@children = args.children ? []
@namespace = args.namespace ? NS_HTML
@parent = args.parent ? null
+ @token = args.token ? null
if args.id?
@id = "#{args.id}+"
else
# 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: ->
+ if @token?
+ @token.flag 'did_self_close'
+ else
+ @flag 'did_self_close', true
+ flag: ->
# fixfull
serialize: (shallow = false, show_ids = false) -> # for unit tests
ret = ''
# http://www.w3.org/TR/html5/infrastructure.html#space-character
space_chars = "\u0009\u000a\u000c\u000d\u0020"
+is_space = (txt) ->
+ return txt.length is 1 and space_chars.indexOf(txt) > -1
+is_space_tok = (t) ->
+ return t.type is TYPE_TEXT && t.text.length is 1 and space_chars.indexOf(t.text) > -1
# https://en.wikipedia.org/wiki/Whitespace_character#Unicode
whitespace_chars = "\u0009\u000a\u000b\u000c\u000d\u0020\u0085\u00a0\u1680\u2000\u2001\u2002\u2003\u2004\u2005\u2006\u2007\u2008\u2009\u200a\u2028\u2029\u202f\u205f\u3000"
}
el_is_special = (e) ->
- return special_elements[e.name]?
- # FIXME it should really be:
- #return special_elements[e.name] is e.namespace
+ return special_elements[e.name] is e.namespace
# decode_named_char_ref()
#
parse_html = (txt, parse_error_cb = null) ->
cur = 0 # index of next char in txt to be parsed
- # declare tree and tokenizer variables so they're in scope below
- tree = null
+ # declare doc and tokenizer variables so they're in scope below
+ doc = null
open_els = null # stack of open elements
afe = null # active formatting elements
template_insertion_modes = null
flag_foster_parenting = null
form_element_pointer = null
temporary_buffer = null
+ pending_table_character_tokens = null
+ head_element_pointer = null
+ flag_fragment_parsing = null
+
+ stop_parsing = ->
+ flag_parsing = false
parse_error = ->
if parse_error_cb?
return false
return false
+ clear_to_table_stopers = {
+ 'table': true
+ 'template': true
+ 'html': true
+ }
+ clear_stack_to_table_context = ->
+ loop
+ if clear_to_table_stopers[open_els[0].name]?
+ break
+ open_els.shift()
+ return
+ clear_to_table_body_stopers = {
+ 'tbody': true
+ 'tfoot': true
+ 'thead': true
+ 'template': true
+ 'html': true
+ }
+ clear_stack_to_table_body_context = ->
+ loop
+ if clear_to_table_body_stopers[open_els[0].name]?
+ break
+ open_els.shift()
+ return
+ clear_to_table_row_stopers = {
+ 'tr': true
+ 'template': true
+ 'html': true
+ }
+ clear_stack_to_table_row_context = ->
+ loop
+ if clear_to_table_row_stopers[open_els[0].name]?
+ break
+ open_els.shift()
+ return
+ clear_afe_to_marker = ->
+ loop
+ el = afe.shift()
+ if el.type is TYPE_AFE_MARKER
+ return
+
# 8.2.3.1 ...
# http://www.w3.org/TR/html5/syntax.html#reset-the-insertion-mode-appropriately
reset_insertion_mode = ->
# http://www.w3.org/TR/html5/syntax.html#unclosed-formatting-elements
adoption_agency = (subject) ->
debug_log "adoption_agency()"
- debug_log "tree: #{serialize_els tree.children, false, true}"
+ debug_log "tree: #{serialize_els doc.children, false, true}"
debug_log "open_els: #{serialize_els open_els, true, true}"
debug_log "afe: #{serialize_els afe, true, true}"
if open_els[0].name is subject
break
node = node_next ? node_above
debug_log "inner loop #{inner}"
- debug_log "tree: #{serialize_els tree.children, false, true}"
+ debug_log "tree: #{serialize_els doc.children, false, true}"
debug_log "open_els: #{serialize_els open_els, true, true}"
debug_log "afe: #{serialize_els afe, true, true}"
debug_log "ca: #{ca.name}##{ca.id} children: #{serialize_els ca.children, true, true}"
# at the appropriate place for inserting a node, but using common
# ancestor as the override target.
- # JASON: In the case where fe is immediately followed by fb:
+ # In the case where fe is immediately followed by fb:
# * inner loop exits out early (node==fe)
# * last_node is fb
# * last_node is still in the tree (not a duplicate)
debug_log "fe: #{fe.name}##{fe.id} children: #{serialize_els fe.children, true, true}"
debug_log "fb: #{fb.name}##{fb.id} children: #{serialize_els fb.children, true, true}"
debug_log "last_node: #{last_node.name}##{last_node.id} children: #{serialize_els last_node.children, true, true}"
- debug_log "tree: #{serialize_els tree.children, false, true}"
+ debug_log "tree: #{serialize_els doc.children, false, true}"
debug_log "insert"
debug_log "fe: #{fe.name}##{fe.id} children: #{serialize_els fe.children, true, true}"
debug_log "fb: #{fb.name}##{fb.id} children: #{serialize_els fb.children, true, true}"
debug_log "last_node: #{last_node.name}##{last_node.id} children: #{serialize_els last_node.children, true, true}"
- debug_log "tree: #{serialize_els tree.children, false, true}"
+ debug_log "tree: #{serialize_els doc.children, false, true}"
# 15. Create an element for the token for which formatting element
# was created, in the HTML namespace, with furthest block as the
break
# 20. Jump back to the step labeled outer loop.
debug_log "done wrapping fb's children. new_element: #{new_element.name}##{new_element.id}"
- debug_log "tree: #{serialize_els tree.children, false, true}"
+ debug_log "tree: #{serialize_els doc.children, false, true}"
debug_log "open_els: #{serialize_els open_els, true, true}"
debug_log "afe: #{serialize_els afe, true, true}"
debug_log "AAA DONE"
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
# http://www.w3.org/TR/html5/syntax.html#insert-a-comment
# position should be [node, index_within_children]
- tree_insert_comment = (t, position = null) ->
+ insert_comment = (t, position = null) ->
position ?= adjusted_insertion_location()
position[0].children.splice position[1], 0, t
while end_tag_implied[open_els[0].name] and open_els[0].name isnt except
open_els.shift()
+ # 8.2.5.4 The rules for parsing tokens in HTML content
+ # http://www.w3.org/TR/html5/syntax.html#parsing-main-inhtml
+
+ # 8.2.5.4.1 The "initial" insertion mode
+ # http://www.w3.org/TR/html5/syntax.html#the-initial-insertion-mode
+ ins_mode_initial = (t) ->
+ if is_space_tok t
+ return
+ if t.type is TYPE_COMMENT
+ # fixfull this is supposed to be "the last child of the document object"
+ doc.children.push t
+ return
+ if t.type is TYPE_DOCTYPE
+ # fixfull
+ t.name = 'html'
+ doc.children.push t
+ insertion_mode = ins_mode_before_html
+ return
+ # Anything else
+ #fixfull (iframe, quirks)
+ insertion_mode = ins_mode_before_html
+ insertion_mode t # reprocess the token
+ return
+
+ # 8.2.5.4.2 http://www.w3.org/TR/html5/syntax.html#the-before-html-insertion-mode
+ ins_mode_before_html = (t) ->
+ if t.type is TYPE_DOCTYPE
+ parse_error()
+ return
+ if t.type is TYPE_COMMENT
+ doc.children.push t
+ return
+ if is_space_tok t
+ return
+ if t.type is TYPE_START_TAG and t.name is 'html'
+ el = token_to_element t, NS_HTML, doc
+ open_els.unshift(el)
+ # fixfull (big paragraph in spec about manifest, fragment, urls, etc)
+ insertion_mode = ins_mode_before_head
+ return
+ if t.type is TYPE_END_TAG
+ if t.name is 'head' or t.name is 'body' or t.name is 'html' or t.name is 'br'
+ # fall through to "anything else"
+ else
+ parse_error()
+ return
+ # Anything else
+ html_tok = new_open_tag 'html'
+ el = token_to_element html_tok, NS_HTML, doc
+ doc.children.push el
+ open_els.unshift el
+ # ?fixfull browsing context
+ insertion_mode = ins_mode_before_head
+ insertion_mode t
+ return
+
+ # 8.2.5.4.3 http://www.w3.org/TR/html5/syntax.html#the-before-head-insertion-mode
+ ins_mode_before_head = (t) ->
+ if is_space_tok 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 'head'
+ el = insert_html_element t
+ head_element_pointer = el
+ insertion_mode = ins_mode_in_head
+ if t.type is TYPE_END_TAG
+ if t.name is 'head' or t.name is 'body' or t.name is 'html' or t.name is 'br'
+ # fall through to Anything else below
+ else
+ parse_error()
+ return
+ # Anything else
+ head_tok = new_open_tag 'head'
+ el = insert_html_element head_tok
+ head_element_pointer = el
+ insertion_mode = ins_mode_in_head
+ insertion_mode t # reprocess current token
+
# 8.2.5.4.4 http://www.w3.org/TR/html5/syntax.html#parsing-main-inhead
ins_mode_in_head_else = (t) -> # factored out for same-as-spec flow control
open_els.shift() # spec says this will be a 'head' node
insert_character t
return
if t.type is TYPE_COMMENT
- tree_insert_comment t
+ insert_comment t
return
if t.type is TYPE_DOCTYPE
parse_error()
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'
else
parse_error()
return
- if (t.type is TYPE_OPEN_TAG and t.name is 'head') or t.type is TYPE_END_TAG
+ if (t.type is TYPE_START_TAG and t.name is 'head') or t.type is TYPE_END_TAG
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'
+ insert_html_element body_tok
+ insertion_mode = ins_mode_in_body
+ insertion_mode t # reprocess token
+ return
+ ins_mode_after_head = (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 'body'
+ insert_html_element t
+ flag_frameset_ok = false
+ insertion_mode = ins_mode_in_body
+ return
+ if t.type is TYPE_START_TAG and t.name is 'frameset'
+ insert_html_element t
+ insertion_mode = ins_mode_in_frameset
+ 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')
+ parse_error()
+ open_els.unshift head_element_pointer
+ ins_mode_in_head t
+ for el, i of open_els
+ if el is head_element_pointer
+ open_els.splice i, 1
+ return
+ console.log "warning: 23904 couldn't find head element in open_els"
+ return
+ if t.type is TYPE_END_TAG and t.name is 'template'
+ ins_mode_in_head t
+ return
+ if t.type is TYPE_END_TAG and (t.name is 'body' or t.name is 'html' or t.name is 'br')
+ ins_mode_after_head_else t
+ return
+ if (t.type is TYPE_START_TAG and t.name is 'head') or t.type is TYPE_END_TAG
+ parse_error()
+ return
+ # Anything else
+ ins_mode_after_head_else t
# 8.2.5.4.7 http://www.w3.org/TR/html5/syntax.html#parsing-main-inbody
in_body_any_other_end_tag = (name) -> # factored out because adoption agency calls it
insert_character t
flag_frameset_ok = false
when TYPE_COMMENT
- tree_insert_comment t
+ insert_comment t
when TYPE_DOCTYPE
parse_error()
when TYPE_START_TAG
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'
flag_foster_parenting = true # FIXME
ins_mode_in_body t
flag_foster_parenting = false
- can_in_table = {
+ can_in_table = { # FIXME do this inline like everywhere else
'table': true
'tbody': true
'tfoot': true
'thead': true
'tr': true
}
- clear_to_table_stopers = {
- 'table': true
- 'template': true
- 'html': true
- }
- clear_stack_to_table_context = ->
- loop
- if clear_to_table_stopers[open_els[0].name]?
- break
- open_els.shift()
- return
- clear_to_table_body_stopers = {
- 'tbody': true
- 'tfoot': true
- 'thead': true
- 'template': true
- 'html': true
- }
- clear_stack_to_table_body_context = ->
- loop
- if clear_to_table_body_stopers[open_els[0].name]?
- break
- open_els.shift()
- return
- clear_to_table_row_stopers = {
- 'tr': true
- 'template': true
- 'html': true
- }
- clear_stack_to_table_row_context = ->
- loop
- if clear_to_table_row_stopers[open_els[0].name]?
- break
- open_els.shift()
- return
- clear_afe_to_marker = ->
- loop
- el = afe.shift()
- if el.type is TYPE_AFE_MARKER
- return
# 8.2.5.4.8 http://www.w3.org/TR/html5/syntax.html#parsing-main-incdata
ins_mode_text = (t) ->
else
ins_mode_in_table_else t
when TYPE_COMMENT
- tree_insert_comment t
+ insert_comment t
when TYPE_DOCTYPE
parse_error()
when TYPE_START_TAG
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?
ins_mode_in_table_else t
- # 8.2.5.4.11 http://www.w3.org/TR/html5/syntax.html#parsing-main-intabletext
+ # 8.2.5.4.10 http://www.w3.org/TR/html5/syntax.html#parsing-main-intabletext
ins_mode_in_table_text = (t) ->
- switch t.type
- when TYPE_TEXT
- switch t.text
- when "\u0000"
- parse_error()
- return
- console.log "unimplemented ins_mode_in_table_text"
- # FIXME CONTINUE
+ if t.type is TYPE_TEXT and t.text is "\u0000"
+ # huh? I thought the tokenizer didn't emit these
+ parse_error()
+ return
+ if t.type is TYPE_TEXT
+ pending_table_character_tokens.push t
+ return
+ # Anything else
+ all_space = true
+ for old in pending_table_character_tokens
+ unless is_space_tok old
+ all_space = false
+ break
+ if all_space
+ for old in pending_table_character_tokens
+ insert_character old
+ else
+ for old in pending_table_character_tokens
+ ins_mode_table_else old
+ pending_table_character_tokens = [] # FIXME test (spec doesn't say this)
+ insertion_mode = original_insertion_mode
+ insertion_mode t
+
+ # 8.2.5.4.11 http://www.w3.org/TR/html5/syntax.html#parsing-main-incaption
+ ins_mode_in_caption = (t) ->
+ if t.type is TYPE_END_TAG and t.name is 'caption'
+ if is_in_table_scope 'caption'
+ generate_implied_end_tags()
+ if open_els[0].name isnt 'caption'
+ parse_error()
+ loop
+ el = open_els.shift()
+ if el.name is 'caption'
+ break
+ clear_afe_to_marker()
+ insertion_mode = in_table
+ else
+ parse_error()
+ # fragment case
+ return
+ if (t.type is TYPE_START_TAG and (t.name is 'caption' or t.name is 'col' or t.name is 'colgroup' or t.name is 'tbody' or t.name is 'td' or t.name is 'tfoot' or t.name is 'th' or t.name is 'thead' or t.name is 'tr')) or t.type is TYPE_END_TAG and t.name is 'table'
+ parse_error()
+ if is_in_table_scope 'caption'
+ loop
+ el = open_els.shift()
+ if el.name is 'caption'
+ break
+ clear_afe_to_marker()
+ insertion_mode = in_table
+ insertion_mode t
+ # else fragment case
+ return
+ if t.type is TYPE_END_TAG and (t.name is 'body' or t.name is 'col' or t.name is 'colgroup' or t.name is 'html' or t.name is 'tbody' or t.name is 'td' or t.name is 'tfoot' or t.name is 'th' or t.name is 'thead' or t.name is 'tr')
+ parse_error()
+ return
+ # Anything else
+ ins_mode_in_body t
+
+ # 8.2.5.4.12 http://www.w3.org/TR/html5/syntax.html#parsing-main-incolgroup
+ ins_mode_in_column_group = (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 'col'
+ el = insert_html_element t
+ open_els.shift()
+ t.acknowledge_self_closing()
+ return
+ if t.type is TYPE_END_TAG and t.name is 'colgroup'
+ if open_els[0].name is 'colgroup'
+ open_els[0].shift()
+ insertion_mode = ins_mode_in_table
+ else
+ parse_error()
+ return
+ if t.type is TYPE_END_TAG and t.name is 'col'
+ parse_error()
+ return
+ if (t.type is TYPE_START_TAG or t.type is TYPE_END_TAG) and t.name is 'template'
+ ins_mode_in_head t
+ return
+ if t.type is TYPE_EOF
+ ins_mode_in_body t
+ return
+ # Anything else
+ if open_els[0].name isnt 'colgroup'
+ parse_error()
+ return
+ open_els.shift()
+ insertion_mode = ins_mode_in_table
+ insertion_mode t
+ return
# 8.2.5.4.13 http://www.w3.org/TR/html5/syntax.html#parsing-main-intbody
ins_mode_in_table_body = (t) ->
if t.type is TYPE_START_TAG and t.name is 'tr'
clear_stack_to_table_body_context()
insert_html_element t
+ insertion_mode = ins_mode_in_row
return
if t.type is TYPE_START_TAG and (t.name is 'th' or t.name is 'td')
parse_error()
# Anything Else
ins_mode_in_body t
+ # 8.2.5.4.16 http://www.w3.org/TR/html5/syntax.html#parsing-main-inselect
+ ins_mode_in_select = (t) ->
+ if t.type is TYPE_TEXT and t.text is "\u0000"
+ parse_error()
+ return
+ if t.type is TYPE_TEXT
+ 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 'option'
+ if open_els[0].name is 'option'
+ open_els.shift()
+ insert_html_element t
+ return
+ if t.type is TYPE_START_TAG and t.name is 'optgroup'
+ if open_els[0].name is 'option'
+ open_els.shift()
+ if open_els[0].name is 'optgroup'
+ open_els.shift()
+ insert_html_element t
+ return
+ if t.type is TYPE_END_TAG and t.name is 'optgroup'
+ if open_els[0].name is 'option' and open_els[1].name is 'optgroup'
+ open_els.shift()
+ if open_els[0].name is 'optgroup'
+ open_els.shift()
+ else
+ parse_error()
+ return
+ if t.type is TYPE_END_TAG and t.name is 'option'
+ if open_els[0].name is 'option'
+ open_els.shift()
+ else
+ parse_error()
+ return
+ if t.type is TYPE_END_TAG and t.name is 'select'
+ if is_in_select_scope 'select'
+ loop
+ el = open_els.shift()
+ if el.name is 'select'
+ break
+ reset_insertion_mode()
+ else
+ parse_error()
+ return
+ if t.type is TYPE_START_TAG and t.name is 'select'
+ parse_error()
+ loop
+ el = open_els.shift()
+ if el.name is 'select'
+ break
+ reset_insertion_mode()
+ # spec says that this is the same as </select> but it doesn't say
+ # to check scope first
+ return
+ if t.type is TYPE_START_TAG and (t.name is 'input' or t.name is 'keygen' or t.name is 'textarea')
+ parse_error()
+ if is_in_select_scope 'select'
+ return
+ loop
+ el = open_els.shift()
+ if el.name is 'select'
+ break
+ reset_insertion_mode()
+ insertion_mode t
+ return
+ if t.type is TYPE_START_TAG and (t.name is 'script' or t.name is 'template')
+ ins_mode_in_head t
+ return
+ if t.type is TYPE_EOF
+ ins_mode_in_body t
+ return
+ # Anything else
+ parse_error()
+ return
+
+ # 8.2.5.4.17 http://www.w3.org/TR/html5/syntax.html#parsing-main-inselectintable
+ ins_mode_in_select_in_table = (t) ->
+ if t.type is TYPE_START_TAG and (t.name is 'caption' or t.name is 'table' or t.name is 'tbody' or t.name is 'tfoot' or t.name is 'thead' or t.name is 'tr' or t.name is 'td' or t.name is 'th')
+ parse_error()
+ loop
+ el = open_els.shift()
+ if el.name is 'select'
+ break
+ reset_insertion_mode()
+ insertion_mode t
+ return
+ if t.type is TYPE_END_TAG and (t.name is 'caption' or t.name is 'table' or t.name is 'tbody' or t.name is 'tfoot' or t.name is 'thead' or t.name is 'tr' or t.name is 'td' or t.name is 'th')
+ parse_error()
+ unless is_in_table_scope t.name, NS_HTML
+ return
+ loop
+ el = open_els.shift()
+ if el.name is 'select'
+ break
+ reset_insertion_mode()
+ insertion_mode t
+ return
+ # Anything else
+ ins_mode_in_select t
+ return
+
+ # 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
+
+
+
+
+
# 8.2.4.1 http://www.w3.org/TR/html5/syntax.html#data-state
tok_state_data = ->
switch c = txt.charAt(cur++)
# tree constructor initialization
# see comments on TYPE_TAG/etc for the structure of this data
- tree = new Node TYPE_TAG, name: 'html', namespace: NS_HTML
- open_els = [tree]
+ doc = new Node TYPE_TAG, name: 'html', namespace: NS_HTML
+ open_els = [doc]
afe = [] # active formatting elements
template_insertion_modes = []
- insertion_mode = ins_mode_in_body
+ insertion_mode = ins_mode_initial
original_insertion_mode = insertion_mode # TODO check spec
flag_scripting = true # TODO might need an extra flag to get <noscript> to parse correctly
flag_frameset_ok = true
flag_foster_parenting = false
form_element_pointer = 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)
# tokenizer initialization
tok_state = tok_state_data
t = tok_state()
if t?
insertion_mode t
- return tree.children
+ # fixfull parse error if has self-closing flag, but it wasn't acknolwedged
+ return doc.children
+test_results = passed: 0, failed: 0
# everything below is tests on the above
test_equals = (description, output, expected_output) ->
if output is expected_output
prev_node_id = 0 # reset counter
parsed = parse_html args.html, errors_cb
serialized = serialize_els parsed, false, false
- if serialized isnt args.expected
+ 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: #{args.expected}"
+ 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"
+ test_results.failed += 1
else
- console.log "passed \"#{args.name}\""
+ #console.log "passed \"#{args.name}\""
+ test_results.passed += 1
+test_summary = ->
+ console.log "Tests passed: #{test_results.passed}"
+ console.log "Tests Failed: #{test_results.failed}"
test_parser name: "empty", \
html: "",
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","g":"hij","klm":"nopqrstuv\\"","autofocus":""},[text:"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&=2&o=3&lt=foo\">bar",
expected: 'text:"foo",tag:"a",{"href":"foo?t=1&=2&o=3<=foo"},[text:"bar"]'
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",{},[]]]]'
+test_summary()
Got questions, comments, patches, etc.? Contact Jason Woofenden