I search sites sometimes, and see the result link is of a non-canonical location of a given document, even though there is a <link rel=canonical... in the page, and the body of that page is identical to the one linked.
One example is my own contact/homepage https://qui.suis.je/, which is served byte-identically by a number of domains (such as https://magnifi.ca/ and https://aaronhamilton.us/) for historical reasons... on Kagi the result is returned three times; twice on non-canonical domains, and once on a bare IP address.
Links to an identical resource with rel=canonical should be combined, both in terms of ranking/filtering and when returning the search result.
