Default robots.txt blocks Twitter Card images

By default, the robots.txt file contains the following entry:

User-agent: *
# ...
Disallow: /u

This configuration blocks the image for Twitter cards. Checking the card of various forums on Twitter’s validator does not display the logo and shows the following warning:

image

The image URL forum.[…].org/uploads/[…].png specified by the ‘twitter:image’ metatag may be restricted by the site’s robots.txt file, which will prevent Twitter from fetching it.

The problem is that the path /uploads also starts with a u and is therefore matched by the robots.txt rule.

I know that the robots.txt file can be overwritten but I think it would be good to change the default. Unfortunately, I don’t understand the code well enough to create a pull request.

When overwriting, keep in mind that Twitter checks the robots.txt file less often than the actual html of the page, so it takes a few hours to get picked up.

1 Like

Encountered the same problem and posted here https://meta.discourse.org/t/metatag-may-be-restricted-twitter-card-validator/181169

Is there a quick fix to try? Thanks!

Yes, at least it makes the Twitter cards work. I hope that it does not have unintended side effects.

You can open the setting “allow index in robots txt” and click “override robots.txt”. Then look up the line Disallow: /u and place a slash in the end, so it becomes Disallow: /u/. It will take Twitter a few hours to pick up the change.

2 Likes

Not sure why the post I linked to was deleted. Will give it a try. Thanks for the tip!

2 Likes

Looks like this was a problem (see this thread) in the past and then got re-introduced.

In this commit, the robots rules with slash got removed because it was thought they are no longer needed. Later, in this commit they were re-introduced but without the slash in the end.

I put together a quick-and-dirty pull request that should fix the default robots.txt file: https://github.com/discourse/discourse/pull/12349

1 Like

What do you think @sam?

1 Like

I can’t really look at the PR without a CLA, but anchoring and changing /u/u/ sounds reasonable and safe.

2 Likes

Sorry - CLA is signed now :slight_smile:

I’ve merged it. Let’s give it a shot.

4 Likes