Skip to content

More conservative number of retry attempts for failed requests #42400

Closed

Description

@desrosj

Description

While testing to see if Core-13779 was still happening in WordPress 6.0, I found that it was, and actually causes potentially greater problems in the block editor.

TLDR of that ticket: When the WordPress files are in a directory and a site is set up to run as such (site.com/wp) and the site is changed to instead be visible from the site's root (site.com), all users who are logged in will not be properly authenticated. This is because the cookie set on login does not exist for the new location.

In the past, this caused issues with the Classic Editor and previewing posts. However, in WordPress 6.0, if you open the block editor, there are repeated requests to the REST API that return 403 responses. After opening the editor with my browser console open, ~500 requests were made in 30 seconds.

Regardless of the root cause of the 403 responses, the editor should not continually retry the same requests infinitely.

Step-by-step reproduction instructions

  1. Create a WordPress install in a subdirectory (/wp for example).
  2. Log in, and follow steps 2-9 found here.
  3. Open the browser's console and click the network tab.
  4. Create a new post or edit an old one.
  5. Observe the repeated requests.

Screenshots, screen recording, code snippet

No response

Environment info

No response

Please confirm that you have searched existing issues in the repo.

Yes

Please confirm that you have tested with all plugins deactivated except Gutenberg.

Yes

Activity

  1. Mamaduka commented on Jul 14, 2022

    @Mamaduka
    Member

    Probably related to #13509.

  2. linked a pull request that will close this issuePrevent autosave when the post is locked #32475on Feb 9, 2024
  3. Mamaduka commented on Dec 11, 2024

    @Mamaduka
    Member

    Created a PR with possible fix #67812.

  4. youknowriad commented on Oct 20, 2026

    @youknowriad
    Contributor

    Thanks @desrosj! As you noted in #13509, this is the same endless retry loop: apiFetch keeps refreshing the nonce and retrying the 403 responses. Carrying over to #13509: another reliable way to reproduce it is moving a WordPress install from a subdirectory (/wp) to the site root, so the existing login cookie no longer matches and the editor sends hundreds of failing requests within seconds. Thanks also to @desaiuditd for #32475, which already references #13509, so it continues there. Closing as a duplicate of #13509.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Assignees

Type

No type

Projects

No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions