v2.0.0

This is a major, breaking change that comes 17 years after v1.2.2. It updates parseUri’s API, adds new features, includes bug/security fixes, adds tests, uses modern JavaScript, becomes a truly universal URI parser, and remains tiny.

Breaking Changes

• Renamed many URI part properties to better mach the built-in URL object and other URI libraries: source➜href, userInfo➜userinfo, host➜hostname, relative➜resource, path➜pathname, file➜filename, anchor➜fragment, queryKey➜queryParams.
• Updated queryParams to be returned as a URLSearchParams object, with support for multiple query keys with the same name.
• Now uses what used to be called strict parsing mode by default.
• The parsing mode is now specified via an argument (options: 'default', 'friendly').
• Friendly parsing mode (formerly called loose mode) is now consistent with default mode in considering '/entry' to include {filename: 'entry'} rather than treating 'entry' as part of directory.
• Removed parseUri.options properties, including the ability to rename URI keys or provide a non-standard query parser.

New Features

• Added six new URI parts to result objects: origin, host, subdomain, domain, tld, suffix.
  • tld includes extensible support for second-level domains that should be treated as part of the top-level domain (ex: co.uk).
• Added support for IPv6 addresses.

Improvements

• Improved handling for web protocol (http, https, ws, wss, ftp) and protocol-relative URLs, for parsing accuracy and security.
• Improved handling for URNs.
• Added an extensive test suite.
• Added a demo page for testing and comparison with v1, the built-in URL constructor, and other libraries.

Bug Fixes

• Fixed an issue where @ in the resource led to incorrect parsing.
• Fixed an ReDoS vulnerability with extremely long URIs that use @ in the resource.
• Fixed numerous additional edge cases with atypical/invalid URIs.