//! Loaded states cached by their heads (hub spec §16.3). A state is a pure

//! function of its heads, so entries never go stale; the cache is only

//! bounded. Shared across threads because `oxy_core::state::State` is

//! `Send + Sync`.

use oxy_core::hash::ChangeHash;

use oxy_core::repo::Repo;

use oxy_core::state::{Snapshot, State};

use std::collections::HashMap;

use std::sync::{Arc, Mutex};

/// A state and its materialised snapshot.

pub struct Loaded {

    pub heads: Vec<ChangeHash>,

    pub state: State,

    pub snapshot: Snapshot,

}

#[derive(Clone, Debug, PartialEq, Eq, Hash)]

struct Key {

    repo: [u8; 16],

    heads: Vec<ChangeHash>,

}

struct Entry {

    loaded: Arc<Loaded>,

    last_used: u64,

    /// Its estimated size in bytes (`weight`).

    weight: usize,

}

/// A loaded state's estimated size (oxygenhub#246): five bytes per byte of

/// text and 384 per line (each file's line graph), and a kilobyte per change.

/// Measured on 2026-10-17 against what loading a state really adds to the

/// process (`measures_a_real_state`): oxygen's states ~85 MiB, oxygenhub's

/// ~70 MiB, which this puts at 88 and 75. The older "text three times over"

/// said 18 and 28, so a cache that thought it held 2.5 GB held 6-8 GB.

fn weight(l: &Loaded) -> usize {

    let (mut text, mut lines) = (0usize, 0usize);

    for f in l.snapshot.files.values() {

        match &f.content {

            oxy_core::state::FileContent::Text { bytes, .. } => {

                text += bytes.len();

                lines += bytes.iter().filter(|b| **b == b'\n').count() + 1;

            }

            _ => text += 64,

        }

    }

    text.saturating_mul(5).saturating_add(lines.saturating_mul(384)).saturating_add(l.state.len().saturating_mul(1024))

}

/// What the state cache may hold, in bytes: OXYHUB_STATE_CACHE_MB, else

/// 3 GiB. A count alone let a few large repositories' states grow the hub

/// to 17 GB.

fn budget() -> usize {

    std::env::var("OXYHUB_STATE_CACHE_MB").ok().and_then(|v| v.parse::<usize>().ok()).unwrap_or(3 * 1024).saturating_mul(1 << 20)

}

/// What the bookmark lists and a change's "in bookmarks" need of a state:

/// small, so many are kept, and computed without filling the state cache.

/// A repository with more bookmarks than the state cache holds otherwise

/// cycled through it on every such page, so nothing stayed cached

/// (metalgear's 128 bookmarks: 7.7 s per change page, and every page of

/// every repository slower after it).

pub struct Summary {

    pub changes: std::collections::HashSet<ChangeHash>,

    pub latest_ms: u64,

    pub conflicted: bool,

}

/// Summaries kept; each is the state's change set, a few hundred kilobytes

/// for a long history.

const SUMMARIES: usize = 2048;

/// Summaries by heads, each with when it was last used, and the clock.

type Summaries = (HashMap<Key, (Arc<Summary>, u64)>, u64);

/// A bounded least-recently-used cache of loaded states.

pub struct StateCache {

    capacity: usize,

    /// Bytes of estimated state it may hold (`budget`).

    budget: usize,

    inner: Mutex<(HashMap<Key, Entry>, u64)>,

    summaries: Mutex<Summaries>,

}

impl StateCache {

    pub fn new(capacity: usize) -> StateCache {

        StateCache::with_budget(capacity, budget())

    }

    pub fn with_budget(capacity: usize, budget: usize) -> StateCache {

        StateCache { capacity: capacity.max(1), budget, inner: Mutex::new((HashMap::new(), 0)), summaries: Mutex::new((HashMap::new(), 0)) }

    }

    /// The summary of the state with these heads: from the state cache when

    /// it holds the state, else loaded and dropped, the state cache left as

    /// it was.

    pub fn summary(&self, repo: &Repo, heads: &[ChangeHash]) -> Result<Arc<Summary>, oxy_core::repo::RepoError> {

        let mut sorted = heads.to_vec();

        sorted.sort();

        sorted.dedup();

        let key = Key { repo: repo.id()?, heads: sorted.clone() };

        {

            let mut guard = self.summaries.lock().unwrap();

            let (map, tick) = &mut *guard;

            *tick += 1;

            if let Some(e) = map.get_mut(&key) {

                e.1 = *tick;

                return Ok(e.0.clone());

            }

        }

        let held = self.inner.lock().unwrap().0.get(&key).map(|e| e.loaded.clone());

        let summary = Arc::new(match held {

            Some(l) => summarise(&l.state, &l.snapshot),

            None => {

                let state = repo.load_state(&sorted)?;

                let snapshot = state.materialize();

                summarise(&state, &snapshot)

            }

        });

        let mut guard = self.summaries.lock().unwrap();

        let (map, tick) = &mut *guard;

        if map.len() >= SUMMARIES {

            if let Some(oldest) = map.iter().min_by_key(|(_, e)| e.1).map(|(k, _)| k.clone()) {

                map.remove(&oldest);

            }

        }

        map.insert(key, (summary.clone(), *tick));

        Ok(summary)

    }

    /// The state with these heads, loaded once per hub process.

    pub fn load(&self, repo: &Repo, heads: &[ChangeHash]) -> Result<Arc<Loaded>, oxy_core::repo::RepoError> {

        let mut sorted = heads.to_vec();

        sorted.sort();

        sorted.dedup();

        let key = Key { repo: repo.id()?, heads: sorted.clone() };

        {

            let mut guard = self.inner.lock().unwrap();

            let (map, tick) = &mut *guard;

            *tick += 1;

            if let Some(e) = map.get_mut(&key) {

                e.last_used = *tick;

                return Ok(e.loaded.clone());

            }

        }

        let state = repo.load_state(&sorted)?;

        let snapshot = state.materialize();

        let loaded = Arc::new(Loaded { heads: sorted, state, snapshot });

        let w = weight(&loaded);

        let mut guard = self.inner.lock().unwrap();

        let (map, tick) = &mut *guard;

        // The oldest go until it fits by count and by size; the new one

        // stays even when it alone is over the budget.

        let budget = self.budget;

        let mut total: usize = map.values().map(|e| e.weight).sum::<usize>() + w;

        while !map.is_empty() && (map.len() >= self.capacity || total > budget) {

            let Some(oldest) = map.iter().min_by_key(|(_, e)| e.last_used).map(|(k, _)| k.clone()) else {

                break;

            };

            if let Some(e) = map.remove(&oldest) {

                total -= e.weight;

            }

        }

        map.insert(key, Entry { loaded: loaded.clone(), last_used: *tick, weight: w });

        Ok(loaded)

    }

    pub fn len(&self) -> usize {

        self.inner.lock().unwrap().0.len()

    }

    pub fn is_empty(&self) -> bool {

        self.len() == 0

    }

}

fn summarise(state: &State, snapshot: &Snapshot) -> Summary {

    Summary {

        changes: state.changes().map(|(h, _)| *h).collect(),

        latest_ms: state.changes().map(|(_, m)| m.hlc.wall_ms).max().unwrap_or(0),

        conflicted: snapshot.is_conflicted(),

    }

}

#[cfg(test)]

mod tests {

    use super::*;

    /// What a loaded state really costs against `weight` (oxygenhub#246):

    /// `OXYHUB_MEASURE_REPO=path cargo test -p oxygenhub --lib

    /// measures_a_real_state -- --ignored --nocapture`.

    #[test]

    #[ignore]

    fn measures_a_real_state() {

        fn rss_kib() -> usize {

            let s = std::fs::read_to_string("/proc/self/status").unwrap();

            s.lines().find(|l| l.starts_with("VmRSS:")).and_then(|l| l.split_whitespace().nth(1)).unwrap().parse().unwrap()

        }

        let path = std::env::var("OXYHUB_MEASURE_REPO").expect("OXYHUB_MEASURE_REPO");

        let repo = Repo::open(std::path::Path::new(&path)).unwrap();

        let view = repo.view().unwrap();

        // Kept alive, so each load's memory is new, not reused.

        let cache = StateCache::with_budget(64, usize::MAX);

        let mut held = Vec::new();

        for bookmark in view.bookmarks.keys().take(6).cloned().collect::<Vec<_>>() {

            let heads = repo.heads_of(&view, &bookmark).unwrap();

            let before = rss_kib();

            let loaded = cache.load(&repo, &heads).unwrap();

            let after = rss_kib();

            let files = loaded.snapshot.files.len();

            let (mut text, mut lines) = (0usize, 0usize);

            for f in loaded.snapshot.files.values() {

                if let oxy_core::state::FileContent::Text { bytes, .. } = &f.content {

                    text += bytes.len();

                    lines += bytes.iter().filter(|b| **b == b'\n').count();

                }

            }

            println!(

                "{bookmark}: {} changes, {files} files, {} KiB text, {lines} lines: weight {} MiB, resident +{} MiB",

                loaded.state.len(),

                text >> 10,

                weight(&loaded) >> 20,

                (after.saturating_sub(before)) >> 10

            );

            held.push(loaded);

        }

    }

    #[test]

    fn caches_by_heads_and_evicts_the_oldest() {

        let dir = tempfile::tempdir().unwrap();

        let repo = Repo::init_bare(dir.path()).unwrap();

        let cache = StateCache::new(2);

        let a = cache.load(&repo, &[]).unwrap();

        let b = cache.load(&repo, &[]).unwrap();

        assert!(Arc::ptr_eq(&a, &b), "same heads, same entry");

        assert_eq!(cache.len(), 1);

        let h1 = oxy_core::hash::ObjectRef([1; 32]);

        let h2 = oxy_core::hash::ObjectRef([2; 32]);

        // Unknown heads fail to load and are not cached.

        assert!(cache.load(&repo, &[h1]).is_err());

        assert_eq!(cache.len(), 1);

        let _ = (h2, dir);

    }

    #[test]

    fn summaries_leave_the_state_cache_alone() {

        let dir = tempfile::tempdir().unwrap();

        let repo = Repo::init_bare(dir.path()).unwrap();

        let cache = StateCache::new(2);

        let a = cache.summary(&repo, &[]).unwrap();

        assert_eq!(cache.len(), 0, "a summary does not fill the state cache");

        assert!(Arc::ptr_eq(&a, &cache.summary(&repo, &[]).unwrap()), "kept");

        assert!(a.changes.is_empty() && !a.conflicted);

        assert!(cache.summary(&repo, &[oxy_core::hash::ObjectRef([3; 32])]).is_err());

    }

    #[test]

    fn a_size_budget_evicts_the_oldest_and_keeps_the_newest() {

        let dir = tempfile::tempdir().unwrap();

        let repo = Repo::init(dir.path()).unwrap();

        std::fs::write(dir.path().join("a.txt"), "a\n".repeat(1000)).unwrap();

        let one = repo.commit("one", &[]).unwrap();

        std::fs::write(dir.path().join("b.txt"), "b\n".repeat(1000)).unwrap();

        let two = repo.commit("two", &[]).unwrap();

        // Room for nothing: each load stays, the one before goes.

        let cache = StateCache::with_budget(64, 0);

        cache.load(&repo, &[one]).unwrap();

        assert_eq!(cache.len(), 1, "kept though over the budget");

        cache.load(&repo, &[two]).unwrap();

        assert_eq!(cache.len(), 1, "the older one went");

        // Room for plenty: both stay.

        let roomy = StateCache::with_budget(64, 1 << 30);

        roomy.load(&repo, &[one]).unwrap();

        roomy.load(&repo, &[two]).unwrap();

        assert_eq!(roomy.len(), 2);

    }

}