0.1.3

Mounted Digital Miner scan speed pass.

• The mounted chunk-section scanner now resolves queueable mining targets directly from the section block state it already read.
• Removed the duplicate per-candidate world block lookup from the scan batch path; queued targets are still revalidated immediately before mining.
• Shared the mounted target rules between scan-time selection and final pre-mine revalidation so filtering stays consistent.
• Raised the default mounted scan budget from 8,192 to 16,384 inspected positions per pass.
• Added a 12 ms soft time budget per mounted scan pass so slower servers can yield before spending a full tick on one miner.
• Raised the default visible mounted target queue from 8,192 to 32,768 targets.
• Stops a scan pass as soon as the remaining queue capacity is filled, avoiding wasted discovery work on dense inverse-filter scans.

Algorithm notes:

• The scanner treats the Digital Miner work area as chunk-section jobs. It skips unloaded sections, air-only sections, and palette-impossible sections before walking block states.
• For sections that can match, it now uses the already-read section BlockState to produce queueable targets. The previous build re-read the world block for each candidate before queueing it.
• The final mining step still revalidates the queued block immediately before breaking, so the fast path does not trust stale scan data for mutations.
• The scan and pre-mine checks now share the same mounted target rules.

Research references used while choosing the scan shape:

• John Amanatides and Andrew Woo, "A Fast Voxel Traversal Algorithm for Ray Tracing" (1987): https://doi.org/10.2312/egtp.19871000
• Samuli Laine and Tero Karras, "Efficient Sparse Voxel Octrees" (2010): https://research.nvidia.com/publication/2010-02_efficient-sparse-voxel-octrees
• Samuli Laine and Tero Karras, "Efficient Sparse Voxel Octrees - Analysis, Extensions, and Implementation" (2010): https://research.nvidia.com/publication/2010-02_efficient-sparse-voxel-octrees-analysis-extensions-and-implementation