gRegorLove.com — little g big R
gRegor Morrill

My name is gRegor Morrill, a.k.a. gRegorLove. I live in San Diego, enjoy tinkering on the web, and try to make people laugh. Yes, “Gregor is a weird name,” and I know gRegor is a weird capitalization. More about me

If you would like to keep up with my posts, check out How to Follow Me.

👻👻👻👻👻👻👻👻👻👻👻👻👻👻👻👻👻👻👻👻👻👻👻👻👻👻👻👻👻👻👻

Recent Articles

❦

Photos

a box that is about 6 inches long, 4 inches wide, and 4 inches deep, with only a small Duracell battery package in it

Efficient packaging. Guess they were out of padded envelopes? 🤣

❦

Notes

New Meute song! I really want to see these guys live.

MEUTE, “Reset”


Added some spoopy to my site with the posting activity bar on the homepage:

screenshot of the posting activity currently on the homepage of my site: a mix of ghost and pumpkin emojis correlating to how (in)frequently I posted


Canonical Links in Google Products

Short version: Adding the rel=canonical URL in your web pages is important for more than just search results. It can also impact link previews in other Google products like Docs.


I recently ran into a weird edge case with link previews in Google Docs. A work client has an online directory and each profile has a unique permalink. Someone pasted their link in a Google Doc and clicked on it, which brought up a card preview showing the title. The title was showing the name of someone else’s profile, though!

My first thought was something with Google’s cache of the link preview, although the other person’s name had never appeared at that link, so that didn’t make much sense. I tried adding the link with a query string at the end: ?v=2026-10-02. With caching, that usually is seen as a unique URL and can avoid the cached response. Docs showed no preview initially, but after a couple minutes, it showed the same (incorrect) title.

My next thought was that these permalinks did not include a rel=canonical tag. The contents of the profile pages are largely similar except for the person’s name and list of credentials. I didn’t think that should be an issue, though, because the permalinks all contain a unique, 32-character identifier and no query strings. The links are very distinct and it did not make sense how Google would be picking another, seemingly random profile.

I was pretty stumped at that point but decided to check Google Search Console. It reported “Page is not indexed. Duplicate without user-selected canonical.” It explained that since we had not specified the canonical URL, they had selected another permalink as the canonical. Sure enough, the one they had chosen was the name that was showing up in the link preview.

That de-duplication process makes sense when it comes to search results, but I didn’t expect Docs to also be relying on that. I expected when you paste a URL, it should always show a link preview for that URL. It appears what they actually do is check the search index and use the canonical for the link preview only. They don’t change the URL itself, so you can still click through and get to the intended location.

This definitely seems like a bug and is not in any Google documentation I could find, which is partly why I am writing this up. A StackOverflow answer also hypothesizes this behavior.

Edit: I forgot to mention, we added the rel=canonical and started the “Validate Fix” process in Search Console. It says it may take a few days to finish the process. I expect it to fix this issue in Docs, but will update this post with the results either way.


Haha, the Christianity Today website 404 pages say “Oops! Looks like this page is formless and void.” Well-played.


New (to me) music: Spiro World by Time Wharp. Cello, woodwind, bit of electronics. Pretty chill.


My Health and Safety Guidelines • Privacy Policy

← Proud member of An IndieWeb Webring 🕸💍 →

❦