///|
/// Linkify URL and email text inside a DOM subtree.
///
/// The transform mutates and returns `node`. By default it skips existing
/// anchors and whitespace-preserving tags (`code`, `pre`, `script`, `style`,
/// and `textarea`), while still processing normal children such as template
/// contents.
pub fn linkify_dom(
node : @dom.Node,
config? : LinkifyConfig,
skip_tags? : Array[String],
) -> @dom.Node {
let config = config.unwrap_or_else(() => LinkifyConfig())
let skip_tags = skip_tags.unwrap_or_else(() => linkify_default_dom_skip_tags())
let skip_set = linkify_dom_skip_tag_set(skip_tags)
match node.kind {
Text => {
let wrapper = @dom.fragment(children=[node])
linkify_dom_children(
wrapper,
config,
skip_set,
false,
then_template=false,
)
if wrapper.children.length() == 1 {
let only = wrapper.children[0]
ignore(wrapper.children.remove(0))
only.parent = None
only
} else {
wrapper
}
}
Document | Fragment | Element => {
let skip = node.kind == Element &&
skip_set.contains(@syn.lower_ascii(node.name))
linkify_dom_children(node, config, skip_set, skip, then_template=true)
node
}
Comment | Doctype => node
}
}
///|
/// Return the default element names whose descendants are not linkified.
///
/// The list includes existing anchors and text contexts where automatic link
/// insertion would change code, raw text, or whitespace-preserving content.
pub fn linkify_default_dom_skip_tags() -> Array[String] {
["a", "code", "pre", "script", "style", "textarea"]
}
///|
fn linkify_dom_skip_tag_set(tags : Array[String]) -> Set[String] {
let out : Set[String] = Default::default()
for tag in tags {
let normalized = @syn.lower_ascii(tag.trim())
if normalized != "" {
out.add(normalized)
}
}
out
}
///|
/// Explicit-stack frame for `linkify_dom_children`: the children of `parent`
/// are scanned from `index` with the inherited `skip` state. When
/// `then_template` is set, `parent`'s template contents are walked (with the
/// same skip state) after its regular children.
priv struct LinkifyDomFrame {
parent : @dom.Node
skip : Bool
then_template : Bool
mut index : Int32
}
///|
/// Linkify text below `root`, then (when `then_template` is set) inside its
/// template contents. Element children are walked children-first, then their
/// template contents. Uses an explicit frame stack instead of recursion so
/// deeply nested trees cannot overflow the call stack.
fn linkify_dom_children(
root : @dom.Node,
config : LinkifyConfig,
skip_tags : Set[String],
skip_linkify : Bool,
then_template~ : Bool,
) -> Unit {
let stack : Array[LinkifyDomFrame] = [
{ parent: root, skip: skip_linkify, then_template, index: 0, },
]
while stack.last() is Some(frame) {
let parent = frame.parent
let index = frame.index
if index >= parent.children.length() {
ignore(stack.pop())
if frame.then_template && parent.template_contents is Some(contents) {
stack.push({
parent: contents,
skip: frame.skip,
then_template: false,
index: 0,
})
}
continue
}
let child = parent.children[index]
match child.kind {
Text if !frame.skip && child.data != "" => {
let inserted = linkify_text_child_at(parent, index, child, config)
if inserted > 0 {
frame.index = index + inserted
} else {
frame.index = index + 1
}
}
Document | Fragment => {
frame.index = index + 1
stack.push({
parent: child,
skip: frame.skip,
then_template: false,
index: 0,
})
}
Element => {
let child_skip = frame.skip ||
skip_tags.contains(@syn.lower_ascii(child.name))
frame.index = index + 1
stack.push({
parent: child,
skip: child_skip,
then_template: true,
index: 0,
})
}
_ => frame.index = index + 1
}
}
}
///|
fn linkify_text_child_at(
parent : @dom.Node,
index : Int32,
child : @dom.Node,
config : LinkifyConfig,
) -> Int32 {
let matches = find_links_with_config(child.data, config)
if matches.is_empty() {
return 0
}
linkify_text_child_at_matches(parent, index, child, matches)
}
///|
/// Replace one text child with text and anchor nodes for precomputed matches.
///
/// `matches` must use UTF-16 offsets into `child.data`. The helper detaches the
/// original child, inserts replacement nodes at `index`, preserves the parent's
/// namespace for generated anchors, and returns the number of inserted nodes.
fn linkify_text_child_at_matches(
parent : @dom.Node,
index : Int32,
child : @dom.Node,
matches : Array[LinkMatch],
) -> Int32 {
if matches.is_empty() {
return 0
}
let replacements : Array[@dom.Node] = []
let mut cursor = 0
for m in matches {
if m.start > cursor {
match child.data.get_view(start=cursor, end=m.start) {
Some(prefix) if !prefix.is_empty() =>
replacements.push(@dom.text(prefix))
_ => ()
}
}
let ns = parent.ns.unwrap_or("html")
let anchor = @dom.element(
"a",
attrs={ "href": Some(m.href) },
children=[@dom.text(m.text)],
ns~,
)
replacements.push(anchor)
cursor = m.end
}
if cursor < child.data.length() {
match child.data.get_view(start=cursor, end=child.data.length()) {
Some(tail) if !tail.is_empty() => replacements.push(@dom.text(tail))
_ => ()
}
}
ignore(parent.children.remove(index))
child.parent = None
let mut inserted = 0
for replacement in replacements {
replacement.parent = Some(parent)
parent.children.insert(index + inserted, replacement)
inserted += 1
}
inserted
}