doslang-mirror/fec/std/map.fe

unit std.map;

// A table from a run of bytes to a value.
//
// A compiler looks names up constantly and a linear scan over a list is the
// wrong shape for that. Keys are copied into one buffer the map owns and each
// slot records where in it the key sits -- the arena-and-handle shape R11 asks
// for, which also means letting go of the map is two frees and not one per
// entry. Both buffers are owned, so R1 does the letting go and no `drop` is
// written here.
//
// Open addressing with linear probing. The table is a power of two so the
// index is a mask rather than a division, and it grows at three quarters full
// because probing gets long well before the table gets full.

pub struct Slot(V) {
    at: usize,
    len: usize,
    used: bool,
    value: V,
}

pub struct Map(V) {
    slots: ^[]mut Slot(V),
    bytes: ^[]mut u8,
    used_bytes: usize,
    count: usize,

    pub fn with_capacity(n: usize) -> !Self {
        var room: usize = 8;
        while room < n * 2 { room = room * 2; }
        let table: ^[]mut Slot(V) = try mem.alloc_slice(Slot(V), room);
        let text: ^[]mut u8 = try mem.alloc_slice(u8, 64);
        var i: usize = 0;
        while i < room {
            table.^[i].used = false;
            i = i + 1;
        }
        return Self{ slots: table, bytes: text, used_bytes: 0, count: 0 };
    }

    pub fn count_of(self: &Self) -> usize { return self.count; }

    pub fn room(self: &Self) -> usize { return self.slots.^.n; }

    /// Where `key` sits in the table: the slot holding it, or the first free
    /// slot it could go in. Probing stops at a free slot, which is why a slot
    /// is never cleared -- only ever filled.
    fn find(self: &Self, key: []u8) -> usize {
        let mask: usize = self.slots.^.n - 1;
        var at: usize = hash(key) & mask;
        while true {
            if not self.slots.^[at].used { return at; }
            if self.same(at, key) { return at; }
            at = (at + 1) & mask;
        }
        return 0;
    }

    fn same(self: &Self, slot: usize, key: []u8) -> bool {
        if self.slots.^[slot].len != key.n { return false; }
        let from: usize = self.slots.^[slot].at;
        var i: usize = 0;
        while i < key.n {
            if self.bytes.^[from + i] != key[i] { return false; }
            i = i + 1;
        }
        return true;
    }

    pub fn has(self: &Self, key: []u8) -> bool {
        return self.slots.^[self.find(key)].used;
    }

    pub fn get(self: &Self, key: []u8, missing: V) -> V {
        let at: usize = self.find(key);
        if self.slots.^[at].used { return self.slots.^[at].value; }
        return missing;
    }

    pub fn put(self: &mut Self, key: []u8, value: V) -> !void {
        if self.count * 4 >= self.slots.^.n * 3 { try self.regrow(); }
        let at: usize = self.find(key);
        if self.slots.^[at].used {
            self.slots.^[at].value = value;
            return;
        }
        try self.keep(key, at);
        self.slots.^[at].value = value;
        self.slots.^[at].used = true;
        self.count = self.count + 1;
        return;
    }

    /// Make sure `need` bytes fit, moving to a bigger buffer if they do not.
    ///
    /// Kept apart from `keep` because the borrow that swaps the buffer in is
    /// a borrow of the whole of `self` -- borrowing is tracked at the root --
    /// and it has to be over before any field is read again.
    fn ensure_room(self: &mut Self, need: usize) -> !void {
        let have: usize = self.bytes.^.n;
        if need <= have { return; }
        var room: usize = have;
        while room < need { room = room * 2; }
        let bigger: ^[]mut u8 = try mem.alloc_slice(u8, room);
        var k: usize = 0;
        let filled: usize = self.used_bytes;
        while k < filled {
            bigger.^[k] = self.bytes.^[k];
            k = k + 1;
        }
        let old: ^[]mut u8 = mem.replace(&mut self.bytes, bigger);
        mem.destroy(old);
        return;
    }

    /// Copy `key` into the byte buffer and point the slot at it.
    fn keep(self: &mut Self, key: []u8, slot: usize) -> !void {
        let at: usize = self.used_bytes;
        let need: usize = at + key.n;
        try self.ensure_room(need);
        var i: usize = 0;
        while i < key.n {
            self.bytes.^[at + i] = key[i];
            i = i + 1;
        }
        self.slots.^[slot].at = at;
        self.slots.^[slot].len = key.n;
        self.used_bytes = at + key.n;
        return;
    }

    /// Twice the slots, everything placed again. The keys do not move: they
    /// live in the byte buffer and the slots only point at them.
    fn regrow(self: &mut Self) -> !void {
        let bigger: ^[]mut Slot(V) = try mem.alloc_slice(Slot(V),
                                                        self.slots.^.n * 2);
        let mask: usize = bigger.^.n - 1;
        var i: usize = 0;
        while i < bigger.^.n {
            bigger.^[i].used = false;
            i = i + 1;
        }
        i = 0;
        while i < self.slots.^.n {
            if self.slots.^[i].used {
                let from: usize = self.slots.^[i].at;
                let len: usize = self.slots.^[i].len;
                var at: usize = hash(self.bytes.^[from..from + len]) & mask;
                while bigger.^[at].used { at = (at + 1) & mask; }
                bigger.^[at] = self.slots.^[i];
            }
            i = i + 1;
        }
        let old: ^[]mut Slot(V) = mem.replace(&mut self.slots, bigger);
        mem.destroy(old);
        return;
    }
}

/// FNV-1a. Small, fast, and good enough for identifiers; nothing here has to
/// resist an adversary choosing the keys.
pub fn hash(key: []u8) -> usize {
    var h: u32 = 2166136261;
    var i: usize = 0;
    while i < key.n {
        h = h ^ (key[i] as u32);
        h = h * 16777619;
        i = i + 1;
    }
    return h as usize;
}