@namespace = args.namespace ? NS_HTML
@parent = args.parent ? null
@token = args.token ? null
+ @flags = args.flags ? {}
if args.id?
@id = "#{args.id}+"
else
@token.flag 'did_self_close'
else
@flag 'did_self_close', true
- flag: ->
- # fixfull
+ 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
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] = el.namespace
+is_html_integration = (el) -> # DON'T PASS A TOKEN
+ if el.namespace is NS_MATHML and 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
+}
+
+# FIXME namespacify
foster_parenting_targets = {
table: true
tbody: true
tr: true
}
+# FIXME namespacify
# all html I presume
end_tag_implied = {
dd: true
el_is_special = (e) ->
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'
+}
+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
+ return
+
# decode_named_char_ref()
#
# The list of named character references is _huge_ so ask the browser to decode
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
+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_insertion_modes = null
- insertion_mode = null
- original_insertion_mode = null
+ 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_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}"
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)
# 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.
# 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
+ 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
+ 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
+ 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
+ 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
+ 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
+ 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
+ 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.
# 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
+ 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
+ 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
+ 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
+ ins_mode = ins_mode_in_frameset
return
# 16. If node is an html element, run these substeps:
if node.name is 'html'
# 2. Otherwise, the head element pointer is not null,
# switch the insertion mode to "after head" and abort these
# steps.
- insertion_mode = ins_mode_after_head
+ 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.
# 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
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 (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
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
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
# 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.
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) ->
+ 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]
parse_generic_raw_text = (t) ->
insert_html_element t
tok_state = tok_state_rawtext
- original_insertion_mode = insertion_mode
- insertion_mode = ins_mode_text
+ 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_insertion_mode = insertion_mode
- insertion_mode = ins_mode_text
+ 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
# FIXME check identifiers, set quirks, etc
# fixfull
doc.children.push t
- insertion_mode = ins_mode_before_html
+ ins_mode = ins_mode_before_html
return
# Anything else
#fixfull (iframe, quirks)
- insertion_mode = ins_mode_before_html
- insertion_mode t # reprocess the token
+ 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
doc.children.push el
open_els.unshift(el)
# fixfull (big paragraph in spec about manifest, fragment, urls, etc)
- insertion_mode = ins_mode_before_head
+ 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'
doc.children.push el
open_els.unshift el
# ?fixfull browsing context
- insertion_mode = ins_mode_before_head
- insertion_mode t
+ 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
if t.type is TYPE_START_TAG and t.name is 'head'
el = insert_html_element t
head_element_pointer = el
- insertion_mode = ins_mode_in_head
+ ins_mode = ins_mode_in_head
if t.type is TYPE_END_TAG
if t.name is 'head' or t.name is 'body' or t.name is 'html' or t.name is 'br'
# fall through to Anything else below
head_tok = new_open_tag 'head'
el = insert_html_element head_tok
head_element_pointer = el
- insertion_mode = ins_mode_in_head
- insertion_mode t # reprocess current token
+ 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
- insertion_mode = ins_mode_after_head
- insertion_mode t
+ 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
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'))
+ if t.type is TYPE_START_TAG and ((t.name is 'noscript' and flag_scripting) or t.name is 'noframes' or t.name is 'style')
parse_generic_raw_text t
return
if t.type is TYPE_START_TAG and t.name is 'noscript' and flag_scripting is false
insert_html_element t
- insertion_mode = ins_mode_in_head_noscript # FIXME implement
+ 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 # FIXME implement
+ 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_insertion_mode = insertion_mode # make sure orig... is defined
- insertion_mode = ins_mode_text # FIXME implement
+ 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
- insertion_mode = ins_mode_after_head
+ 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
insert_html_element t
afe_push_marker()
flag_frameset_ok = false
- insertion_mode = ins_mode_in_template
- template_insertion_modes.unshift ins_mode_in_template # FIXME implement
+ 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()
if el.name is 'template'
break
clear_afe_to_marker()
- template_insertion_modes.shift()
- reset_insertion_mode()
+ template_ins_modes.shift()
+ reset_ins_mode()
else
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) ->
- # FIXME ?fixfull
- console.log "ins_mode_in_head_noscript unimplemented"
+ 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
- insertion_mode = ins_mode_in_body
- insertion_mode t # reprocess token
+ ins_mode = ins_mode_in_body
+ process_token t
return
ins_mode_after_head = (t) ->
if is_space_tok t
if t.type is TYPE_START_TAG and t.name is 'body'
insert_html_element t
flag_frameset_ok = false
- insertion_mode = ins_mode_in_body
+ ins_mode = ins_mode_in_body
return
if t.type is TYPE_START_TAG and t.name is 'frameset'
insert_html_element t
- insertion_mode = ins_mode_in_frameset
+ 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()
# 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.namespace is NS_HTML and el.name is name
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()
- insert_character t
- else
- reconstruct_active_formatting_elements()
- insert_character t
- flag_frameset_ok = false
- when TYPE_COMMENT
- insert_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.ns is NS_HTML
+ return unless second.name is 'body'
+ return if template_tag_is_open()
+ frameset_ok_flag = 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.ns is NS_HTML
+ return unless second.name is 'body'
+ flag_frameset_ok = false
+ 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'
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 </template> (end tag)
- return ins_mode_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'
+ 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]] is NS_HTML
+ 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()
+ 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
- # TODO stack of template insertion modes thing
- 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
+ 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
+ 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
+ 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
- ins_mode_in_table_else = (t) ->
- parse_error()
- flag_foster_parenting = true # FIXME
- ins_mode_in_body t
- flag_foster_parenting = false
- can_in_table = { # FIXME do this inline like everywhere else
- 'table': true
- 'tbody': true
- 'tfoot': true
- 'thead': true
- 'tr': true
- }
-
# 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
if open_els[0].name is 'script'
open_els[0].flag 'already started', true
open_els.shift()
- insertion_mode = original_insertion_mode
- insertion_mode t
+ ins_mode = original_ins_mode
+ process_token t
return
if t.type is TYPE_END_TAG and t.name is 'script'
open_els.shift()
- insertion_mode = original_insertion_mode
+ 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()
- insertion_mode = original_insertion_mode
+ ins_mode = original_ins_mode
return
console.log 'warning: end of ins_mode_text reached'
# 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
+ ins_mode_in_body t
+ flag_foster_parenting = false
+ return
+ can_in_table = { # FIXME do this inline like everywhere else
+ 'table': true
+ 'tbody': true
+ 'tfoot': true
+ 'thead': true
+ 'tr': true
+ }
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
+ 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
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'
el = open_els.shift()
if el.name is 'table'
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 is_input_hidden_tok t
+ unless is_input_hidden_tok t
ins_mode_in_table_else t
else
parse_error()
el = open_els.shift()
if el.name is 'table'
break
- reset_insertion_mode()
+ reset_ins_mode()
else
parse_error
when 'body', 'caption', 'col', 'colgroup', 'html', 'tbody', 'td', 'tfoot', 'th', 'thead', 'tr'
for old in pending_table_character_tokens
ins_mode_table_else old
pending_table_character_tokens = [] # FIXME test (spec doesn't say this)
- insertion_mode = original_insertion_mode
- insertion_mode t
+ 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 el.name is 'caption'
break
clear_afe_to_marker()
- insertion_mode = ins_mode_in_table
+ ins_mode = ins_mode_in_table
else
parse_error()
# fragment case
if el.name is 'caption'
break
clear_afe_to_marker()
- insertion_mode = ins_mode_in_table
- insertion_mode t
+ 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')
if t.type is TYPE_END_TAG and t.name is 'colgroup'
if open_els[0].name is 'colgroup'
open_els.shift()
- insertion_mode = ins_mode_in_table
+ ins_mode = ins_mode_in_table
else
parse_error()
return
parse_error()
return
open_els.shift()
- insertion_mode = ins_mode_in_table
- insertion_mode t
+ 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
if t.type is TYPE_START_TAG and t.name is 'tr'
clear_stack_to_table_body_context()
insert_html_element t
- insertion_mode = ins_mode_in_row
+ 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
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
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()
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'
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 is_in_table_scope 'tr'
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 is_in_table_scope 'tr'
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 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
# 8.2.5.4.15 http://www.w3.org/TR/html5/syntax.html#parsing-main-intd
ins_mode_in_cell = (t) ->
if el.name is t.name
break
clear_afe_to_marker()
- insertion_mode = ins_mode_in_row
+ ins_mode = ins_mode_in_row
else
parse_error()
return
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()
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
close_the_cell()
- insertion_mode t
+ process_token t
else
parse_error()
return
el = open_els.shift()
if el.name is 'select'
break
- reset_insertion_mode()
+ reset_ins_mode()
else
parse_error()
return
el = open_els.shift()
if el.name is 'select'
break
- reset_insertion_mode()
+ reset_ins_mode()
# spec says that this is the same as </select> but it doesn't say
# to check scope first
return
el = open_els.shift()
if el.name is 'select'
break
- reset_insertion_mode()
- insertion_mode t
+ 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
el = open_els.shift()
if el.name is 'select'
break
- reset_insertion_mode()
- insertion_mode t
+ 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()
el = open_els.shift()
if el.name is 'select'
break
- reset_insertion_mode()
- insertion_mode t
+ reset_ins_mode()
+ process_token t
return
# Anything else
ins_mode_in_select t
ins_mode_in_head t
return
if t.type is TYPE_START_TAG and (t.name is 'caption' or t.name is 'colgroup' or t.name is 'tbody' or t.name is 'tfoot' or t.name is 'thead')
- template_insertion_modes.shift()
- template_insertion_modes.unshift ins_mode_in_table
- insertion_mode = ins_mode_in_table
- insertion_mode t
+ 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_insertion_modes.shift()
- template_insertion_modes.unshift ins_mode_in_column_group
- insertion_mode = ins_mode_in_column_group
- insertion_mode t
+ 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_insertion_modes.shift()
- template_insertion_modes.unshift ins_mode_in_table_body
- insertion_mode = ins_mode_in_table_body
- insertion_mode t
+ 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_insertion_modes.shift()
- template_insertion_modes.unshift ins_mode_in_row
- insertion_mode = ins_mode_in_row
- insertion_mode t
+ 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_insertion_modes.shift()
- template_insertion_modes.unshift ins_mode_in_body
- insertion_mode = ins_mode_in_body
- insertion_mode t
+ 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()
if el.name is 'template' # fixfull check namespace
break
clear_afe_to_marker()
- template_insertion_modes.shift()
- reset_insertion_mode()
- insertion_mode t
+ 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) ->
return
if t.type is TYPE_END_TAG and t.name is 'html'
# fixfull fragment case
- insertion_mode = ins_mode_after_after_body
+ ins_mode = ins_mode_after_after_body
return
if t.type is TYPE_EOF
stop_parsing()
return
# Anything ELse
parse_error()
- insertion_mode = ins_mode_in_body
- insertion_mode t
+ 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) ->
return # fragment case
open_els.shift()
if flag_fragment_parsing is false and open_els[0].name isnt 'frameset'
- insertion_mode = ins_mode_after_frameset
+ ins_mode = ins_mode_after_frameset
return
if t.type is TYPE_START_TAG and t.name is 'frame'
insert_html_element t
return
# Anything else
parse_error()
- insertion_mode = ins_mode_in_body
+ ins_mode = ins_mode_in_body
return
# 8.2.5.4.23 http://www.w3.org/TR/html5/syntax.html#the-after-after-frameset-insertion-mode
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()
+ 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()
+ cn = open_els[0]
+ if is_mathml_text_integration_point(cn) or is_html_integration(cn) or cn.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
+ if open_els[0].name.toLowerCase() isnt t.name
+ parse_error()
+ for node in open_els
+ 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
+ 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
is_appropriate_end_tag = (t) ->
# spec says to check against "the tag name of the last start tag to
# have been emitted from this tokenizer", but this is only called from
- # the various "raw" states, which I'm pretty sure all push the start
- # token onto open_els. TODO: verify this after the script data states
- # are implemented
+ # 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
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
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 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 is_uc_alpha(c)
- tok_cur_tag.attrs_a[0][0] = c.toLowerCase()
+ tok_cur_tag.attrs_a[0][0] += c.toLowerCase()
else
tok_cur_tag.attrs_a[0][0] += c
return null
tok_state = tok_state_comment_start_dash
when "\u0000"
parse_error()
+ tok_state = tok_state_comment
return new_character_token "\ufffd"
when '>'
parse_error()
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
# 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
+ val = val.replace(new RegExp("\u0000", 'g'), "\ufffd") # fixfull spec doesn't say this
+ val = val.replace(new RegExp("\r\n", 'g'), "\n") # fixfull spec doesn't say this
+ val = val.replace(new RegExp("\r", 'g'), "\n") # fixfull spec doesn't say this
+ 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
# tree constructor initialization
# see comments on TYPE_TAG/etc for the structure of this data
+ txt = args.html
+ cur = 0
doc = new Node TYPE_TAG, name: 'html', namespace: NS_HTML
open_els = []
afe = [] # active formatting elements
- template_insertion_modes = []
- insertion_mode = ins_mode_initial
- original_insertion_mode = insertion_mode # TODO check spec
- flag_scripting = true # TODO might need an extra flag to get <noscript> to parse correctly
+ 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 <noscript> to parse correctly
flag_frameset_ok = true
flag_parsing = true
flag_foster_parenting = false
tok_state = tok_state_data
# proccess input
+ # http://www.w3.org/TR/html5/syntax.html#tree-construction
while flag_parsing
t = tok_state()
if t?
- insertion_mode t
+ process_token t
# fixfull parse error if has self-closing flag, but it wasn't acknolwedged
return doc.children
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
Got questions, comments, patches, etc.? Contact Jason Woofenden