//! Repository integrity check (spec §21.1): `oxy fsck`.
//!
//! Reads everything and changes nothing. Four passes:
//!
//! 1. every stored object: header, length, hash, and that its body decodes
//!    as its type says;
//! 2. the operation log: every operation reachable from the head is
//!    present and decodes, and clocks never run backwards;
//! 3. references: the current view's bookmarks, tags, remote bookmarks,
//!    workspaces and policy, then the dependency closure of everything
//!    they name, with each change's author identity, blobs and chunks,
//!    and each change's signatures or seals;
//! 4. derived data: every bookmark's state validates, every checkpoint
//!    restores exactly the state it names, and reference and tag
//!    attestations carry good signatures.
//!
//! Missing blobs are warnings, not errors, in a partial clone
//! (`filter.max-blob` or a sparse filter), and changes beyond a shallow
//! boundary are expected to be absent.

use super::*;
use std::collections::{HashSet, VecDeque};

#[derive(Clone, Copy, Debug, PartialEq, Eq, PartialOrd, Ord)]
pub enum Severity {
    /// Worth knowing; the repository is still sound.
    Warning,
    /// Data is missing, corrupt or fails verification.
    Error,
}

#[derive(Clone, Debug, PartialEq, Eq)]
pub struct Problem {
    pub severity: Severity,
    /// The object the problem is about, when there is one.
    pub object: Option<ObjectRef>,
    pub what: String,
}

#[derive(Clone, Debug, Default, PartialEq, Eq)]
pub struct FsckReport {
    pub objects: usize,
    pub ops: usize,
    /// Changes checked, working-copy changes included.
    pub changes: usize,
    /// Committed changes authenticated by their own signature or by a seal.
    pub signed: usize,
    pub unsigned: usize,
    pub checkpoints: usize,
    pub attestations: usize,
    pub problems: Vec<Problem>,
}

impl FsckReport {
    pub fn errors(&self) -> usize {
        self.problems
            .iter()
            .filter(|p| p.severity == Severity::Error)
            .count()
    }

    pub fn warnings(&self) -> usize {
        self.problems
            .iter()
            .filter(|p| p.severity == Severity::Warning)
            .count()
    }

    fn error(&mut self, object: Option<ObjectRef>, what: impl Into<String>) {
        self.problems.push(Problem {
            severity: Severity::Error,
            object,
            what: what.into(),
        });
    }

    fn warn(&mut self, object: Option<ObjectRef>, what: impl Into<String>) {
        self.problems.push(Problem {
            severity: Severity::Warning,
            object,
            what: what.into(),
        });
    }
}

/// Whether `body` decodes as an object of type `ty`.
fn decodes(ty: ObjectType, body: &[u8]) -> std::result::Result<(), String> {
    let r = match ty {
        ObjectType::Chunk => Ok(()),
        ObjectType::Blob => Blob::decode(body).map(|_| ()),
        // Structure only: the text blobs are checked with the change graph.
        ObjectType::Change => {
            Change::decode_with(body, &mut |_, _| Ok(vec![Vec::new()])).map(|_| ())
        }
        ObjectType::Checkpoint => crate::objects::Checkpoint::decode(body).map(|_| ()),
        ObjectType::Op => match Op::decode(body) {
            Ok(_) => Ok(()),
            Err(e) => Op::decode_format1(body).map(|_| ()).map_err(|_| e),
        },
        ObjectType::Attest => Attest::decode(body).map(|_| ()),
        ObjectType::Identity => Identity::decode(body).map(|_| ()),
        ObjectType::Policy => crate::policy::Policy::decode(body).map(|_| ()),
        ObjectType::Receipt => crate::receipt::Receipt::decode(body).map(|_| ()),
        ObjectType::State => crate::objects::StateObject::decode(body).map(|_| ()),
    };
    r.map_err(|e| e.to_string())
}

impl Repo {
    /// Check the whole repository; see the module documentation.
    pub fn fsck(&self) -> Result<FsckReport> {
        let mut rep = FsckReport::default();
        let bad = {
            let _t = crate::timing::phase("fsck: objects");
            self.fsck_objects(&mut rep)?
        };
        {
            let _t = crate::timing::phase("fsck: operations");
            self.fsck_ops(&mut rep, &bad);
        }
        {
            let _t = crate::timing::phase("fsck: references");
            self.fsck_references(&mut rep, &bad)?;
        }
        {
            let _t = crate::timing::phase("fsck: states and checkpoints");
            self.fsck_derived(&mut rep)?;
        }
        rep.problems.sort_by_key(|p| std::cmp::Reverse(p.severity));
        Ok(rep)
    }

    /// Pass 1. Returns the objects that failed, so later passes report
    /// them once rather than at every reference.
    fn fsck_objects(&self, rep: &mut FsckReport) -> Result<HashSet<ObjectRef>> {
        let mut bad = HashSet::new();
        for r in self.store.list()? {
            rep.objects += 1;
            match self.store.get(&r) {
                Ok((ty, body)) => {
                    if let Err(e) = decodes(ty, &body) {
                        rep.error(
                            Some(r),
                            format!("{} object does not decode: {e}", ty.name()),
                        );
                        bad.insert(r);
                    }
                }
                Err(e) => {
                    rep.error(Some(r), format!("unreadable object: {e}"));
                    bad.insert(r);
                }
            }
        }
        Ok(bad)
    }

    /// Pass 2.
    fn fsck_ops(&self, rep: &mut FsckReport, bad: &HashSet<ObjectRef>) {
        let head = match fs::read_to_string(self.oxy.join("ops").join("head")) {
            Ok(h) => h,
            Err(e) => {
                rep.error(None, format!("cannot read .oxy/ops/head: {e}"));
                return;
            }
        };
        let Some(head) = ObjectRef::from_hex(head.trim()) else {
            rep.error(None, ".oxy/ops/head does not hold an object reference");
            return;
        };
        let mut seen = HashSet::new();
        let mut queue: VecDeque<(ObjectRef, Option<(ObjectRef, u64)>)> =
            VecDeque::from([(head, None)]);
        while let Some((r, child)) = queue.pop_front() {
            if !seen.insert(r) {
                continue;
            }
            if bad.contains(&r) {
                continue;
            }
            let body = match self.store.get_typed(&r, ObjectType::Op) {
                Ok(b) => b,
                Err(StoreError::NotFound(_)) => {
                    let by = child
                        .map(|(c, _)| format!(", the parent of operation {}", c.short(12)))
                        .unwrap_or_default();
                    rep.error(
                        Some(r),
                        format!("operation missing{by}; the log is cut here"),
                    );
                    continue;
                }
                Err(e) => {
                    rep.error(Some(r), format!("operation unreadable: {e}"));
                    continue;
                }
            };
            rep.ops += 1;
            let (parents, wall_ms) = match Op::decode(&body) {
                Ok(op) => (op.parents, op.hlc.wall_ms),
                Err(_) => match Op::decode_format1(&body) {
                    Ok(op) => (op.parents, op.hlc.wall_ms),
                    Err(e) => {
                        rep.error(Some(r), format!("operation does not decode: {e}"));
                        continue;
                    }
                },
            };
            if let Some((c, child_ms)) = child {
                if wall_ms > child_ms {
                    rep.warn(
                        Some(c),
                        format!(
                            "operation is timed before its parent {} (a clock was wrong)",
                            r.short(12)
                        ),
                    );
                }
            }
            queue.extend(parents.into_iter().map(|p| (p, Some((r, wall_ms)))));
        }
    }

    /// Pass 3.
    fn fsck_references(&self, rep: &mut FsckReport, bad: &HashSet<ObjectRef>) -> Result<()> {
        let view = match self.view() {
            Ok(v) => v,
            Err(e) => {
                rep.error(None, format!("cannot read the current view: {e}"));
                return Ok(());
            }
        };
        let boundary = self.boundary();
        let partial = self.config.max_blob.is_some() || !self.config.sparse_paths.is_empty();

        let mut roots: Vec<(ChangeHash, String)> = Vec::new();
        let mut states: Vec<(ObjectRef, String)> = Vec::new();
        states.extend(
            view.bookmarks
                .iter()
                .map(|(n, s)| (*s, format!("bookmark {n}"))),
        );
        states.extend(view.tags.iter().map(|(n, s)| (*s, format!("tag {n}"))));
        for (remote, bms) in &view.remotes {
            states.extend(bms.iter().map(|(n, s)| (*s, format!("{remote}/{n}"))));
        }
        for (st, what) in states {
            match self.store.get_state(&st) {
                Ok(s) => roots.extend(s.heads.into_iter().map(|h| (h, what.clone()))),
                Err(e) if !bad.contains(&st) => {
                    rep.error(Some(st), format!("state of {what} unusable: {e}"))
                }
                Err(_) => {}
            }
        }
        for (ws, h) in &view.workspaces {
            roots.push((*h, format!("workspace {}", ws.to_hex())));
        }
        if let Some(p) = view.policy {
            match self.store.get_typed(&p, ObjectType::Policy) {
                Ok(bytes) => {
                    if let Ok(policy) = crate::policy::Policy::decode(&bytes) {
                        for id in &policy.identities {
                            if !self.store.has(&id.hash()) {
                                rep.warn(
                                    Some(id.hash()),
                                    "an identity the policy names is not stored",
                                );
                            }
                        }
                    }
                }
                Err(e) if !bad.contains(&p) => {
                    rep.error(Some(p), format!("the policy is unusable: {e}"))
                }
                Err(_) => {}
            }
        }

        // Working-copy changes are never signed; they are checked but not tallied.
        let working: HashSet<ChangeHash> = view.workspaces.values().copied().collect();
        let mut seen: HashSet<ChangeHash> = HashSet::new();
        let mut blobs_seen: HashSet<ObjectRef> = HashSet::new();
        let mut stack = roots;
        while let Some((h, needed_by)) = stack.pop() {
            if !seen.insert(h) {
                continue;
            }
            // The skeleton: the text is checked as the blobs it is stored
            // in, not decoded (the objects pass checked every object).
            let (c, texts) = match self.store.get_change_lazy(&h) {
                Ok(c) => c,
                Err(StoreError::NotFound(_)) if boundary.contains(&h) => continue,
                Err(StoreError::NotFound(_)) => {
                    rep.error(Some(h), format!("change missing, needed by {needed_by}"));
                    continue;
                }
                Err(e) => {
                    if !bad.contains(&h) {
                        rep.error(Some(h), format!("change unusable: {e}"));
                    }
                    continue;
                }
            };
            rep.changes += 1;
            let label = format!("change {}", h.short(12));
            stack.extend(c.deps.iter().map(|d| (*d, label.clone())));
            if !c.author.is_zero() && !self.store.has(&c.author) {
                rep.warn(
                    Some(h),
                    format!("author identity {} is not stored", c.author.short(12)),
                );
            }
            for t in texts {
                if let crate::objects::TextRef::Blob(b, _, _) = t {
                    if blobs_seen.insert(b) {
                        // A change cannot be read without its text: never partial.
                        self.fsck_blob(rep, b, &label, false, bad);
                    }
                }
            }
            for e in &c.edits {
                let Edit::Replace { old, new, .. } = e else {
                    continue;
                };
                for b in [old, new].into_iter().flatten() {
                    if !blobs_seen.insert(*b) {
                        continue;
                    }
                    self.fsck_blob(rep, *b, &label, partial, bad);
                }
            }
            if working.contains(&h) {
                continue;
            }
            match self.authenticate_change(&h) {
                Ok(SignatureStatus::Valid { .. }) => rep.signed += 1,
                Ok(SignatureStatus::Unsigned) => rep.unsigned += 1,
                Ok(SignatureStatus::Invalid(why)) => {
                    rep.error(Some(h), format!("bad signature: {why}"))
                }
                Err(e) => rep.error(Some(h), format!("cannot check signatures: {e}")),
            }
        }
        Ok(())
    }

    fn fsck_blob(
        &self,
        rep: &mut FsckReport,
        b: ObjectRef,
        used_by: &str,
        partial: bool,
        bad: &HashSet<ObjectRef>,
    ) {
        match self.store.get_typed(&b, ObjectType::Blob) {
            Ok(bytes) => {
                let Ok(blob) = Blob::decode(&bytes) else {
                    return;
                };
                for ch in blob.chunks {
                    if !self.store.has(&ch) {
                        rep.error(
                            Some(ch),
                            format!("chunk missing from blob {} ({used_by})", b.short(12)),
                        );
                    }
                }
            }
            Err(StoreError::NotFound(_)) if partial => rep.warn(
                Some(b),
                format!("blob not fetched ({used_by}); expected in a partial clone"),
            ),
            Err(StoreError::NotFound(_)) => rep.error(Some(b), format!("blob missing ({used_by})")),
            Err(e) if !bad.contains(&b) => {
                rep.error(Some(b), format!("blob unusable ({used_by}): {e}"))
            }
            Err(_) => {}
        }
    }

    /// Whether `heads` form a valid state: its history loads and every file
    /// graph builds. Graphs are built a batch at a time and dropped, so the
    /// check needs the memory of a batch, not of the whole tree.
    fn check_state(&self, heads: &[ChangeHash]) -> Result<()> {
        let mut st = self.load_state_history(heads)?;
        let nodes = st.unbuilt_nodes();
        for batch in nodes.chunks(2048) {
            self.load_graphs_of(&mut st, batch)?;
            st.drop_graphs(batch);
        }
        Ok(())
    }

    /// Pass 4.
    fn fsck_derived(&self, rep: &mut FsckReport) -> Result<()> {
        let Ok(view) = self.view() else { return Ok(()) };
        let shallow = self.is_shallow();
        if !shallow {
            let mut checked: HashSet<ObjectRef> = HashSet::new();
            for (name, st) in &view.bookmarks {
                if !checked.insert(*st) {
                    continue;
                }
                let Ok(s) = self.store.get_state(st) else {
                    continue;
                };
                if let Err(e) = self.check_state(&s.heads) {
                    rep.error(
                        Some(*st),
                        format!("bookmark {name} does not form a valid state: {e}"),
                    );
                }
            }
        }
        for r in self.checkpoint_refs()? {
            rep.checkpoints += 1;
            let cp = match self.store.get_typed(&r, ObjectType::Checkpoint) {
                Ok(b) => match crate::objects::Checkpoint::decode(&b) {
                    Ok(cp) => cp,
                    Err(_) => continue,
                },
                Err(StoreError::NotFound(_)) => {
                    rep.warn(
                        Some(r),
                        "checkpoint pointer names a missing checkpoint; it will be rewritten",
                    );
                    continue;
                }
                Err(_) => continue,
            };
            let payload = match self.read_blob(&cp.payload) {
                Ok(p) => p,
                Err(e) => {
                    rep.error(Some(r), format!("checkpoint payload unreadable: {e}"));
                    continue;
                }
            };
            if let Err(e) = self.state_from_checkpoint(&payload) {
                rep.error(Some(r), format!("checkpoint payload does not decode: {e}"));
                continue;
            }
            // The payload must be what the named changes produce, in its own version.
            let version = crate::cbor::decode(&payload).ok().and_then(|v| {
                crate::objects::map(&v, "payload").ok().and_then(|m| {
                    m.get(&0)
                        .and_then(|x| crate::objects::uint(x, "version").ok())
                })
            });
            let expected = |full: &State| {
                if version == Some(crate::state::CHECKPOINT_INDEX_VERSION) {
                    full.checkpoint_index(&full.checkpoint_file_graphs(), Repo::blob_ref)
                } else {
                    full.checkpoint_payload()
                }
            };
            if !shallow {
                match self.load_state_full(&cp.heads) {
                    Ok(full) if expected(&full) != payload => {
                        rep.error(Some(r), "checkpoint disagrees with the changes it names")
                    }
                    Ok(_) => {}
                    Err(e) => rep.error(
                        Some(r),
                        format!("checkpoint names changes that do not form a state: {e}"),
                    ),
                }
            }
        }
        let mut attests = Vec::new();
        for name in view.bookmarks.keys() {
            if let Ok(Some((r, _))) = self.latest_ref_attest(name) {
                attests.push((r, format!("bookmark {name}")));
            }
        }
        for name in view.tags.keys() {
            if let Ok(Some((r, _))) = self.tag_attest(name) {
                attests.push((r, format!("tag {name}")));
            }
        }
        for (r, what) in attests {
            rep.attestations += 1;
            let sigs = self.store.signatures(&r)?;
            if sigs.is_empty() {
                rep.warn(Some(r), format!("attestation for {what} is unsigned"));
            }
            for sig in sigs {
                match self.store.get_identity(&sig.key) {
                    Ok(id) => {
                        if let Err(e) = verify_signature(&id, &sig, Domain::Attest, &r) {
                            rep.error(
                                Some(r),
                                format!("attestation for {what} has a bad signature: {e}"),
                            );
                        }
                    }
                    Err(_) => rep.warn(
                        Some(r),
                        format!(
                            "attestation for {what} is signed by {}, whose identity is not stored",
                            sig.key.short(12)
                        ),
                    ),
                }
            }
        }
        Ok(())
    }
}