HTML5::Parser

Section 12.2.5

add_child adds a child node n to the top element, and pushes n onto the stack of open elements if it is an element node.

add_element adds a child element based on the current token.

Section 12.2.4.3

add_text adds text to the preceding node if it is a Text Node, or else it calls add_child with a new Text Node

Section 12.2.4.2

Section 12.2.4.3.

clear_stack_to_context pops elements off the stack of open elements until a scope-defined element is found.

doc is the document root element

element_in_scope is like pop_until, except that it doesn't modify the stack of open elements

foster_parent adds a child node according to the foster parenting rules. Section 12.2.6.1, "foster parenting"

fragment is whether the parser is parsing an HTML fragment.

generate_implied_end_tags pops nodes off the stack of open elements as long as the top node has a tag name of dd, dt, li, optgroup, option, p, rb, rp, rt or rtc. If exceptions are specified, nodes with that name will not be popped off.

Self-closing tags like

---

are treated as start tags, except that has_self_closing_token is set while they are being processed

Self-closing tags like

---

are treated as start tags, except that has_self_closing_token is set while they are being processed

TODO: this is a fairly literal line-by-line translation of that algorithm. Once the code successfully parses the comprehensive test suite, we should refactor this code to be more idiomatic.

performs the "any other end tag" algorithm for in_body_im. "Any other end tag" handling from 12.2.6.5 The rules for parsing tokens in foreign content https://html.spec.whatwg.org/multipage/syntax.html#parsing-main-inforeign

Section 12.2.6

index_of_element_in_scope returns the index of in @oe of the highest element whose tag is in match_tags that is in scope. If no matching element is in scope, it returns -1

runs the current token through the parsing routines until it is consumed.

parse_generic_raw_text_elements implements the generic raw text element parsing algorithm defined in 12.2.6.2. https://html.spec.whatwg.org/multipage/parsing.html#parsing-elements-that-contain-only-text TODO: Since both RAWTEXT and RCDATA states are treated as tokenizer's part officially, need to make tokenizer consider both states.

parses a token as thoug it had appeared in the parser's input

pop_until pops the stack of open elements at the highest element whose tag is in matchTags, provided there is no higher element in the scope's stop tags (as defined in section 12.2.4.2). It returns whether or not there was such an element. If there was not, popUntil leaves the stack unchanged.

For example, the set of stop tags for table scope is: "html", "table". If the stack was: ["html", "body", "font", "table", "b", "i", "u"] then pop_until(tableScope, "font") would return false, but pop_until(tableScope, "i") would return true and the stack would become: ["html", "body", "font", "table", "b"]

If an element's tag is in both the stop tags and match_tags, then the stack will be popped and the function returns true (provided, of course, there was no higher element in the stack that was also in the stop tags). For example, pop_until(tableScope, "table") returns true and leaves: ["html", "body", "font"]

Section 12.2.4.3.

Section 12.2.4.1, "reset the insertion mode".

set_original_im sets the insertion mode to return to after completing a text or inTableText insertion mode. Section 12.2.4.1, "using the rules for".

should_foster_parent returns whether the next node to be added should be foster parented.