Initial scan is very slow on some cloud-mounted drives (confirmed on Google Drive mounts)

[StippieDot]

Summary

Initial scanning can be extremely slow on some cloud-mounted drives, even when thumbnail generation and metadata extraction are not the bottleneck.

This is confirmed on mounted Google Drive setups. It is likely also true for other cloud storage services, depending on how the mount software and remote service cache directory listings.

Observed behavior

• The slowdown happens during the initial recursive folder scan/loading phase.
• After a reboot, scanning a large cloud-mounted drive can be very slow.
• If VideoCull is closed and reopened, scanning is fast again up to roughly the point the previous scan had already reached.
• This strongly suggests the current scan path benefits from temporary OS / mount-layer caching, but VideoCull does not yet maintain its own persistent scan index.

What does not seem to be the main problem

• Thumbnail generation is not the main bottleneck here.
• Metadata extraction is not the main bottleneck here.
• The expensive part is the initial recursive discovery/loading of directories and video files.

Scope notes

• Confirmed problematic: mounted Google Drive.
• Likely also affects: other cloud-mounted services.
• SFTP, WebDAV, and SMB can behave normally, but actual scan speed still depends on how quickly the remote server and mount layer allow recursive directory reads.

Expected behavior

VideoCull should be able to show previously discovered directories/files immediately from a local persistent index, then continue scanning in the background for unknown, stale, pending, or failed directories.

Potential direction

Add a persistent local scan index so scan progress can survive app restarts and cold remote-filesystem caches.