Black Friday: 50% lifetime discount until December 16th.Check it out

5.1.3

by

This update adds primary term selection to quick/bulk edit, option visibility toggles, comprehensive Avada theme compatibility, improved Elementor handling, and various bug fixes.

We also deleted all deprecated code from before TSF v5.0. If you added custom snippets from before 2 years ago without resolving the deprecations, you may experience issues. Feel free to reach out to us in our support forums if you need help.

Detailed log

> View code changes.
> View closed GitHub issues.

For everyone

  • Upgraded: Now uses TSF database version 5130.
  • Added:
    • You can now select the primary term in the quick edit and bulk edit interfaces on post overview pages.
    • You can now hide TSF’s term, user, and quick/bulk-edit interfaces via the “SEO Settings > General Settings > Layout.”
  • Compatibility:
    • Theme: Adava:
      • Added comprehensive compatibility with the Avada theme to prevent SEO functionality conflicts.
        1. We now hide Avada SEO settings from their post edit meta box.
        2. We now hide most Avada’s SEO settings from their theme options.
          • Except for “Rich Snippets FAQ” (used by their fusion_faq shortcode), for which “Rich Snippets” should also be allowed.
        3. We now disable all hidden Avada SEO settings on the front-end to prevent conflicts.
          • This includes Open Graph and structured data for Title, Author, and Date output.
    • Plugin: Elementor:
      • We now again output The SEO Framework’s metadata on Elementor’s “Landing Pages” post type (e-landing-page) and “Templates” elementor_library.
      • With that, we now force the “noindex” directive on these post types.
        • This is something that they should’ve done a long time ago, and it’s negatively impacting millions of websites’ ranking, including yours – unless you use TSF.
      • We added Elementor’s “Floating Buttons” post type (e-floating-buttons) to a bespoke “Elementor’s dumb post type” list.
        • This means that The SEO Framework will also hide its interface and force the “noindex” directive on this post type.
      • Note: We considered adding a dynamic “headless” mode for Elementor’s dumb post types, so that TSF’s custom post meta fields could be ignored from these, but this adds needless complexity for a corner case that should never have existed in the first place.
  • Fixed:
    • Resolved an issue where abbreviations at the start of sentences weren’t considered by the description generator.
    • Resolved an issue where the breadcrumb generator would include a broken ancestor if a parent post is deleted without purging the child post’s post_parent field. Now skips the broken ancestor (theoretical fix; WordPress returns 404 for such posts anyway).
    • Resolved an issue where the breadcrumb generator would add an “Archives” link pointing to the current post if a term ancestor is deleted. Now skips the broken ancestor.
    • Resolved an issue where the more information about sitemaps-link wasn’t outputted on the sitemap stylesheet.
    • Resolved an issue where image type warnings that were meant for social sharing were also shown for other contexts, such as logos and structured data.
    • Resolved an issue where the tooltip content would overflow its container in WordPress 6.9 (because they changed the ID of a common element AGAIN without considering other developers).
    • Resolved an issue where tooltips were clipped on WordPress 6.9 due to needlessly added overflow: hidden on the meta box containers.
  • Removed:
    • Compatibility with the Headway theme has been removed. The theme is no longer maintained since 2017 and the developer’s website is down.

For translators

  • Updated:
    • POT translation file.
    • Various sentences have been updated for clarity.

For developers

  • PHP API notes:
    • Added:
      • The_SEO_Framework\Meta\URI\Utils::get_site_path() ( tsf()->uri()->utils()->get_site_path()) is now available. The value should be equivalent to $wp_rewrite->front, but then filtered via get_home_url().
    • Changed:
      • The_SEO_Framework\Data\Filter\Sanitize::metadata_content() ( tsf()->sanitize()->metadata_content()) now decodes HTML entities at the end of its process.
      • The_SEO_Framework\Meta\Title::get_custom_title() and get_generated_title() ( tsf()->title()->get_custom_title(), tsf()->title()->get_generated_title()) now run their output through Sanitize::metadata_content().
      • The_SEO_Framework\Meta\Image::get_first_image_url(), get_first_custom_image_url(), and get_first_generated_image_url() ( tsf()->image()->get_first_image_url(), tsf()->image()->get_first_custom_image_url(), tsf()->image()->get_first_generated_image_url()) now have null as the default value for the first argument.
      • The_SEO_Framework\Data\Term::get_term_parents() ( tsf()->data()->term()->get_term_parents()) no longer uses memoization to cache results.
        • This is now in line with The_SEO_Framework\Data\Post::get_post_parents() ( tsf()->data()->post()->get_post_parents()), which never used memoization.
        • We probably added this because of this unresolved caching issue, but we never invoked the memoization anyway.
    • Fixed:
      • The_SEO_Framework\Helper\Format\Markdown ( tsf()->format()->markdown()->convert()) no longer grabs closing parentheses as part of its link when two or more are followed. Parentheses in links are still supported, but they must be balanced with an opening parentheses.
      • The_SEO_Framework\Data\Post::get_post_parents() ( tsf()->data()->post()->get_post_parents()) now filters out deleted posts (broken ancestors). Next to the user-facing breadcrumb issue, this resolves some PHP warnings that could occur when fetching deleted parent posts for canonical URL generation.
      • The_SEO_Framework\Helper\Format\HTML::strip_tags_cs() ( tsf()->format()->html()->strip_tags_cs()):
        1. Added ‘body’ and ‘style’ to the phrase elements.
        2. Added conditional “NO_JIT” modifier for huge inputs to prevent abortion due to suspected memory issues.
        3. Improved regex pattern to ignore bitwise operators (<<) encountered in scripts. Also prevents recursive lookup loops when encountering these operators in elements.
      • Resolved an issue where pools tsf()->escape() and tsf()->sanitize() were incorrectly marked to be from pool tsf()->filter()->escape() and tsf()->filter()->sanitize() respectively.
      • Resolved an issue in The_SEO_Framework\Traits\Internal\Static_Deprecator trait where dynamic property access used incorrect syntax (self instead of static::class and self::$name instead of static::$$name).
    • Other:
      • We now use The_SEO_Framework\Data\User::get_userdata() ( tsf()->data()->user()->get_userdata()) instead of the expensive get_userdata() directly to fetch user data. This improves output performance at the expense of a slight memory overhead.
      • Converted all static calls to self in final classes. This improves performance marginally by avoiding late static binding runtime resolution. We measured less than 1% performance improvement in our test suite; that’s likely because most of our code is procedural instead of object-oriented.
  • JS API notes:
    • Changed:
      • tsfCanonical.rewrite is no longer available. This was not meant to be publicly accessible.
      • tsfCanonical.rootUrl is no longer available, without deprecation. This was not meant to be publicly accessible.
      • tsfCanonicalL10n.rootUrl is now an array of scheme, host, port, and path properties, instead of just a plain string of the host.
      • tsfCanonicalL10n.allowCanonicalURLNotationTool is renamed to tsfCanonicalL10n.allowCanonicalURLNotationTracker, which is more consistent with the rest of the codebase.
        • This change is not backward compatible; however, the property was marked with the comment “TEMP: […]”, as it was a quick workaround for a compatibility issue with multilingual plugins.
      • tsfMediaL10n.warning.warnedTypes and tsfMediaL10n.warning.forbiddenTypes are now context-aware objects instead of flat arrays.
        • warnedTypes now has a social property containing image types that trigger warnings for social images (e.g., webp, heic).
        • forbiddenTypes now has an all property containing universally forbidden image types (e.g., apng, bmp, svg).
        • This is a semi-breaking change for the JS API. We highly doubt anyone used these properties externally, as they were introduced in v5.1.0.
        • The image warning system now checks for context-specific warnings first, then falls back to universal warnings, making it more extensible for future image contexts.
  • Option notes:
    • Of option autodescription-site-settings (constant THE_SEO_FRAMEWORK_SITE_OPTIONS, pool tsf()->data()->plugin(), or legacy API tsf()->get_options()):
      • Added index display_list_edit_options. Default 1.
      • Added index display_term_edit_options. Default 1.
      • Added index display_user_edit_options. Default 1.
  • Filter notes:
    • Fixed:
      • For the_seo_framework_extract_content_strip_args, when adjusting the space or clear indexes in such a manner that empty void, clear, or space queries are created, the resulting Context-Sensitive tag stripper (The_SEO_Framework\Helper\Format\strip_tags_cs()) now correctly ignores those empty queries instead of halting the context-sensitive stripping process.
  • Other:
    • We updated our coding standards, so the code is slightly altered.
    • We now properly capitalize the proper noun Boolean.
    • We purged all deprecated methods, properties, classes, files, folders, and filters that were marked for removal in v5.0.0.
      • We removed about 7 classes, 388 methods, 13 properties, and 35 filters.
      • For a full list, please refer to the v5.0.0 detailed changelog.