zig/lib/compiler/aro/aro/Value.zig
Ryan Liptak 76fb2b685b std: Convert deprecated aliases to compile errors and fix usages
Deprecated aliases that are now compile errors:

- `std.fs.MAX_PATH_BYTES` (renamed to `std.fs.max_path_bytes`)
- `std.mem.tokenize` (split into `tokenizeAny`, `tokenizeSequence`, `tokenizeScalar`)
- `std.mem.split` (split into `splitSequence`, `splitAny`, `splitScalar`)
- `std.mem.splitBackwards` (split into `splitBackwardsSequence`, `splitBackwardsAny`, `splitBackwardsScalar`)
- `std.unicode`
  + `utf16leToUtf8Alloc`, `utf16leToUtf8AllocZ`, `utf16leToUtf8`, `fmtUtf16le` (all renamed to have capitalized `Le`)
  + `utf8ToUtf16LeWithNull` (renamed to `utf8ToUtf16LeAllocZ`)
- `std.zig.CrossTarget` (moved to `std.Target.Query`)

Deprecated `lib/std/std.zig` decls were deleted instead of made a `@compileError` because the `refAllDecls` in the test block would trigger the `@compileError`. The deleted top-level `std` namespaces are:

- `std.rand` (renamed to `std.Random`)
- `std.TailQueue` (renamed to `std.DoublyLinkedList`)
- `std.ChildProcess` (renamed/moved to `std.process.Child`)

This is not exhaustive. Deprecated aliases that I didn't touch:
  + `std.io.*`
  + `std.Build.*`
  + `std.builtin.Mode`
  + `std.zig.c_translation.CIntLiteralRadix`
  + anything in `src/`
2024-06-13 10:18:59 -04:00

729 lines
28 KiB
Zig
Vendored

const std = @import("std");
const assert = std.debug.assert;
const BigIntConst = std.math.big.int.Const;
const BigIntMutable = std.math.big.int.Mutable;
const backend = @import("../backend.zig");
const Interner = backend.Interner;
const BigIntSpace = Interner.Tag.Int.BigIntSpace;
const Compilation = @import("Compilation.zig");
const Type = @import("Type.zig");
const target_util = @import("target.zig");
const Value = @This();
opt_ref: Interner.OptRef = .none,
pub const zero = Value{ .opt_ref = .zero };
pub const one = Value{ .opt_ref = .one };
pub const @"null" = Value{ .opt_ref = .null };
pub fn intern(comp: *Compilation, k: Interner.Key) !Value {
const r = try comp.interner.put(comp.gpa, k);
return .{ .opt_ref = @enumFromInt(@intFromEnum(r)) };
}
pub fn int(i: anytype, comp: *Compilation) !Value {
const info = @typeInfo(@TypeOf(i));
if (info == .ComptimeInt or info.Int.signedness == .unsigned) {
return intern(comp, .{ .int = .{ .u64 = i } });
} else {
return intern(comp, .{ .int = .{ .i64 = i } });
}
}
pub fn ref(v: Value) Interner.Ref {
std.debug.assert(v.opt_ref != .none);
return @enumFromInt(@intFromEnum(v.opt_ref));
}
pub fn is(v: Value, tag: std.meta.Tag(Interner.Key), comp: *const Compilation) bool {
if (v.opt_ref == .none) return false;
return comp.interner.get(v.ref()) == tag;
}
/// Number of bits needed to hold `v`.
/// Asserts that `v` is not negative
pub fn minUnsignedBits(v: Value, comp: *const Compilation) usize {
var space: BigIntSpace = undefined;
const big = v.toBigInt(&space, comp);
assert(big.positive);
return big.bitCountAbs();
}
test "minUnsignedBits" {
const Test = struct {
fn checkIntBits(comp: *Compilation, v: u64, expected: usize) !void {
const val = try intern(comp, .{ .int = .{ .u64 = v } });
try std.testing.expectEqual(expected, val.minUnsignedBits(comp));
}
};
var comp = Compilation.init(std.testing.allocator);
defer comp.deinit();
const target_query = try std.Target.Query.parse(.{ .arch_os_abi = "x86_64-linux-gnu" });
comp.target = try std.zig.system.resolveTargetQuery(target_query);
try Test.checkIntBits(&comp, 0, 0);
try Test.checkIntBits(&comp, 1, 1);
try Test.checkIntBits(&comp, 2, 2);
try Test.checkIntBits(&comp, std.math.maxInt(i8), 7);
try Test.checkIntBits(&comp, std.math.maxInt(u8), 8);
try Test.checkIntBits(&comp, std.math.maxInt(i16), 15);
try Test.checkIntBits(&comp, std.math.maxInt(u16), 16);
try Test.checkIntBits(&comp, std.math.maxInt(i32), 31);
try Test.checkIntBits(&comp, std.math.maxInt(u32), 32);
try Test.checkIntBits(&comp, std.math.maxInt(i64), 63);
try Test.checkIntBits(&comp, std.math.maxInt(u64), 64);
}
/// Minimum number of bits needed to represent `v` in 2's complement notation
/// Asserts that `v` is negative.
pub fn minSignedBits(v: Value, comp: *const Compilation) usize {
var space: BigIntSpace = undefined;
const big = v.toBigInt(&space, comp);
assert(!big.positive);
return big.bitCountTwosComp();
}
test "minSignedBits" {
const Test = struct {
fn checkIntBits(comp: *Compilation, v: i64, expected: usize) !void {
const val = try intern(comp, .{ .int = .{ .i64 = v } });
try std.testing.expectEqual(expected, val.minSignedBits(comp));
}
};
var comp = Compilation.init(std.testing.allocator);
defer comp.deinit();
const target_query = try std.Target.Query.parse(.{ .arch_os_abi = "x86_64-linux-gnu" });
comp.target = try std.zig.system.resolveTargetQuery(target_query);
try Test.checkIntBits(&comp, -1, 1);
try Test.checkIntBits(&comp, -2, 2);
try Test.checkIntBits(&comp, -10, 5);
try Test.checkIntBits(&comp, -101, 8);
try Test.checkIntBits(&comp, std.math.minInt(i8), 8);
try Test.checkIntBits(&comp, std.math.minInt(i16), 16);
try Test.checkIntBits(&comp, std.math.minInt(i32), 32);
try Test.checkIntBits(&comp, std.math.minInt(i64), 64);
}
pub const FloatToIntChangeKind = enum {
/// value did not change
none,
/// floating point number too small or large for destination integer type
out_of_range,
/// tried to convert a NaN or Infinity
overflow,
/// fractional value was converted to zero
nonzero_to_zero,
/// fractional part truncated
value_changed,
};
/// Converts the stored value from a float to an integer.
/// `.none` value remains unchanged.
pub fn floatToInt(v: *Value, dest_ty: Type, comp: *Compilation) !FloatToIntChangeKind {
if (v.opt_ref == .none) return .none;
const float_val = v.toFloat(f128, comp);
const was_zero = float_val == 0;
if (dest_ty.is(.bool)) {
const was_one = float_val == 1.0;
v.* = fromBool(!was_zero);
if (was_zero or was_one) return .none;
return .value_changed;
} else if (dest_ty.isUnsignedInt(comp) and v.compare(.lt, zero, comp)) {
v.* = zero;
return .out_of_range;
}
const had_fraction = @rem(float_val, 1) != 0;
const is_negative = std.math.signbit(float_val);
const floored = @floor(@abs(float_val));
var rational = try std.math.big.Rational.init(comp.gpa);
defer rational.deinit();
rational.setFloat(f128, floored) catch |err| switch (err) {
error.NonFiniteFloat => {
v.* = .{};
return .overflow;
},
error.OutOfMemory => return error.OutOfMemory,
};
// The float is reduced in rational.setFloat, so we assert that denominator is equal to one
const big_one = std.math.big.int.Const{ .limbs = &.{1}, .positive = true };
assert(rational.q.toConst().eqlAbs(big_one));
if (is_negative) {
rational.negate();
}
const signedness = dest_ty.signedness(comp);
const bits: usize = @intCast(dest_ty.bitSizeof(comp).?);
// rational.p.truncate(rational.p.toConst(), signedness: Signedness, bit_count: usize)
const fits = rational.p.fitsInTwosComp(signedness, bits);
v.* = try intern(comp, .{ .int = .{ .big_int = rational.p.toConst() } });
try rational.p.truncate(&rational.p, signedness, bits);
if (!was_zero and v.isZero(comp)) return .nonzero_to_zero;
if (!fits) return .out_of_range;
if (had_fraction) return .value_changed;
return .none;
}
/// Converts the stored value from an integer to a float.
/// `.none` value remains unchanged.
pub fn intToFloat(v: *Value, dest_ty: Type, comp: *Compilation) !void {
if (v.opt_ref == .none) return;
const bits = dest_ty.bitSizeof(comp).?;
return switch (comp.interner.get(v.ref()).int) {
inline .u64, .i64 => |data| {
const f: Interner.Key.Float = switch (bits) {
16 => .{ .f16 = @floatFromInt(data) },
32 => .{ .f32 = @floatFromInt(data) },
64 => .{ .f64 = @floatFromInt(data) },
80 => .{ .f80 = @floatFromInt(data) },
128 => .{ .f128 = @floatFromInt(data) },
else => unreachable,
};
v.* = try intern(comp, .{ .float = f });
},
.big_int => |data| {
const big_f = bigIntToFloat(data.limbs, data.positive);
const f: Interner.Key.Float = switch (bits) {
16 => .{ .f16 = @floatCast(big_f) },
32 => .{ .f32 = @floatCast(big_f) },
64 => .{ .f64 = @floatCast(big_f) },
80 => .{ .f80 = @floatCast(big_f) },
128 => .{ .f128 = @floatCast(big_f) },
else => unreachable,
};
v.* = try intern(comp, .{ .float = f });
},
};
}
/// Truncates or extends bits based on type.
/// `.none` value remains unchanged.
pub fn intCast(v: *Value, dest_ty: Type, comp: *Compilation) !void {
if (v.opt_ref == .none) return;
const bits: usize = @intCast(dest_ty.bitSizeof(comp).?);
var space: BigIntSpace = undefined;
const big = v.toBigInt(&space, comp);
const limbs = try comp.gpa.alloc(
std.math.big.Limb,
std.math.big.int.calcTwosCompLimbCount(@max(big.bitCountTwosComp(), bits)),
);
defer comp.gpa.free(limbs);
var result_bigint = std.math.big.int.Mutable{ .limbs = limbs, .positive = undefined, .len = undefined };
result_bigint.truncate(big, dest_ty.signedness(comp), bits);
v.* = try intern(comp, .{ .int = .{ .big_int = result_bigint.toConst() } });
}
/// Converts the stored value to a float of the specified type
/// `.none` value remains unchanged.
pub fn floatCast(v: *Value, dest_ty: Type, comp: *Compilation) !void {
if (v.opt_ref == .none) return;
// TODO complex values
const bits = dest_ty.makeReal().bitSizeof(comp).?;
const f: Interner.Key.Float = switch (bits) {
16 => .{ .f16 = v.toFloat(f16, comp) },
32 => .{ .f32 = v.toFloat(f32, comp) },
64 => .{ .f64 = v.toFloat(f64, comp) },
80 => .{ .f80 = v.toFloat(f80, comp) },
128 => .{ .f128 = v.toFloat(f128, comp) },
else => unreachable,
};
v.* = try intern(comp, .{ .float = f });
}
pub fn toFloat(v: Value, comptime T: type, comp: *const Compilation) T {
return switch (comp.interner.get(v.ref())) {
.int => |repr| switch (repr) {
inline .u64, .i64 => |data| @floatFromInt(data),
.big_int => |data| @floatCast(bigIntToFloat(data.limbs, data.positive)),
},
.float => |repr| switch (repr) {
inline else => |data| @floatCast(data),
},
else => unreachable,
};
}
fn bigIntToFloat(limbs: []const std.math.big.Limb, positive: bool) f128 {
if (limbs.len == 0) return 0;
const base = std.math.maxInt(std.math.big.Limb) + 1;
var result: f128 = 0;
var i: usize = limbs.len;
while (i != 0) {
i -= 1;
const limb: f128 = @as(f128, @floatFromInt(limbs[i]));
result = @mulAdd(f128, base, result, limb);
}
if (positive) {
return result;
} else {
return -result;
}
}
pub fn toBigInt(val: Value, space: *BigIntSpace, comp: *const Compilation) BigIntConst {
return switch (comp.interner.get(val.ref()).int) {
inline .u64, .i64 => |x| BigIntMutable.init(&space.limbs, x).toConst(),
.big_int => |b| b,
};
}
pub fn isZero(v: Value, comp: *const Compilation) bool {
if (v.opt_ref == .none) return false;
switch (v.ref()) {
.zero => return true,
.one => return false,
.null => return target_util.nullRepr(comp.target) == 0,
else => {},
}
const key = comp.interner.get(v.ref());
switch (key) {
.float => |repr| switch (repr) {
inline else => |data| return data == 0,
},
.int => |repr| switch (repr) {
inline .i64, .u64 => |data| return data == 0,
.big_int => |data| return data.eqlZero(),
},
.bytes => return false,
else => unreachable,
}
}
/// Converts value to zero or one;
/// `.none` value remains unchanged.
pub fn boolCast(v: *Value, comp: *const Compilation) void {
if (v.opt_ref == .none) return;
v.* = fromBool(v.toBool(comp));
}
pub fn fromBool(b: bool) Value {
return if (b) one else zero;
}
pub fn toBool(v: Value, comp: *const Compilation) bool {
return !v.isZero(comp);
}
pub fn toInt(v: Value, comptime T: type, comp: *const Compilation) ?T {
if (v.opt_ref == .none) return null;
if (comp.interner.get(v.ref()) != .int) return null;
var space: BigIntSpace = undefined;
const big_int = v.toBigInt(&space, comp);
return big_int.to(T) catch null;
}
pub fn add(res: *Value, lhs: Value, rhs: Value, ty: Type, comp: *Compilation) !bool {
const bits: usize = @intCast(ty.bitSizeof(comp).?);
if (ty.isFloat()) {
const f: Interner.Key.Float = switch (bits) {
16 => .{ .f16 = lhs.toFloat(f16, comp) + rhs.toFloat(f16, comp) },
32 => .{ .f32 = lhs.toFloat(f32, comp) + rhs.toFloat(f32, comp) },
64 => .{ .f64 = lhs.toFloat(f64, comp) + rhs.toFloat(f64, comp) },
80 => .{ .f80 = lhs.toFloat(f80, comp) + rhs.toFloat(f80, comp) },
128 => .{ .f128 = lhs.toFloat(f128, comp) + rhs.toFloat(f128, comp) },
else => unreachable,
};
res.* = try intern(comp, .{ .float = f });
return false;
} else {
var lhs_space: BigIntSpace = undefined;
var rhs_space: BigIntSpace = undefined;
const lhs_bigint = lhs.toBigInt(&lhs_space, comp);
const rhs_bigint = rhs.toBigInt(&rhs_space, comp);
const limbs = try comp.gpa.alloc(
std.math.big.Limb,
std.math.big.int.calcTwosCompLimbCount(bits),
);
defer comp.gpa.free(limbs);
var result_bigint = std.math.big.int.Mutable{ .limbs = limbs, .positive = undefined, .len = undefined };
const overflowed = result_bigint.addWrap(lhs_bigint, rhs_bigint, ty.signedness(comp), bits);
res.* = try intern(comp, .{ .int = .{ .big_int = result_bigint.toConst() } });
return overflowed;
}
}
pub fn sub(res: *Value, lhs: Value, rhs: Value, ty: Type, comp: *Compilation) !bool {
const bits: usize = @intCast(ty.bitSizeof(comp).?);
if (ty.isFloat()) {
const f: Interner.Key.Float = switch (bits) {
16 => .{ .f16 = lhs.toFloat(f16, comp) - rhs.toFloat(f16, comp) },
32 => .{ .f32 = lhs.toFloat(f32, comp) - rhs.toFloat(f32, comp) },
64 => .{ .f64 = lhs.toFloat(f64, comp) - rhs.toFloat(f64, comp) },
80 => .{ .f80 = lhs.toFloat(f80, comp) - rhs.toFloat(f80, comp) },
128 => .{ .f128 = lhs.toFloat(f128, comp) - rhs.toFloat(f128, comp) },
else => unreachable,
};
res.* = try intern(comp, .{ .float = f });
return false;
} else {
var lhs_space: BigIntSpace = undefined;
var rhs_space: BigIntSpace = undefined;
const lhs_bigint = lhs.toBigInt(&lhs_space, comp);
const rhs_bigint = rhs.toBigInt(&rhs_space, comp);
const limbs = try comp.gpa.alloc(
std.math.big.Limb,
std.math.big.int.calcTwosCompLimbCount(bits),
);
defer comp.gpa.free(limbs);
var result_bigint = std.math.big.int.Mutable{ .limbs = limbs, .positive = undefined, .len = undefined };
const overflowed = result_bigint.subWrap(lhs_bigint, rhs_bigint, ty.signedness(comp), bits);
res.* = try intern(comp, .{ .int = .{ .big_int = result_bigint.toConst() } });
return overflowed;
}
}
pub fn mul(res: *Value, lhs: Value, rhs: Value, ty: Type, comp: *Compilation) !bool {
const bits: usize = @intCast(ty.bitSizeof(comp).?);
if (ty.isFloat()) {
const f: Interner.Key.Float = switch (bits) {
16 => .{ .f16 = lhs.toFloat(f16, comp) * rhs.toFloat(f16, comp) },
32 => .{ .f32 = lhs.toFloat(f32, comp) * rhs.toFloat(f32, comp) },
64 => .{ .f64 = lhs.toFloat(f64, comp) * rhs.toFloat(f64, comp) },
80 => .{ .f80 = lhs.toFloat(f80, comp) * rhs.toFloat(f80, comp) },
128 => .{ .f128 = lhs.toFloat(f128, comp) * rhs.toFloat(f128, comp) },
else => unreachable,
};
res.* = try intern(comp, .{ .float = f });
return false;
} else {
var lhs_space: BigIntSpace = undefined;
var rhs_space: BigIntSpace = undefined;
const lhs_bigint = lhs.toBigInt(&lhs_space, comp);
const rhs_bigint = rhs.toBigInt(&rhs_space, comp);
const limbs = try comp.gpa.alloc(
std.math.big.Limb,
lhs_bigint.limbs.len + rhs_bigint.limbs.len,
);
defer comp.gpa.free(limbs);
var result_bigint = BigIntMutable{ .limbs = limbs, .positive = undefined, .len = undefined };
const limbs_buffer = try comp.gpa.alloc(
std.math.big.Limb,
std.math.big.int.calcMulLimbsBufferLen(lhs_bigint.limbs.len, rhs_bigint.limbs.len, 1),
);
defer comp.gpa.free(limbs_buffer);
result_bigint.mul(lhs_bigint, rhs_bigint, limbs_buffer, comp.gpa);
const signedness = ty.signedness(comp);
const overflowed = !result_bigint.toConst().fitsInTwosComp(signedness, bits);
if (overflowed) {
result_bigint.truncate(result_bigint.toConst(), signedness, bits);
}
res.* = try intern(comp, .{ .int = .{ .big_int = result_bigint.toConst() } });
return overflowed;
}
}
/// caller guarantees rhs != 0
pub fn div(res: *Value, lhs: Value, rhs: Value, ty: Type, comp: *Compilation) !bool {
const bits: usize = @intCast(ty.bitSizeof(comp).?);
if (ty.isFloat()) {
const f: Interner.Key.Float = switch (bits) {
16 => .{ .f16 = lhs.toFloat(f16, comp) / rhs.toFloat(f16, comp) },
32 => .{ .f32 = lhs.toFloat(f32, comp) / rhs.toFloat(f32, comp) },
64 => .{ .f64 = lhs.toFloat(f64, comp) / rhs.toFloat(f64, comp) },
80 => .{ .f80 = lhs.toFloat(f80, comp) / rhs.toFloat(f80, comp) },
128 => .{ .f128 = lhs.toFloat(f128, comp) / rhs.toFloat(f128, comp) },
else => unreachable,
};
res.* = try intern(comp, .{ .float = f });
return false;
} else {
var lhs_space: BigIntSpace = undefined;
var rhs_space: BigIntSpace = undefined;
const lhs_bigint = lhs.toBigInt(&lhs_space, comp);
const rhs_bigint = rhs.toBigInt(&rhs_space, comp);
const limbs_q = try comp.gpa.alloc(
std.math.big.Limb,
lhs_bigint.limbs.len,
);
defer comp.gpa.free(limbs_q);
var result_q = BigIntMutable{ .limbs = limbs_q, .positive = undefined, .len = undefined };
const limbs_r = try comp.gpa.alloc(
std.math.big.Limb,
rhs_bigint.limbs.len,
);
defer comp.gpa.free(limbs_r);
var result_r = BigIntMutable{ .limbs = limbs_r, .positive = undefined, .len = undefined };
const limbs_buffer = try comp.gpa.alloc(
std.math.big.Limb,
std.math.big.int.calcDivLimbsBufferLen(lhs_bigint.limbs.len, rhs_bigint.limbs.len),
);
defer comp.gpa.free(limbs_buffer);
result_q.divTrunc(&result_r, lhs_bigint, rhs_bigint, limbs_buffer);
res.* = try intern(comp, .{ .int = .{ .big_int = result_q.toConst() } });
return !result_q.toConst().fitsInTwosComp(ty.signedness(comp), bits);
}
}
/// caller guarantees rhs != 0
/// caller guarantees lhs != std.math.minInt(T) OR rhs != -1
pub fn rem(lhs: Value, rhs: Value, ty: Type, comp: *Compilation) !Value {
var lhs_space: BigIntSpace = undefined;
var rhs_space: BigIntSpace = undefined;
const lhs_bigint = lhs.toBigInt(&lhs_space, comp);
const rhs_bigint = rhs.toBigInt(&rhs_space, comp);
const signedness = ty.signedness(comp);
if (signedness == .signed) {
var spaces: [3]BigIntSpace = undefined;
const min_val = BigIntMutable.init(&spaces[0].limbs, ty.minInt(comp)).toConst();
const negative = BigIntMutable.init(&spaces[1].limbs, -1).toConst();
const big_one = BigIntMutable.init(&spaces[2].limbs, 1).toConst();
if (lhs_bigint.eql(min_val) and rhs_bigint.eql(negative)) {
return .{};
} else if (rhs_bigint.order(big_one).compare(.lt)) {
// lhs - @divTrunc(lhs, rhs) * rhs
var tmp: Value = undefined;
_ = try tmp.div(lhs, rhs, ty, comp);
_ = try tmp.mul(tmp, rhs, ty, comp);
_ = try tmp.sub(lhs, tmp, ty, comp);
return tmp;
}
}
const limbs_q = try comp.gpa.alloc(
std.math.big.Limb,
lhs_bigint.limbs.len,
);
defer comp.gpa.free(limbs_q);
var result_q = BigIntMutable{ .limbs = limbs_q, .positive = undefined, .len = undefined };
const limbs_r = try comp.gpa.alloc(
std.math.big.Limb,
rhs_bigint.limbs.len,
);
defer comp.gpa.free(limbs_r);
var result_r = BigIntMutable{ .limbs = limbs_r, .positive = undefined, .len = undefined };
const limbs_buffer = try comp.gpa.alloc(
std.math.big.Limb,
std.math.big.int.calcDivLimbsBufferLen(lhs_bigint.limbs.len, rhs_bigint.limbs.len),
);
defer comp.gpa.free(limbs_buffer);
result_q.divTrunc(&result_r, lhs_bigint, rhs_bigint, limbs_buffer);
return intern(comp, .{ .int = .{ .big_int = result_r.toConst() } });
}
pub fn bitOr(lhs: Value, rhs: Value, comp: *Compilation) !Value {
var lhs_space: BigIntSpace = undefined;
var rhs_space: BigIntSpace = undefined;
const lhs_bigint = lhs.toBigInt(&lhs_space, comp);
const rhs_bigint = rhs.toBigInt(&rhs_space, comp);
const limbs = try comp.gpa.alloc(
std.math.big.Limb,
@max(lhs_bigint.limbs.len, rhs_bigint.limbs.len),
);
defer comp.gpa.free(limbs);
var result_bigint = std.math.big.int.Mutable{ .limbs = limbs, .positive = undefined, .len = undefined };
result_bigint.bitOr(lhs_bigint, rhs_bigint);
return intern(comp, .{ .int = .{ .big_int = result_bigint.toConst() } });
}
pub fn bitXor(lhs: Value, rhs: Value, comp: *Compilation) !Value {
var lhs_space: BigIntSpace = undefined;
var rhs_space: BigIntSpace = undefined;
const lhs_bigint = lhs.toBigInt(&lhs_space, comp);
const rhs_bigint = rhs.toBigInt(&rhs_space, comp);
const limbs = try comp.gpa.alloc(
std.math.big.Limb,
@max(lhs_bigint.limbs.len, rhs_bigint.limbs.len),
);
defer comp.gpa.free(limbs);
var result_bigint = std.math.big.int.Mutable{ .limbs = limbs, .positive = undefined, .len = undefined };
result_bigint.bitXor(lhs_bigint, rhs_bigint);
return intern(comp, .{ .int = .{ .big_int = result_bigint.toConst() } });
}
pub fn bitAnd(lhs: Value, rhs: Value, comp: *Compilation) !Value {
var lhs_space: BigIntSpace = undefined;
var rhs_space: BigIntSpace = undefined;
const lhs_bigint = lhs.toBigInt(&lhs_space, comp);
const rhs_bigint = rhs.toBigInt(&rhs_space, comp);
const limbs = try comp.gpa.alloc(
std.math.big.Limb,
@max(lhs_bigint.limbs.len, rhs_bigint.limbs.len),
);
defer comp.gpa.free(limbs);
var result_bigint = std.math.big.int.Mutable{ .limbs = limbs, .positive = undefined, .len = undefined };
result_bigint.bitAnd(lhs_bigint, rhs_bigint);
return intern(comp, .{ .int = .{ .big_int = result_bigint.toConst() } });
}
pub fn bitNot(val: Value, ty: Type, comp: *Compilation) !Value {
const bits: usize = @intCast(ty.bitSizeof(comp).?);
var val_space: Value.BigIntSpace = undefined;
const val_bigint = val.toBigInt(&val_space, comp);
const limbs = try comp.gpa.alloc(
std.math.big.Limb,
std.math.big.int.calcTwosCompLimbCount(bits),
);
defer comp.gpa.free(limbs);
var result_bigint = std.math.big.int.Mutable{ .limbs = limbs, .positive = undefined, .len = undefined };
result_bigint.bitNotWrap(val_bigint, ty.signedness(comp), bits);
return intern(comp, .{ .int = .{ .big_int = result_bigint.toConst() } });
}
pub fn shl(res: *Value, lhs: Value, rhs: Value, ty: Type, comp: *Compilation) !bool {
var lhs_space: Value.BigIntSpace = undefined;
const lhs_bigint = lhs.toBigInt(&lhs_space, comp);
const shift = rhs.toInt(usize, comp) orelse std.math.maxInt(usize);
const bits: usize = @intCast(ty.bitSizeof(comp).?);
if (shift > bits) {
if (lhs_bigint.positive) {
res.* = try intern(comp, .{ .int = .{ .u64 = ty.maxInt(comp) } });
} else {
res.* = try intern(comp, .{ .int = .{ .i64 = ty.minInt(comp) } });
}
return true;
}
const limbs = try comp.gpa.alloc(
std.math.big.Limb,
lhs_bigint.limbs.len + (shift / (@sizeOf(std.math.big.Limb) * 8)) + 1,
);
defer comp.gpa.free(limbs);
var result_bigint = std.math.big.int.Mutable{ .limbs = limbs, .positive = undefined, .len = undefined };
result_bigint.shiftLeft(lhs_bigint, shift);
const signedness = ty.signedness(comp);
const overflowed = !result_bigint.toConst().fitsInTwosComp(signedness, bits);
if (overflowed) {
result_bigint.truncate(result_bigint.toConst(), signedness, bits);
}
res.* = try intern(comp, .{ .int = .{ .big_int = result_bigint.toConst() } });
return overflowed;
}
pub fn shr(lhs: Value, rhs: Value, ty: Type, comp: *Compilation) !Value {
var lhs_space: Value.BigIntSpace = undefined;
const lhs_bigint = lhs.toBigInt(&lhs_space, comp);
const shift = rhs.toInt(usize, comp) orelse return zero;
const result_limbs = lhs_bigint.limbs.len -| (shift / (@sizeOf(std.math.big.Limb) * 8));
if (result_limbs == 0) {
// The shift is enough to remove all the bits from the number, which means the
// result is 0 or -1 depending on the sign.
if (lhs_bigint.positive) {
return zero;
} else {
return intern(comp, .{ .int = .{ .i64 = -1 } });
}
}
const bits: usize = @intCast(ty.bitSizeof(comp).?);
const limbs = try comp.gpa.alloc(
std.math.big.Limb,
std.math.big.int.calcTwosCompLimbCount(bits),
);
defer comp.gpa.free(limbs);
var result_bigint = std.math.big.int.Mutable{ .limbs = limbs, .positive = undefined, .len = undefined };
result_bigint.shiftRight(lhs_bigint, shift);
return intern(comp, .{ .int = .{ .big_int = result_bigint.toConst() } });
}
pub fn compare(lhs: Value, op: std.math.CompareOperator, rhs: Value, comp: *const Compilation) bool {
if (op == .eq) {
return lhs.opt_ref == rhs.opt_ref;
} else if (lhs.opt_ref == rhs.opt_ref) {
return std.math.Order.eq.compare(op);
}
const lhs_key = comp.interner.get(lhs.ref());
const rhs_key = comp.interner.get(rhs.ref());
if (lhs_key == .float or rhs_key == .float) {
const lhs_f128 = lhs.toFloat(f128, comp);
const rhs_f128 = rhs.toFloat(f128, comp);
return std.math.compare(lhs_f128, op, rhs_f128);
}
var lhs_bigint_space: BigIntSpace = undefined;
var rhs_bigint_space: BigIntSpace = undefined;
const lhs_bigint = lhs.toBigInt(&lhs_bigint_space, comp);
const rhs_bigint = rhs.toBigInt(&rhs_bigint_space, comp);
return lhs_bigint.order(rhs_bigint).compare(op);
}
pub fn print(v: Value, ty: Type, comp: *const Compilation, w: anytype) @TypeOf(w).Error!void {
if (ty.is(.bool)) {
return w.writeAll(if (v.isZero(comp)) "false" else "true");
}
const key = comp.interner.get(v.ref());
switch (key) {
.null => return w.writeAll("nullptr_t"),
.int => |repr| switch (repr) {
inline else => |x| return w.print("{d}", .{x}),
},
.float => |repr| switch (repr) {
.f16 => |x| return w.print("{d}", .{@round(@as(f64, @floatCast(x)) * 1000) / 1000}),
.f32 => |x| return w.print("{d}", .{@round(@as(f64, @floatCast(x)) * 1000000) / 1000000}),
inline else => |x| return w.print("{d}", .{@as(f64, @floatCast(x))}),
},
.bytes => |b| return printString(b, ty, comp, w),
else => unreachable, // not a value
}
}
pub fn printString(bytes: []const u8, ty: Type, comp: *const Compilation, w: anytype) @TypeOf(w).Error!void {
const size: Compilation.CharUnitSize = @enumFromInt(ty.elemType().sizeof(comp).?);
const without_null = bytes[0 .. bytes.len - @intFromEnum(size)];
switch (size) {
inline .@"1", .@"2" => |sz| {
const data_slice: []const sz.Type() = @alignCast(std.mem.bytesAsSlice(sz.Type(), without_null));
const formatter = if (sz == .@"1") std.zig.fmtEscapes(data_slice) else std.unicode.fmtUtf16Le(data_slice);
try w.print("\"{}\"", .{formatter});
},
.@"4" => {
try w.writeByte('"');
const data_slice = std.mem.bytesAsSlice(u32, without_null);
var buf: [4]u8 = undefined;
for (data_slice) |item| {
if (item <= std.math.maxInt(u21) and std.unicode.utf8ValidCodepoint(@intCast(item))) {
const codepoint: u21 = @intCast(item);
const written = std.unicode.utf8Encode(codepoint, &buf) catch unreachable;
try w.print("{s}", .{buf[0..written]});
} else {
try w.print("\\x{x}", .{item});
}
}
try w.writeByte('"');
},
}
}