Struct DecodedPageCache 

pub struct DecodedPageCache { /* private fields */ }

Bounded LRU/CLOCK cache of decoded columnar pages (Phase 15.4). The PageCache above holds raw (ciphertext) page bytes; this second layer caches the post-decompress, post-decrypt typed page so a repeat scan skips decode entirely. Pages are immutable per (run_id, column_id, page_seq) identity (keyed via [crate::sorted_run::page_cache_key]), so there is no MVCC/invalidation concern — runs never change, and a rewritten page lives in a different run (different id) and simply misses here.

Governor attachment (with_governor) behaves exactly as on PageCache: live bytes are reported to the governor and evict_reclaimable is the governor-driven entry point (S1E-003 step 2).

Implementations

impl DecodedPageCache

pub fn new(capacity_bytes: u64) -> Self

pub fn with_governor(self, governor: MemoryGovernor, class: MemoryClass) -> Self

Attach a crate::memory::MemoryGovernor (S1E-003): the cache keeps one reservation sized to its live bytes and sheds entries when the governor denies growth. See PageCache::with_governor.

pub fn stats(&self) -> CacheStats

Cumulative hit/miss counts since construction (or the last reset_stats).

pub fn reset_stats(&self)

Zero the hit/miss counters.

pub fn try_get(&self, key: &[u8; 32]) -> Option<Arc<NativeColumn>>

Non-blocking probe used by the parallel scan path (&self, no write lock) — returns the cached decoded page on hit, None on miss.

pub fn insert(&mut self, key: [u8; 32], col: Arc<NativeColumn>)

Insert a decoded page (replaces any entry with the same key). Evicts cold entries (CLOCK) when over capacity.

pub fn evict_reclaimable(&mut self, budget: u64) -> u64

Governor entry point (S1E-003 step 2): evict at least budget bytes of reclaimable entries, returning the bytes actually freed.

pub fn len(&self) -> usize

pub fn is_empty(&self) -> bool

pub fn used_bytes(&self) -> u64