X-Git-Url: https://jasonwoof.com/gitweb/?a=blobdiff_plain;f=parse-html.coffee;h=f64c734e13e7dab898248150aad813cacce83041;hb=adc7477c34f3a2aa480e7f2af5ea954d2421d000;hp=aaecff348a733319abf14a41aa35cbb0d47e613d;hpb=5be4d9791f4d11e80a608da2b6e122734628bf5c;p=peach-html5-editor.git diff --git a/parse-html.coffee b/parse-html.coffee index aaecff3..f64c734 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 @@ -93,7 +100,14 @@ 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: -> + 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 = '' switch @type @@ -105,8 +119,17 @@ class Node ret += "##{@id}," if shallow break - ret += JSON.stringify @attrs - ret += ',[' + attr_keys = [] + for k of @attrs + attr_keys.push k + attr_keys.sort() + ret += '{' + sep = '' + for k in attr_keys + ret += sep + sep = ',' + ret += "#{JSON.stringify k}:#{JSON.stringify @attrs[k]}" + ret += '},[' sep = '' for c in @children ret += sep @@ -140,8 +163,11 @@ new_element = (name) -> return new Node TYPE_TAG, name: name new_text_node = (txt) -> return new Node TYPE_TEXT, text: txt -new_comment_node = (txt) -> +new_character_token = new_text_node +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 = -> @@ -149,17 +175,35 @@ new_afe_marker = -> new_aaa_bookmark = -> return new Node TYPE_AAA_BOOKMARK -lc_alpha = "abcdefghijklmnopqrstuvwxqz" -uc_alpha = "ABCDEFGHIJKLMNOPQRSTUVWXQZ" +lc_alpha = "abcdefghijklmnopqrstuvwxyz" +uc_alpha = "ABCDEFGHIJKLMNOPQRSTUVWXYZ" digits = "0123456789" alnum = lc_alpha + uc_alpha + digits hex_chars = digits + "abcdefABCDEF" +is_uc_alpha = (str) -> + return str.length is 1 and uc_alpha.indexOf(str) > -1 +is_lc_alpha = (str) -> + return str.length is 1 and lc_alpha.indexOf(str) > -1 + # some SVG elements have dashes in them tag_name_chars = alnum + "-" # 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 + +is_input_hidden_tok = (t) -> + return unless t.type is TYPE_START_TAG + for a of t.attrs_a + if a[0] is 'type' + if a[1].toLowerCase() is 'hidden' + return true + return false + return false # 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" @@ -301,9 +345,7 @@ end_tag_implied = { } 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() # @@ -329,17 +371,28 @@ decode_named_char_ref = (txt) -> 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 - open_els = [] # stack of open elements + # 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 insertion_mode = null + original_insertion_mode = null tok_state = null tok_cur_tag = null # partially parsed tag + flag_scripting = null flag_frameset_ok = null flag_parsing = null flag_foster_parenting = null form_element_pointer = null - afe = [] # active formatting elements + 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? @@ -347,6 +400,21 @@ parse_html = (txt, parse_error_cb = null) -> else console.log "Parse error at character #{cur} of #{txt.length}" + afe_push = (new_el) -> + matches = 0 + for el, i in afe + if el.name is new_el.name and el.namespace is new_el.namespace + for k, v of el.attrs + continue unless new_el.attrs[k] is v + for k, v of new_el.attrs + continue unless el.attrs[k] is v + matches += 1 + if matches is 3 + afe.splice i, 1 + break + afe.unshift new_el + afe_push_marker = -> + afe.unshift new_afe_marker() # the functions below impliment the Tree Contstruction algorithm # http://www.w3.org/TR/html5/syntax.html#tree-construction @@ -396,7 +464,7 @@ parse_html = (txt, parse_error_cb = null) -> for t in open_els if t.name is tag_name and (namespace is null or namespace is t.namespace) return true - if t.ns isnt NS_HTML t.name isnt 'optgroup' and t.name isnt 'option' + if t.ns isnt NS_HTML and t.name isnt 'optgroup' and t.name isnt 'option' return false return false # this checks for a particular element, not by name @@ -408,6 +476,49 @@ parse_html = (txt, parse_error_cb = null) -> 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 + return unless afe.length > 0 # this happens in fragment case, ?spec error + el = afe.shift() + if el.type is TYPE_AFE_MARKER + return + return + # 8.2.3.1 ... # http://www.w3.org/TR/html5/syntax.html#reset-the-insertion-mode-appropriately reset_insertion_mode = -> @@ -531,6 +642,14 @@ 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. @@ -563,7 +682,7 @@ parse_html = (txt, parse_error_cb = null) -> # 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 @@ -675,7 +794,7 @@ parse_html = (txt, parse_error_cb = null) -> 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}" @@ -769,7 +888,7 @@ parse_html = (txt, parse_error_cb = null) -> # 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) @@ -786,7 +905,7 @@ parse_html = (txt, parse_error_cb = null) -> 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" @@ -802,7 +921,7 @@ parse_html = (txt, parse_error_cb = null) -> 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 @@ -842,13 +961,12 @@ parse_html = (txt, parse_error_cb = null) -> 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" # http://www.w3.org/TR/html5/syntax.html#close-a-p-element - # FIXME test this (particularly emplied end tags) close_p_element = -> generate_implied_end_tags 'p' # arg is exception if open_els[0].name isnt 'p' @@ -859,10 +977,11 @@ parse_html = (txt, parse_error_cb = null) -> return close_p_if_in_button_scope = -> if is_in_button_scope 'p' - close_a_p_element() + close_p_element() # http://www.w3.org/TR/html5/syntax.html#insert-a-character - tree_insert_text = (t) -> + # aka insert_a_character = (t) -> + insert_character = (t) -> dest = adjusted_insertion_location() # fixfull check for Document node if dest[1] > 0 @@ -915,7 +1034,7 @@ parse_html = (txt, parse_error_cb = null) -> # last template's template contents, after its last child (if # any), and abort these substeps. if last_template and (last_table is null or last_template_i < last_table_i) - target = template # fixfull should be it's contents + target = last_template # fixfull should be it's contents target_i = target.children.length break # 4. If there is no last table, then let adjusted insertion @@ -973,7 +1092,7 @@ 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 @@ -1022,17 +1141,257 @@ parse_html = (txt, parse_error_cb = null) -> # http://www.w3.org/TR/html5/syntax.html#insert-a-comment # position should be [node, index_within_children] - tree_insert_a_comment = (t, position = null) -> + insert_comment = (t, position = null) -> position ?= adjusted_insertion_location() position[0].children.splice position[1], 0, t + # 8.2.5.2 + # http://www.w3.org/TR/html5/syntax.html#generic-raw-text-element-parsing-algorithm + parse_generic_raw_text = (t) -> + insert_html_element t + tok_state = tok_state_rawtext + original_insertion_mode = insertion_mode + insertion_mode = ins_mode_text + parse_generic_rcdata_text = (t) -> + insert_html_element t + tok_state = tok_state_rcdata + original_insertion_mode = insertion_mode + insertion_mode = ins_mode_text + # 8.2.5.3 http://www.w3.org/TR/html5/syntax.html#closing-elements-that-have-implied-end-tags # http://www.w3.org/TR/html5/syntax.html#generate-implied-end-tags generate_implied_end_tags = (except = null) -> while end_tag_implied[open_els[0].name] and open_els[0].name isnt except open_els.shift() - # 8.2.5.4 http://www.w3.org/TR/html5/syntax.html#parsing-main-inbody + # 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 + doc.children.push t + return + if t.type is TYPE_DOCTYPE + # FIXME check identifiers, set quirks, etc + # fixfull + 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 + doc.children.push el + 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 + insertion_mode = ins_mode_after_head + insertion_mode t + ins_mode_in_head = (t) -> + if t.type is TYPE_TEXT and (t.text is "\t" or t.text is "\n" or t.text is "\u000c" or t.text is ' ') + 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 'base' or t.name is 'basefont' or t.name is 'bgsound' or t.name is 'link') + el = insert_html_element t + open_els.shift() + 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() + t.acknowledge_self_closing() + # fixfull encoding stuff + return + if t.type is TYPE_START_TAG and t.name is 'title' + 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 + return + if t.type is TYPE_START_TAG and t.name is 'noscript' and flag_scripting is false + insert_html_element t + insertion_mode = ins_mode_in_head_noscript # FIXME implement + return + 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 + # fixfull frament case + ail[0].children.splice ail[1], 0, el + open_els.unshift el + tok_state = tok_state_script_data + original_insertion_mode = insertion_mode # make sure orig... is defined + insertion_mode = ins_mode_text # FIXME implement + return + if t.type is TYPE_END_TAG and t.name is 'head' + open_els.shift() # will be a head element... spec says so + insertion_mode = ins_mode_after_head + 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_in_head_else t + return + if t.type is TYPE_START_TAG and t.name is 'template' + insert_html_element t + afe_push_marker() + flag_frameset_ok = false + insertion_mode = ins_mode_in_template + template_insertion_modes.unshift ins_mode_in_template # FIXME implement + return + if t.type is TYPE_END_TAG and t.name is 'template' + if template_tag_is_open() + generate_implied_end_tags + if open_els[0].name isnt 'template' + parse_error() + loop + el = open_els.shift() + if el.name is 'template' + break + clear_afe_to_marker() + template_insertion_modes.shift() + reset_insertion_mode() + else + parse_error() + return + 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 for node, i in open_els if node.name is name # FIXME check namespace too @@ -1053,13 +1412,13 @@ parse_html = (txt, parse_error_cb = null) -> parse_error() when "\t", "\u000a", "\u000c", "\u000d", ' ' reconstruct_active_formatting_elements() - tree_insert_text t + insert_character t else reconstruct_active_formatting_elements() - tree_insert_text t + insert_character t flag_frameset_ok = false when TYPE_COMMENT - tree_insert_a_comment t + insert_comment t when TYPE_DOCTYPE parse_error() when TYPE_START_TAG @@ -1072,7 +1431,7 @@ parse_html = (txt, parse_error_cb = null) -> root_attrs[k] = v unless root_attrs[k]? when 'base', 'basefont', 'bgsound', 'link', 'meta', 'noframes', 'script', 'style', 'template', 'title' # FIXME also do this for (end tag) - return tree_in_head t + return ins_mode_in_head t when 'body' parse_error() # TODO @@ -1117,11 +1476,11 @@ parse_html = (txt, parse_error_cb = null) -> open_els.splice i, 1 reconstruct_active_formatting_elements() el = insert_html_element t - afe.unshift el + afe_push el when 'b', 'big', 'code', 'em', 'font', 'i', 's', 'small', 'strike', 'strong', 'tt', 'u' reconstruct_active_formatting_elements() el = insert_html_element t - afe.unshift el + afe_push el when 'table' # fixfull quirksmode thing close_p_if_in_button_scope() @@ -1141,7 +1500,7 @@ 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' @@ -1184,53 +1543,43 @@ parse_html = (txt, parse_error_cb = null) -> 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 + + # 8.2.5.4.8 http://www.w3.org/TR/html5/syntax.html#parsing-main-incdata + ins_mode_text = (t) -> + if t.type is TYPE_TEXT + insert_character t + return + if t.type is TYPE_EOF + parse_error() + if open_els[0].name is 'script' + open_els[0].flag 'already started', true 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 + insertion_mode = original_insertion_mode + insertion_mode t + return + if t.type is TYPE_END_TAG and t.name is 'script' 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 + insertion_mode = original_insertion_mode + # fixfull the spec seems to assume that I'm going to run the script + # http://www.w3.org/TR/html5/syntax.html#scriptEndTag + return + if t.type is TYPE_END_TAG open_els.shift() - return - clear_afe_to_marker = -> - loop - el = afe.shift() - if el.type is TYPE_AFE_MARKER - return + insertion_mode = original_insertion_mode + return + console.log 'warning: end of ins_mode_text reached' + + # the functions below implement the tokenizer stats described here: + # http://www.w3.org/TR/html5/syntax.html#tokenization + + # 8.2.5.4.9 http://www.w3.org/TR/html5/syntax.html#parsing-main-intable ins_mode_in_table = (t) -> switch t.type when TYPE_TEXT @@ -1241,14 +1590,14 @@ parse_html = (txt, parse_error_cb = null) -> else ins_mode_in_table_else t when TYPE_COMMENT - tree_insert_a_comment t + insert_comment t when TYPE_DOCTYPE parse_error() when TYPE_START_TAG switch t.name when 'caption' clear_stack_to_table_context() - afe.unshift new_afe_marker() + afe_push_marker() insert_html_element t insertion_mode = ins_mode_in_caption when 'colgroup' @@ -1281,13 +1630,13 @@ parse_html = (txt, parse_error_cb = null) -> when 'style', 'script', 'template' ins_mode_in_head t when 'input' - if token_is_input_hidden t + if is_input_hidden_tok t ins_mode_in_table_else t else parse_error() - insert_html_element t + el = insert_html_element t open_els.shift() - # fixfull acknowledge sef-closing flag + t.acknowledge_self_closing() when 'form' parse_error() if form_element_pointer? @@ -1321,20 +1670,116 @@ parse_html = (txt, parse_error_cb = null) -> ins_mode_in_table_else t + # 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 = ins_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 = ins_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.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() @@ -1373,12 +1818,13 @@ parse_html = (txt, parse_error_cb = null) -> # Anything else ins_mode_in_table t + # 8.2.5.4.14 http://www.w3.org/TR/html5/syntax.html#parsing-main-intr ins_mode_in_row = (t) -> if t.type is TYPE_START_TAG and (t.name is 'th' or t.name is 'td') clear_stack_to_table_row_context() insert_html_element t insertion_mode = ins_mode_in_cell - afe.unshift new_afe_marker() + afe_push_marker() return if t.type is TYPE_END_TAG and t.name is 'tr' if is_in_table_scope 'tr' @@ -1425,7 +1871,7 @@ parse_html = (txt, parse_error_cb = null) -> clear_afe_to_marker() insertion_mode = ins_mode_in_row - # http://www.w3.org/TR/html5/syntax.html#parsing-main-intd + # 8.2.5.4.15 http://www.w3.org/TR/html5/syntax.html#parsing-main-intd ins_mode_in_cell = (t) -> if t.type is TYPE_END_TAG and (t.name is 'td' or t.name is 'th') if is_in_table_scope t.name @@ -1468,30 +1914,387 @@ parse_html = (txt, parse_error_cb = null) -> # 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 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 TYPE_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++) + when '&' + return new_text_node parse_character_reference() + when '<' + tok_state = tok_state_tag_open + when "\u0000" + parse_error() + return new_text_node c + when '' # EOF + return new_eof_token() + else + return new_text_node c + return null + + # 8.2.4.2 http://www.w3.org/TR/html5/syntax.html#character-reference-in-data-state + # not needed: tok_state_character_reference_in_data = -> + # just call parse_character_reference() + + # 8.2.4.3 http://www.w3.org/TR/html5/syntax.html#rcdata-state + tok_state_rcdata = -> + switch c = txt.charAt(cur++) + when '&' + return new_text_node parse_character_reference() + when '<' + tok_state = tok_state_rcdata_less_than_sign + when "\u0000" + parse_error() + return new_character_token "\ufffd" + when '' # EOF + return new_eof_token() + else + return new_character_token c + return null + + # 8.2.4.4 http://www.w3.org/TR/html5/syntax.html#character-reference-in-rcdata-state + # not needed: tok_state_character_reference_in_rcdata = -> + # just call parse_character_reference() + + # 8.2.4.5 http://www.w3.org/TR/html5/syntax.html#rawtext-state + tok_state_rawtext = -> + switch c = txt.charAt(cur++) + when '<' + tok_state = tok_state_rawtext_less_than_sign + when "\u0000" + parse_error() + return new_character_token "\ufffd" + when '' # EOF + return new_eof_token() + else + return new_character_token c + return null + + # 8.2.4.6 http://www.w3.org/TR/html5/syntax.html#script-data-state + tok_state_script_data = -> + switch c = txt.charAt(cur++) + when '<' + tok_state = tok_state_script_data_less_than_sign + when "\u0000" + parse_error() + return new_character_token "\ufffd" + when '' # EOF + return new_eof_token() + else + return new_character_token c + return null + + # 8.2.4.7 http://www.w3.org/TR/html5/syntax.html#plaintext-state + tok_state_plaintext = -> + switch c = txt.charAt(cur++) + when "\u0000" + parse_error() + return new_character_token "\ufffd" + when '' # EOF + return new_eof_token() + else + return new_character_token c + return null + - # the functions below implement the tokenizer stats described here: - # http://www.w3.org/TR/html5/syntax.html#tokenization - - # 8.2.4.1 http://www.w3.org/TR/html5/syntax.html#data-state - tok_state_data = -> - switch c = txt.charAt(cur++) - when '&' - return new_text_node tokenize_character_reference() - when '<' - tok_state = tok_state_tag_open - when "\u0000" - parse_error() - return new_text_node c - when '' # EOF - return new_eof_token() - else - return new_text_node c - return null - - # 8.2.4.2 http://www.w3.org/TR/html5/syntax.html#character-reference-in-data-state - # not needed: tok_state_character_reference_in_data = -> - # just call tok_state_character_reference_in_data() - # 8.2.4.8 http://www.w3.org/TR/html5/syntax.html#tag-open-state tok_state_tag_open = -> switch c = txt.charAt(cur++) @@ -1501,12 +2304,13 @@ 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 + if is_lc_alpha(c) tok_cur_tag = new_open_tag c tok_state = tok_state_tag_name - else if uc_alpha.indexOf(c) > -1 + else if is_uc_alpha(c) tok_cur_tag = new_open_tag c.toLowerCase() tok_state = tok_state_tag_name else @@ -1527,14 +2331,15 @@ parse_html = (txt, parse_error_cb = null) -> tok_state = tok_state_data return new_text_node ' -1 + if is_uc_alpha(c) tok_cur_tag = new_end_tag c.toLowerCase() tok_state = tok_state_tag_name - else if lc_alpha.indexOf(c) > -1 + else if is_lc_alpha(c) tok_cur_tag = new_end_tag c tok_state = tok_state_tag_name else parse_error() + tok_cur_tag = new_comment_token '/' tok_state = tok_state_bogus_comment return null @@ -1557,11 +2362,472 @@ parse_html = (txt, parse_error_cb = null) -> parse_error() tok_state = tok_state_data else - if uc_alpha.indexOf(c) > -1 - tok_cur_tag.name += c.toLowerCase() - else - tok_cur_tag.name += c - return null + if is_uc_alpha(c) + tok_cur_tag.name += c.toLowerCase() + else + tok_cur_tag.name += c + return null + + # 8.2.4.11 http://www.w3.org/TR/html5/syntax.html#rcdata-less-than-sign-state + tok_state_rcdata_less_than_sign = -> + c = txt.charAt(cur++) + if c is '/' + temporary_buffer = '' + tok_state = tok_state_rcdata_end_tag_open + return null + # Anything else + tok_state = tok_state_rcdata + cur -= 1 # reconsume the input character + return new_character_token '<' + + # 8.2.4.12 http://www.w3.org/TR/html5/syntax.html#rcdata-end-tag-open-state + tok_state_rcdata_end_tag_open = -> + c = txt.charAt(cur++) + if is_uc_alpha(c) + tok_cur_tag = new_end_tag c.toLowerCase() + temporary_buffer += c + tok_state = tok_state_rcdata_end_tag_name + return null + if is_lc_alpha(c) + tok_cur_tag = new_end_tag c + temporary_buffer += c + tok_state = tok_state_rcdata_end_tag_name + return null + # Anything else + tok_state = tok_state_rcdata + cur -= 1 # reconsume the input character + return new_character_token " + # spec says to check against "the tag name of the last start tag to + # have been emitted from this tokenizer", but this is only called from + # the various "raw" states, which I'm pretty sure all push the start + # token onto open_els. TODO: verify this after the script data states + # are implemented + debug_log "#{t.type}, #{t.name} open_els: #{serialize_els open_els, true, true}" + return t.type is TYPE_END_TAG and t.name is open_els[0].name + + # 8.2.4.13 http://www.w3.org/TR/html5/syntax.html#rcdata-end-tag-name-state + tok_state_rcdata_end_tag_name = -> + c = txt.charAt(cur++) + if c is "\t" or c is "\n" or c is "\u000c" or c is ' ' + if is_appropriate_end_tag tok_cur_tag + tok_state = tok_state_before_attribute_name + return + # else fall through to "Anything else" + if c is '/' + if is_appropriate_end_tag tok_cur_tag + tok_state = tok_state_self_closing_start_tag # FIXME spec typo? + return + # else fall through to "Anything else" + if c is '>' + if is_appropriate_end_tag tok_cur_tag + tok_state = tok_state_data + return tok_cur_tag + # else fall through to "Anything else" + if is_uc_alpha(c) + tok_cur_tag.name += c.toLowerCase() + temporary_buffer += c + return null + if is_lc_alpha(c) + tok_cur_tag.name += c + temporary_buffer += c + return null + # Anything else + tok_state = tok_state_rcdata + cur -= 1 # reconsume the input character + return new_character_token ' + c = txt.charAt(cur++) + if c is '/' + temporary_buffer = '' + tok_state = tok_state_rawtext_end_tag_open + return null + # Anything else + tok_state = tok_state_rawtext + cur -= 1 # reconsume the input character + return new_character_token '<' + + # 8.2.4.15 http://www.w3.org/TR/html5/syntax.html#rawtext-end-tag-open-state + tok_state_rawtext_end_tag_open = -> + c = txt.charAt(cur++) + if is_uc_alpha(c) + tok_cur_tag = new_end_tag c.toLowerCase() + temporary_buffer += c + tok_state = tok_state_rawtext_end_tag_name + return null + if is_lc_alpha(c) + tok_cur_tag = new_end_tag c + temporary_buffer += c + tok_state = tok_state_rawtext_end_tag_name + return null + # Anything else + tok_state = tok_state_rawtext + cur -= 1 # reconsume the input character + return new_character_token " + c = txt.charAt(cur++) + if c is "\t" or c is "\n" or c is "\u000c" or c is ' ' + if is_appropriate_end_tag tok_cur_tag + tok_state = tok_state_before_attribute_name + return + # else fall through to "Anything else" + if c is '/' + if is_appropriate_end_tag tok_cur_tag + tok_state = tok_state_self_closing_start_tag + return + # else fall through to "Anything else" + if c is '>' + if is_appropriate_end_tag tok_cur_tag + tok_state = tok_state_data + return tok_cur_tag + # else fall through to "Anything else" + if is_uc_alpha(c) + tok_cur_tag.name += c.toLowerCase() + temporary_buffer += c + return null + if is_lc_alpha(c) + tok_cur_tag.name += c + temporary_buffer += c + return null + # Anything else + tok_state = tok_state_rawtext + cur -= 1 # reconsume the input character + return new_character_token ' + c = txt.charAt(cur++) + if c is '/' + temporary_buffer = '' + tok_state = tok_state_script_data_end_tag_open + return + if c is '!' + tok_state = tok_state_script_data_escape_start + return new_character_token ' + c = txt.charAt(cur++) + if is_uc_alpha(c) + tok_cur_tag = new_end_tag c.toLowerCase() + temporary_buffer += c + tok_state = tok_state_script_data_end_tag_name + return + if is_lc_alpha(c) + tok_cur_tag = new_end_tag c + temporary_buffer += c + tok_state = tok_state_script_data_end_tag_name + return + # Anything else + tok_state = tok_state_script_data + cur -= 1 # Reconsume + return new_character_token ' + c = txt.charAt(cur++) + if c is "\t" or c is "\n" or c is "\u000c" or c is ' ' + if is_appropriate_end_tag tok_cur_tag + tok_state = tok_state_before_attribute_name + return + # fall through + if c is '/' + if is_appropriate_end_tag tok_cur_tag + tok_state = tok_state_self_closing_start_tag + return + # fall through + if is_uc_alpha(c) + tok_cur_tag.name += c.toLowerCase() + temporary_buffer += c + return + if is_lc_alpha(c) + tok_cur_tag.name += c + temporary_buffer += c + return + # Anything else + tok_state = tok_state_script_data + cur -= 1 # Reconsume + return new_character_token " + c = txt.charAt(cur++) + if c is '-' + tok_state = tok_state_script_data_escape_start_dash + return new_character_token '-' + # Anything else + tok_state = tok_state_script_data + cur -= 1 # Reconsume + return + + # 8.2.4.21 http://www.w3.org/TR/html5/syntax.html#script-data-escape-start-dash-state + tok_state_script_data_escape_start_dash = -> + c = txt.charAt(cur++) + if c is '-' + tok_state = tok_state_script_data_escaped_dash_dash + return new_character_token '-' + # Anything else + tok_state = tok_state_script_data + cur -= 1 # Reconsume + return + + # 8.2.4.22 http://www.w3.org/TR/html5/syntax.html#script-data-escaped-state + tok_state_script_data_escaped = -> + c = txt.charAt(cur++) + if c is '-' + tok_state = tok_state_script_data_escaped_dash + return new_character_token '-' + if c is '<' + tok_state = tok_state_script_data_escaped_less_than_sign + return + if c is "\u0000" + parse_error() + return new_character_token "\ufffd" + if c is '' # EOF + tok_state = tok_state_data + parse_error() + cur -= 1 # Reconsume + return + # Anything else + return new_character_token c + + # 8.2.4.23 http://www.w3.org/TR/html5/syntax.html#script-data-escaped-dash-state + tok_state_script_data_escaped_dash = -> + c = txt.charAt(cur++) + if c is '-' + tok_state = tok_state_script_data_escaped_dash_dash + return new_character_token '-' + if c is '<' + tok_state = tok_state_script_data_escaped_less_than_sign + return + if c is "\u0000" + parse_error() + tok_state = tok_state_script_data_escaped + return new_character_token "\ufffd" + if c is '' # EOF + tok_state = tok_state_data + parse_error() + cur -= 1 # Reconsume + return + # Anything else + tok_state = tok_state_script_data_escaped + return new_character_token c + + # 8.2.4.24 http://www.w3.org/TR/html5/syntax.html#script-data-escaped-dash-dash-state + tok_state_script_data_escaped_dash_dash = -> + c = txt.charAt(cur++) + if c is '-' + return new_character_token '-' + if c is '<' + tok_state = tok_state_script_data_escaped_less_than_sign + return + if c is '>' + tok_state = tok_state_script_data + return new_character_token '>' + if c is "\u0000" + parse_error() + tok_state = tok_state_script_data_escaped + return new_character_token "\ufffd" + if c is '' # EOF + parse_error() + tok_state = tok_state_data + cur -= 1 # Reconsume + return + # Anything else + tok_state = tok_state_script_data_escaped + return new_character_token c + + # 8.2.4.25 http://www.w3.org/TR/html5/syntax.html#script-data-escaped-less-than-sign-state + tok_state_script_data_escaped_less_than_sign = -> + c = txt.charAt(cur++) + if c is '/' + temporary_buffer = '' + tok_state = tok_state_script_data_escaped_end_tag_open + return + if is_uc_alpha(c) + temporary_buffer = c.toLowerCase() # yes, really + tok_state = tok_state_script_data_double_escape_start + return new_character_token "<#{c}" # fixfull split + if is_lc_alpha(c) + temporary_buffer = c + tok_state = tok_state_script_data_double_escape_start + return new_character_token "<#{c}" # fixfull split + # Anything else + tok_state = tok_state_script_data_escaped + cur -= 1 # Reconsume + return new_character_token c + + # 8.2.4.26 http://www.w3.org/TR/html5/syntax.html#script-data-escaped-end-tag-open-state + tok_state_script_data_escaped_end_tag_open = -> + c = txt.charAt(cur++) + if is_uc_alpha(c) + tok_cur_tag = new_end_tag c.toLowerCase() + temporary_buffer += c + tok_state = tok_state_script_data_escaped_end_tag_name + return + if is_lc_alpha(c) + tok_cur_tag = new_end_tag c + temporary_buffer += c + tok_state = tok_state_script_data_escaped_end_tag_name + return + # Anything else + tok_state = tok_state_script_data_escaped + cur -= 1 # Reconsume + return new_character_token ' + c = txt.charAt(cur++) + if c is "\t" or c is "\u000a" or c is "\u000c" or c is ' ' + if is_appropriate_end_tag tok_cur_tag + tok_state = tok_state_before_attribute_name + return + # fall through + if c is '/' + if is_appropriate_end_tag tok_cur_tag + tok_state = tok_state_self_closing_start_tag + return + # fall through + if is_uc_alpha(c) + tok_cur_tag.name += c.toLowerCase() + temporary_buffer += c.toLowerCase() + return + if is_lc_alpha(c) + tok_cur_tag.name += c + temporary_buffer += c.toLowerCase() + return + # Anything else + tok_state = tok_state_script_data_escaped + cur -= 1 # Reconsume + return new_character_token " + c = txt.charAt(cur++) + if c is "\t" or c is "\u000a" or c is "\u000c" or c is ' ' or c is '/' or c is '>' + if temporary_buffer is 'script' + tok_state = tok_state_script_data_double_escaped + else + tok_state = tok_state_script_data_escaped + return new_character_token c + if is_uc_alpha(c) + temporary_buffer += c.toLowerCase() # yes, really lowercase + return new_character_token c + if is_lc_alpha(c) + temporary_buffer += c + return new_character_token c + # Anything else + tok_state = tok_state_script_data_escaped + cur -= 1 # Reconsume + return + + # 8.2.4.29 http://www.w3.org/TR/html5/syntax.html#script-data-double-escaped-state + tok_state_script_data_double_escaped = -> + c = txt.charAt(cur++) + if c is '-' + tok_state = tok_state_script_data_double_escaped_dash + return new_character_token '-' + if c is '<' + tok_state = tok_state_script_data_double_escaped_less_than_sign + return new_character_token '<' + if c is "\u0000" + parse_error() + return new_character_token "\ufffd" + if c is '' # EOF + parse_error() + tok_state = tok_state_data + cur -= 1 # Reconsume + return + # Anything else + return new_character_token c + + # 8.2.4.30 http://www.w3.org/TR/html5/syntax.html#script-data-double-escaped-dash-state + tok_state_script_data_double_escaped_dash = -> + c = txt.charAt(cur++) + if c is '-' + tok_state = tok_state_script_data_double_escaped_dash_dash + return new_character_token '-' + if c is '<' + tok_state = tok_state_script_data_double_escaped_less_than_sign + return new_character_token '<' + if c is "\u0000" + parse_error() + tok_state = tok_state_script_data_double_escaped + return new_character_token "\ufffd" + if c is '' # EOF + parse_error() + tok_state = tok_state_data + cur -= 1 # Reconsume + return + # Anything else + tok_state = tok_state_script_data_double_escaped + return new_character_token c + + # 8.2.4.31 http://www.w3.org/TR/html5/syntax.html#script-data-double-escaped-dash-dash-state + tok_state_script_data_double_escaped_dash_dash = -> + c = txt.charAt(cur++) + if c is '-' + return new_character_token '-' + if c is '<' + tok_state = tok_state_script_data_double_escaped_less_than_sign + return new_character_token '<' + if c is '>' + tok_state = tok_state_script_data + return new_character_token '>' + if c is "\u0000" + parse_error() + tok_state = tok_state_script_data_double_escaped + return new_character_token "\ufffd" + if c is '' # EOF + parse_error() + tok_state = tok_state_data + cur -= 1 # Reconsume + return + # Anything else + tok_state = tok_state_script_data_double_escaped + return new_character_token c + + # 8.2.4.32 http://www.w3.org/TR/html5/syntax.html#script-data-double-escaped-less-than-sign-state + tok_state_script_data_double_escaped_less_than_sign = -> + c = txt.charAt(cur++) + if c is '/' + temporary_buffer = '' + tok_state = tok_state_script_data_double_escape_end + return new_character_token '/' + # Anything else + tok_state = tok_state_script_data_double_escaped + cur -= 1 # Reconsume + return + + # 8.2.4.33 http://www.w3.org/TR/html5/syntax.html#script-data-double-escape-end-state + tok_state_script_data_double_escape_end = -> + c = txt.charAt(cur++) + if c is "\t" or c is "\u000a" or c is "\u000c" or c is ' ' or c is '/' or c is '>' + if temporary_buffer is 'script' + tok_state = tok_state_script_data_escaped + else + tok_state = tok_state_script_data_double_escaped + return new_character_token c + if is_uc_alpha(c) + temporary_buffer += c.toLowerCase() # yes, really lowercase + return new_character_token c + if is_lc_alpha(c) + temporary_buffer += c + return new_character_token c + # Anything else + tok_state = tok_state_script_data_double_escaped + cur -= 1 # Reconsume + return # 8.2.4.34 http://www.w3.org/TR/html5/syntax.html#before-attribute-name-state tok_state_before_attribute_name = -> @@ -1587,7 +2853,7 @@ parse_html = (txt, parse_error_cb = null) -> parse_error() tok_state = tok_state_data else - if uc_alpha.indexOf(c) > -1 + if is_uc_alpha(c) attr_name = c.toLowerCase() else attr_name = c @@ -1620,12 +2886,47 @@ parse_html = (txt, parse_error_cb = null) -> parse_error() tok_state = tok_state_data else - if uc_alpha.indexOf(c) > -1 + if is_uc_alpha(c) tok_cur_tag.attrs_a[0][0] = c.toLowerCase() else tok_cur_tag.attrs_a[0][0] += c return null + # 8.2.4.36 http://www.w3.org/TR/html5/syntax.html#after-attribute-name-state + tok_state_after_attribute_name = -> + c = txt.charAt(cur++) + if c is "\t" or c is "\n" or c is "\u000c" or c is ' ' + return + if c is '/' + tok_state = tok_state_self_closing_start_tag + return + if c is '=' + tok_state = tok_state_before_attribute_value + return + if c is '>' + tok_state = tok_state_data + return + if is_uc_alpha(c) + tok_cur_tag.attrs_a.unshift [c.toLowerCase(), ''] + tok_state = tok_state_attribute_name + return + if c is "\u0000" + parse_error() + tok_cur_tag.attrs_a.unshift ["\ufffd", ''] + tok_state = tok_state_attribute_name + return + if c is '' # EOF + parse_error() + tok_state = tok_state_data + cur -= 1 # reconsume + return + if c is '"' or c is "'" or c is '<' + parse_error() + # fall through to Anything else + # Anything else + tok_cur_tag.attrs_a.unshift [c, ''] + tok_state = tok_state_attribute_name + # 8.2.4.37 http://www.w3.org/TR/html5/syntax.html#before-attribute-value-state tok_state_before_attribute_value = -> switch c = txt.charAt(cur++) @@ -1662,7 +2963,7 @@ parse_html = (txt, parse_error_cb = null) -> when '"' tok_state = tok_state_after_attribute_value_quoted when '&' - tok_cur_tag.attrs_a[0][1] += tokenize_character_reference '"', true + tok_cur_tag.attrs_a[0][1] += parse_character_reference '"', true when "\u0000" # Parse error tok_cur_tag.attrs_a[0][1] += "\ufffd" @@ -1679,7 +2980,7 @@ parse_html = (txt, parse_error_cb = null) -> when "'" tok_state = tok_state_after_attribute_value_quoted when '&' - tok_cur_tag.attrs_a[0][1] += tokenize_character_reference "'", true + tok_cur_tag.attrs_a[0][1] += parse_character_reference "'", true when "\u0000" # Parse error tok_cur_tag.attrs_a[0][1] += "\ufffd" @@ -1696,7 +2997,7 @@ parse_html = (txt, parse_error_cb = null) -> when "\t", "\n", "\u000c", ' ' tok_state = tok_state_before_attribute_name when '&' - tok_cur_tag.attrs_a[0][1] += tokenize_character_reference '>', true + tok_cur_tag.attrs_a[0][1] += parse_character_reference '>', true when '>' tok_state = tok_state_data tmp = tok_cur_tag @@ -1733,10 +3034,623 @@ parse_html = (txt, parse_error_cb = null) -> cur -= 1 # we didn't handle that char return null + # 8.2.4.43 http://www.w3.org/TR/html5/syntax.html#self-closing-start-tag-state + tok_state_self_closing_start_tag = -> + c = txt.charAt(cur++) + if c is '>' + tok_cur_tag.flag 'self-closing' + tok_state = tok_state_data + return tok_cur_tag + if c is '' + parse_error() + tok_state = tok_state_data + cur -= 1 # Reconsume + return + # Anything else + parse_error() + tok_state = tok_state_before_attribute_name + cur -= 1 # Reconsume + return + + # 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 txt.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 is_uc_alpha(c) + 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 is_uc_alpha(c) + 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) - tokenize_character_reference = (allowed_char = null, in_attr = false) -> + parse_character_reference = (allowed_char = null, in_attr = false) -> if cur >= txt.length return '&' switch c = txt.charAt(cur) @@ -1811,14 +3725,22 @@ parse_html = (txt, parse_error_cb = null) -> # 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] - insertion_mode = ins_mode_in_body + doc = new Node TYPE_TAG, name: 'html', namespace: NS_HTML + open_els = [] + afe = [] # active formatting elements + template_insertion_modes = [] + insertion_mode = ins_mode_initial + original_insertion_mode = insertion_mode # TODO check spec + flag_scripting = true # TODO might need an extra flag to get