feat(richtext-lexical): update validation of custom URLs to include relative and anchor links (#6525)

Updates the regex to allow relative and anchor links as well. Manually
tested all common variations of absolute, relative and anchor links with
a combination

## Type of change

- [x] New feature (non-breaking change which adds functionality)

packages/richtext-lexical/src/field/lexical/utils/url.ts

@@ -14,8 +14,16 @@ export function sanitizeUrl(url: string): string {
 }
 
 // Source: https://stackoverflow.com/a/8234912/2013580
-const urlRegExp =
-  /((([A-Za-z]{3,9}:(?:\/\/)?)(?:[-;:&=+$,\w]+@)?[A-Za-z\d.-]+|(?:www.|[-;:&=+$,\w]+@)[A-Za-z\d.-]+)((?:\/[+~%/.\w-]*)?\??[-+=&;%@.\w]*#?\w*)?)/
+const absoluteRegExp =
+  /(?:[A-Za-z]{3,9}:(?:\/\/)?(?:[-;:&=+$,\w]+@)?[A-Za-z\d.-]+|(?:www.|[-;:&=+$,\w]+@)[A-Za-z\d.-]+)(?:\/[+~%/.\w-]*)?\??[-+=&;%@.\w]*#?\w*/
+
+/**
+ * This regex checks for relative URLs starting with / or anchor links starting with # in a string. Tested for the following use cases:
+ * - /privacy-policy
+ * - /privacy-policy#primary-terms
+ * - #primary-terms
+ *  */
+const relativeOrAnchorRegExp = /^[\w\-./]*(?:#\w[\w-]*)?$/
 
 // Do not keep validateUrl function too loose. This is run when pasting in text, to determine if links are in that text and if it should create AutoLinkNodes.
 // This is why we do not allow stuff like anchors here, as we don't want copied anchors to be turned into AutoLinkNodes.
@@ -28,7 +36,10 @@ export function validateUrl(url: string): boolean {
   if (url === 'https://') return true
 
   // This makes sure URLs starting with www. instead of https are valid too
-  if (urlRegExp.test(url)) return true
+  if (absoluteRegExp.test(url)) return true
+
+  // Check relative or anchor links
+  if (relativeOrAnchorRegExp.test(url)) return true
 
   // While this doesn't allow URLs starting with www (which is why we use the regex above), it does properly handle tel: URLs
   try {