parse comments (13 passing tests)

diff --git a/parse-html.coffee b/parse-html.coffee
index 385a2a0..4c1d3cc 100644 (file)
--- a/parse-html.coffee
+++ b/parse-html.coffee
@@ -164,7 +164,7 @@ new_element = (name) ->
 new_text_node = (txt) ->
        return new Node TYPE_TEXT, text: txt
 new_character_token = new_text_node
-new_comment_node = (txt) ->
+new_comment_token = (txt) ->
        return new Node TYPE_COMMENT, text: txt
 new_eof_token = ->
        return new Node TYPE_EOF
@@ -373,6 +373,7 @@ parse_html = (txt, parse_error_cb = null) ->
        pending_table_character_tokens = null
        head_element_pointer = null
        flag_fragment_parsing = null
+       context_element = null
 
        stop_parsing = ->
                flag_parsing = false
@@ -623,6 +624,14 @@ parse_html = (txt, parse_error_cb = null) ->
                        node = open_els[node_i]
                        # 19. Return to the step labeled loop.
 
+       # 8.2.3.2
+
+       # http://www.w3.org/TR/html5/syntax.html#adjusted-current-node
+       adjusted_current_node = ->
+               if open_els.length is 1 and flag_fragment_parsing
+                       return context_element
+               return open_els[0]
+
        # http://www.w3.org/TR/html5/syntax.html#reconstruct-the-active-formatting-elements
        # this implementation is structured (mostly) as described at the link above.
        # capitalized comments are the "labels" described at the link above.
@@ -1306,12 +1315,12 @@ parse_html = (txt, parse_error_cb = null) ->
                        parse_error()
                        return
                ins_mode_in_head_else t
-       
+
        # 8.2.5.4.5 http://www.w3.org/TR/html5/syntax.html#parsing-main-inheadnoscript
        ins_mode_in_head_noscript = (t) ->
                # FIXME ?fixfull
                console.log "ins_mode_in_head_noscript unimplemented"
-       
+
        # 8.2.5.4.6 http://www.w3.org/TR/html5/syntax.html#the-after-head-insertion-mode
        ins_mode_after_head_else = (t) ->
                body_tok = new_open_tag 'body'
@@ -2276,6 +2285,7 @@ parse_html = (txt, parse_error_cb = null) ->
                                tok_state = tok_state_end_tag_open
                        when '?'
                                parse_error()
+                               tok_cur_tag = new_comment_token '?'
                                tok_state = tok_state_bogus_comment
                        else
                                if lc_alpha.indexOf(c) > -1
@@ -2310,6 +2320,7 @@ parse_html = (txt, parse_error_cb = null) ->
                                        tok_state = tok_state_tag_name
                                else
                                        parse_error()
+                                       tok_cur_tag = new_comment_token '/'
                                        tok_state = tok_state_bogus_comment
                return null
 
@@ -2677,6 +2688,174 @@ parse_html = (txt, parse_error_cb = null) ->
                                cur -= 1 # we didn't handle that char
                return null
 
+       # 8.2.4.44 http://www.w3.org/TR/html5/syntax.html#bogus-comment-state
+       # WARNING: put a comment token in tok_cur_tag before setting this state
+       tok_state_bogus_comment = ->
+               next_gt = txt.indexOf '>', cur
+               if next_gt is -1
+                       val = txt.substr cur
+                       cur = txt.length
+               else
+                       val = txt.substr cur, (next_gt - cur)
+                       cur = next_gt + 1
+               val = val.replace "\u0000", "\ufffd"
+               tok_cur_tag.text += val
+               tok_state = tok_state_data
+               return tok_cur_tag
+
+       # 8.2.4.45 http://www.w3.org/TR/html5/syntax.html#markup-declaration-open-state
+       tok_state_markup_declaration_open = ->
+               if txt.substr(cur, 2) is '--'
+                       cur += 2
+                       tok_cur_tag = new_comment_token ''
+                       tok_state = tok_state_comment_start
+                       return
+               if txt.substr(cur, 7).toLowerCase() is 'doctype'
+                       cur += 7
+                       tok_state = tok_state_doctype
+                       return
+               acn = adjusted_current_node()
+               if acn and acn.namespace isnt NS_HTML and text.substr(cur, 7) is '[CDATA['
+                       cur += 7
+                       tok_state = tok_state_cdata_section
+                       return
+               # Otherwise
+               parse_errer()
+               tok_cur_tag = new_comment_token '!' # TODO test ("!" right?)
+               tok_state = tok_state_bogus_comment
+               return
+
+       # 8.2.4.46 http://www.w3.org/TR/html5/syntax.html#comment-start-state
+       tok_state_comment_start = ->
+               switch c = txt.charAt(cur++)
+                       when '-'
+                               tok_state = tok_state_comment_start_dash
+                       when "\u0000"
+                               parse_error()
+                               return new_character_token "\ufffd"
+                       when '>'
+                               parse_error()
+                               tok_state = tok_state_data
+                               return tok_cur_tag
+                       when '' # EOF
+                               parse_error()
+                               tok_state = tok_state_data
+                               cur -= 1 # Reconsume
+                               return tok_cur_tag
+                       else
+                               tok_cur_tag.text += c
+               return null
+
+       # 8.2.4.47 http://www.w3.org/TR/html5/syntax.html#comment-start-dash-state
+       tok_state_comment_start_dash = ->
+               switch c = txt.charAt(cur++)
+                       when '-'
+                               tok_state = tok_state_comment_end
+                       when "\u0000"
+                               parse_error()
+                               tok_cur_tag.text += "-\ufffd"
+                               tok_state = tok_state_comment
+                       when '>'
+                               parse_error()
+                               tok_state = tok_state_data
+                               return tok_cur_tag
+                       when '' # EOF
+                               parse_error()
+                               tok_state = tok_state_data
+                               cur -= 1 # Reconsume
+                               return tok_cur_tag
+                       else
+                               tok_cur_tag.text += "-#{c}"
+                               tok_state = tok_state_comment
+               return null
+
+       # 8.2.4.48 http://www.w3.org/TR/html5/syntax.html#comment-state
+       tok_state_comment = ->
+               switch c = txt.charAt(cur++)
+                       when '-'
+                               tok_state = tok_state_comment_end_dash
+                       when "\u0000"
+                               parse_error()
+                               tok_cur_tag.text += "\ufffd"
+                       when '' # EOF
+                               parse_error()
+                               tok_state = tok_state_data
+                               cur -= 1 # Reconsume
+                               return tok_cur_tag
+                       else
+                               tok_cur_tag.text += c
+               return null
+
+       # 8.2.4.49 http://www.w3.org/TR/html5/syntax.html#comment-end-dash-state
+       tok_state_comment_end_dash = ->
+               switch c = txt.charAt(cur++)
+                       when '-'
+                               tok_state = tok_state_comment_end
+                       when "\u0000"
+                               parse_error()
+                               tok_cur_tag.text += "-\ufffd"
+                               tok_state = tok_state_comment
+                       when '' # EOF
+                               parse_error()
+                               tok_state = tok_state_data
+                               cur -= 1 # Reconsume
+                               return tok_cur_tag
+                       else
+                               tok_cur_tag.text += "-#{c}"
+                               tok_state = tok_state_comment
+               return null
+
+       # 8.2.4.50 http://www.w3.org/TR/html5/syntax.html#comment-end-state
+       tok_state_comment_end = ->
+               switch c = txt.charAt(cur++)
+                       when '>'
+                               tok_state = tok_state_data
+                               return tok_cur_tag
+                       when "\u0000"
+                               parse_error()
+                               tok_cur_tag.text += "--\ufffd"
+                               tok_state = tok_state_comment
+                       when '!'
+                               parse_error()
+                               tok_state = tok_state_comment_end_bang
+                       when '-'
+                               parse_error()
+                               tok_cur_tag.text += '-'
+                       when '' # EOF
+                               parse_error()
+                               tok_state = tok_state_data
+                               cur -= 1 # Reconsume
+                               return tok_cur_tag
+                       else
+                               parse_error()
+                               tok_cur_tag.text += "--#{c}"
+                               tok_state = tok_state_comment
+               return null
+
+       # 8.2.4.51 http://www.w3.org/TR/html5/syntax.html#comment-end-bang-state
+       tok_state_comment_end_bang = ->
+               switch c = txt.charAt(cur++)
+                       when '-'
+                               tok_cur_tag.text += "--!#{c}"
+                               tok_state = tok_state_comment_end_dash
+                       when '>'
+                               tok_state = tok_state_data
+                               return tok_cur_tag
+                       when "\u0000"
+                               parse_error()
+                               tok_cur_tag.text += "--!\ufffd"
+                               tok_state = tok_state_comment
+                       when '' # EOF
+                               parse_error()
+                               tok_state = tok_state_data
+                               cur -= 1 # Reconsume
+                               return tok_cur_tag
+                       else
+                               tok_cur_tag.text += "--!#{c}"
+                               tok_state = tok_state_comment
+               return null
+
+
        # 8.2.4.69 http://www.w3.org/TR/html5/syntax.html#consume-a-character-reference
        # Don't set this as a state, just call it
        # returns a string (NOT a text node)
@@ -2770,6 +2949,7 @@ parse_html = (txt, parse_error_cb = null) ->
        pending_table_character_tokens = []
        head_element_pointer = null
        flag_fragment_parsing = false # parser originally created as part of the html fragment parsing algorithm (fragment case)
+       context_element = null # FIXME initialize from args.fragment
 
        # tokenizer initialization
        tok_state = tok_state_data

diff --git a/test.coffee b/test.coffee
index 1c45b68..baea4e7 100644 (file)
--- a/test.coffee
+++ b/test.coffee
@@ -7807,7 +7807,7 @@ serialize_els = (els, prefix = '| ') ->
                        when wheic.TYPE_TEXT
                                ret += "#{prefix}\"#{el.text}\"\n"
                        when wheic.TYPE_COMMENT
-                               ret += "#{prefix}<!--#{el.text}-->\n"
+                               ret += "#{prefix}<!-- #{el.text} -->\n"
                        when wheic.TYPE_DOCTYPE
                                ret += "#{prefix}<!DOCTYPE#{el.name}>\n" # FIXME add ids
                        else