#bogus-comment-state
+	# WARNING: put a comment token in tok_cur_tag before setting this state
+	tok_state_bogus_comment = ->
+		next_gt = txt.indexOf '>', cur
+		if next_gt is -1
+			val = txt.substr cur
+			cur = txt.length
+		else
+			val = txt.substr cur, (next_gt - cur)
+			cur = next_gt + 1
+		val = val.replace "\u0000", "\ufffd"
+		tok_cur_tag.text += val
+		tok_state = tok_state_data
+		return tok_cur_tag
+
+	# 8.2.4.45 http://www.w3.org/TR/html5/syntax.html#markup-declaration-open-state
+	tok_state_markup_declaration_open = ->
+		if txt.substr(cur, 2) is '--'
+			cur += 2
+			tok_cur_tag = new_comment_token ''
+			tok_state = tok_state_comment_start
+			return
+		if txt.substr(cur, 7).toLowerCase() is 'doctype'
+			cur += 7
+			tok_state = tok_state_doctype
+			return
+		acn = adjusted_current_node()
+		if acn and acn.namespace isnt NS_HTML and text.substr(cur, 7) is '[CDATA['
+			cur += 7
+			tok_state = tok_state_cdata_section
+			return
+		# Otherwise
+		parse_error()
+		tok_cur_tag = new_comment_token '!' # TODO test ("!" right?)
+		tok_state = tok_state_bogus_comment
+		return
+
+	# 8.2.4.46 http://www.w3.org/TR/html5/syntax.html#comment-start-state
+	tok_state_comment_start = ->
+		switch c = txt.charAt(cur++)
+			when '-'
+				tok_state = tok_state_comment_start_dash
+			when "\u0000"
+				parse_error()
+				return new_character_token "\ufffd"
+			when '>'
+				parse_error()
+				tok_state = tok_state_data
+				return tok_cur_tag
+			when '' # EOF
+				parse_error()
+				tok_state = tok_state_data
+				cur -= 1 # Reconsume
+				return tok_cur_tag
+			else
+				tok_cur_tag.text += c
+		return null
+
+	# 8.2.4.47 http://www.w3.org/TR/html5/syntax.html#comment-start-dash-state
+	tok_state_comment_start_dash = ->
+		switch c = txt.charAt(cur++)
+			when '-'
+				tok_state = tok_state_comment_end
+			when "\u0000"
+				parse_error()
+				tok_cur_tag.text += "-\ufffd"
+				tok_state = tok_state_comment
+			when '>'
+				parse_error()
+				tok_state = tok_state_data
+				return tok_cur_tag
+			when '' # EOF
+				parse_error()
+				tok_state = tok_state_data
+				cur -= 1 # Reconsume
+				return tok_cur_tag
+			else
+				tok_cur_tag.text += "-#{c}"
+				tok_state = tok_state_comment
+		return null
+
+	# 8.2.4.48 http://www.w3.org/TR/html5/syntax.html#comment-state
+	tok_state_comment = ->
+		switch c = txt.charAt(cur++)
+			when '-'
+				tok_state = tok_state_comment_end_dash
+			when "\u0000"
+				parse_error()
+				tok_cur_tag.text += "\ufffd"
+			when '' # EOF
+				parse_error()
+				tok_state = tok_state_data
+				cur -= 1 # Reconsume
+				return tok_cur_tag
+			else
+				tok_cur_tag.text += c
+		return null
+
+	# 8.2.4.49 http://www.w3.org/TR/html5/syntax.html#comment-end-dash-state
+	tok_state_comment_end_dash = ->
+		switch c = txt.charAt(cur++)
+			when '-'
+				tok_state = tok_state_comment_end
+			when "\u0000"
+				parse_error()
+				tok_cur_tag.text += "-\ufffd"
+				tok_state = tok_state_comment
+			when '' # EOF
+				parse_error()
+				tok_state = tok_state_data
+				cur -= 1 # Reconsume
+				return tok_cur_tag
+			else
+				tok_cur_tag.text += "-#{c}"
+				tok_state = tok_state_comment
+		return null
+
+	# 8.2.4.50 http://www.w3.org/TR/html5/syntax.html#comment-end-state
+	tok_state_comment_end = ->
+		switch c = txt.charAt(cur++)
+			when '>'
+				tok_state = tok_state_data
+				return tok_cur_tag
+			when "\u0000"
+				parse_error()
+				tok_cur_tag.text += "--\ufffd"
+				tok_state = tok_state_comment
+			when '!'
+				parse_error()
+				tok_state = tok_state_comment_end_bang
+			when '-'
+				parse_error()
+				tok_cur_tag.text += '-'
+			when '' # EOF
+				parse_error()
+				tok_state = tok_state_data
+				cur -= 1 # Reconsume
+				return tok_cur_tag
+			else
+				parse_error()
+				tok_cur_tag.text += "--#{c}"
+				tok_state = tok_state_comment
+		return null
+
+	# 8.2.4.51 http://www.w3.org/TR/html5/syntax.html#comment-end-bang-state
+	tok_state_comment_end_bang = ->
+		switch c = txt.charAt(cur++)
+			when '-'
+				tok_cur_tag.text += "--!#{c}"
+				tok_state = tok_state_comment_end_dash
+			when '>'
+				tok_state = tok_state_data
+				return tok_cur_tag
+			when "\u0000"
+				parse_error()
+				tok_cur_tag.text += "--!\ufffd"
+				tok_state = tok_state_comment
+			when '' # EOF
+				parse_error()
+				tok_state = tok_state_data
+				cur -= 1 # Reconsume
+				return tok_cur_tag
+			else
+				tok_cur_tag.text += "--!#{c}"
+				tok_state = tok_state_comment
+		return null
+
+	# 8.2.4.52 http://www.w3.org/TR/html5/syntax.html#doctype-state
+	tok_state_doctype = ->
+		switch c = txt.charAt(cur++)
+			when "\t", "\u000a", "\u000c", ' '
+				tok_state = tok_state_before_doctype_name
+			when '' # EOF
+				parse_error()
+				tok_state = tok_state_data
+				el = new_doctype_token ''
+				el.flag 'force-quirks', true
+				cur -= 1 # Reconsume
+				return el
+			else
+				parse_error()
+				tok_state = tok_state_before_doctype_name
+				cur -= 1 # Reconsume
+		return null
+
+	# 8.2.4.52 http://www.w3.org/TR/html5/syntax.html#doctype-state
+	tok_state_before_doctype_name = ->
+		c = txt.charAt(cur++)
+		if c is "\t" or c is "\u000a" or c is "\u000c" or c is ' '
+			return
+		if uc_alpha.indexOf(c) > -1
+			tok_cur_tag = new_doctype_token c.toLowerCase()
+			tok_state = tok_state_doctype_name
+			return
+		if c is "\u0000"
+			parse_error()
+			tok_cur_tag = new_doctype_token "\ufffd"
+			tok_state = tok_state_doctype_name
+			return
+		if c is '>'
+			parse_error()
+			el = new_doctype_token ''
+			el.flag 'force-quirks', true
+			tok_state = tok_state_data
+			return el
+		if c is '' # EOF
+			parse_error()
+			tok_state = tok_state_data
+			el = new_doctype_token ''
+			el.flag 'force-quirks', true
+			cur -= 1 # Reconsume
+			return el
+		# Anything else
+		tok_cur_tag = new_doctype_token c
+		tok_state = tok_state_doctype_name
+		return null
+
+	# 8.2.4.54 http://www.w3.org/TR/html5/syntax.html#doctype-name-state
+	tok_state_doctype_name = ->
+		c = txt.charAt(cur++)
+		if c is "\t" or c is "\u000a" or c is "\u000c" or c is ' '
+			tok_state = tok_state_after_doctype_name
+			return
+		if c is '>'
+			tok_state = tok_state_data
+			return tok_cur_tag
+		if uc_alpha.indexOf(c) > -1
+			tok_cur_tag.name += c.toLowerCase()
+			return
+		if c is "\u0000"
+			parse_error()
+			tok_cur_tag.name += "\ufffd"
+			return
+		if c is '' # EOF
+			parse_error()
+			tok_state = tok_state_data
+			tok_cur_tag.flag 'force-quirks', true
+			cur -= 1 # Reconsume
+			return tok_cur_tag
+		# Anything else
+		tok_cur_tag.name += c
+		return null
+
+	# 8.2.4.55 http://www.w3.org/TR/html5/syntax.html#after-doctype-name-state
+	tok_state_after_doctype_name = ->
+		c = txt.charAt(cur++)
+		if c is "\t" or c is "\u000a" or c is "\u000c" or c is ' '
+			return
+		if c is '>'
+			tok_state = tok_state_data
+			return tok_cur_tag
+		if c is '' # EOF
+			parse_error()
+			tok_state = tok_state_data
+			tok_cur_tag.flag 'force-quirks', true
+			cur -= 1 # Reconsume
+			return tok_cur_tag
+		# Anything else
+		if txt.substr(cur - 1, 6).toLowerCase() is 'public'
+			cur += 5
+			tok_state = tok_state_after_doctype_public_keyword
+			return
+		if txt.substr(cur - 1, 6).toLowerCase() is 'system'
+			cur += 5
+			tok_state = tok_state_after_doctype_system_keyword
+			return
+		parse_error()
+		tok_cur_tag.flag 'force-quirks', true
+		tok_state = tok_state_bogus_doctype
+		return null
+
+	# 8.2.4.56 http://www.w3.org/TR/html5/syntax.html#after-doctype-public-keyword-state
+	tok_state_after_doctype_public_keyword = ->
+		c = txt.charAt(cur++)
+		if c is "\t" or c is "\u000a" or c is "\u000c" or c is ' '
+			tok_state = tok_state_before_doctype_public_identifier
+			return
+		if c is '"'
+			parse_error()
+			tok_cur_tag.public_identifier = '' # FIXME should this go in @attrs or @text?
+			tok_state = tok_state_doctype_public_identifier_double_quoted
+			return
+		if c is "'"
+			parse_error()
+			tok_cur_tag.public_identifier = '' # FIXME should this go in @attrs or @text?
+			tok_state = tok_state_doctype_public_identifier_single_quoted
+			return
+		if c is '>'
+			parse_error()
+			tok_cur_tag.flag 'force-quirks', true
+			tok_state = tok_state_data
+			return tok_cur_tag
+		if c is '' # EOF
+			parse_error()
+			tok_state = tok_state_data
+			tok_cur_tag.flag 'force-quirks', true
+			cur -= 1 # Reconsume
+			return tok_cur_tag
+		# Anything else
+		parse_error()
+		tok_cur_tag.flag 'force-quirks', true
+		tok_state = tok_state_bogus_doctype
+		return null
+
+	# 8.2.4.57 http://www.w3.org/TR/html5/syntax.html#before-doctype-public-identifier-state
+	tok_state_before_doctype_public_identifier = ->
+		c = txt.charAt(cur++)
+		if c is "\t" or c is "\u000a" or c is "\u000c" or c is ' '
+			return
+		if c is '"'
+			parse_error()
+			tok_cur_tag.public_identifier = '' # FIXME should this go in @attrs or @text?
+			tok_state = tok_state_doctype_public_identifier_double_quoted
+			return
+		if c is "'"
+			parse_error()
+			tok_cur_tag.public_identifier = '' # FIXME should this go in @attrs or @text?
+			tok_state = tok_state_doctype_public_identifier_single_quoted
+			return
+		if c is '>'
+			parse_error()
+			tok_cur_tag.flag 'force-quirks', true
+			tok_state = tok_state_data
+			return tok_cur_tag
+		if c is '' # EOF
+			parse_error()
+			tok_state = tok_state_data
+			tok_cur_tag.flag 'force-quirks', true
+			cur -= 1 # Reconsume
+			return tok_cur_tag
+		# Anything else
+		parse_error()
+		tok_cur_tag.flag 'force-quirks', true
+		tok_state = tok_state_bogus_doctype
+		return null
+
+
+	# 8.2.4.58 http://www.w3.org/TR/html5/syntax.html#doctype-public-identifier-(double-quoted)-state
+	tok_state_doctype_public_identifier_double_quoted = ->
+		c = txt.charAt(cur++)
+		if c is '"'
+			tok_state = tok_state_after_doctype_public_identifier
+			return
+		if c is "\u0000"
+			parse_error()
+			tok_cur_tag.public_identifier += "\ufffd"
+			return
+		if c is '>'
+			parse_error()
+			tok_cur_tag.flag 'force-quirks', true
+			tok_state = tok_state_data
+			return tok_cur_tag
+		if c is '' # EOF
+			parse_error()
+			tok_state = tok_state_data
+			tok_cur_tag.flag 'force-quirks', true
+			cur -= 1 # Reconsume
+			return tok_cur_tag
+		# Anything else
+		tok_cur_tag.public_identifier += c
+		return null
+
+	# 8.2.4.59 http://www.w3.org/TR/html5/syntax.html#doctype-public-identifier-(single-quoted)-state
+	tok_state_doctype_public_identifier_single_quoted = ->
+		c = txt.charAt(cur++)
+		if c is "'"
+			tok_state = tok_state_after_doctype_public_identifier
+			return
+		if c is "\u0000"
+			parse_error()
+			tok_cur_tag.public_identifier += "\ufffd"
+			return
+		if c is '>'
+			parse_error()
+			tok_cur_tag.flag 'force-quirks', true
+			tok_state = tok_state_data
+			return tok_cur_tag
+		if c is '' # EOF
+			parse_error()
+			tok_state = tok_state_data
+			tok_cur_tag.flag 'force-quirks', true
+			cur -= 1 # Reconsume
+			return tok_cur_tag
+		# Anything else
+		tok_cur_tag.public_identifier += c
+		return null
+
+	# 8.2.4.60 http://www.w3.org/TR/html5/syntax.html#after-doctype-public-identifier-state
+	tok_state_after_doctype_public_identifier = ->
+		c = txt.charAt(cur++)
+		if c is "\t" or c is "\u000a" or c is "\u000c" or c is ' '
+			tok_state = tok_state_between_doctype_public_and_system_identifiers
+			return
+		if c is '>'
+			tok_state = tok_state_data
+			return tok_cur_tag
+		if c is '"'
+			parse_error()
+			tok_cur_tag.system_identifier = ''
+			tok_state = tok_state_doctype_system_identifier_double_quoted
+			return
+		if c is "'"
+			parse_error()
+			tok_cur_tag.system_identifier = ''
+			tok_state = tok_state_doctype_system_identifier_single_quoted
+			return
+		if c is '' # EOF
+			parse_error()
+			tok_state = tok_state_data
+			tok_cur_tag.flag 'force-quirks', true
+			cur -= 1 # Reconsume
+			return tok_cur_tag
+		# Anything else
+		parse_error()
+		tok_cur_tag.flag 'force-quirks', true
+		tok_state = tok_state_bogus_doctype
+		return null
+
+	# 8.2.4.61 http://www.w3.org/TR/html5/syntax.html#between-doctype-public-and-system-identifiers-state
+	tok_state_between_doctype_public_and_system_identifiers = ->
+		c = txt.charAt(cur++)
+		if c is "\t" or c is "\u000a" or c is "\u000c" or c is ' '
+			return
+		if c is '>'
+			tok_state = tok_state_data
+			return tok_cur_tag
+		if c is '"'
+			parse_error()
+			tok_cur_tag.system_identifier = ''
+			tok_state = tok_state_doctype_system_identifier_double_quoted
+			return
+		if c is "'"
+			parse_error()
+			tok_cur_tag.system_identifier = ''
+			tok_state = tok_state_doctype_system_identifier_single_quoted
+			return
+		if c is '' # EOF
+			parse_error()
+			tok_state = tok_state_data
+			tok_cur_tag.flag 'force-quirks', true
+			cur -= 1 # Reconsume
+			return tok_cur_tag
+		# Anything else
+		parse_error()
+		tok_cur_tag.flag 'force-quirks', true
+		tok_state = tok_state_bogus_doctype
+		return null
+
+	# 8.2.4.62 http://www.w3.org/TR/html5/syntax.html#after-doctype-system-keyword-state
+	tok_state_after_doctype_system_keyword = ->
+		c = txt.charAt(cur++)
+		if c is "\t" or c is "\u000a" or c is "\u000c" or c is ' '
+			tok_state = tok_state_before_doctype_system_identifier
+			return
+		if c is '"'
+			parse_error()
+			tok_cur_tag.system_identifier = ''
+			tok_state = tok_state_doctype_system_identifier_double_quoted
+			return
+		if c is "'"
+			parse_error()
+			tok_cur_tag.system_identifier = ''
+			tok_state = tok_state_doctype_system_identifier_single_quoted
+			return
+		if c is '>'
+			parse_error()
+			tok_cur_tag.flag 'force-quirks', true
+			tok_state = tok_state_data
+			return tok_cur_tag
+		if c is '' # EOF
+			parse_error()
+			tok_state = tok_state_data
+			tok_cur_tag.flag 'force-quirks', true
+			cur -= 1 # Reconsume
+			return tok_cur_tag
+		# Anything else
+		parse_error()
+		tok_cur_tag.flag 'force-quirks', true
+		tok_state = tok_state_bogus_doctype
+		return null
+
+	# 8.2.4.63 http://www.w3.org/TR/html5/syntax.html#before-doctype-system-identifier-state
+	tok_state_before_doctype_system_identifier = ->
+		c = txt.charAt(cur++)
+		if c is "\t" or c is "\u000a" or c is "\u000c" or c is ' '
+			return
+		if c is '"'
+			tok_cur_tag.system_identifier = ''
+			tok_state = tok_state_doctype_system_identifier_double_quoted
+			return
+		if c is "'"
+			tok_cur_tag.system_identifier = ''
+			tok_state = tok_state_doctype_system_identifier_single_quoted
+			return
+		if c is '>'
+			parse_error()
+			tok_cur_tag.flag 'force-quirks', true
+			tok_state = tok_state_data
+			return tok_cur_tag
+		if c is '' # EOF
+			parse_error()
+			tok_state = tok_state_data
+			tok_cur_tag.flag 'force-quirks', true
+			cur -= 1 # Reconsume
+			return tok_cur_tag
+		# Anything else
+		parse_error()
+		tok_cur_tag.flag 'force-quirks', true
+		tok_state = tok_state_bogus_doctype
+		return null
+
+	# 8.2.4.64 http://www.w3.org/TR/html5/syntax.html#doctype-system-identifier-(double-quoted)-state
+	tok_state_doctype_system_identifier_double_quoted = ->
+		c = txt.charAt(cur++)
+		if c is '"'
+			tok_state = tok_state_after_doctype_system_identifier
+			return
+		if c is "\u0000"
+			parse_error()
+			tok_cur_tag.system_identifier += "\ufffd"
+			return
+		if c is '>'
+			parse_error()
+			tok_cur_tag.flag 'force-quirks', true
+			tok_state = tok_state_data
+			return tok_cur_tag
+		if c is '' # EOF
+			parse_error()
+			tok_state = tok_state_data
+			tok_cur_tag.flag 'force-quirks', true
+			cur -= 1 # Reconsume
+			return tok_cur_tag
+		# Anything else
+		tok_cur_tag.system_identifier += c
+		return null
+
+	# 8.2.4.65 http://www.w3.org/TR/html5/syntax.html#doctype-system-identifier-(single-quoted)-state
+	tok_state_doctype_system_identifier_single_quoted = ->
+		c = txt.charAt(cur++)
+		if c is "'"
+			tok_state = tok_state_after_doctype_system_identifier
+			return
+		if c is "\u0000"
+			parse_error()
+			tok_cur_tag.system_identifier += "\ufffd"
+			return
+		if c is '>'
+			parse_error()
+			tok_cur_tag.flag 'force-quirks', true
+			tok_state = tok_state_data
+			return tok_cur_tag
+		if c is '' # EOF
+			parse_error()
+			tok_state = tok_state_data
+			tok_cur_tag.flag 'force-quirks', true
+			cur -= 1 # Reconsume
+			return tok_cur_tag
+		# Anything else
+		tok_cur_tag.system_identifier += c
+		return null
+
+	# 8.2.4.66 http://www.w3.org/TR/html5/syntax.html#after-doctype-system-identifier-state
+	tok_state_after_doctype_system_identifier = ->
+		c = txt.charAt(cur++)
+		if c is "\t" or c is "\u000a" or c is "\u000c" or c is ' '
+			return
+		if c is '>'
+			tok_state = tok_state_data
+			return tok_cur_tag
+		if c is '' # EOF
+			parse_error()
+			tok_state = tok_state_data
+			tok_cur_tag.flag 'force-quirks', true
+			cur -= 1 # Reconsume
+			return tok_cur_tag
+		# Anything else
+		parse_error()
+		# do _not_ tok_cur_tag.flag 'force-quirks', true
+		tok_state = tok_state_bogus_doctype
+		return null
+
+	# 8.2.4.67 http://www.w3.org/TR/html5/syntax.html#bogus-doctype-state
+	tok_state_bogus_doctype = ->
+		c = txt.charAt(cur++)
+		if c is '>'
+			tok_state = tok_state_data
+			return tok_cur_tag
+		if c is '' # EOF
+			tok_state = tok_state_data
+			cur -= 1 # Reconsume
+			return tok_cur_tag
+		# Anything else
+		return null
+
+
 	# 8.2.4.69 http://www.w3.org/TR/html5/syntax.html#consume-a-character-reference
 	# Don't set this as a state, just call it
 	# returns a string (NOT a text node)
@@ -2578,6 +3377,8 @@ parse_html = (txt, parse_error_cb = null) ->
 	temporary_buffer = null
 	pending_table_character_tokens = []
 	head_element_pointer = null
+	flag_fragment_parsing = false # parser originally created as part of the html fragment parsing algorithm (fragment case)
+	context_element = null # FIXME initialize from args.fragment
 
 	# tokenizer initialization
 	tok_state = tok_state_data
@@ -2587,16 +3388,9 @@ parse_html = (txt, parse_error_cb = null) ->
 		t = tok_state()
 		if t?
 			insertion_mode t
+			# fixfull parse error if has self-closing flag, but it wasn't acknolwedged
 	return doc.children
 
-# everything below is tests on the above
-test_equals = (description, output, expected_output) ->
-	if output is expected_output
-		console.log "passed." # don't say name, so smart consoles can merge all of these
-	else
-		console.log "FAILED: \"#{description}\""
-		console.log "   Expected: #{expected_output}"
-		console.log "     Actual: #{output}"
 serialize_els = (els, shallow, show_ids) ->
 	serialized = ''
 	sep = ''
@@ -2605,199 +3399,12 @@ serialize_els = (els, shallow, show_ids) ->
 		sep = ','
 		serialized += t.serialize shallow, show_ids
 	return serialized
-test_parser = (args) ->
-	debug_log_reset()
-	parse_errors = []
-	errors_cb = (i) ->
-		parse_errors.push i
-	prev_node_id = 0 # reset counter
-	parsed = parse_html args.html, errors_cb
-	serialized = serialize_els parsed, false, false
-	expected = 'tag:"html",{},[tag:"head",{},[],tag:"body",{},[' + args.expected + ']]'
-	if serialized isnt expected
-		debug_log_each (str) ->
-			console.log str
-		console.log "FAILED: \"#{args.name}\""
-		console.log "      Input: #{args.html}"
-		console.log "    Correct: #{expected}"
-		console.log "     Output: #{serialized}"
-		if parse_errors.length > 0
-			console.log " parse errs: #{JSON.stringify parse_errors}"
-		else
-			console.log "   No parse errors"
-	else
-		console.log "passed \"#{args.name}\""
-
-test_parser name: "empty", \
-	html: "",
-	expected: ''
-test_parser name: "just text", \
-	html: "abc",
-	expected: 'text:"abc"'
-test_parser name: "named entity", \
-	html: "a&1234",
-	expected: 'text:"a&1234"'
-test_parser name: "broken named character references", \
-	html: "1&amp2&&3&aabbcc;",
-	expected: 'text:"1&2&&3&aabbcc;"'
-test_parser name: "numbered entity overrides", \
-	html: "1&#X80€ &#x83",
-	expected: 'text:"1€€ ƒ"'
-test_parser name: "open tag", \
-	html: "foo<span>bar",
-	expected: 'text:"foo",tag:"span",{},[text:"bar"]'
-test_parser name: "open tag with attributes", \
-	html: "foo<span style=\"foo: bar\" title=\"hi\">bar",
-	expected: 'text:"foo",tag:"span",{"style":"foo: bar","title":"hi"},[text:"bar"]'
-test_parser name: "open tag with attributes of various quotings", \
-	html: "foo<span abc=\"def\" g=hij klm='nopqrstuv\"' autofocus>bar",
-	expected: 'text:"foo",tag:"span",{"abc":"def","autofocus":"","g":"hij","klm":"nopqrstuv\\""},[text:"bar"]'
-test_parser name: "attribute entity exceptions dq", \
-	html: "foo<a href=\"foo?t=1&amp=2&ampo=3&amp;lt=foo\">bar",
-	expected: 'text:"foo",tag:"a",{"href":"foo?t=1&amp=2&ampo=3&lt=foo"},[text:"bar"]'
-test_parser name: "attribute entity exceptions sq", \
-	html: "foo<a href='foo?t=1&amp=2&ampo=3&amp;lt=foo'>bar",
-	expected: 'text:"foo",tag:"a",{"href":"foo?t=1&amp=2&ampo=3&lt=foo"},[text:"bar"]'
-test_parser name: "attribute entity exceptions uq", \
-	html: "foo<a href=foo?t=1&amp=2&ampo=3&amp;lt=foo>bar",
-	expected: 'text:"foo",tag:"a",{"href":"foo?t=1&amp=2&ampo=3&lt=foo"},[text:"bar"]'
-test_parser name: "matching closing tags", \
-	html: "foo<a href=\"hi\">hi</a><div>1<div>foo</div>2</div>bar",
-	expected: 'text:"foo",tag:"a",{"href":"hi"},[text:"hi"],tag:"div",{},[text:"1",tag:"div",{},[text:"foo"],text:"2"],text:"bar"'
-test_parser name: "missing closing tag inside", \
-	html: "foo<div>bar<span>baz</div>qux",
-	expected: 'text:"foo",tag:"div",{},[text:"bar",tag:"span",{},[text:"baz"]],text:"qux"'
-test_parser name: "mis-matched closing tags", \
-	html: "<span>12<div>34</span>56</div>78",
-	expected: 'tag:"span",{},[text:"12",tag:"div",{},[text:"3456"],text:"78"]'
-test_parser name: "mis-matched formatting elements", \
-	html: "12<b>34<i>56</b>78</i>90",
-	expected: 'text:"12",tag:"b",{},[text:"34",tag:"i",{},[text:"56"]],tag:"i",{},[text:"78"],text:"90"'
-test_parser name: "8.2.8.1 Misnested tags: <b><i></b></i>", \
-	html: '<p>1<b>2<i>3</b>4</i>5</p>',
-	expected: 'tag:"p",{},[text:"1",tag:"b",{},[text:"2",tag:"i",{},[text:"3"]],tag:"i",{},[text:"4"],text:"5"]'
-test_parser name: "8.2.8.2 Misnested tags: <b><p></b></p>", \
-	html: '<b>1<p>2</b>3</p>',
-	expected: 'tag:"b",{},[text:"1"],tag:"p",{},[tag:"b",{},[text:"2"],text:"3"]'
-test_parser name: "crazy formatting elements test", \
-	html: "<b><i><a><s><tt><div></b>first</b></div></tt></s></a>second</i>",
-	# chrome does this: expected: 'tag:"b",{},[tag:"i",{},[tag:"a",{},[tag:"s",{},[tag:"tt",{},[]]],text:"second"]],tag:"a",{},[tag:"s",{},[tag:"tt",{},[tag:"div",{},[tag:"b",{},[],text:"first"]]]]'
-	# firefox does this:
-	expected: 'tag:"b",{},[tag:"i",{},[tag:"a",{},[tag:"s",{},[tag:"tt",{},[]]]]],tag:"a",{},[tag:"s",{},[tag:"tt",{},[tag:"div",{},[tag:"b",{},[],text:"first"]]]],text:"second"'
-# tests from https://github.com/html5lib/html5lib-tests/blob/master/tree-construction/adoption01.dat
-test_parser name: "html5lib aaa 1", \
-	html: '<a><p></a></p>',
-	expected: 'tag:"a",{},[],tag:"p",{},[tag:"a",{},[]]'
-test_parser name: "html5lib aaa 2", \
-	html: '<a>1<p>2</a>3</p>',
-	expected: 'tag:"a",{},[text:"1"],tag:"p",{},[tag:"a",{},[text:"2"],text:"3"]'
-test_parser name: "html5lib aaa 3", \
-	html: '<a>1<button>2</a>3</button>',
-	expected: 'tag:"a",{},[text:"1"],tag:"button",{},[tag:"a",{},[text:"2"],text:"3"]'
-test_parser name: "html5lib aaa 4", \
-	html: '<a>1<b>2</a>3</b>',
-	expected: 'tag:"a",{},[text:"1",tag:"b",{},[text:"2"]],tag:"b",{},[text:"3"]'
-test_parser name: "html5lib aaa 5 (two divs deep)", \
-	html: '<a>1<div>2<div>3</a>4</div>5</div>',
-	expected: 'tag:"a",{},[text:"1"],tag:"div",{},[tag:"a",{},[text:"2"],tag:"div",{},[tag:"a",{},[text:"3"],text:"4"],text:"5"]'
-test_parser name: "html5lib aaa 6 (foster parenting)", \
-	html: '<table><a>1<p>2</a>3</p>',
-	expected: 'tag:"a",{},[text:"1"],tag:"p",{},[tag:"a",{},[text:"2"],text:"3"],tag:"table",{},[]'
-test_parser name: "html5lib aaa 7 (aaa, eof) 1", \
-	html: '<b><b><a><p></a>',
-	expected: 'tag:"b",{},[tag:"b",{},[tag:"a",{},[],tag:"p",{},[tag:"a",{},[]]]]'
-test_parser name: "html5lib aaa 8 (aaa, eof) 2", \
-	html: '<b><a><b><p></a>',
-	expected: 'tag:"b",{},[tag:"a",{},[tag:"b",{},[]],tag:"b",{},[tag:"p",{},[tag:"a",{},[]]]]'
-test_parser name: "html5lib aaa 9 (aaa, eof) 3", \
-	html: '<a><b><b><p></a>',
-	expected: 'tag:"a",{},[tag:"b",{},[tag:"b",{},[]]],tag:"b",{},[tag:"b",{},[tag:"p",{},[tag:"a",{},[]]]]'
-test_parser name: "html5lib aaa 10 (formatting, nesting, attrs, aaa)", \
-	html: '<p>1<s id="A">2<b id="B">3</p>4</s>5</b>',
-	expected: 'tag:"p",{},[text:"1",tag:"s",{"id":"A"},[text:"2",tag:"b",{"id":"B"},[text:"3"]]],tag:"s",{"id":"A"},[tag:"b",{"id":"B"},[text:"4"]],tag:"b",{"id":"B"},[text:"5"]'
-test_parser name: "html5lib aaa 11 (table with foster parenting, formatting el and td)", \
-	html: '<table><a>1<td>2</td>3</table>',
-	expected: 'tag:"a",{},[text:"1"],tag:"a",{},[text:"3"],tag:"table",{},[tag:"tbody",{},[tag:"tr",{},[tag:"td",{},[text:"2"]]]]'
-test_parser name: "html5lib aaa 12 (table with foster parenting, split text)", \
-	html: '<table>A<td>B</td>C</table>',
-	expected: 'text:"AC",tag:"table",{},[tag:"tbody",{},[tag:"tr",{},[tag:"td",{},[text:"B"]]]]'
-# TODO implement svg and namespacing
-#test_parser name: "html5lib aaa 13 (svg tr input)", \
-#	html: '<a><svg><tr><input></a>',
-#	expected: 'tag:"a",{},[svg:"svg",{},[svg:"tr",{},[svg:"input"]]]'
-test_parser name: "html5lib aaa 14 (deep ?outer aaa)", \
-	html: '<div><a><b><div><div><div><div><div><div><div><div><div><div></a>',
-	expected: 'tag:"div",{},[tag:"a",{},[tag:"b",{},[]],tag:"b",{},[tag:"div",{},[tag:"a",{},[],tag:"div",{},[tag:"a",{},[],tag:"div",{},[tag:"a",{},[],tag:"div",{},[tag:"a",{},[],tag:"div",{},[tag:"a",{},[],tag:"div",{},[tag:"a",{},[],tag:"div",{},[tag:"a",{},[],tag:"div",{},[tag:"a",{},[tag:"div",{},[tag:"div",{},[]]]]]]]]]]]]]'
-test_parser name: "html5lib aaa 15 (deep ?inner aaa)", \
-	html: '<div><a><b><u><i><code><div></a>',
-	expected: 'tag:"div",{},[tag:"a",{},[tag:"b",{},[tag:"u",{},[tag:"i",{},[tag:"code",{},[]]]]],tag:"u",{},[tag:"i",{},[tag:"code",{},[tag:"div",{},[tag:"a",{},[]]]]]]'
-test_parser name: "html5lib aaa 16 (correctly nested 4b)", \
-	html: '<b><b><b><b>x</b></b></b></b>y',
-	expected: 'tag:"b",{},[tag:"b",{},[tag:"b",{},[tag:"b",{},[text:"x"]]]],text:"y"'
-test_parser name: "html5lib aaa 17 (formatting, implied /p, noah's ark)", \
-	html: '<p><b><b><b><b><p>x',
-	expected: 'tag:"p",{},[tag:"b",{},[tag:"b",{},[tag:"b",{},[tag:"b",{},[]]]]],tag:"p",{},[tag:"b",{},[tag:"b",{},[tag:"b",{},[text:"x"]]]]'
-test_parser name: "variation on html5lib aaa 17 (with attributes in various orders)", \
-	html: '<p><b c="d" e="f"><b e="f" c="d"><b e="f" c="d"><b c="d" e="f"><p>x',
-	expected: 'tag:"p",{},[tag:"b",{"c":"d","e":"f"},[tag:"b",{"c":"d","e":"f"},[tag:"b",{"c":"d","e":"f"},[tag:"b",{"c":"d","e":"f"},[]]]]],tag:"p",{},[tag:"b",{"c":"d","e":"f"},[tag:"b",{"c":"d","e":"f"},[tag:"b",{"c":"d","e":"f"},[text:"x"]]]]'
-test_parser name: "junk after attribute close-quote", \
-	html: '<p><b c="d", e="f">foo<p>x',
-	expected: 'tag:"p",{},[tag:"b",{",":"","c":"d","e":"f"},[text:"foo"]],tag:"p",{},[tag:"b",{",":"","c":"d","e":"f"},[text:"x"]]'
-test_parser name: "html5lib aaa02 1", \
-	html: '<b>1<i>2<p>3</b>4',
-	expected: 'tag:"b",{},[text:"1",tag:"i",{},[text:"2"]],tag:"i",{},[tag:"p",{},[tag:"b",{},[text:"3"],text:"4"]]'
-test_parser name: "html5lib aaa02 2", \
-	html: '<a><div><style></style><address><a>',
-	expected: 'tag:"a",{},[],tag:"div",{},[tag:"a",{},[tag:"style",{},[]],tag:"address",{},[tag:"a",{},[],tag:"a",{},[]]]'
-test_parser name: "html5lib tables 1", \
-	html: '<table><th>',
-	expected: 'tag:"table",{},[tag:"tbody",{},[tag:"tr",{},[tag:"th",{},[]]]]'
-test_parser name: "html5lib tables 2", \
-	html: '<table><td>',
-	expected: 'tag:"table",{},[tag:"tbody",{},[tag:"tr",{},[tag:"td",{},[]]]]'
-test_parser name: "html5lib tables 3", \
-	html: "<table><col foo='bar'>",
-	expected: 'tag:"table",{},[tag:"colgroup",{},[tag:"col",{"foo":"bar"},[]]]'
-test_parser name: "html5lib tables 4", \
-	html: '<table><colgroup></html>foo',
-	expected: 'text:"foo",tag:"table",{},[tag:"colgroup",{},[]]'
-test_parser name: "html5lib tables 5", \
-	html: '<table></table><p>foo',
-	expected: 'tag:"table",{},[],tag:"p",{},[text:"foo"]'
-test_parser name: "html5lib tables 6", \
-	html: '<table></body></caption></col></colgroup></html></tbody></td></tfoot></th></thead></tr><td>',
-	expected: 'tag:"table",{},[tag:"tbody",{},[tag:"tr",{},[tag:"td",{},[]]]]'
-test_parser name: "html5lib tables 7", \
-	html: '<table><select><option>3</select></table>',
-	expected: 'tag:"select",{},[tag:"option",{},[text:"3"]],tag:"table",{},[]'
-test_parser name: "html5lib tables 8", \
-	html: '<table><select><table></table></select></table>',
-	expected: 'tag:"select",{},[],tag:"table",{},[],tag:"table",{},[]'
-test_parser name: "html5lib tables 9", \
-	html: '<table><select></table>',
-	expected: 'tag:"select",{},[],tag:"table",{},[]'
-test_parser name: "html5lib tables 10", \
-	html: '<table><select><option>A<tr><td>B</td></tr></table>',
-	expected: 'tag:"select",{},[tag:"option",{},[text:"A"]],tag:"table",{},[tag:"tbody",{},[tag:"tr",{},[tag:"td",{},[text:"B"]]]]'
-test_parser name: "html5lib tables 11", \
-	html: '<table><td></body></caption></col></colgroup></html>foo',
-	expected: 'tag:"table",{},[tag:"tbody",{},[tag:"tr",{},[tag:"td",{},[text:"foo"]]]]'
-test_parser name: "html5lib tables 12", \
-	html: '<table><td>A</table>B',
-	expected: 'tag:"table",{},[tag:"tbody",{},[tag:"tr",{},[tag:"td",{},[text:"A"]]]],text:"B"'
-test_parser name: "html5lib tables 13", \
-	html: '<table><tr><caption>',
-	expected: 'tag:"table",{},[tag:"tbody",{},[tag:"tr",{},[]],tag:"caption",{},[]]'
-test_parser name: "html5lib tables 14", \
-	html: '<table><tr></body></caption></col></colgroup></html></td></th><td>foo',
-	expected: 'tag:"table",{},[tag:"tbody",{},[tag:"tr",{},[tag:"td",{},[text:"foo"]]]]'
-test_parser name: "html5lib tables 15", \
-	html: '<table><td><tr>',
-	expected: 'tag:"table",{},[tag:"tbody",{},[tag:"tr",{},[tag:"td",{},[]],tag:"tr",{},[]]]'
-test_parser name: "html5lib tables 16", \
-	html: '<table><td><button><td>',
-	expected: 'tag:"table",{},[tag:"tbody",{},[tag:"tr",{},[tag:"td",{},[tag:"button",{},[]],tag:"td",{},[]]]]'
-# TODO implement svg parsing
-#test_parser name: "html5lib tables 17", \
-#	html: '<table><tr><td><svg><desc><td>',
-#	expected: 'tag:"table",{},[tag:"tbody",{},[tag:"tr",{},[tag:"td",{},[svg:"svg",{},[svg:"desc",{},[]]],tag:"td",{},[]]]]'
+
+# TODO export TYPE_*
+module.exports.parse_html = parse_html
+module.exports.debug_log_reset = debug_log_reset
+module.exports.debug_log_each = debug_log_each
+module.exports.TYPE_TAG = TYPE_TAG
+module.exports.TYPE_TEXT = TYPE_TEXT
+module.exports.TYPE_COMMENT = TYPE_COMMENT
+module.exports.TYPE_DOCTYPE = TYPE_DOCTYPE