Blender Image Editor ↔ External File System: Cache Consistency When Source Images Change

Answer to the Core Questions

Is the current cache invalidation behavior documented as a deliberate design choice across all supported Blender versions?

Yes. The Blender Manual (Help → Manual → "Image Editor → Caching") states that image data is cached in memory on first load and is not automatically refreshed when the source file changes. This is presented as a performance‑over‑consistency trade‑off that has been stable since the 2.5 release series.

Are there version‑independent preferences or official operators that guarantee automatic reloading of images when the underlying file changes?

No. The only built‑in mechanisms are:

• Manual reload via the Image → Reload menu or the Reload button in the Image Editor header.
• The bpy.ops.image.reload() Python operator, which must be called in a context where an image is active.
• The Auto Refresh checkbox in the Image Editor sidebar (N‑panel → Image). This option only triggers a refresh when the area gains focus or a frame change occurs; it does not monitor the filesystem.

There is no single, version‑independent preference that guarantees automatic invalidation on external file change.

How do render engines that read directly from disk interact with this caching policy, and does that affect the need for a unified cache‑invalidation strategy?

Cycles and Eevee both use Blender’s central image cache during rendering. If the cache contains stale data, the render will display the outdated image until a manual reload is performed. Add‑on render engines that bypass the cache and read directly from disk will see the latest file, but this behavior is engine‑specific and not part of Blender’s core policy. Because all built‑in engines rely on the cache, a unified strategy is required only if you rely on the default rendering pipeline.

Practical Steps for Keeping Images Fresh

1. Open the image in the Image Editor.

2. After editing the file externally, click Image → Reload or press Ctrl‑R (default shortcut).

3. To automate for a single session, enable the Auto Refresh checkbox. Remember it will not catch rapid changes while the editor is unfocused.

4. For scripts or add‑ons that need to reload programmatically, use:

   import bpy
   bpy.ops.image.reload()  # Works when an image is active in the context

5. Before rendering, run a pre‑render script that iterates over all images and calls bpy.ops.image.reload() to ensure the cache is up to date.

Diagnostic Detail (if needed)

If you find that the Auto Refresh option is still not catching changes, confirm that the Image Editor area is the active context and that the file modification time is newer than the cached timestamp. This can be verified by printing image.filepath and image.filepath_user in a Python console.