Adaptive Media continues serving stale adapted images after URL-only image replacement

Feedback Bug Reports
1

Summary

When Adaptive Media (AM) is enabled, updating an image by changing only the image URL can cause a critical issue:
previously generated Adaptive Media variants continue to be served even after the image is updated.

This is more serious than Adaptive Media simply not applying.
It results in stale and incorrect images being delivered to end users, without any error, warning, or validation.

This behavior represents a Feature / UX defect at the product design level, not a usage or documentation issue.

Official Documentation Reference

All observations below are based on the official documentation:

https://learn.liferay.com/w/dxp/digital-asset-management/publishing-and-sharing/using-adaptive-media/using-adapted-images-in-site-content

Relevant statements from the documentation:

  • “Each adapted image is identified in the HTML with a data-fileentryid attribute.”

  • “For web content articles, Adaptive Media works only with images added from the file selector’s Documents and Media tab.”

  • “Adaptive Media doesn’t work with images added from the URL tab… providing no image file for Adaptive Media to copy.”

Problem Description

Expected Design (per documentation)

  • Adaptive Media relies on Document Library FileEntry

  • Adapted images are resolved using FileEntry ID (data-fileentryid)

  • URL-based images are explicitly unsupported

Actual Behavior

  1. Adaptive Media is enabled.

  2. An image is initially added correctly via Documents and Media.

  3. Adaptive Media variants are generated and served.

  4. Later, the image is updated by changing only the image URL (for example via HTML editing or an equivalent UI operation).

  5. The content is saved successfully.

Result (Critical Issue)

  • The image URL (src) is updated to a new image.

  • The FileEntry reference (data-fileentryid) remains unchanged.

  • Adaptive Media continues to serve the old adapted images generated from the previous FileEntry.

  • End users receive outdated or incorrect images, even though the content has been updated.

  • No validation error, warning, or cache invalidation occurs.

Depending on client conditions and caches, pages may:

  • Display old images instead of the updated ones, or

  • Mix new original images with old Adaptive Media variants

Why This Is a Serious Feature / UX Defect

This issue is particularly severe because:

  • It delivers incorrect content, not just unoptimized content.

  • The failure is silent and persistent.

  • Content editors cannot reliably detect the issue.

  • Cache and CDN behavior makes it non-deterministic.

  • Users reasonably expect that updating an image invalidates previous variants.

This cannot be mitigated by documentation or user training because:

  • The operation is possible through standard UI editing flows.

  • The resulting behavior violates content correctness expectations.

  • Recovery requires non-obvious actions (reselecting the image via Documents and Media).

In short:

A valid UI operation causes stale Adaptive Media content to be served without detection.

Expected / Desired Behavior

When Adaptive Media is enabled, the product should ensure at least one of the following:

  • Prevent URL-only image replacement for AM-managed images, or

  • Invalidate or regenerate Adaptive Media variants when the image source changes, or

  • Block saving or explicitly warn when the image URL and FileEntry reference are inconsistent

Regardless of implementation, serving stale Adaptive Media after content updates should not be possible.

Impact

  • Incorrect images delivered to end users

  • Loss of trust in content updates

  • High operational and troubleshooting cost

  • Significant risk in enterprise and public-facing sites