# This file implements a parser for html snippets, meant to be used by a
-# WYSIWYG editor. Hence it does not attempt to parse doctypes, <html>, <head>
-# or <body> 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
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]
TYPE_TEXT = 1 # "text"
return t.type is TYPE_TEXT && t.text.length is 1 and space_chars.indexOf(t.text) > -1
is_input_hidden_tok = (t) ->
- return unless t.type is TYPE_START_TAG
- for a of t.attrs_a
+ 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
# 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 = {
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,
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
+ return mathml_text_integration[el.name] is 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
+ 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'
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
- tfoot: true
- thead: true
- tr: true
+ table: NS_HTML
+ tbody: NS_HTML
+ tfoot: NS_HTML
+ thead: NS_HTML
+ tr: NS_HTML
}
-# FIXME namespacify
-# 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) ->
# 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) ->
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 and 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
open_els.shift()
return
clear_to_table_body_stopers = {
- 'tbody': true
- 'tfoot': true
- 'thead': true
- 'template': true
- 'html': true
+ 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]?
+ if clear_to_table_body_stopers[open_els[0].name] is open_els[0].namespace
break
open_els.shift()
return
# 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.
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'
+ 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.
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
+ 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'
+ 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'
+ 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'
+ 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'
+ 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'
+ 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
+ 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
+ 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'
+ 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'
+ 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)
if head_element_pointer is null
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)
# http://www.w3.org/TR/html5/syntax.html#close-a-p-element
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
# 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.
# 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
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
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
parse_error()
loop
el = open_els.shift()
- if el.name is 'template'
+ if el.name is 'template' and el.namespace is NS_HTML
break
clear_afe_to_marker()
template_ins_modes.shift()
if t.type is TYPE_DOCTYPE
parse_error()
return
- if t.type is TYPE_START_TAG
+ 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 (t.type is TYPE_TEXT and (t.text is "\t" or t.text is "\u000a" or t.text is "\u000c" or t.text is "\u000d" or t.text is ' ')) 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'))
+ 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'
# 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 el, i in open_els
- if el.namespace is NS_HTML and el.name is name
+ 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
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.namespace is NS_HTML
return unless second.name is 'body'
return if template_tag_is_open()
- frameset_ok_flag = false
+ flag_frameset_ok = false
for a of t.attrs_a
second.attrs[a[0]] = a[1] unless second.attrs[a[0]]?
return
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.namespace is NS_HTML
return unless second.name is 'body'
- flag_frameset_ok = false
+ if flag_frameset_ok is false
+ return
if second.parent?
for el, i in second.parent.children
if el is second
stop_parsing()
return
if t.type is TYPE_END_TAG and t.name is 'body'
- unless is_in_scope 'body'
+ unless is_in_scope 'body', NS_HTML
parse_error()
return
ok_tags = {
ins_mode = ins_mode_after_body
return
if t.type is TYPE_END_TAG and t.name is 'html'
- unless is_in_scope 'body'
+ unless is_in_scope 'body', NS_HTML
parse_error()
return
ok_tags = {
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
+ if h_tags[open_els[0].name] is open_els[0].namespace
parse_error()
open_els.shift()
insert_html_element t
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')
+# 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 'rt'
+ 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' # arg is exception
+ 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
return
if t.type is TYPE_EOF
parse_error()
- if open_els[0].name is 'script'
+ 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
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]
+ 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
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_ins_mode()
process_token t
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_ins_mode()
else
- parse_error
+ parse_error()
when 'body', 'caption', 'col', 'colgroup', 'html', 'tbody', 'td', 'tfoot', 'th', 'thead', 'tr'
parse_error()
when 'template'
# 8.2.5.4.10 http://www.w3.org/TR/html5/syntax.html#parsing-main-intabletext
ins_mode_in_table_text = (t) ->
if t.type is TYPE_TEXT and t.text is "\u0000"
- # huh? I thought the tokenizer didn't emit these
+ # from javascript?
parse_error()
return
if t.type is TYPE_TEXT
insert_character old
else
for old in pending_table_character_tokens
- ins_mode_table_else old
- pending_table_character_tokens = [] # FIXME test (spec doesn't say this)
+ 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'
+ 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'
+ if el.name is 'caption' and el.namespace is NS_HTML
break
clear_afe_to_marker()
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 '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'
+ if is_in_table_scope 'caption', NS_HTML
loop
el = open_els.shift()
- if el.name is 'caption'
+ if el.name is 'caption' and el.namespace is NS_HTML
break
clear_afe_to_marker()
ins_mode = ins_mode_in_table
t.acknowledge_self_closing()
return
if t.type is TYPE_END_TAG and t.name is 'colgroup'
- if open_els[0].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
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()
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()
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()
ins_mode = ins_mode_in_table_body
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()
ins_mode = ins_mode_in_table_body
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()
ins_mode = ins_mode_in_table_body
# 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()
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 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()
ins_mode = ins_mode_in_row
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()
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()
process_token t
else
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'
+ 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'
+ 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'
+ 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[1].name is 'optgroup'
- open_els.shift()
- if open_els[0].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'
+ 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'
+ if is_in_select_scope 'select', NS_HTML
loop
el = open_els.shift()
- if el.name is 'select'
+ if el.name is 'select' and el.namespace is NS_HTML
break
reset_ins_mode()
else
parse_error()
loop
el = open_els.shift()
- if el.name is 'select'
+ if el.name is 'select' and el.namespace is NS_HTML
break
reset_ins_mode()
# spec says that this is the same as </select> but it doesn't say
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'
+ if is_in_select_scope 'select', NS_HTML
return
loop
el = open_els.shift()
- if el.name is 'select'
+ if el.name is 'select' and el.namespace is NS_HTML
break
reset_ins_mode()
process_token t
parse_error()
loop
el = open_els.shift()
- if el.name is 'select'
+ if el.name is 'select' and el.namespace is NS_HTML
break
reset_ins_mode()
process_token t
return
loop
el = open_els.shift()
- if el.name is 'select'
+ if el.name is 'select' and el.namespace is NS_HTML
break
reset_ins_mode()
process_token t
parse_error()
loop
el = open_els.shift()
- if el.name is 'template' # fixfull check namespace
+ if el.name is 'template' and el.namespace is NS_HTML
break
clear_afe_to_marker()
template_ins_modes.shift()
insert_html_element t
return
if t.type is TYPE_END_TAG and t.name is 'frameset'
- # TODO ?correct for: "if the current node is the root html element"
if open_els.length is 1
parse_error()
return # fragment case
ins_mode_in_head t
return
if t.type is TYPE_EOF
- # TODO ?correct for: "if the current node is not the root html element"
if open_els.length isnt 1
parse_error()
stop_parsing()
tok_state = tok_state_tag_open
when "\u0000"
parse_error()
- return new_text_node c
+ return new_text_node "\ufffd"
when '' # EOF
return new_eof_token()
else
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()
else
val = txt.substr cur, (next_gt - cur)
cur = next_gt + 1
- val = val.replace "\u0000", "\ufffd"
+ val = val.replace(new RegExp("\u0000", 'g'), "\ufffd")
tok_cur_tag.text += val
tok_state = tok_state_data
return tok_cur_tag
return
# Otherwise
parse_error()
- tok_cur_tag = new_comment_token '!' # TODO test ("!" right?)
+ tok_cur_tag = new_comment_token ''
tok_state = tok_state_bogus_comment
return
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
return
if c is '"'
parse_error()
- tok_cur_tag.public_identifier = '' # FIXME should this go in @attrs or @text?
+ 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 = '' # FIXME should this go in @attrs or @text?
+ tok_cur_tag.public_identifier = ''
tok_state = tok_state_doctype_public_identifier_single_quoted
return
if c is '>'
return
if c is '"'
parse_error()
- tok_cur_tag.public_identifier = '' # FIXME should this go in @attrs or @text?
+ 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 = '' # FIXME should this go in @attrs or @text?
+ tok_cur_tag.public_identifier = ''
tok_state = tok_state_doctype_public_identifier_single_quoted
return
if c is '>'
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
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
# tokenizer initialization
tok_state = tok_state_data
+ # text pre-processing
+ # FIXME http://www.w3.org/TR/html5/syntax.html#preprocessing-the-input-stream
+ txt = txt.replace(new RegExp("\u0000", 'g'), "\ufffd") # fixfull spec doesn't say this
+ txt = txt.replace(new RegExp("\r\n", 'g'), "\n") # fixfull spec doesn't say this
+ txt = txt.replace(new RegExp("\r", 'g'), "\n") # fixfull spec doesn't say this
+
+ if args.name is "plain-text-unsafe.dat #4"
+ console.log "hi"
# proccess input
# http://www.w3.org/TR/html5/syntax.html#tree-construction
while flag_parsing
serialized += t.serialize shallow, show_ids
return serialized
-# TODO export TYPE_*
module.exports.parse_html = parse_html
module.exports.debug_log_reset = debug_log_reset
module.exports.debug_log_each = debug_log_each
Got questions, comments, patches, etc.? Contact Jason Woofenden