Broken Links

This is simple specification used for finding broken internal links. External links are not followed. It’s based on Domen Kožar’s gist. [1]

import quickstrom;

// Only links beginning with a slash and ending
// with .html are followed. We could also use
// an absolute base URL, e.g:
//
//     a[href^='https://example.com']
//
action ~navigate! =
  click!(`a[href^='/'][href$='.html']`);

// We're interested in finding links that lead
// to pages rendered with these status codes in
// the heading.
let patterns = [ "404", "500" ];

// In our system's error pages, status codes
// are rendered in an <h1> element.
let ~heading =
  let h1 = first(`h1`);
  h1.textContent when h1 != null;

// Check if the heading has any of the error
// codes in the text.
let ~hasErrorCode = 
  exists p in patterns {
    contains(p, heading)
  };

// This is the safety property. At no point,
// following only internal links, should we
// see error codes.
let ~proposition = always (not hasErrorCode);

check proposition with * when loaded?;

Tweak the patterns and the predicate to fit your use case. You might not have status codes rendered, but texts like “Page not found” or “Access denied”.