getLink()function

getLink(href: Nullish<PossibleLink>, url?: ImmutableURL, root: ImmutableURL | undefined = url): ImmutableURI | undefined

Param	Type
`href`	`Nullish<PossibleLink>`	The link to resolve. Strings are classified by shape; `URL` instances pass through. required
`url`	`ImmutableURL`	The current page URL — base for relative refs and scheme-prefixed URIs.
`.href`	`URLString`	A URL string has a protocol and a //. required readonly
`.origin`	`URLString`	A URL string has a protocol and a //. required readonly
`.pathname`	`AbsolutePath`	Absolute path string starting with a / slash. required readonly
`root`	`ImmutableURL`	The site root URL — base for absolute paths. Defaults to `url`. Defaults to `url`
`.href`	`URLString`	A URL string has a protocol and a //. required readonly
`.origin`	`URLString`	A URL string has a protocol and a //. required readonly
`.pathname`	`AbsolutePath`	Absolute path string starting with a / slash. required readonly

Return
`ImmutableURI \| undefined`	An absolute URI object, or `undefined` if `link` is missing, not a string/URL, or cannot be resolved.

Resolve a possible link to an absolute URI string, or return undefined if resolution fails.

Classification:

URL instance → returns its .href directly.
Absolute path (single leading /, e.g. /schema) → resolved against root with a dot prefix so the resolution honors root's subfolder — /schema against https://x.com/app/ becomes https://x.com/app/schema, not https://x.com/schema.
Anything else — relative ref (./foo, ../foo, foo, #anchor, ?q), protocol-relative URL (//host/path), or scheme-prefixed URI (mailto:a@b, tel:123, https://x.com/y) → fed to getBasedURI() with url as the base. Self-contained URIs ignore the base (their own scheme makes them absolute); protocol-relative and relative refs resolve against the page URL.

Defaults:

root defaults to url, so passing only url makes absolute paths resolve under the page URL's directory (same coordinate space the page lives in).
When both are omitted, all branches return undefined (no base to resolve against).

Bases are passed through to getURL() / getBasedURI() lazily — neither url nor root is materialised into a normalised base until the chosen branch needs it.

Examples

getLink("/schema", pageURL, siteRoot) // → "https://x.com/app/schema" when siteRoot is "https://x.com/app/"

getLink("./db", new URL("https://x.com/app/schema/")) // → "https://x.com/app/schema/db"

getLink("mailto:a@b") // → "mailto:a@b"