Use macOS Available space to learn how much more the disk can hold, CoreRobin's space map to locate path-based allocated usage, and Finder for an individual file's logical content size. Compare the same path at the same time and let the scan finish first.
First separate three measurements
The length of data represented by a file. A sparse file can report a large length without allocating blocks for every empty region.
Disk blocks allocated by the file system. Compression, sparsity, and shared data can make this differ from logical size.
Space the whole volume can currently provide, potentially including content the system can reclaim. It is not the sum of folder sizes.
Six common sources of differences
-
1
Compressed and sparse files
Content length and allocated blocks differ. Virtual-machine disks, databases, and some caches commonly behave this way.
-
2
Hard links and shared content
Multiple paths can point to the same underlying data. Adding every path in full double-counts storage, so a scanner must identify the file object.
-
3
APFS clones, snapshots, and reclaimable space
Shared blocks cannot always be assigned to one path. Snapshots and system-reclaimable content affect volume availability without necessarily appearing in an ordinary directory walk.
-
4
iCloud and other cloud placeholders
Finder can show a cloud item's logical size even when the content is not fully downloaded. CoreRobin skips cloud file-provider roots to avoid triggering large downloads merely to measure them.
-
5
Permissions and unreadable locations
Without Full Disk Access, Mail, Messages, and other app data may be unreadable. CoreRobin preserves the gap instead of turning unknown content into an estimated total.
-
6
Files change during the scan
Browser caches, logs, downloads, and virtual-machine disks can change while being scanned. Finder, System Settings, and CoreRobin will differ if sampled at different times.
How CoreRobin calculates usage
- On macOS and Linux, it uses file-system allocated blocks rather than only file length.
- Within a scan root, it identifies hard links by device and file identity so the same storage is not counted twice.
- It does not follow symbolic links or cross onto another volume, preventing loops, boundary escapes, and external storage from being counted as the system disk.
- The result is a snapshot at completion. Before deletion, CoreRobin rechecks the path, file identity, and changes, and stops if they no longer match.
Which number should you use?
Do not delete unfamiliar files merely to make two totals match. The goal is to find content you understand and no longer need, then verify the outcome using volume-level available space.