Repository navigation
Keep page-builder shortcode text out of the video sitemap description - #2235
Merged
Merged
Conversation
The sitemap is served at parse_request priority 1, before WPBakery and Divi register their shortcodes, so core's strip_shortcodes() left tags such as [vc_row] in the fallback description taken from the page text. Add the tag names found in the text to strip_shortcodes()'s own list so unregistered shortcodes are removed the same way registered ones are, which gives the same description in the sitemap and on the front end. The strip_shortcodes() test stub removed every [tag], which hid the bug. It now follows core: only registered shortcodes are removed.
…ck description Review of the first version found that it also removed ordinary bracketed text ([sic], [1], [b]bold[/b]) from the front-end description, that one very long bracketed word made core's pattern too large and the description empty, that a tag another filter had taken off the strip list was re-added, and that the filter stayed registered if stripping threw. Only names that look like a builder tag are added now (a letter first, an underscore or hyphen in the name, 64 characters at most), tags that are registered are left to core, and the filter is removed in a finally block. New tests pin the cap, the length limit, the filter removal and the ordinary bracketed text.
🔍 WordPress Plugin Check Report
📊 Report
|
| 📍 Line | 🔖 Check | 💬 Message |
|---|---|---|
0 |
mismatched_plugin_name | Plugin name "GoDAM - Organize WordPress Media Library & File Manager with Unlimited Folders for Images, Videos & more" is different from the name declared in plugin header "GoDAM". |
0 |
trademarked_term | The plugin name includes a restricted term. Your chosen plugin name - "GoDAM - Organize WordPress Media Library & File Manager with Unlimited Folders for Images, Videos & more" - contains the restricted term "wordpress" which cannot be used at all in your plugin name. |
📁 assets/build/css/main.css (1 warning)
| 📍 Line | 🔖 Check | 💬 Message |
|---|---|---|
0 |
EnqueuedStylesScope | This style is being loaded in all contexts. |
📁 assets/src/libs/analytics.min.js (5 warnings)
| 📍 Line | 🔖 Check | 💬 Message |
|---|---|---|
0 |
EnqueuedScriptsScope | This script is being loaded in all frontend contexts. |
0 |
NonBlockingScripts.NoStrategy | This script on http://localhost:8880 (with handle analytics-library) is loaded in the footer. Consider a defer or async script loading strategy instead. |
0 |
NonBlockingScripts.NoStrategy | This script on http://localhost:8880/2026/10/21/hello-world/ (with handle analytics-library) is loaded in the footer. Consider a defer or async script loading strategy instead. |
0 |
NonBlockingScripts.NoStrategy | This script on http://localhost:8880/sample-page/ (with handle analytics-library) is loaded in the footer. Consider a defer or async script loading strategy instead. |
0 |
NonBlockingScripts.NoStrategy | This script on http://localhost:8880/demo-attachment-post/ (with handle analytics-library) is loaded in the footer. Consider a defer or async script loading strategy instead. |
📁 assets/build/js/main.min.js (5 warnings)
| 📍 Line | 🔖 Check | 💬 Message |
|---|---|---|
0 |
EnqueuedScriptsScope | This script is being loaded in all frontend contexts. |
0 |
NonBlockingScripts.NoStrategy | This script on http://localhost:8880 (with handle rtgodam-script) is loaded in the footer. Consider a defer or async script loading strategy instead. |
0 |
NonBlockingScripts.NoStrategy | This script on http://localhost:8880/2026/10/21/hello-world/ (with handle rtgodam-script) is loaded in the footer. Consider a defer or async script loading strategy instead. |
0 |
NonBlockingScripts.NoStrategy | This script on http://localhost:8880/sample-page/ (with handle rtgodam-script) is loaded in the footer. Consider a defer or async script loading strategy instead. |
0 |
NonBlockingScripts.NoStrategy | This script on http://localhost:8880/demo-attachment-post/ (with handle rtgodam-script) is loaded in the footer. Consider a defer or async script loading strategy instead. |
🤖 Generated by WordPress Plugin Check Action • Learn more about Plugin Check
Member
|
Tested on my local site: Reproduced, and fixed by this PR Video with no description and no transcript, on three pages:
Note for anyone retesting: the video needs a thumbnail. A video without one is left out of the sitemap, so the page doesn't appear at all. I also checked the builder-tag pattern against sample names: Minor, non-blocking
Not tested: real Divi (not installed), PHP 7.4. |
rudrakshi-gupta
approved these changes
Oct 21, 2026
Collaborator
Author
|
Thank you for testing this, and for checking the pattern against sample names. Both notes are right, and I made no code change for them:
|
subodhr258
added a commit
that referenced
this pull request
Oct 22, 2026
* Stop the admin menu jumping when a GoDAM page loads (#2207) * Fix: stop the admin menu jumping when a GoDAM page loads pages/style.css restores Tailwind's border reset for every element on the page. It is injected by page-css.min.js after the page has painted, and core's a { transition: border .05s } then animates each admin menu link's border from 3px to 0: every item grows from 34px to 40px and the whole menu jumps by about 186px, then snaps back. Keep the reset off #adminmenumain and #wpadminbar. * Fix: open the Media Editor with the admin menu already folded The editor folds the menu from React after the page has painted, so the menu showed at full width and then collapsed. Add the folded body class on the server when the editor opens a file. * Keep the reset off the menu and admin bar containers too; test the editor fold - pages/style.css also excludes #adminmenumain and #wpadminbar themselves, not only their descendants. - MediaEditorMenuFoldTest: the Media Editor with an id gets 'folded'; its file picker, other GoDAM pages, the Media Library and no screen do not. * Regenerate the POT: 12 references into class-pages.php moved down Line numbers only; no strings added or changed. Built like CI: Node 22.14.0, npm ci, npm run build:prod, composer pot. * feat(seo): fix VideoObject schema and add video sitemap (#2186) * feat(seo): fix VideoObject schema and add video sitemap (#1622 v1) Make GoDAM's video schema indexable by Google and add a video sitemap that plugs into Yoast SEO, Rank Math, All in One SEO or the WordPress core sitemap. No new public URLs are created. - Schema contentUrl now uses the HLS (.m3u8) stream, falling back to the original file; the DASH .mpd is never output. Resolved at output time, so pages cached before this change are fixed without a re-save. - Add embedUrl pointing at GoDAM's existing video-embed player page. - Never output an empty description: fall back to a 160-char transcript excerpt (fetched on save only, cached in _godam_seo_transcript_excerpt), then the page excerpt, otherwise omit the property. - Output isFamilyFriendly as a real JSON boolean (string values such as "false" are parsed correctly) and omit an empty uploadDate. - Add Video_Sitemap: /godam-video-sitemap.xml with one <url> per published page and one <video:video> per video, built from the cached video SEO post meta via the shared schema builder so both always agree. - Exclude drafts, private and password-protected posts, pages noindexed in Yoast / Rank Math / AIOSEO, and videos the editor excludes. - List the sitemap in the Yoast, Rank Math and AIOSEO indexes, or via a core sitemap provider when none is active; paginate at 1000 URLs, cache per site with version-based invalidation. - Add GoDAM > Settings > SEO with an "Enable video sitemap" toggle (off by default) and a per-video "Exclude from video sitemap" toggle in Gutenberg, Elementor and WPBakery. - Warn editors in all three editors when a video title looks like a raw file name. - Prefer the HLS URL when prefilling SEO contentUrl in the editors. - Add new filters: godam_video_seo_content_url, godam_video_seo_embed_url, godam_video_seo_description, godam_video_sitemap_enabled, godam_video_sitemap_entries_per_page, godam_video_sitemap_post_types, godam_video_sitemap_exclude_post, godam_video_sitemap_entry. - Add PHPUnit coverage (VideoSeoSchemaTest, VideoSitemapTest) and Jest tests for the filename-title check. * Resolve feedback * Resolve feedbacks and regenrate POT file * Revert the POT file * Resolve Privacy Issue * feat(seo): add key moments to video schema from chapters (#1597) (#2194) Output each GoDAM video chapter as a schema.org Clip in the VideoObject's hasPart, so Google can show "key moments" (jump-to links) for the video in search results. - Add Seo::get_key_moment_clips(): reads the chapters from the attachment's rtgodam_meta at output time, so editing chapters updates the schema without re-saving the page. - Each Clip has name, startOffset, endOffset and a deep link of the form <page>?t=<seconds>#godam-video-<attachment id>, which the GoDAM player already uses to seek on page load. - A clip ends at the chapter's own endTime, else the next chapter's start, else the video duration (clamped to the duration when known). - Skip chapters with no title, an unreadable start, a start at or past the end of the video, or zero length; drop duplicate start times; strip HTML from chapter names. - Parse chapter times as plain seconds, mm:ss or hh:mm:ss(.ms), falling back to originalTime when startTime is missing. - Build clip URLs without add_query_arg(): use & on plain permalinks and drop any existing fragment from the page URL. - Only add hasPart when the video has an attachment and a host page (render-time output on archives/404 has nothing to deep-link to). - Add filters godam_video_seo_key_moments_enabled and godam_video_seo_key_moments. - Move ISO 8601 duration parsing to Seo::iso8601_to_seconds(); Video_Sitemap::iso8601_duration_to_seconds() now delegates to it. - Add VideoSeoKeyMomentsTest (27 tests) and a get_permalink() test stub. * Match WPBakery SEO Settings to Gutenberg Model (#2209) * ci: run the POT check only on PRs into main (#2188) * ci: run the POT check only on the develop → main merge Feature PRs into develop no longer have to regenerate languages/godam.pot. The release PR from develop into main still fails if the committed POT is stale. PRs into main from any other branch skip the job. * ci: run the POT check on every PR into main Patch releases reach main from fix and hotfix branches, not only from develop, so they need the check too. The branches: [main] trigger already keeps it off feature PRs into develop. * Pass only a numeric form ID to the form layer preview (#2223) Cast the layer id to an integer in the preview template and in the Ninja Forms and MetForm callbacks, and add unit tests for it. * Keep the Gravity Forms recorder working after an AJAX page change (#2217) * Keep the Gravity Forms recorder working after an AJAX page change Gravity Forms replaces a page's markup on every AJAX page change or failed submit. The recorder kept pointing at the replaced container, file input and button, and Uppy had bound its dashboard trigger once to the old button, so Start Recording opened nothing on the new page. On every gform_post_render, an uploader whose markup left the page now finds the new container by its id, reads the input and button from it, binds the new button once and shows the recording it holds again. * Regenerate the translation template for the recorder's moved lines * Stop sending the signed-in viewer's email with video analytics (#2213) * Stop sending the signed-in viewer's email with video analytics The page params that GoDAM prints on every front-end page no longer carry the current user's email address (`emailId`), and the analytics request body no longer has an `email` field. The WordPress user id is still sent. Tests: the request-body helper and the page params are both checked for the absence of the email, including a page cached by an older version. * Remove the viewer's email from the analytics data listed in the readme The analytics.godam.io entry under External services no longer lists Email as data sent. * Keep the signed-in viewer's email out of the player's page data The player script's page data (godamData.currentLoggedInUserData) carried the current user's email address, which nothing in the browser reads. It now holds the display name and whether someone is signed in. The comments routes still read the email on the server. * Keep the fixture lines untouched in the JS test The explanation moves into the test, so the fixture lines this branch leaves alone cannot conflict with the sibling change. * List the page's user fields instead of removing one rtgodam_get_current_user_data_for_page() now returns only the display name and the sign-in type, so a field added later to the server-side user data is not printed into the page by accident. * Regenerate the POT for the new line numbers The new helper in custom-functions.php moves five strings further down the file, so their source references change. Nothing else in the file differs from develop. * Clean up video share popup: remove GoDAM Central/Embed links, use Page link, show share button for all videos (#2211) * Clean up video share popup links and share button gating - Remove the redundant GoDAM Central and Embed copy fields from the share modal footer, keeping only the Page Link field. - Point the social share links at the WordPress page link instead of the GoDAM Central URL, which is private by default and shows a login prompt. - Show the share button for all videos when sharing is enabled in the block settings, instead of only transcoded videos. Closes #2210 * Address review: keep share anchor and gate button on an ID - Encode the share link with encodeURIComponent so the #godam-video-{id} anchor and any query param travel with the social share URL instead of splitting off (encodeURI left #, ? and & intact). - Hide the share button for players with neither a WordPress attachment ID nor a transcoding job ID ("Insert from URL" embeds), since their share link could not target the video. - Remove stale README references to the removed GoDAM Central/Embed links and correct the ID described for share links. --------- Co-authored-by: KMchaudhary <kuldipkumar.chaudhary@rtcamp.com> * Save converted MP3 audio as audio/mpeg, and fix files already saved as audio/mp3 (#2208) * Fix: store transcoded MP3 audio as audio/mpeg, not audio/mp3 WordPress registers MP3 as audio/mpeg. Attachments saved as audio/mp3 were left out of REST media_type=audio, and REST ignored a mime_type=audio/mp3 filter. New conversions are saved as audio/mpeg, and a one-time migration (2.4.1 batch) fixes the ones saved before. * Test the saved type, satisfy the hook-integrity check, retry a failed SELECT - TranscodedAudioMimeTypeTest calls add_transcoded_files() and checks that converted audio is saved as audio/mpeg (video stays video/mp4). With the old audio/mp3 it fails. - add_transcoded_files() starts $transcoded_files as array(), not false; PHP 8.1+ reported every conversion as a deprecated false-to-array use. - The migration's two direct queries carry godam-coverage-ignore: the runner already runs it once per site. - A failed SELECT now returns false so the runner retries, instead of marking the site done with nothing changed. - The runner test checks batch order instead of which batch is last. * Run the MP3 type migration in the 2.5.0 batch 2.4.1 ships without this change but still stores 2.4.1 as each site's migration version, and the runner skips batches at or below the stored version. Keyed 2.4.1, the migration would never run on sites that update through 2.4.1; keyed 2.5.0 it runs for both 2.4.0 and 2.4.1 sites. * Release prep v2.5.0: version bump, changelog and POT refresh (#2231) - Version 2.5.0 in godam.php, readme.txt, README.md, package.json and package-lock.json. - v2.5.0 changelog in readme.txt and CHANGELOG.md; v2.3.2 removed from readme.txt to keep the last three. - Every @SInCE n.e.x.t replaced with 2.5.0 (51 tags in 6 files). - languages/godam.pot regenerated from a production build. * Fix the two new Plugin Check warnings in the Video SEO sitemap code (#2232) - class-video-sitemap.php: the phpcs:ignore named slow_query_meta_key, which is not a sniff, so it ignored nothing. Use slow_db_query_meta_key. - class-video-sitemap-core-provider.php: ignore NonPrefixedHooknameFound on WordPress core's own wp_sitemaps_index_entry filter. * Keep page-builder shortcode text out of the video sitemap description (#2235) * Keep page-builder shortcode text out of the video sitemap description The sitemap is served at parse_request priority 1, before WPBakery and Divi register their shortcodes, so core's strip_shortcodes() left tags such as [vc_row] in the fallback description taken from the page text. Add the tag names found in the text to strip_shortcodes()'s own list so unregistered shortcodes are removed the same way registered ones are, which gives the same description in the sitemap and on the front end. The strip_shortcodes() test stub removed every [tag], which hid the bug. It now follows core: only registered shortcodes are removed. * Only strip builder-style unregistered shortcode names from the fallback description Review of the first version found that it also removed ordinary bracketed text ([sic], [1], [b]bold[/b]) from the front-end description, that one very long bracketed word made core's pattern too large and the description empty, that a tag another filter had taken off the strip list was re-added, and that the filter stayed registered if stripping threw. Only names that look like a builder tag are added now (a letter first, an underscore or hyphen in the name, 64 characters at most), tags that are registered are left to core, and the filter is removed in a finally block. New tests pin the cap, the length limit, the filter removal and the ordinary bracketed text. * Set the 2.5.0 changelog date to October 22, 2026 (#2236) 2.5.0 releases on 22 Oct (release ticket rtCamp/godam-plugin-wp#60), but readme.txt and CHANGELOG.md dated it October 21. --------- Co-authored-by: Kuldip Chaudhary <64731232+KMchaudhary@users.noreply.github.com> Co-authored-by: Rudrakshi Gupta <109859631+rudrakshi-gupta@users.noreply.github.com> Co-authored-by: Vishal Kumar <vishnshiv3@gmail.com> Co-authored-by: KMchaudhary <kuldipkumar.chaudhary@rtcamp.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Review packet
In one line: On a page built with a page builder (WPBakery, Divi), the video sitemap no longer shows the builder's raw shortcodes (
[vc_row][vc_column]...) in<video:description>; the description now matches what the front-end JSON-LD gives for the same page. Reported on #1622 (the Video SEO epic that #2186 is part of): the bug report. This PR does not close the epic.1. Criteria and proof
<video:description>for the WPBakery page from the bug report has no[vc_...]text (it falls back to the video title, as the front end gives no description)VideoSitemapTest::test_wpbakery_page_does_not_leak_shortcode_text_into_the_description; recordings V1 (before) and V2 (after)VideoSitemapTest::test_page_text_outside_the_builder_shortcodes_is_used_as_the_description; recordings V1 (before) and V2 (after)VideoSeoSchemaTest::test_page_fallback_is_the_same_before_and_after_the_builder_registers_its_shortcodesdiffof the saved JSON-LD on godam-dev (section 2); and 9 sample texts give the same description with and without the fix when the builder is registered[sic],[1],[b]bold[/b]) is not changedVideoSeoSchemaTest::test_page_fallback_leaves_ordinary_bracketed_text_alone2. Before and after
Same three pages on godam-dev (WPBakery 9.0 beta active, video sitemap switched on, a video with no description of its own), read logged out as a crawler would, first on
developand then on this branch. The sitemap file differs in exactly three lines, the three descriptions below; every other entry is identical.<video:description>beforedescriptionbefore[vc_row][vc_column][vc_column_text]Learn to brew filter coffee at home, step by step.[/vc_column_text][godam_video id="265"][/vc_column][/vc_row][vc_row][vc_column][vc_column_text]Learn to brew filter coffee at home, step by step.[/vc_column_text][/vc_column][/vc_row]Filter-Coffee-Tutorial-1-original)A short guide to brewing filter coffee at home. [vc_row][vc_column][godam_video id="265"][/vc_column][/vc_row]A short guide to brewing filter coffee at home. [vc_row][vc_column][/vc_column][/vc_row]A short guide to brewing filter coffee at home.A short guide to brewing filter coffee at home.Intro line. [et_pb_section][et_pb_row][et_pb_text]Divi copy.[/et_pb_text][/et_pb_row][/et_pb_section] [godam_video id="265"] Outro line.(Divi is not installed there, so these tags are never registered, not even on the front end)Intro line. [et_pb_section]...[/et_pb_section] Outro line.Intro line. Outro line.Intro line. Outro line.B1: the sitemap on
develop, the leaking description highlightedS01: the sitemap on this branch, the same entry highlighted
V1: BEFORE, on
develop. Two panes: the page's own front-end JSON-LD on the left, the crawler's sitemap on the right; every step has a caption and a check. The sitemap shows the raw[vc_row]...text.V2: AFTER, on this branch, the same steps. The sitemap shows the video title for the first page and the sentence for the second, as the front end does.
3. Tests
tests/php/VideoSeoSchemaTest.php(11 tests): builder tags not registered yet leave no[; same description before and after the builder registers; Diviet_pb_*tags; a self-closing unregistered tag; a hyphenated builder tag (contact-form-7); a tag another filter took off the strip list is left alone; ordinary bracketed text is left alone; 20,000 distinct builder-style tags still give a description; one 70,000-character tag name still gives a description; the filter is not left registered, also when another filter throws.tests/php/VideoSitemapTest.php(2 tests): the WPBakery page from the bug report, and a page with text outside the builder tags, read from the generated sitemap XML.tests/stubs/video-seo-functions.php. Thestrip_shortcodes()stub removed every[tag], which hid this bug (it never looked at which shortcodes are registered). It now follows core: only registered shortcodes are removed, with the text they enclose, andstrip_shortcodes_tagnamesis applied.tests/stubs/hooks.phpgets aremove_filter()stub.inc/classes/class-seo.phpput back todevelop, 7 tests fail, for example "Failed asserting that two arrays are identical" with the actual value'[vc_row][vc_column][vc_column_text]Learn to brew filter coffee at home, step by step.[/vc_column_text][/vc_column][/vc_row]'. The other new tests are safety checks and pass on both. Each safeguard (the cap, the name length limit,finally,remove_filter, the registered-name check, the name pattern) was also broken on purpose in turn, and a test failed each time.bin/hook-check) pass.4. Risk flags
No risk flags fired on #2235. Nothing in the diff touched the watched areas.
5. Independent review
A fresh reviewer (Sonnet 5.5) saw only the problem statement and the diff. It ran the change against WordPress's real shortcode code, not just the test stubs. Its findings, each checked against the code:
class-seo.php: any bracketed word ([sic],[1],[b]..[/b],[Read more]) was also removed from the front-end description, so it did change for ordinary pagesstrip_shortcodes()threwremove_filter()runs in afinallyblock. New testremove_filter()call[vc_column][vc_column]..), a closing tag with no opener, an unterminated tag and an escaped[[vc_row]]still show some bracket text[gallery]before this change). Sitemap pages are cached for 24 hours. Not changedwp_strip_all_tagsafter decoding, only one caller, three random-order full runs)6. Not verified
[et_pb_*]content was tested (unit tests, and a page on godam-dev where the tags are never registered), not the plugin itself./godam-video-sitemap.xmlwas read on the site.strip_shortcodes_tagnamesfilter exists since WordPress 4.7.7. Scope
class-seo.phpand its tests: the two stub changes in section 3.[et_pb_*]text and now shows clean text. The Divi-shaped test page in section 2 shows it. Pages whose shortcodes are all registered are unchanged.[my_note]) is treated like a builder tag in the fallback description, and when a closing tag follows, the text between them is removed too. It only changes the fallback description (used when a video has no description of its own and no transcript). Plain bracketed words such as[sic]or[1]are untouched. Noted in review.@since 2.5.0is used, like the code around it, because the 2.5.0 version bump has already run and this ships in 2.5.0.