X-Git-Url: https://jasonwoof.com/gitweb/?a=blobdiff_plain;f=parse-html.coffee;h=2b8fda76e1a2488f8c4f082003e6d54f6d72c038;hb=cbb22f2cf485ae661e0b6649b1a751c43d2ded3e;hp=86d0136257d03a06aa195ca0b1ad4512dc792fda;hpb=12ed8923d45bd72d534bfadb0dc5a644c97f3bc8;p=peach-html5-editor.git diff --git a/parse-html.coffee b/parse-html.coffee index 86d0136..2b8fda7 100644 --- a/parse-html.coffee +++ b/parse-html.coffee @@ -16,13 +16,17 @@ # This file implements a parser for html snippets, meant to be used by a -# WYSIWYG editor. Hence it does not attempt to parse doctypes, , -# or tags, nor does it produce the top level "document" node in the dom -# tree, nor nodes for html, head or body. Comments containing "fixfull" -# indicate places where additional code is needed for full HTML document -# parsing. +# WYSIWYG editor. + +# The implementation is a pretty direct implementation of the parsing algorithm +# described here: +# http://www.w3.org/TR/html5/syntax.html#preprocessing-the-input-stream +# +# Deviations from that spec: # -# Instead, the data structure produced by this parser is an array of Nodes. +# Purposeful: search this file for "WTAG" +# +# Not finished yet: search this file for "fixfull", "TODO" and "FIXME" # stacks/lists @@ -47,6 +51,21 @@ # 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 + +from_code_point = (x) -> + if String.fromCodePoint? + return String.fromCodePoint x + else + if x <= 0xffff + return String.fromCharCode x + x -= 0x10000 + return String.fromCharCode((x >> 10) + 0xd800, (x % 0x400) + 0xdc00) # Each node is an obect of the Node class. Here are the Node types: TYPE_TAG = 0 # name, {attributes}, [children] @@ -85,15 +104,22 @@ class Node @children = args.children ? [] @namespace = args.namespace ? NS_HTML @parent = args.parent ? null + @token = args.token ? null + @flags = args.flags ? {} if args.id? @id = "#{args.id}+" else @id = "#{++prev_node_id}" - shallow_clone: -> # return a new node that's the same except without the children or parent - # 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 + acknowledge_self_closing: -> + if @token? + @token.flag 'did_self_close', true + else + @flag 'did_self_close', true + flag: (key, value = null) -> + if value? + @flags[key] = value + else + return @flags[key] serialize: (shallow = false, show_ids = false) -> # for unit tests ret = '' switch @type @@ -129,8 +155,7 @@ class Node ret += 'comment:' ret += JSON.stringify @text when TYPE_DOCTYPE - ret += 'doctype' - # FIXME + ret += "doctype:#{@name},#{JSON.stringify(@public_identifier ? '')},#{JSON.stringify(@system_identifier ? '')}" when TYPE_AFE_MARKER ret += 'marker' when TYPE_AAA_BOOKMARK @@ -149,8 +174,11 @@ 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 = -> @@ -158,21 +186,69 @@ 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 false unless t.type is TYPE_START_TAG + for a in 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" +unicode_fixes = {} +unicode_fixes[0x00] = "\uFFFD" +unicode_fixes[0x80] = "\u20AC" +unicode_fixes[0x82] = "\u201A" +unicode_fixes[0x83] = "\u0192" +unicode_fixes[0x84] = "\u201E" +unicode_fixes[0x85] = "\u2026" +unicode_fixes[0x86] = "\u2020" +unicode_fixes[0x87] = "\u2021" +unicode_fixes[0x88] = "\u02C6" +unicode_fixes[0x89] = "\u2030" +unicode_fixes[0x8A] = "\u0160" +unicode_fixes[0x8B] = "\u2039" +unicode_fixes[0x8C] = "\u0152" +unicode_fixes[0x8E] = "\u017D" +unicode_fixes[0x91] = "\u2018" +unicode_fixes[0x92] = "\u2019" +unicode_fixes[0x93] = "\u201C" +unicode_fixes[0x94] = "\u201D" +unicode_fixes[0x95] = "\u2022" +unicode_fixes[0x96] = "\u2013" +unicode_fixes[0x97] = "\u2014" +unicode_fixes[0x98] = "\u02DC" +unicode_fixes[0x99] = "\u2122" +unicode_fixes[0x9A] = "\u0161" +unicode_fixes[0x9B] = "\u203A" +unicode_fixes[0x9C] = "\u0153" +unicode_fixes[0x9E] = "\u017E" +unicode_fixes[0x9F] = "\u0178" + # These are the character references that don't need a terminating semicolon # min length: 2, max: 6, none are a prefix of any other. legacy_char_refs = { @@ -264,14 +340,17 @@ special_elements = { h2:NS_HTML, h3:NS_HTML, h4:NS_HTML, h5:NS_HTML, h6:NS_HTML, head:NS_HTML, header:NS_HTML, hgroup:NS_HTML, hr:NS_HTML, html:NS_HTML, iframe:NS_HTML, img:NS_HTML, input:NS_HTML, isindex:NS_HTML, li:NS_HTML, link:NS_HTML, - listing:NS_HTML, main:NS_HTML, marquee:NS_HTML, meta:NS_HTML, nav:NS_HTML, - noembed:NS_HTML, noframes:NS_HTML, noscript:NS_HTML, object:NS_HTML, - ol:NS_HTML, p:NS_HTML, param:NS_HTML, plaintext:NS_HTML, pre:NS_HTML, - script:NS_HTML, section:NS_HTML, select:NS_HTML, source:NS_HTML, - style:NS_HTML, summary:NS_HTML, table:NS_HTML, tbody:NS_HTML, td:NS_HTML, - template:NS_HTML, textarea:NS_HTML, tfoot:NS_HTML, th:NS_HTML, - thead:NS_HTML, title:NS_HTML, tr:NS_HTML, track:NS_HTML, ul:NS_HTML, - wbr:NS_HTML, xmp:NS_HTML, + listing:NS_HTML, main:NS_HTML, marquee:NS_HTML, + + menu:NS_HTML,menuitem:NS_HTML, # WATWG adds these + + meta:NS_HTML, nav:NS_HTML, noembed:NS_HTML, noframes:NS_HTML, + noscript:NS_HTML, object:NS_HTML, ol:NS_HTML, p:NS_HTML, param:NS_HTML, + plaintext:NS_HTML, pre:NS_HTML, script:NS_HTML, section:NS_HTML, + select:NS_HTML, source:NS_HTML, style:NS_HTML, summary:NS_HTML, + table:NS_HTML, tbody:NS_HTML, td:NS_HTML, template:NS_HTML, + textarea:NS_HTML, tfoot:NS_HTML, th:NS_HTML, thead:NS_HTML, title:NS_HTML, + tr:NS_HTML, track:NS_HTML, ul:NS_HTML, wbr:NS_HTML, xmp:NS_HTML, # MathML: mi:NS_MATHML, mo:NS_MATHML, mn:NS_MATHML, ms:NS_MATHML, mtext:NS_MATHML, @@ -287,32 +366,190 @@ formatting_elements = { u: true } +mathml_text_integration = { + mi: NS_MATHML, mo: NS_MATHML, mn: NS_MATHML, ms: NS_MATHML, mtext: NS_MATHML +} +is_mathml_text_integration_point = (el) -> + return mathml_text_integration[el.name] is el.namespace +is_html_integration = (el) -> # DON'T PASS A TOKEN + if el.namespace is NS_MATHML + if el.name is 'annotation-xml' + if el.attrs.encoding? + if el.attrs.encoding.toLowerCase() is 'text/html' + return true + if el.attrs.encoding.toLowerCase() is 'application/xhtml+xml' + return true + return false + if el.namespace is NS_SVG + if el.name is 'foreignObject' or el.name is 'desc' or el.name is 'title' + return true + return false + +h_tags = { + h1:NS_HTML, h2:NS_HTML, h3:NS_HTML, h4:NS_HTML, h5:NS_HTML, h6:NS_HTML +} + foster_parenting_targets = { - table: true - tbody: true - tfoot: true - thead: true - tr: true + table: NS_HTML + tbody: NS_HTML + tfoot: NS_HTML + thead: NS_HTML + tr: NS_HTML } -# all html I presume end_tag_implied = { - dd: true - dt: true - li: true - option: true - optgroup: true - p: true - rb: true - rp: true - rt: true - rtc: true + dd: NS_HTML + dt: NS_HTML + li: NS_HTML + option: NS_HTML + optgroup: NS_HTML + p: NS_HTML + rb: NS_HTML + rp: NS_HTML + rt: NS_HTML + rtc: NS_HTML } 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 + +adp_els = { address: NS_HTML, div: NS_HTML, p: NS_HTML } +el_is_special_not_adp = (el) -> + return special_elements[el.name] is el.namespace and adp_els[el.name] isnt el.namespace + +svg_name_fixes = { + altglyph: 'altGlyph' + altglyphdef: 'altGlyphDef' + altglyphitem: 'altGlyphItem' + animatecolor: 'animateColor' + animatemotion: 'animateMotion' + animatetransform: 'animateTransform' + clippath: 'clipPath' + feblend: 'feBlend' + fecolormatrix: 'feColorMatrix' + fecomponenttransfer: 'feComponentTransfer' + fecomposite: 'feComposite' + feconvolvematrix: 'feConvolveMatrix' + fediffuselighting: 'feDiffuseLighting' + fedisplacementmap: 'feDisplacementMap' + fedistantlight: 'feDistantLight' + fedropshadow: 'feDropShadow' + feflood: 'feFlood' + fefunca: 'feFuncA' + fefuncb: 'feFuncB' + fefuncg: 'feFuncG' + fefuncr: 'feFuncR' + fegaussianblur: 'feGaussianBlur' + feimage: 'feImage' + femerge: 'feMerge' + femergenode: 'feMergeNode' + femorphology: 'feMorphology' + feoffset: 'feOffset' + fepointlight: 'fePointLight' + fespecularlighting: 'feSpecularLighting' + fespotlight: 'feSpotLight' + fetile: 'feTile' + feturbulence: 'feTurbulence' + foreignobject: 'foreignObject' + glyphref: 'glyphRef' + lineargradient: 'linearGradient' + radialgradient: 'radialGradient' + textpath: 'textPath' +} +svg_attribute_fixes = { + attributename: 'attributeName' + attributetype: 'attributeType' + basefrequency: 'baseFrequency' + baseprofile: 'baseProfile' + calcmode: 'calcMode' + clippathunits: 'clipPathUnits' + contentscripttype: 'contentScriptType' + contentstyletype: 'contentStyleType' + diffuseconstant: 'diffuseConstant' + edgemode: 'edgeMode' + externalresourcesrequired: 'externalResourcesRequired' + filterres: 'filterRes' + filterunits: 'filterUnits' + glyphref: 'glyphRef' + gradienttransform: 'gradientTransform' + gradientunits: 'gradientUnits' + kernelmatrix: 'kernelMatrix' + kernelunitlength: 'kernelUnitLength' + keypoints: 'keyPoints' + keysplines: 'keySplines' + keytimes: 'keyTimes' + lengthadjust: 'lengthAdjust' + limitingconeangle: 'limitingConeAngle' + markerheight: 'markerHeight' + markerunits: 'markerUnits' + markerwidth: 'markerWidth' + maskcontentunits: 'maskContentUnits' + maskunits: 'maskUnits' + numoctaves: 'numOctaves' + pathlength: 'pathLength' + patterncontentunits: 'patternContentUnits' + patterntransform: 'patternTransform' + patternunits: 'patternUnits' + pointsatx: 'pointsAtX' + pointsaty: 'pointsAtY' + pointsatz: 'pointsAtZ' + preservealpha: 'preserveAlpha' + preserveaspectratio: 'preserveAspectRatio' + primitiveunits: 'primitiveUnits' + refx: 'refX' + refy: 'refY' + repeatcount: 'repeatCount' + repeatdur: 'repeatDur' + requiredextensions: 'requiredExtensions' + requiredfeatures: 'requiredFeatures' + specularconstant: 'specularConstant' + specularexponent: 'specularExponent' + spreadmethod: 'spreadMethod' + startoffset: 'startOffset' + stddeviation: 'stdDeviation' + stitchtiles: 'stitchTiles' + surfacescale: 'surfaceScale' + systemlanguage: 'systemLanguage' + tablevalues: 'tableValues' + targetx: 'targetX' + targety: 'targetY' + textlength: 'textLength' + viewbox: 'viewBox' + viewtarget: 'viewTarget' + xchannelselector: 'xChannelSelector' + ychannelselector: 'yChannelSelector' + zoomandpan: 'zoomAndPan' +} +foreign_attr_fixes = { + 'xlink:actuate': 'xlink actuate' + 'xlink:arcrole': 'xlink arcrole' + 'xlink:href': 'xlink href' + 'xlink:role': 'xlink role' + 'xlink:show': 'xlink show' + 'xlink:title': 'xlink title' + 'xlink:type': 'xlink type' + 'xml:base': 'xml base' + 'xml:lang': 'xml lang' + 'xml:space': 'xml space' + 'xmlns': 'xmlns' + 'xmlns:xlink': 'xmlns xlink' +} +adjust_mathml_attributes = (t) -> + for a in t.attrs_a + if a[0] is 'definitionurl' + a[0] = 'definitionURL' + return +adjust_svg_attributes = (t) -> + for a in t.attrs_a + if svg_attribute_fixes[a[0]]? + a[0] = svg_attribute_fixes[a[0]] + return +adjust_foreign_attributes = (t) -> + # fixfull + for a in t.attrs_a + if foreign_attr_fixes[a[0]]? + a[0] = foreign_attr_fixes[a[0]] + return # decode_named_char_ref() # @@ -336,23 +573,35 @@ decode_named_char_ref = (txt) -> return null if decoded is txt return g_dncr.cache[txt] = decoded -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 - insertion_mode = null +parse_html = (args) -> + txt = null + cur = null # index of next char in txt to be parsed + # 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_ins_modes = null + ins_mode = null + original_ins_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? - parse_error_cb cur + if args.error_cb? + args.error_cb cur else console.log "Parse error at character #{cur} of #{txt.length}" @@ -378,7 +627,7 @@ parse_html = (txt, parse_error_cb = null) -> # But first... the helpers template_tag_is_open = -> for t in open_els - if t.name is 'template' # maybe should also check: and t.namespace is 'html' + if t.name is 'template' and t.namespace is NS_HTML return true return false is_in_scope_x = (tag_name, scope, namespace) -> @@ -397,13 +646,13 @@ parse_html = (txt, parse_error_cb = null) -> if scope2[t.name] is t.namespace return false return false - standard_scopers = { # FIXME these are supposed to be namespace specific + standard_scopers = { applet: NS_HTML, caption: NS_HTML, html: NS_HTML, table: NS_HTML, td: NS_HTML, th: NS_HTML, marquee: NS_HTML, object: NS_HTML, - template: NS_HTML, mi: NS_MATHML, + template: NS_HTML, - mo: NS_MATHML, mn: NS_MATHML, ms: NS_MATHML, mtext: NS_MATHML, - 'annotation-xml': NS_MATHML, + mi: NS_MATHML, mo: NS_MATHML, mn: NS_MATHML, ms: NS_MATHML, + mtext: NS_MATHML, 'annotation-xml': NS_MATHML, foreignObject: NS_SVG, desc: NS_SVG, title: NS_SVG } @@ -416,25 +665,72 @@ parse_html = (txt, parse_error_cb = null) -> return is_in_scope_x_y tag_name, standard_scopers, button_scopers, namespace is_in_table_scope = (tag_name, namespace = null) -> return is_in_scope_x tag_name, table_scopers, namespace + # aka is_in_list_item_scope + is_in_li_scope = (tag_name, namespace = null) -> + return is_in_scope_x_y tag_name, standard_scopers, li_scopers, namespace is_in_select_scope = (tag_name, namespace = 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.namespace 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 - el_is_in_scope = (el) -> - for t in open_els - if t is el + # this requires a namespace match + el_is_in_scope = (needle) -> + for el in open_els + if el is needle return true - if standard_scopers[t.name] is t.namespace + if standard_scopers[el.name] is el.namespace 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: NS_HTML + tfoot: NS_HTML + thead: NS_HTML + template: NS_HTML + html: NS_HTML + } + clear_stack_to_table_body_context = -> + loop + if clear_to_table_body_stopers[open_els[0].name] is open_els[0].namespace + 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 = -> + reset_ins_mode = -> # 1. Let last be false. last = false # 2. Let node be the last node in the stack of open elements. @@ -450,7 +746,7 @@ parse_html = (txt, parse_error_cb = null) -> # fixfull (fragment case) # 4. If node is a select element, run these substeps: - if node.name is 'select' + if node.name is 'select' and node.namespace is NS_HTML # 1. If last is true, jump to the step below labeled done. unless last # 2. Let ancestor be node. @@ -467,87 +763,90 @@ parse_html = (txt, parse_error_cb = null) -> ancestor = open_els[ancestor_i] # 5. If ancestor is a template node, jump to the step below # labeled done. - if ancestor.name is 'template' + if ancestor.name is 'template' and ancestor.namespace is NS_HTML break # 6. If ancestor is a table node, switch the insertion mode # to "in select in table" and abort these steps. - if ancestor.name is 'table' - insertion_mode = ins_mode_in_select_in_table + if ancestor.name is 'table' and ancestor.namespace is NS_HTML + ins_mode = ins_mode_in_select_in_table return # 7. Jump back to the step labeled loop. # 8. Done: Switch the insertion mode to "in select" and abort # these steps. - insertion_mode = ins_mode_in_select + ins_mode = ins_mode_in_select return # 5. If node is a td or th element and last is false, then switch # the insertion mode to "in cell" and abort these steps. - if (node.name is 'td' or node.name is 'th') and last is false - insertion_mode = ins_mode_in_cell + if (node.name is 'td' or node.name is 'th') and node.namespace is NS_HTML and last is false + ins_mode = ins_mode_in_cell return # 6. If node is a tr element, then switch the insertion mode to "in # row" and abort these steps. - if node.name is 'tr' - insertion_mode = ins_mode_in_row + if node.name is 'tr' and node.namespace is NS_HTML + ins_mode = ins_mode_in_row return # 7. If node is a tbody, thead, or tfoot element, then switch the # insertion mode to "in table body" and abort these steps. - if node.name is 'tbody' or node.name is 'thead' or node.name is 'tfoot' - insertion_mode = ins_mode_in_table_body + if (node.name is 'tbody' or node.name is 'thead' or node.name is 'tfoot') and node.namespace is NS_HTML + ins_mode = ins_mode_in_table_body return # 8. If node is a caption element, then switch the insertion mode # to "in caption" and abort these steps. - if node.name is 'caption' - insertion_mode = ins_mode_in_caption + if node.name is 'caption' and node.namespace is NS_HTML + ins_mode = ins_mode_in_caption return # 9. If node is a colgroup element, then switch the insertion mode # to "in column group" and abort these steps. - if node.name is 'colgroup' - insertion_mode = ins_mode_in_column_group + if node.name is 'colgroup' and node.namespace is NS_HTML + ins_mode = ins_mode_in_column_group return # 10. If node is a table element, then switch the insertion mode to # "in table" and abort these steps. - if node.name is 'table' - insertion_mode = ins_mode_in_table + if node.name is 'table' and node.namespace is NS_HTML + ins_mode = ins_mode_in_table return # 11. If node is a template element, then switch the insertion mode # to the current template insertion mode and abort these steps. - # fixfull (template insertion mode stack) - + if node.name is 'template' and node.namespace is NS_HTML + ins_mode = template_ins_modes[0] + return # 12. If node is a head element and last is true, then switch the # insertion mode to "in body" ("in body"! not "in head"!) and abort # these steps. (fragment case) - if node.name is 'head' and last - insertion_mode = ins_mode_in_body + if node.name is 'head' and node.namespace is NS_HTML and last + ins_mode = ins_mode_in_body return # 13. If node is a head element and last is false, then switch the # insertion mode to "in head" and abort these steps. - if node.name is 'head' and last is false - insertion_mode = ins_mode_in_head + if node.name is 'head' and node.namespace is NS_HTML and last is false + ins_mode = ins_mode_in_head return # 14. If node is a body element, then switch the insertion mode to # "in body" and abort these steps. - if node.name is 'body' - insertion_mode = ins_mode_in_body + if node.name is 'body' and node.namespace is NS_HTML + ins_mode = ins_mode_in_body return # 15. If node is a frameset element, then switch the insertion mode # to "in frameset" and abort these steps. (fragment case) - if node.name is 'frameset' - insertion_mode = ins_mode_in_frameset + if node.name is 'frameset' and node.namespace is NS_HTML + ins_mode = ins_mode_in_frameset return # 16. If node is an html element, run these substeps: - if node.name is 'html' + if node.name is 'html' and node.namespace is NS_HTML # 1. If the head element pointer is null, switch the insertion # mode to "before head" and abort these steps. (fragment case) - # fixfull (fragment case) - - # 2. Otherwise, the head element pointer is not null, switch - # the insertion mode to "after head" and abort these steps. - insertion_mode = ins_mode_in_body # FIXME fixfull + if head_element_pointer is null + ins_mode = ins_mode_before_head + else + # 2. Otherwise, the head element pointer is not null, + # switch the insertion mode to "after head" and abort these + # steps. + ins_mode = ins_mode_after_head return # 17. If last is true, then switch the insertion mode to "in body" # and abort these steps. (fragment case) if last - insertion_mode = ins_mode_in_body + ins_mode = ins_mode_in_body return # 18. Let node now be the node before node in the stack of open # elements. @@ -555,10 +854,18 @@ 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. - reconstruct_active_formatting_elements = -> + reconstruct_afe = -> return if afe.length is 0 if afe[0].type is TYPE_AFE_MARKER or afe[0] in open_els return @@ -573,8 +880,7 @@ parse_html = (txt, parse_error_cb = null) -> break # Create loop - el = afe[i].shallow_clone() - tree_insert_element el + el = insert_html_element afe[i].token afe[i] = el break if i is 0 i -= 1 # Advance @@ -587,10 +893,10 @@ 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 + if open_els[0].name is subject and open_els[0].namespace is NS_HTML el = open_els[0] open_els.shift() # remove it from the list of active formatting elements (if found) @@ -699,7 +1005,7 @@ 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}" @@ -745,7 +1051,7 @@ parse_html = (txt, parse_error_cb = null) -> # element, replace the entry for node in the stack of open # elements with an entry for the new element, and let node be # the new element. - new_node = node.shallow_clone() + new_node = token_to_element node.token, NS_HTML, ca for t, i in afe if t is node afe[i] = new_node @@ -793,7 +1099,7 @@ 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) @@ -810,7 +1116,7 @@ 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" @@ -826,12 +1132,12 @@ 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 # intended parent. - new_element = fe.shallow_clone() # FIXME intended parent thing + new_element = token_to_element fe.token, NS_HTML, fb # 16. Take all of the child nodes of furthest block and append them # to the element created in the last step. while fb.children.length @@ -866,27 +1172,27 @@ 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' + unless open_els[0].name is 'p' and open_els[0].namespace is NS_HTML parse_error() while open_els.length > 1 # just in case el = open_els.shift() - if el.name is 'p' + if el.name is 'p' and el.namespace is NS_HTML return close_p_if_in_button_scope = -> - if is_in_button_scope 'p' + if is_in_button_scope 'p', NS_HTML 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 @@ -896,6 +1202,36 @@ parse_html = (txt, parse_error_cb = null) -> return dest[0].children.splice dest[1], 0, t + + # 8.2.5 http://www.w3.org/TR/html5/syntax.html#tree-construction + process_token = (t) -> + acn = adjusted_current_node() + unless acn? + ins_mode t + return + if acn.namespace is NS_HTML + ins_mode t + return + if is_mathml_text_integration_point(acn) + if t.type is TYPE_START_TAG and not (t.name is 'mglyph' or t.name is 'malignmark') + ins_mode t + return + if t.type is TYPE_TEXT + ins_mode t + return + if acn.namespace is NS_MATHML and acn.name is 'annotation-xml' and t.type is TYPE_START_TAG and t.name is 'svg' + ins_mode t + return + if is_html_integration acn + if t.type is TYPE_START_TAG or t.type is TYPE_TEXT + ins_mode t + return + if t.type is TYPE_EOF + ins_mode t + return + in_foreign_content t + return + # 8.2.5.1 # http://www.w3.org/TR/html5/syntax.html#creating-and-inserting-nodes # http://www.w3.org/TR/html5/syntax.html#appropriate-place-for-inserting-a-node @@ -912,14 +1248,14 @@ parse_html = (txt, parse_error_cb = null) -> # If foster parenting is enabled and target is a table, tbody, tfoot, # thead, or tr element Foster parenting happens when content is # misnested in tables. - if flag_foster_parenting and foster_parenting_targets[target.name] + if flag_foster_parenting and foster_parenting_targets[target.name] is target.namespace loop # once. this is here so we can ``break`` to "abort these substeps" # 1. Let last template be the last template element in the # stack of open elements, if any. last_template = null last_template_i = null for el, i in open_els - if el.name is 'template' + if el.name is 'template' and el.namespace is NS_HTML last_template = el last_template_i = i break @@ -928,7 +1264,7 @@ parse_html = (txt, parse_error_cb = null) -> last_table = null last_table_i for el, i in open_els - if el.name is 'table' + if el.name is 'table' and el.namespace is NS_HTML last_table = el last_table_i = i break @@ -939,7 +1275,7 @@ 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 @@ -950,6 +1286,7 @@ parse_html = (txt, parse_error_cb = null) -> # this is odd target = open_els[open_els.length - 1] target_i = target.children.length + break # 5. If last table has a parent element, then let adjusted # insertion location be inside last table's parent element, # immediately before last table, and abort these substeps. @@ -991,13 +1328,11 @@ parse_html = (txt, parse_error_cb = null) -> # http://www.w3.org/TR/html5/syntax.html#create-an-element-for-the-token # aka create_an_element_for_token token_to_element = (t, namespace, intended_parent) -> - t.type = TYPE_TAG # not TYPE_START_TAG # convert attributes into a hash attrs = {} - while t.attrs_a.length - a = t.attrs_a.pop() + for a in t.attrs_a 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 @@ -1021,251 +1356,881 @@ parse_html = (txt, parse_error_cb = null) -> open_els.unshift el return el # http://www.w3.org/TR/html5/syntax.html#insert-an-html-element - insert_html_element = insert_foreign_element # (token, namespace) -> - - # FIXME read implement "foster parenting" part - # FIXME read spec, do this right - # FIXME implement the override target thing - # note: this assumes it's an open tag - # FIXME what part of the spec is this? - # TODO look through all callers of this, and see what they should really be doing. - # eg probably insert_html_element for tokens - tree_insert_element = (el, override_target = null, namespace = null) -> - if namespace? - el.namespace = namespace - dest = adjusted_insertion_location override_target - if el.type is TYPE_START_TAG # means it's a "token" - el = token_to_element el, namespace, dest[0] - unless el.namespace? - namespace = dest.namespace - # fixfull: Document nodes sometimes can't accept more chidren - dest[0].children.splice dest[1], 0, el - el.parent = dest[0] - open_els.unshift el - return el + insert_html_element = (token) -> + insert_foreign_element token, NS_HTML # 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_ins_mode = ins_mode + ins_mode = ins_mode_text + parse_generic_rcdata_text = (t) -> + insert_html_element t + tok_state = tok_state_rcdata + original_ins_mode = ins_mode + ins_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 + while end_tag_implied[open_els[0].name] is open_els[0].namespace 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 + doc.children.push t + return + if t.type is TYPE_DOCTYPE + # FIXME check identifiers, set quirks, etc + # fixfull + doc.children.push t + ins_mode = ins_mode_before_html + return + # Anything else + #fixfull (iframe, quirks) + ins_mode = ins_mode_before_html + process_token t + 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) + ins_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 + ins_mode = ins_mode_before_head + process_token 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 + ins_mode = ins_mode_in_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 below + else + parse_error() + return + # Anything else + head_tok = new_open_tag 'head' + el = insert_html_element head_tok + head_element_pointer = el + ins_mode = ins_mode_in_head + process_token t + + # 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 + ins_mode = ins_mode_after_head + process_token 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 + ins_mode = ins_mode_in_head_noscript + 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 + # fixfull frament case + ail[0].children.splice ail[1], 0, el + open_els.unshift el + tok_state = tok_state_script_data + original_ins_mode = ins_mode # make sure orig... is defined + ins_mode = ins_mode_text + return + if t.type is TYPE_END_TAG and t.name is 'head' + open_els.shift() # will be a head element... spec says so + ins_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 + ins_mode = ins_mode_in_template + template_ins_modes.unshift ins_mode_in_template + 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' and el.namespace is NS_HTML + break + clear_afe_to_marker() + template_ins_modes.shift() + reset_ins_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_else = (t) -> + parse_error() + open_els.shift() + ins_mode = ins_mode_in_head + process_token t + ins_mode_in_head_noscript = (t) -> + 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 'noscript' open_els.shift() + ins_mode = ins_mode_in_head + return + if is_space_tok(t) or t.type is TYPE_COMMENT or (t.type is TYPE_START_TAG and (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 'style')) + ins_mode_in_head t + return + if t.type is TYPE_END_TAG and t.name is 'br' + ins_mode_in_head_noscript_else t + return + if (t.type is TYPE_START_TAG and (t.name is 'head' or t.name is 'noscript')) or t.type is TYPE_END_TAG + parse_error() + return + # Anything else + ins_mode_in_head_noscript_else t + return + + + + # 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 + ins_mode = ins_mode_in_body + process_token t + 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 + ins_mode = ins_mode_in_body + return + if t.type is TYPE_START_TAG and t.name is 'frameset' + insert_html_element t + ins_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 http://www.w3.org/TR/html5/syntax.html#parsing-main-inbody + # 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 + for el, i in open_els + if el.name is name and el.namespace is NS_HTML generate_implied_end_tags name # arg is exception parse_error() unless i is 0 while i >= 0 open_els.shift() i -= 1 return - if special_elements[node.name]? # FIXME check namespac too + if special_elements[el.name] is el.namespace parse_error() return + return ins_mode_in_body = (t) -> - switch t.type - when TYPE_TEXT - switch t.text - when "\u0000" - parse_error() - when "\t", "\u000a", "\u000c", "\u000d", ' ' - reconstruct_active_formatting_elements() - tree_insert_text t - else - reconstruct_active_formatting_elements() - tree_insert_text t - flag_frameset_ok = false - when TYPE_COMMENT - tree_insert_a_comment t - when TYPE_DOCTYPE + if t.type is TYPE_TEXT and t.text is "\u0000" + parse_error() + return + if is_space_tok t + reconstruct_afe() + insert_character t + return + if t.type is TYPE_TEXT + reconstruct_afe() + insert_character t + flag_frameset_ok = false + 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' + parse_error() + return if template_tag_is_open() + root_attrs = open_els[open_els.length - 1].attrs + for a of t.attrs_a + root_attrs[a[0]] = a[1] unless root_attrs[a[0]]? + 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 'body' + parse_error() + return if open_els.length < 2 + second = open_els[open_els.length - 2] + return unless second.namespace is NS_HTML + return unless second.name is 'body' + return if template_tag_is_open() + flag_frameset_ok = false + for a of t.attrs_a + second.attrs[a[0]] = a[1] unless second.attrs[a[0]]? + return + if t.type is TYPE_START_TAG and t.name is 'frameset' + parse_error() + return if open_els.length < 2 + second_i = open_els.length - 2 + second = open_els[second_i] + return unless second.namespace is NS_HTML + return unless second.name is 'body' + if flag_frameset_ok is false + return + if second.parent? + for el, i in second.parent.children + if el is second + second.parent.children.splice i, 1 + break + open_els.splice second_i, 1 + # pop everything except the "root html element" + while open_els.length > 1 + open_els.shift() + insert_html_element t + ins_mode = ins_mode_in_frameset + return + if t.type is TYPE_EOF + ok_tags = { + dd:NS_HTML, dt:NS_HTML, li:NS_HTML, p:NS_HTML, tbody:NS_HTML, + td:NS_HTML, tfoot:NS_HTML, th:NS_HTML, thead:NS_HTML, + tr:NS_HTML, body:NS_HTML, html:NS_HTML, + } + for el in open_els + unless ok_tags[t.name] is el.namespace + parse_error() + break + if template_ins_modes.length > 0 + ins_mode_in_template t + else + stop_parsing() + return + if t.type is TYPE_END_TAG and t.name is 'body' + unless is_in_scope 'body', NS_HTML parse_error() - when TYPE_START_TAG - switch t.name - when 'html' - parse_error() - return if template_tag_is_open() - root_attrs = open_els[open_els.length - 1].attrs - for k, v of t.attrs - 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 - when 'body' + return + ok_tags = { + dd:NS_HTML, dt:NS_HTML, li:NS_HTML, optgroup:NS_HTML, + option:NS_HTML, p:NS_HTML, rb:NS_HTML, rp:NS_HTML, rt:NS_HTML, + rtc:NS_HTML, tbody:NS_HTML, td:NS_HTML, tfoot:NS_HTML, + th:NS_HTML, thead:NS_HTML, tr:NS_HTML, body:NS_HTML, + html:NS_HTML + } + for el in open_els + unless ok_tags[t.name] is el.namespace + parse_error() + break + ins_mode = ins_mode_after_body + return + if t.type is TYPE_END_TAG and t.name is 'html' + unless is_in_scope 'body', NS_HTML + parse_error() + return + ok_tags = { + dd:NS_HTML, dt:NS_HTML, li:NS_HTML, optgroup:NS_HTML, + option:NS_HTML, p:NS_HTML, rb:NS_HTML, rp:NS_HTML, rt:NS_HTML, + rtc:NS_HTML, tbody:NS_HTML, td:NS_HTML, tfoot:NS_HTML, + th:NS_HTML, thead:NS_HTML, tr:NS_HTML, body:NS_HTML, + html:NS_HTML + } + for el in open_els + unless ok_tags[t.name] is el.namespace + parse_error() + break + ins_mode = ins_mode_after_body + process_token t + return + if t.type is TYPE_START_TAG and (t.name is 'address' or t.name is 'article' or t.name is 'aside' or t.name is 'blockquote' or t.name is 'center' or t.name is 'details' or t.name is 'dialog' or t.name is 'dir' or t.name is 'div' or t.name is 'dl' or t.name is 'fieldset' or t.name is 'figcaption' or t.name is 'figure' or t.name is 'footer' or t.name is 'header' or t.name is 'hgroup' or t.name is 'main' or t.name is 'nav' or t.name is 'ol' or t.name is 'p' or t.name is 'section' or t.name is 'summary' or t.name is 'ul') + close_p_if_in_button_scope() + insert_html_element t + return + if t.type is TYPE_START_TAG and h_tags[t.name]? + close_p_if_in_button_scope() + if h_tags[open_els[0].name] is open_els[0].namespace + parse_error() + open_els.shift() + insert_html_element t + return + if t.type is TYPE_START_TAG and (t.name is 'pre' or t.name is 'listing') + close_p_if_in_button_scope() + insert_html_element t + # spec: If the next token is a "LF" (U+000A) character token, then + # ignore that token and move on to the next one. (Newlines at the + # start of pre blocks are ignored as an authoring convenience.) + if txt.charAt(cur) is "\u000a" # FIXME check for crlf? + cur += 1 + flag_frameset_ok = false + return + if t.type is TYPE_START_TAG and t.name is 'form' + unless form_element_pointer is null or template_tag_is_open() + parse_error() + return + close_p_if_in_button_scope() + el = insert_html_element t + unless template_tag_is_open() + form_element_pointer = el + return + if t.type is TYPE_START_TAG and t.name is 'li' + flag_frameset_ok = false + for node in open_els + if node.name is 'li' and node.namespace is NS_HTML + generate_implied_end_tags 'li' # arg is exception + if open_els[0].name isnt 'li' or open_els[0].namespace isnt NS_HTML parse_error() - # TODO - when 'frameset' + loop + el = open_els.shift() + if el.name is 'li' and el.namespace is NS_HTML + break + break + if el_is_special_not_adp node + break + close_p_if_in_button_scope() + insert_html_element t + return + if t.type is TYPE_START_TAG and (t.name is 'dd' or t.name is 'dt') + flag_frameset_ok = false + for node in open_els + if node.name is 'dd' and node.namespace is NS_HTML + generate_implied_end_tags 'dd' # arg is exception + if open_els[0].name isnt 'dd' or open_els[0].namespace isnt NS_HTML parse_error() - # TODO - when 'address', 'article', 'aside', 'blockquote', 'center', 'details', 'dialog', 'dir', 'div', 'dl', 'fieldset', 'figcaption', 'figure', 'footer', 'header', 'hgroup', 'main', 'nav', 'ol', 'p', 'section', 'summary', 'ul' - close_p_if_in_button_scope() - insert_html_element t - when 'h1', 'h2', 'h3', 'h4', 'h5', 'h6' - close_p_if_in_button_scope() - if open_els[0].name in ['h1', 'h2', 'h3', 'h4', 'h5', 'h6'] - parse_error() - open_els.shift() - insert_html_element t - # TODO lots more to implement here - when 'a' - # If the list of active formatting elements - # contains an a element between the end of the list and - # the last marker on the list (or the start of the list - # if there is no marker on the list), then this is a - # parse error; run the adoption agency algorithm for - # the tag name "a", then remove that element from the - # list of active formatting elements and the stack of - # open elements if the adoption agency algorithm didn't - # already remove it (it might not have if the element - # is not in table scope). - found = false - for el in afe - if el.type is TYPE_AFE_MARKER - break - if el.name is 'a' - found = el - if found? - parse_error() - adoption_agency 'a' - for el, i in afe - if el is found - afe.splice i, 1 - for el, i in open_els - if el is found - open_els.splice i, 1 - reconstruct_active_formatting_elements() - el = insert_html_element t - 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_push el - when 'table' - # fixfull quirksmode thing - close_p_if_in_button_scope() - insert_html_element t - insertion_mode = ins_mode_in_table - # TODO lots more to implement here - else # any other start tag - reconstruct_active_formatting_elements() - insert_html_element t - when TYPE_EOF - ok_tags = { - dd: true, dt: true, li: true, p: true, tbody: true, td: true, - tfoot: true, th: true, thead: true, tr: true, body: true, html: true, - } - for t in open_els - unless ok_tags[t.name]? + loop + el = open_els.shift() + if el.name is 'dd' and el.namespace is NS_HTML + break + break + if node.name is 'dt' and node.namespace is NS_HTML + generate_implied_end_tags 'dt' # arg is exception + if open_els[0].name isnt 'dt' or open_els[0].namespace isnt NS_HTML parse_error() - break - # TODO stack of template insertion modes thing - flag_parsing = false # stop parsing - when TYPE_END_TAG - switch t.name - when 'body' - unless is_in_scope 'body' - parse_error() - return - # TODO implement parse error and move to tree_after_body - when 'html' - unless is_in_scope 'body' # weird, but it's what the spec says - parse_error() - return - # TODO implement parse error and move to tree_after_body, reprocess - when 'address', 'article', 'aside', 'blockquote', 'button', 'center', 'details', 'dialog', 'dir', 'div', 'dl', 'fieldset', 'figcaption', 'figure', 'footer', 'header', 'hgroup', 'listing', 'main', 'nav', 'ol', 'pre', 'section', 'summary', 'ul' - unless is_in_scope t.name, NS_HTML - parse_error() - return - generate_implied_end_tags() - unless open_els[0].name is t.name and open_els[0].namespace is NS_HTML - parse_error() - loop - el = open_els.shift() - if el.name is t.name and el.namespace is NS_HTML - return - # TODO lots more close tags to implement here - when 'p' - unless is_in_button_scope 'p' - parse_error() - insert_html_element new_open_tag 'p' - close_p_element() - # TODO lots more close tags to implement here - when 'a', 'b', 'big', 'code', 'em', 'font', 'i', 'nobr', 's', 'small', 'strike', 'strong', 'tt', 'u' - adoption_agency t.name - # TODO lots more close tags to implement here - else - in_body_any_other_end_tag t.name + loop + el = open_els.shift() + if el.name is 'dt' and el.namespace is NS_HTML + break + break + if el_is_special_not_adp node + break + close_p_if_in_button_scope() + insert_html_element t + return + if t.type is TYPE_START_TAG and t.name is 'plaintext' + close_p_if_in_button_scope() + insert_html_element t + tok_state = tok_state_plaintext + return + if t.type is TYPE_START_TAG and t.name is 'button' + if is_in_scope 'button', NS_HTML + parse_error() + generate_implied_end_tags() + loop + el = open_els.shift() + if el.name is 'button' and el.namespace is NS_HTML + break + reconstruct_afe() + insert_html_element t + flag_frameset_ok = false + return + if t.type is TYPE_END_TAG and (t.name is 'address' or t.name is 'article' or t.name is 'aside' or t.name is 'blockquote' or t.name is 'button' or t.name is 'center' or t.name is 'details' or t.name is 'dialog' or t.name is 'dir' or t.name is 'div' or t.name is 'dl' or t.name is 'fieldset' or t.name is 'figcaption' or t.name is 'figure' or t.name is 'footer' or t.name is 'header' or t.name is 'hgroup' or t.name is 'listing' or t.name is 'main' or t.name is 'nav' or t.name is 'ol' or t.name is 'pre' or t.name is 'section' or t.name is 'summary' or t.name is 'ul') + unless is_in_scope t.name, NS_HTML + parse_error() + return + generate_implied_end_tags() + unless open_els[0].name is t.name and open_els[0].namespace is NS_HTML + parse_error() + loop + el = open_els.shift() + if el.name is t.name and el.namespace is NS_HTML + return + return + if t.type is TYPE_END_TAG and t.name is 'form' + unless template_tag_is_open() + node = form_element_pointer + form_element_pointer = null + if node is null or not el_is_in_scope node + parse_error() + return + generate_implied_end_tags() + if open_els[0] isnt node + parse_error() + for el, i in open_els + if el is node + open_els.splice i, 1 + break + else + unless is_in_scope 'form', NS_HTML + parse_error() + return + generate_implied_end_tags() + if open_els[0].name isnt 'form' or open_els[0].namespace isnt NS_HTML + parse_error() + loop + el = open_els.shift() + if el.name is 'form' and el.namespace is NS_HTML + break + return + if t.type is TYPE_END_TAG and t.name is 'p' + unless is_in_button_scope 'p', NS_HTML + parse_error() + insert_html_element new_open_tag 'p' + close_p_element() + return + if t.type is TYPE_END_TAG and t.name is 'li' + unless is_in_li_scope 'li', NS_HTML + parse_error() + return + generate_implied_end_tags 'li' # arg is exception + if open_els[0].name isnt 'li' or open_els[0].namespace isnt NS_HTML + parse_error() + loop + el = open_els.shift() + if el.name is 'li' and el.namespace is NS_HTML + break + return + if t.type is TYPE_END_TAG and (t.name is 'dd' or t.name is 'dt') + unless is_in_scope t.name, NS_HTML + parse_error() + return + generate_implied_end_tags t.name # arg is exception + if open_els[0].name isnt t.name or open_els[0].namespace isnt NS_HTML + parse_error() + loop + el = open_els.shift() + if el.name is t.name and el.namespace is NS_HTML + break + return + if t.type is TYPE_END_TAG and h_tags[t.name]? + h_in_scope = false + for el in open_els + if h_tags[el.name] is el.namespace + h_in_scope = true + break + if standard_scopers[el.name] is el.namespace + break + unless h_in_scope + parse_error() + return + generate_implied_end_tags() + if open_els[0].name isnt t.name or open_els[0].namespace isnt NS_HTML + parse_error() + loop + el = open_els.shift() + if h_tags[el.name] is el.namespace + break + return + # deep breath! + if t.type is TYPE_START_TAG and t.name is 'a' + # If the list of active formatting elements contains an a element + # between the end of the list and the last marker on the list (or + # the start of the list if there is no marker on the list), then + # this is a parse error; run the adoption agency algorithm for the + # tag name "a", then remove that element from the list of active + # formatting elements and the stack of open elements if the + # adoption agency algorithm didn't already remove it (it might not + # have if the element is not in table scope). + found = false + for el in afe + if el.type is TYPE_AFE_MARKER + break + if el.name is 'a' and el.namespace is NS_HTML + found = el + if found? + parse_error() + adoption_agency 'a' + for el, i in afe + if el is found + afe.splice i, 1 + for el, i in open_els + if el is found + open_els.splice i, 1 + reconstruct_afe() + el = insert_html_element t + afe_push el + return + if t.type is TYPE_START_TAG and (t.name is 'b' or t.name is 'big' or t.name is 'code' or t.name is 'em' or t.name is 'font' or t.name is 'i' or t.name is 's' or t.name is 'small' or t.name is 'strike' or t.name is 'strong' or t.name is 'tt' or t.name is 'u') + reconstruct_afe() + el = insert_html_element t + afe_push el + return + if t.type is TYPE_START_TAG and t.name is 'nobr' + reconstruct_afe() + el = insert_html_element t + afe_push el + return + if t.type is TYPE_END_TAG and (t.name is 'a' or t.name is 'b' or t.name is 'big' or t.name is 'code' or t.name is 'em' or t.name is 'font' or t.name is 'i' or t.name is 'nobr' or t.name is 's' or t.name is 'small' or t.name is 'strike' or t.name is 'strong' or t.name is 'tt' or t.name is 'u') + adoption_agency t.name + return + if t.type is TYPE_START_TAG and (t.name is 'applet' or t.name is 'marquee' or t.name is 'object') + reconstruct_afe() + insert_html_element t + afe_push_marker() + flag_frameset_ok = false + return + if t.type is TYPE_END_TAG and (t.name is 'applet' or t.name is 'marquee' or t.name is 'object') + unless is_in_scope t.name, NS_HTML + parse_error() + return + generate_implied_end_tags() + if open_els[0].name isnt t.name or open_els[0].namespace isnt NS_HTML + parse_error() + loop + el = open_els.shift() + if el.name is t.name and el.namespace is NS_HTML + break + clear_afe_to_marker() + return + if t.type is TYPE_START_TAG and t.name is 'table' + close_p_if_in_button_scope() # fixfull quirksmode thing + insert_html_element t + flag_frameset_ok = false + ins_mode = ins_mode_in_table + return + if t.type is TYPE_END_TAG and t.name is 'br' + parse_error() + t.type is TYPE_START_TAG + # fall through + if t.type is TYPE_START_TAG and (t.name is 'area' or t.name is 'br' or t.name is 'embed' or t.name is 'img' or t.name is 'keygen' or t.name is 'wbr') + reconstruct_afe() + insert_html_element t + open_els.shift() + t.acknowledge_self_closing() + flag_frameset_ok = false + return + if t.type is TYPE_START_TAG and t.name is 'input' + reconstruct_afe() + insert_html_element t + open_els.shift() + t.acknowledge_self_closing() + unless is_input_hidden_tok t + flag_frameset_ok = false + return + if t.type is TYPE_START_TAG and (t.name is 'param' or t.name is 'source' or t.name is 'track') + insert_html_element t + open_els.shift() + t.acknowledge_self_closing() + return + if t.type is TYPE_START_TAG and t.name is 'hr' + close_p_if_in_button_scope() + insert_html_element t + open_els.shift() + t.acknowledge_self_closing() + flag_frameset_ok = false + return + if t.type is TYPE_START_TAG and t.name is 'image' + parse_error() + t.name = 'img' + process_token t + return + if t.type is TYPE_START_TAG and t.name is 'isindex' + parse_error() + if template_tag_is_open() is false and form_element_pointer isnt null + return + t.acknowledge_self_closing() + flag_frameset_ok = false + close_p_if_in_button_scope() + el = insert_html_element new_open_tag 'form' + unless template_tag_is_open() + form_element_pointer = el + for a in t.attrs_a + if a[0] is 'action' + el.attrs['action'] = a[1] + break + insert_html_element new_open_tag 'hr' + open_els.shift() + reconstruct_afe() + insert_html_element new_open_tag 'label' + # note: this is a little out-of-spec-order so we only have to scan t.attrs_a once + input_el = new_open_tag 'input' + prompt = null + for a in t.attrs_a + if a[0] is 'prompt' + prompt = a[1] + if a[0] isnt 'name' and a[0] isnt 'action' and a[0] isnt 'prompt' + input_el.attrs_a.push [a[0], a[1]] + input_el.attrs_a.push ['name', 'isindex'] + # fixfull this next bit is in english... internationalize? + prompt ?= "This is a searchable index. Enter search keywords: " + insert_character new_character_token prompt # fixfull split + # TODO submit typo "balue" in spec + insert_html_element input_el + open_els.shift() + # insert_character '' # you can put chars here if promt attr missing + open_els.shift() + insert_html_element new_open_tag 'hr' + open_els.shift() + open_els.shift() + unless template_tag_is_open() + form_element_pointer = null + return + if t.type is TYPE_START_TAG and t.name is 'textarea' + insert_html_element t + if txt.charAt(cur) is "\u000a" # FIXME check for crlf? + cur += 1 + tok_state = tok_state_rcdata + original_ins_mode = ins_mode + flag_frameset_ok = false + ins_mode = ins_mode_text + return + if t.type is TYPE_START_TAG and t.name is 'xmp' + close_p_if_in_button_scope() + reconstruct_afe() + flag_frameset_ok = false + parse_generic_raw_text t + return + if t.type is TYPE_START_TAG and t.name is 'iframe' + flag_frameset_ok = false + parse_generic_raw_text t + return + if t.type is TYPE_START_TAG and (t.name is 'noembed' or (t.name is 'noscript' and flag_scripting)) + parse_generic_raw_text t + return + if t.type is TYPE_START_TAG and t.name is 'select' + reconstruct_afe() + insert_html_element t + flag_frameset_ok = false + if ins_mode is ins_mode_in_table or ins_mode is ins_mode_in_caption or ins_mode is ins_mode_in_table_body or ins_mode is ins_mode_in_row or ins_mode is ins_mode_in_cell + ins_mode = ins_mode_in_select_in_table + else + ins_mode = ins_mode_in_select + return + if t.type is TYPE_START_TAG and (t.name is 'optgroup' or t.name is 'option') + if open_els[0].name is 'option' and open_els[0].namespace is NS_HTML + open_els.shift() + reconstruct_afe() + insert_html_element t + return +# this comment block implements the W3C spec +# if t.type is TYPE_START_TAG and (t.name is 'rb' or t.name is 'rp' or t.name is 'rtc') +# if is_in_scope 'ruby', NS_HTML +# generate_implied_end_tags() +# unless open_els[0].name is 'ruby' and open_els[0].namespace is NS_HTML +# parse_error() +# insert_html_element t +# return +# if t.type is TYPE_START_TAG and t.name is 'rt' +# if is_in_scope 'ruby', NS_HTML +# generate_implied_end_tags 'rtc' # arg is exception +# unless (open_els[0].name is 'ruby' or open_els[0].name is 'rtc') and open_els[0].namespace is NS_HTML +# parse_error() +# insert_html_element t +# return +# below implements the WATWG spec https://html.spec.whatwg.org/multipage/syntax.html#parsing-main-inbody + if t.type is TYPE_START_TAG and (t.name is 'rb' or t.name is 'rtc') + if is_in_scope 'ruby', NS_HTML + generate_implied_end_tags() + unless open_els[0].name is 'ruby' and open_els[0].namespace is NS_HTML + parse_error() + insert_html_element t + return + if t.type is TYPE_START_TAG and (t.name is 'rp' or t.name is 'rt') + if is_in_scope 'ruby', NS_HTML + generate_implied_end_tags 'rtc' + unless (open_els[0].name is 'ruby' or open_els[0].name is 'rtc') and open_els[0].namespace is NS_HTML + parse_error() + insert_html_element t + return +# end WATWG chunk + if t.type is TYPE_START_TAG and t.name is 'math' + reconstruct_afe() + adjust_mathml_attributes t + adjust_foreign_attributes t + insert_foreign_element t, NS_MATHML + if t.flag 'self-closing' + open_els.shift() + t.acknowledge_self_closing() + return + if t.type is TYPE_START_TAG and t.name is 'svg' + reconstruct_afe() + adjust_svg_attributes t + adjust_foreign_attributes t + insert_foreign_element t, NS_SVG + if t.flag 'self-closing' + open_els.shift() + t.acknowledge_self_closing() + 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 'frame' or t.name is 'head' 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 + if t.type is TYPE_START_TAG # any other start tag + reconstruct_afe() + insert_html_element t + return + if t.type is TYPE_END_TAG # any other end tag + in_body_any_other_end_tag t.name + return return + # 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' and open_els[0].namespace is NS_HTML + open_els[0].flag 'already started', true + open_els.shift() + ins_mode = original_ins_mode + process_token t + return + if t.type is TYPE_END_TAG and t.name is 'script' + open_els.shift() + ins_mode = original_ins_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() + ins_mode = original_ins_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_else = (t) -> parse_error() - flag_foster_parenting = true # FIXME + flag_foster_parenting = true ins_mode_in_body t flag_foster_parenting = false - can_in_table = { - '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 ins_mode_in_table = (t) -> switch t.type when TYPE_TEXT - if can_in_table[t.name] - original_insertion_mode = insertion_mode - insertion_mode = ins_mode_in_table_text - insertion_mode t + if (open_els[0].name is 'table' or open_els[0].name is 'tbody' or open_els[0].name is 'tfoot' or open_els[0].name is 'thead' or open_els[0].name is 'tr') and open_els[0].namespace is NS_HTML + pending_table_character_tokens = [] + original_ins_mode = ins_mode + ins_mode = ins_mode_in_table_text + process_token t 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 @@ -1274,44 +2239,44 @@ parse_html = (txt, parse_error_cb = null) -> clear_stack_to_table_context() afe_push_marker() insert_html_element t - insertion_mode = ins_mode_in_caption + ins_mode = ins_mode_in_caption when 'colgroup' clear_stack_to_table_context() insert_html_element t - insertion_mode = ins_mode_in_column_group + ins_mode = ins_mode_in_column_group when 'col' clear_stack_to_table_context() insert_html_element new_open_tag 'colgroup' - insertion_mode = ins_mode_in_column_group - insertion_mode t + ins_mode = ins_mode_in_column_group + process_token t when 'tbody', 'tfoot', 'thead' clear_stack_to_table_context() insert_html_element t - insertion_mode = ins_mode_in_table_body + ins_mode = ins_mode_in_table_body when 'td', 'th', 'tr' clear_stack_to_table_context() insert_html_element new_open_tag 'tbody' - insertion_mode = ins_mode_in_table_body - insertion_mode t + ins_mode = ins_mode_in_table_body + process_token t when 'table' parse_error() - if is_in_table_scope 'table' + if is_in_table_scope 'table', NS_HTML loop el = open_els.shift() - if el.name is 'table' + if el.name is 'table' and el.namespace is NS_HTML break - reset_insertion_mode() - insertion_mode t + reset_ins_mode() + process_token t when 'style', 'script', 'template' ins_mode_in_head t when 'input' - if token_is_input_hidden t + unless 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? @@ -1325,14 +2290,14 @@ parse_html = (txt, parse_error_cb = null) -> when TYPE_END_TAG switch t.name when 'table' - if is_in_table_scope 'table' + if is_in_table_scope 'table', NS_HTML loop el = open_els.shift() - if el.name is 'table' + if el.name is 'table' and el.namespace is NS_HTML break - reset_insertion_mode() + reset_ins_mode() else - parse_error + parse_error() when 'body', 'caption', 'col', 'colgroup', 'html', 'tbody', 'td', 'tfoot', 'th', 'thead', 'tr' parse_error() when 'template' @@ -1345,51 +2310,147 @@ 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" + # from javascript? + 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_in_table_else old + pending_table_character_tokens = [] + ins_mode = original_ins_mode + process_token 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', NS_HTML + generate_implied_end_tags() + if open_els[0].name isnt 'caption' + parse_error() + loop + el = open_els.shift() + if el.name is 'caption' and el.namespace is NS_HTML + break + clear_afe_to_marker() + ins_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', NS_HTML + loop + el = open_els.shift() + if el.name is 'caption' and el.namespace is NS_HTML + break + clear_afe_to_marker() + ins_mode = ins_mode_in_table + process_token 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' and open_els.namespace is NS_HTML + open_els.shift() + ins_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() + ins_mode = ins_mode_in_table + process_token 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 + ins_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() clear_stack_to_table_body_context() insert_html_element new_open_tag 'tr' - insertion_mode = ins_mode_in_row - insertion_mode t + ins_mode = ins_mode_in_row + process_token t return if t.type is TYPE_END_TAG and (t.name is 'tbody' or t.name is 'tfoot' or t.name is 'thead') - unless is_in_table_scope t.name # fixfull check namespace + unless is_in_table_scope t.name, NS_HTML parse_error() return clear_stack_to_table_body_context() open_els.shift() - insertion_mode = ins_mode_in_table + ins_mode = ins_mode_in_table 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 'tfoot' or t.name is 'thead')) or (t.type is TYPE_END_TAG and t.name is 'table') has = false for el in open_els - if el.name is 'tbody' or el.name is 'tfoot' or el.name is 'thead' + if el.namespace is NS_HTML and (el.name is 'tbody' or el.name is 'tfoot' or el.name is 'thead') has = true break - if table_scopers[el.name] + if table_scopers[el.name] is el.namespace break if !has parse_error() return clear_stack_to_table_body_context() open_els.shift() - insertion_mode = ins_mode_in_table - insertion_mode t + ins_mode = ins_mode_in_table + process_token t return if t.type is TYPE_END_TAG and (t.name is 'body' or t.name is 'caption' or t.name is 'col' or t.name is 'colgroup' or t.name is 'html' or t.name is 'td' or t.name is 'th' or t.name is 'tr') parse_error() @@ -1397,37 +2458,38 @@ 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 + ins_mode = ins_mode_in_cell afe_push_marker() return if t.type is TYPE_END_TAG and t.name is 'tr' - if is_in_table_scope 'tr' + if is_in_table_scope 'tr', NS_HTML clear_stack_to_table_row_context() open_els.shift() - insertion_mode = ins_mode_in_table_body + ins_mode = ins_mode_in_table_body else parse_error() 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 'tfoot' or t.name is 'thead' or t.name is 'tr')) or t.type is TYPE_END_TAG and t.name is 'table' - if is_in_table_scope 'tr' + if is_in_table_scope 'tr', NS_HTML clear_stack_to_table_row_context() open_els.shift() - insertion_mode = ins_mode_in_table_body - insertion_mode t + ins_mode = ins_mode_in_table_body + process_token t else parse_error() return if t.type is TYPE_END_TAG and (t.name is 'tbody' or t.name is 'tfoot' or t.name is 'thead') - if is_in_table_scope t.name # fixfull namespace - if is_in_table_scope 'tr' + if is_in_table_scope t.name, NS_HTML + if is_in_table_scope 'tr', NS_HTML clear_stack_to_table_row_context() open_els.shift() - insertion_mode = ins_mode_in_table_body - insertion_mode t + ins_mode = ins_mode_in_table_body + process_token t else parse_error() return @@ -1440,81 +2502,520 @@ parse_html = (txt, parse_error_cb = null) -> # http://www.w3.org/TR/html5/syntax.html#close-the-cell close_the_cell = -> generate_implied_end_tags() - unless open_els[0].name is 'td' or open_els[0] is 'th' + unless (open_els[0].name is 'td' or open_els[0] is 'th') and open_els[0].namespace is NS_HTML parse_error() loop el = open_els.shift() - if el.name is 'td' or el.name is 'th' + if el.namespace is NS_HTML and (el.name is 'td' or el.name is 'th') break clear_afe_to_marker() - insertion_mode = ins_mode_in_row + ins_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 + if is_in_table_scope t.name, NS_HTML generate_implied_end_tags() - if open_els[0].name isnt t.name - parse_error + unless (open_els[0].name is t.name) and open_els[0].namespace is NS_HTML + parse_error() loop el = open_els.shift() - if el.name is t.name + if el.name is t.name and el.namespace is NS_HTML break clear_afe_to_marker() - insertion_mode = ins_mode_in_row + ins_mode = ins_mode_in_row else parse_error() 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') has = false for el in open_els - if el.name is 'td' or el.name is 'th' + if el.namespace is NS_HTML and (el.name is 'td' or el.name is 'th') has = true break - if table_scopers[el.name] + if table_scopers[el.name] is el.namespace break if !has parse_error() return close_the_cell() - insertion_mode t + process_token t return if t.type is TYPE_END_TAG and (t.name is 'body' or t.name is 'caption' or t.name is 'col' or t.name is 'colgroup' or t.name is 'html') parse_error() return if t.type is TYPE_END_TAG and (t.name is 'table' or t.name is 'tbody' or t.name is 'tfoot' or t.name is 'thead' or t.name is 'tr') - if is_in_table_scope t.name # fixfull namespace + if is_in_table_scope t.name, NS_HTML close_the_cell() - insertion_mode t + process_token t else parse_error() return # Anything Else ins_mode_in_body t - - # 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 + # 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' and open_els[0].namespace is NS_HTML + 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' and open_els[0].namespace is NS_HTML + open_els.shift() + if open_els[0].name is 'optgroup' and open_els[0].namespace is NS_HTML + 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[0].namespace in NS_HTML + if open_els[1].name is 'optgroup' and open_els[0].namespace is NS_HTML + open_els.shift() + if open_els[0].name is 'optgroup' and open_els[0].namespace is NS_HTML + 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' and open_els[0].namespace is NS_HTML + 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', NS_HTML + loop + el = open_els.shift() + if el.name is 'select' and el.namespace is NS_HTML + break + reset_ins_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' and el.namespace is NS_HTML + break + reset_ins_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', NS_HTML + return + loop + el = open_els.shift() + if el.name is 'select' and el.namespace is NS_HTML + break + reset_ins_mode() + process_token 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' and el.namespace is NS_HTML + break + reset_ins_mode() + process_token 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' and el.namespace is NS_HTML + break + reset_ins_mode() + process_token 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_ins_modes.shift() + template_ins_modes.unshift ins_mode_in_table + ins_mode = ins_mode_in_table + process_token t + return + if t.type is TYPE_START_TAG and t.name is 'col' + template_ins_modes.shift() + template_ins_modes.unshift ins_mode_in_column_group + ins_mode = ins_mode_in_column_group + process_token t + return + if t.type is TYPE_START_TAG and t.name is 'tr' + template_ins_modes.shift() + template_ins_modes.unshift ins_mode_in_table_body + ins_mode = ins_mode_in_table_body + process_token t + return + if t.type is TYPE_START_TAG and (t.name is 'td' or t.name is 'th') + template_ins_modes.shift() + template_ins_modes.unshift ins_mode_in_row + ins_mode = ins_mode_in_row + process_token t + return + if t.type is TYPE_START_TAG + template_ins_modes.shift() + template_ins_modes.unshift ins_mode_in_body + ins_mode = ins_mode_in_body + process_token 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' and el.namespace is NS_HTML + break + clear_afe_to_marker() + template_ins_modes.shift() + reset_ins_mode() + process_token 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' + if flag_fragment_parsing + parse_error() + return + ins_mode = ins_mode_after_after_body + return + if t.type is TYPE_EOF + stop_parsing() + return + # Anything ELse + parse_error() + ins_mode = ins_mode_in_body + process_token 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' + 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' + ins_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 + 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() + ins_mode = ins_mode_in_body + process_token t + 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 + + # 8.2.5.5 http://www.w3.org/TR/html5/syntax.html#parsing-main-inforeign + has_color_face_or_size = (t) -> + for a in t.attrs_a + if a[0] is 'color' or a[0] is 'face' or a[0] is 'size' + return true + return false + in_foreign_content_end_script = -> + open_els.shift() + # fixfull + return + in_foreign_content_other_start = (t) -> + acn = adjusted_current_node() + if acn.namespace is NS_MATHML + adjust_mathml_attributes t + if acn.namespace is NS_SVG and svg_name_fixes[t.name]? + t.name = svg_name_fixes[t.name] + if acn.namespace is NS_SVG + adjust_svg_attributes t + adjust_foreign_attributes t + insert_foreign_element t, acn.namespace + if t.flag 'self-closing' + if t.name is 'script' + t.acknowledge_self_closing() + in_foreign_content_end_script() + # fixfull + else + open_els.shift() + t.acknowledge_self_closing() + return + in_foreign_content = (t) -> + if t.type is TYPE_TEXT and t.text is "\u0000" + parse_error() + insert_character new_character_token "\ufffd" + return + if is_space_tok t + insert_character t + return + if t.type is TYPE_TEXT + flag_frameset_ok = false + 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 'b' or t.name is 'big' or t.name is 'blockquote' or t.name is 'body' or t.name is 'br' or t.name is 'center' or t.name is 'code' or t.name is 'dd' or t.name is 'div' or t.name is 'dl' or t.name is 'dt' or t.name is 'em' or t.name is 'embed' or t.name is 'h1' or t.name is 'h2' or t.name is 'h3' or t.name is 'h4' or t.name is 'h5' or t.name is 'h6' or t.name is 'head' or t.name is 'hr' or t.name is 'i' or t.name is 'img' or t.name is 'li' or t.name is 'listing' or t.name is 'main' or t.name is 'meta' or t.name is 'nobr' or t.name is 'ol' or t.name is 'p' or t.name is 'pre' or t.name is 'ruby' or t.name is 's' or t.name is 'small' or t.name is 'span' or t.name is 'strong' or t.name is 'strike' or t.name is 'sub' or t.name is 'sup' or t.name is 'table' or t.name is 'tt' or t.name is 'u' or t.name is 'ul' or t.name is 'var' or (t.name is 'font' and has_color_face_or_size(t))) + parse_error() + if flag_fragment_parsing + in_foreign_content_other_start t + return + loop # is this safe? + open_els.shift() + if is_mathml_text_integration_point(open_els[0]) or is_html_integration(open_els[0]) or open_els[0].namespace is NS_HTML + break + process_token t + return + if t.type is TYPE_START_TAG + in_foreign_content_other_start t + return + if t.type is TYPE_END_TAG and t.name is 'script' and open_els[0].name is 'script' and open_els[0].namespace is NS_SVG + in_foreign_content_end_script() + return + if t.type is TYPE_END_TAG + i = 0 + node = open_els[i] + if node.name.toLowerCase() isnt t.name + parse_error() + loop + if node is open_els[open_els.length - 1] + return + if node.name.toLowerCase() is t.name + loop + el = open_els.shift() + if el is node + return + i += 1 + node = open_els[i] + if node.namespace is NS_HTML + break + ins_mode t # explicitly call HTML insertion mode + + + # 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 "\ufffd" + 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() + # 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 + # 8.2.4.8 http://www.w3.org/TR/html5/syntax.html#tag-open-state tok_state_tag_open = -> @@ -1525,12 +3026,13 @@ 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 @@ -1551,14 +3053,15 @@ 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 @@ -1581,46 +3084,517 @@ 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.name += c.toLowerCase() else tok_cur_tag.name += c return null - # 8.2.4.34 http://www.w3.org/TR/html5/syntax.html#before-attribute-name-state - tok_state_before_attribute_name = -> - attr_name = null - switch c = txt.charAt(cur++) - when "\t", "\n", "\u000c", ' ' - return null - when '/' - tok_state = tok_state_self_closing_start_tag - return null - when '>' - tok_state = tok_state_data - tmp = tok_cur_tag - tok_cur_tag = null - return tmp - when "\u0000" - parse_error() - attr_name = "\ufffd" - when '"', "'", '<', '=' - parse_error() - attr_name = c - when '' # EOF - parse_error() + # 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, so it's hopefully ok to assume that + # open_els[0].name will work instead 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 - else - if uc_alpha.indexOf(c) > -1 - attr_name = c.toLowerCase() - else - attr_name = c - if attr_name? - tok_cur_tag.attrs_a.unshift [attr_name, ''] - tok_state = tok_state_attribute_name - return null + 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 c is '>' + if is_appropriate_end_tag tok_cur_tag + tok_state = tok_state_data + return tok_cur_tag + # 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 c is '>' + if is_appropriate_end_tag tok_cur_tag + tok_state = tok_state_data + return tok_cur_tag + # 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 = -> + attr_name = null + switch c = txt.charAt(cur++) + when "\t", "\n", "\u000c", ' ' + return null + when '/' + tok_state = tok_state_self_closing_start_tag + return null + when '>' + tok_state = tok_state_data + tmp = tok_cur_tag + tok_cur_tag = null + return tmp + when "\u0000" + parse_error() + attr_name = "\ufffd" + when '"', "'", '<', '=' + parse_error() + attr_name = c + when '' # EOF + parse_error() + tok_state = tok_state_data + else + if is_uc_alpha(c) + attr_name = c.toLowerCase() + else + attr_name = c + if attr_name? + tok_cur_tag.attrs_a.unshift [attr_name, ''] + tok_state = tok_state_attribute_name + return null + + # 8.2.4.35 http://www.w3.org/TR/html5/syntax.html#attribute-name-state tok_state_attribute_name = -> switch c = txt.charAt(cur++) when "\t", "\n", "\u000c", ' ' @@ -1636,166 +3610,792 @@ parse_html = (txt, parse_error_cb = null) -> return tmp when "\u0000" parse_error() - tok_cur_tag.attrs_a[0][0] = "\ufffd" + tok_cur_tag.attrs_a[0][0] += "\ufffd" when '"', "'", '<' parse_error() - tok_cur_tag.attrs_a[0][0] = c + tok_cur_tag.attrs_a[0][0] += c when '' # EOF parse_error() tok_state = tok_state_data else - if uc_alpha.indexOf(c) > -1 - tok_cur_tag.attrs_a[0][0] = c.toLowerCase() + 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 = -> + # 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++) + when "\t", "\n", "\u000c", ' ' + return null + when '"' + tok_state = tok_state_attribute_value_double_quoted + when '&' + tok_state = tok_state_attribute_value_unquoted + cur -= 1 + when "'" + tok_state = tok_state_attribute_value_single_quoted + when "\u0000" + # Parse error + tok_cur_tag.attrs_a[0][1] += "\ufffd" + tok_state = tok_state_attribute_value_unquoted + when '>' + # Parse error + tok_state = tok_state_data + tmp = tok_cur_tag + tok_cur_tag = null + return tmp + when '' # EOF + parse_error() + tok_state = tok_state_data + else + tok_cur_tag.attrs_a[0][1] += c + tok_state = tok_state_attribute_value_unquoted + return null + + # 8.2.4.38 http://www.w3.org/TR/html5/syntax.html#attribute-value-(double-quoted)-state + tok_state_attribute_value_double_quoted = -> + switch c = txt.charAt(cur++) + when '"' + tok_state = tok_state_after_attribute_value_quoted + when '&' + tok_cur_tag.attrs_a[0][1] += parse_character_reference '"', true + when "\u0000" + # Parse error + tok_cur_tag.attrs_a[0][1] += "\ufffd" + when '' # EOF + parse_error() + tok_state = tok_state_data + else + tok_cur_tag.attrs_a[0][1] += c + return null + + # 8.2.4.39 http://www.w3.org/TR/html5/syntax.html#attribute-value-(single-quoted)-state + tok_state_attribute_value_single_quoted = -> + switch c = txt.charAt(cur++) + when "'" + tok_state = tok_state_after_attribute_value_quoted + when '&' + tok_cur_tag.attrs_a[0][1] += parse_character_reference "'", true + when "\u0000" + # Parse error + tok_cur_tag.attrs_a[0][1] += "\ufffd" + when '' # EOF + parse_error() + tok_state = tok_state_data + else + tok_cur_tag.attrs_a[0][1] += c + return null + + # 8.2.4.40 http://www.w3.org/TR/html5/syntax.html#attribute-value-(unquoted)-state + tok_state_attribute_value_unquoted = -> + switch c = txt.charAt(cur++) + when "\t", "\n", "\u000c", ' ' + tok_state = tok_state_before_attribute_name + when '&' + tok_cur_tag.attrs_a[0][1] += parse_character_reference '>', true + when '>' + tok_state = tok_state_data + tmp = tok_cur_tag + tok_cur_tag = null + return tmp + when "\u0000" + tok_cur_tag.attrs_a[0][1] += "\ufffd" + when '' # EOF + parse_error() + tok_state = tok_state_data + else + # Parse Error if ', <, = or ` (backtick) + tok_cur_tag.attrs_a[0][1] += c + return null + + # 8.2.4.42 http://www.w3.org/TR/html5/syntax.html#after-attribute-value-(quoted)-state + tok_state_after_attribute_value_quoted = -> + switch c = txt.charAt(cur++) + when "\t", "\n", "\u000c", ' ' + tok_state = tok_state_before_attribute_name + when '/' + tok_state = tok_state_self_closing_start_tag + when '>' + tok_state = tok_state_data + tmp = tok_cur_tag + tok_cur_tag = null + return tmp + when '' # EOF + parse_error() + tok_state = tok_state_data + else + # Parse Error + tok_state = tok_state_before_attribute_name + 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', true + 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(new RegExp("\u0000", 'g'), "\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 '' + 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() + tok_state = tok_state_comment + 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 + tok_state = tok_state_comment + 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 = '' + tok_state = tok_state_doctype_public_identifier_double_quoted + return + if c is "'" + parse_error() + tok_cur_tag.public_identifier = '' + 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 "\n" or c is "\u000c" or c is ' ' + if c is "\t" or c is "\u000a" or c is "\u000c" or c is ' ' return - if c is '/' - tok_state = tok_state_self_closing_start_tag + if c is '"' + parse_error() + tok_cur_tag.public_identifier = '' + tok_state = tok_state_doctype_public_identifier_double_quoted return - if c is '=' - tok_state = tok_state_before_attribute_value + if c is "'" + parse_error() + tok_cur_tag.public_identifier = '' + 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 uc_alpha.indexOf(c) > -1 - tok_cur_tag.attrs_a.unshift [c.toLowerCase(), ''] - tok_state = tok_state_attribute_name + 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.attrs_a.unshift ["\ufffd", ''] - tok_state = tok_state_attribute_name + 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 - cur -= 1 # reconsume + 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 '"' or c is "'" or c is '<' + if c is '>' + tok_state = tok_state_data + return tok_cur_tag + if c is '"' parse_error() - # fall through to Anything else + 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 - tok_cur_tag.attrs_a.unshift [c, ''] - tok_state = tok_state_attribute_name + parse_error() + tok_cur_tag.flag 'force-quirks', true + tok_state = tok_state_bogus_doctype + return null - # 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++) - when "\t", "\n", "\u000c", ' ' - return null - when '"' - tok_state = tok_state_attribute_value_double_quoted - when '&' - tok_state = tok_state_attribute_value_unquoted - cur -= 1 - when "'" - tok_state = tok_state_attribute_value_single_quoted - when "\u0000" - # Parse error - tok_cur_tag.attrs_a[0][1] += "\ufffd" - tok_state = tok_state_attribute_value_unquoted - when '>' - # Parse error - tok_state = tok_state_data - tmp = tok_cur_tag - tok_cur_tag = null - return tmp - when '' # EOF - parse_error() - tok_state = tok_state_data - else - tok_cur_tag.attrs_a[0][1] += c - tok_state = tok_state_attribute_value_unquoted + # 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 - # 8.2.4.38 http://www.w3.org/TR/html5/syntax.html#attribute-value-(double-quoted)-state - tok_state_attribute_value_double_quoted = -> - switch c = txt.charAt(cur++) - when '"' - tok_state = tok_state_after_attribute_value_quoted - when '&' - tok_cur_tag.attrs_a[0][1] += tokenize_character_reference '"', true - when "\u0000" - # Parse error - tok_cur_tag.attrs_a[0][1] += "\ufffd" - when '' # EOF - parse_error() - tok_state = tok_state_data - else - tok_cur_tag.attrs_a[0][1] += c + # 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 - # 8.2.4.39 http://www.w3.org/TR/html5/syntax.html#attribute-value-(single-quoted)-state - tok_state_attribute_value_single_quoted = -> - switch c = txt.charAt(cur++) - when "'" - tok_state = tok_state_after_attribute_value_quoted - when '&' - tok_cur_tag.attrs_a[0][1] += tokenize_character_reference "'", true - when "\u0000" - # Parse error - tok_cur_tag.attrs_a[0][1] += "\ufffd" - when '' # EOF - parse_error() - tok_state = tok_state_data - else - tok_cur_tag.attrs_a[0][1] += c + # 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 - # 8.2.4.40 http://www.w3.org/TR/html5/syntax.html#attribute-value-(unquoted)-state - tok_state_attribute_value_unquoted = -> - switch c = txt.charAt(cur++) - when "\t", "\n", "\u000c", ' ' - tok_state = tok_state_before_attribute_name - when '&' - tok_cur_tag.attrs_a[0][1] += tokenize_character_reference '>', true - when '>' - tok_state = tok_state_data - tmp = tok_cur_tag - tok_cur_tag = null - return tmp - when "\u0000" - tok_cur_tag.attrs_a[0][1] += "\ufffd" - when '' # EOF - parse_error() - tok_state = tok_state_data - else - # Parse Error if ', <, = or ` (backtick) - tok_cur_tag.attrs_a[0][1] += c + # 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 - # 8.2.4.42 http://www.w3.org/TR/html5/syntax.html#after-attribute-value-(quoted)-state - tok_state_after_attribute_value_quoted = -> - switch c = txt.charAt(cur++) - when "\t", "\n", "\u000c", ' ' - tok_state = tok_state_before_attribute_name - when '/' - tok_state = tok_state_self_closing_start_tag - when '>' - tok_state = tok_state_data - tmp = tok_cur_tag - tok_cur_tag = null - return tmp - when '' # EOF - parse_error() - tok_state = tok_state_data - else - # Parse Error - tok_state = tok_state_before_attribute_name - cur -= 1 # we didn't handle that char + # 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 + # 8.2.4.68 http://www.w3.org/TR/html5/syntax.html#cdata-section-state + tok_state_cdata_section = -> + tok_state = tok_state_data + 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 + 3 + return new_character_token val # fixfull split + # 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) @@ -1809,26 +4409,39 @@ parse_html = (txt, parse_error_cb = null) -> if cur + 1 >= txt.length return '&' if txt.charAt(cur + 1).toLowerCase() is 'x' - prefix = '#x' + base = 16 charset = hex_chars start = cur + 2 else charset = digits start = cur + 1 - prefix = '#' + base = 10 i = 0 while start + i < txt.length and charset.indexOf(txt.charAt(start + i)) > -1 i += 1 if i is 0 return '&' + cur = start + i if txt.charAt(start + i) is ';' - i += 1 - # FIXME This is supposed to generate parse errors for some chars - decoded = decode_named_char_ref(prefix + txt.substr(start, i).toLowerCase()) - if decoded? - cur = start + i - return decoded - return '&' + cur += 1 + else + parse_error() + code_point = txt.substr(start, i) + while code_point.charAt(0) is '0' and code_point.length > 1 + code_point = code_point.substr 1 + code_point = parseInt(code_point, base) + if unicode_fixes[code_point]? + parse_error() + return unicode_fixes[code_point] + else + if (code_point >= 0xd800 and code_point <= 0xdfff) or code_point > 0x10ffff + parse_error() + return "\ufffd" + else + if (code_point >= 0x0001 and code_point <= 0x0008) or (code_point >= 0x000D and code_point <= 0x001F) or (code_point >= 0x007F and code_point <= 0x009F) or (code_point >= 0xFDD0 and code_point <= 0xFDEF) or code_point is 0x000B or code_point is 0xFFFE or code_point is 0xFFFF or code_point is 0x1FFFE or code_point is 0x1FFFF or code_point is 0x2FFFE or code_point is 0x2FFFF or code_point is 0x3FFFE or code_point is 0x3FFFF or code_point is 0x4FFFE or code_point is 0x4FFFF or code_point is 0x5FFFE or code_point is 0x5FFFF or code_point is 0x6FFFE or code_point is 0x6FFFF or code_point is 0x7FFFE or code_point is 0x7FFFF or code_point is 0x8FFFE or code_point is 0x8FFFF or code_point is 0x9FFFE or code_point is 0x9FFFF or code_point is 0xAFFFE or code_point is 0xAFFFF or code_point is 0xBFFFE or code_point is 0xBFFFF or code_point is 0xCFFFE or code_point is 0xCFFFF or code_point is 0xDFFFE or code_point is 0xDFFFF or code_point is 0xEFFFE or code_point is 0xEFFFF or code_point is 0xFFFFE or code_point is 0xFFFFF or code_point is 0x10FFFE or code_point is 0x10FFFF + parse_error() + return from_code_point code_point + return else for i in [0...31] if alnum.indexOf(txt.charAt(cur + i)) is -1 @@ -1870,33 +4483,45 @@ 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 + txt = args.html + cur = 0 + doc = new Node TYPE_TAG, name: 'html', namespace: NS_HTML + open_els = [] + afe = [] # active formatting elements + template_ins_modes = [] + ins_mode = ins_mode_initial + original_ins_mode = ins_mode # TODO check spec + flag_scripting = args.scripting ? true # TODO might need an extra flag to get

bar", - expected: 'text:"foo",tag:"span",{},[text:"bar"]' -test_parser name: "open tag with attributes", \ - html: "foobar", - expected: 'text:"foo",tag:"span",{"style":"foo: bar","title":"hi"},[text:"bar"]' -test_parser name: "open tag with attributes of various quotings", \ - html: "foobar", - expected: 'text:"foo",tag:"span",{"abc":"def","g":"hij","klm":"nopqrstuv\\"","autofocus":""},[text:"bar"]' -test_parser name: "attribute entity exceptions dq", \ - html: "foobar", - expected: 'text:"foo",tag:"a",{"href":"foo?t=1&=2&o=3<=foo"},[text:"bar"]' -test_parser name: "attribute entity exceptions sq", \ - html: "foo bar", - expected: 'text:"foo",tag:"a",{"href":"foo?t=1&=2&o=3<=foo"},[text:"bar"]' -test_parser name: "attribute entity exceptions uq", \ - html: "foo bar", - expected: 'text:"foo",tag:"a",{"href":"foo?t=1&=2&o=3<=foo"},[text:"bar"]' -test_parser name: "matching closing tags", \ - html: "foo hi

foo

bar", - expected: 'text:"foo",tag:"a",{"href":"hi"},[text:"hi"],tag:"div",{},[text:"1",tag:"div",{},[text:"foo"],text:"2"],text:"bar"' -test_parser name: "missing closing tag inside", \ - html: "foo

barbaz

qux", - expected: 'text:"foo",tag:"div",{},[text:"bar",tag:"span",{},[text:"baz"]],text:"qux"' -test_parser name: "mis-matched closing tags", \ - html: "12

3456

78", - expected: 'tag:"span",{},[text:"12",tag:"div",{},[text:"3456"],text:"78"]' -test_parser name: "mis-matched formatting elements", \ - html: "1234567890", - expected: 'text:"12",tag:"b",{},[text:"34",tag:"i",{},[text:"56"]],tag:"i",{},[text:"78"],text:"90"' -test_parser name: "8.2.8.1 Misnested tags: ", \ - html: '

12345

', - expected: 'tag:"p",{},[text:"1",tag:"b",{},[text:"2",tag:"i",{},[text:"3"]],tag:"i",{},[text:"4"],text:"5"]' -test_parser name: "8.2.8.2 Misnested tags:

", \ - html: '1

', - expected: 'tag:"b",{},[text:"1"],tag:"p",{},[tag:"b",{},[text:"2"],text:"3"]' -test_parser name: "crazy formatting elements test", \ - html: "~~first~~second", - # chrome does this: expected: 'tag:"b",{},[tag:"i",{},[tag:"a",{},[tag:"s",{},[tag:"tt",{},[]]],text:"second"]],tag:"a",{},[tag:"s",{},[tag:"tt",{},[tag:"div",{},[tag:"b",{},[],text:"first"]]]]' - # firefox does this: - expected: 'tag:"b",{},[tag:"i",{},[tag:"a",{},[tag:"s",{},[tag:"tt",{},[]]]]],tag:"a",{},[tag:"s",{},[tag:"tt",{},[tag:"div",{},[tag:"b",{},[],text:"first"]]]],text:"second"' -# tests from https://github.com/html5lib/html5lib-tests/blob/master/tree-construction/adoption01.dat -test_parser name: "html5lib aaa 1", \ - html: '

', - expected: 'tag:"a",{},[],tag:"p",{},[tag:"a",{},[]]' -test_parser name: "html5lib aaa 2", \ - html: '1

', - expected: 'tag:"a",{},[text:"1"],tag:"p",{},[tag:"a",{},[text:"2"],text:"3"]' -test_parser name: "html5lib aaa 3", \ - html: '1', - expected: 'tag:"a",{},[text:"1"],tag:"button",{},[tag:"a",{},[text:"2"],text:"3"]' -test_parser name: "html5lib aaa 4", \ - html: '123', - expected: 'tag:"a",{},[text:"1",tag:"b",{},[text:"2"]],tag:"b",{},[text:"3"]' -test_parser name: "html5lib aaa 5 (two divs deep)", \ - html: '1

', - expected: 'tag:"a",{},[text:"1"],tag:"div",{},[tag:"a",{},[text:"2"],tag:"div",{},[tag:"a",{},[text:"3"],text:"4"],text:"5"]' -test_parser name: "html5lib aaa 6 (foster parenting)", \ - html: '1

', - expected: 'tag:"a",{},[text:"1"],tag:"p",{},[tag:"a",{},[text:"2"],text:"3"],tag:"table",{},[]' -test_parser name: "html5lib aaa 7 (aaa, eof) 1", \ - html: '', - expected: 'tag:"b",{},[tag:"b",{},[tag:"a",{},[],tag:"p",{},[tag:"a",{},[]]]]' -test_parser name: "html5lib aaa 8 (aaa, eof) 2", \ - html: '
', - expected: 'tag:"b",{},[tag:"a",{},[tag:"b",{},[]],tag:"b",{},[tag:"p",{},[tag:"a",{},[]]]]' -test_parser name: "html5lib aaa 9 (aaa, eof) 3", \ - html: '
', - expected: 'tag:"a",{},[tag:"b",{},[tag:"b",{},[]]],tag:"b",{},[tag:"b",{},[tag:"p",{},[tag:"a",{},[]]]]' -test_parser name: "html5lib aaa 10 (formatting, nesting, attrs, aaa)", \ - html: '
123
45', - expected: 'tag:"p",{},[text:"1",tag:"s",{"id":"A"},[text:"2",tag:"b",{"id":"B"},[text:"3"]]],tag:"s",{"id":"A"},[tag:"b",{"id":"B"},[text:"4"]],tag:"b",{"id":"B"},[text:"5"]' -test_parser name: "html5lib aaa 11 (table with foster parenting, formatting el and td)", \ - html: '

', - expected: 'tag:"a",{},[text:"1"],tag:"a",{},[text:"3"],tag:"table",{},[tag:"tbody",{},[tag:"tr",{},[tag:"td",{},[text:"2"]]]]' -test_parser name: "html5lib aaa 12 (table with foster parenting, split text)", \ - html: 'AC
B
', - expected: 'text:"AC",tag:"table",{},[tag:"tbody",{},[tag:"tr",{},[tag:"td",{},[text:"B"]]]]' -# TODO implement svg and namespacing -#test_parser name: "html5lib aaa 13 (svg tr input)", \ -# html: '', -# expected: 'tag:"a",{},[svg:"svg",{},[svg:"tr",{},[svg:"input"]]]' -test_parser name: "html5lib aaa 14 (deep ?outer aaa)", \ - html: '
', - expected: 'tag:"div",{},[tag:"a",{},[tag:"b",{},[]],tag:"b",{},[tag:"div",{},[tag:"a",{},[],tag:"div",{},[tag:"a",{},[],tag:"div",{},[tag:"a",{},[],tag:"div",{},[tag:"a",{},[],tag:"div",{},[tag:"a",{},[],tag:"div",{},[tag:"a",{},[],tag:"div",{},[tag:"a",{},[],tag:"div",{},[tag:"a",{},[tag:"div",{},[tag:"div",{},[]]]]]]]]]]]]]' -test_parser name: "html5lib aaa 15 (deep ?inner aaa)", \ - html: '
', - expected: 'tag:"div",{},[tag:"a",{},[tag:"b",{},[tag:"u",{},[tag:"i",{},[tag:"code",{},[]]]]],tag:"u",{},[tag:"i",{},[tag:"code",{},[tag:"div",{},[tag:"a",{},[]]]]]]' -test_parser name: "html5lib aaa 16 (correctly nested 4b)", \ - html: 'xy', - expected: 'tag:"b",{},[tag:"b",{},[tag:"b",{},[tag:"b",{},[text:"x"]]]],text:"y"' -test_parser name: "html5lib aaa 17 (formatting, implied /p, noah's ark)", \ - html: ' x', - expected: 'tag:"p",{},[tag:"b",{},[tag:"b",{},[tag:"b",{},[tag:"b",{},[]]]]],tag:"p",{},[tag:"b",{},[tag:"b",{},[tag:"b",{},[text:"x"]]]]' -test_parser name: "variation on html5lib aaa 17 (with attributes in various orders)", \ - html: ' x', - expected: 'tag:"p",{},[tag:"b",{"c":"d","e":"f"},[tag:"b",{"c":"d","e":"f"},[tag:"b",{"c":"d","e":"f"},[tag:"b",{"c":"d","e":"f"},[]]]]],tag:"p",{},[tag:"b",{"c":"d","e":"f"},[tag:"b",{"c":"d","e":"f"},[tag:"b",{"c":"d","e":"f"},[text:"x"]]]]' -test_parser name: "junk after attribute close-quote", \ - html: ' foo x', - expected: 'tag:"p",{},[tag:"b",{",":"","c":"d","e":"f"},[text:"foo"]],tag:"p",{},[tag:"b",{",":"","c":"d","e":"f"},[text:"x"]]' + +module.exports.parse_html = parse_html +module.exports.debug_log_reset = debug_log_reset +module.exports.debug_log_each = debug_log_each +module.exports.TYPE_TAG = TYPE_TAG +module.exports.TYPE_TEXT = TYPE_TEXT +module.exports.TYPE_COMMENT = TYPE_COMMENT +module.exports.TYPE_DOCTYPE = TYPE_DOCTYPE +module.exports.NS_HTML = NS_HTML +module.exports.NS_MATHML = NS_MATHML +module.exports.NS_SVG = NS_SVG