1454 lines
54 KiB
Zig
1454 lines
54 KiB
Zig
const std = @import("std.zig");
|
|
const builtin = @import("builtin");
|
|
const Os = builtin.Os;
|
|
const c = std.c;
|
|
|
|
const math = std.math;
|
|
const debug = std.debug;
|
|
const assert = debug.assert;
|
|
const os = std.os;
|
|
const mem = std.mem;
|
|
const meta = std.meta;
|
|
const trait = meta.trait;
|
|
const Buffer = std.Buffer;
|
|
const fmt = std.fmt;
|
|
const File = std.os.File;
|
|
const testing = std.testing;
|
|
|
|
const is_posix = builtin.os != builtin.Os.windows;
|
|
const is_windows = builtin.os == builtin.Os.windows;
|
|
|
|
const GetStdIoErrs = os.WindowsGetStdHandleErrs;
|
|
|
|
pub fn getStdErr() GetStdIoErrs!File {
|
|
const handle = if (is_windows) try os.windowsGetStdHandle(os.windows.STD_ERROR_HANDLE) else if (is_posix) os.posix.STDERR_FILENO else unreachable;
|
|
return File.openHandle(handle);
|
|
}
|
|
|
|
pub fn getStdOut() GetStdIoErrs!File {
|
|
const handle = if (is_windows) try os.windowsGetStdHandle(os.windows.STD_OUTPUT_HANDLE) else if (is_posix) os.posix.STDOUT_FILENO else unreachable;
|
|
return File.openHandle(handle);
|
|
}
|
|
|
|
pub fn getStdIn() GetStdIoErrs!File {
|
|
const handle = if (is_windows) try os.windowsGetStdHandle(os.windows.STD_INPUT_HANDLE) else if (is_posix) os.posix.STDIN_FILENO else unreachable;
|
|
return File.openHandle(handle);
|
|
}
|
|
|
|
pub const SeekableStream = @import("io/seekable_stream.zig").SeekableStream;
|
|
|
|
pub fn InStream(comptime ReadError: type) type {
|
|
return struct {
|
|
const Self = @This();
|
|
pub const Error = ReadError;
|
|
|
|
/// Return the number of bytes read. If the number read is smaller than buf.len, it
|
|
/// means the stream reached the end. Reaching the end of a stream is not an error
|
|
/// condition.
|
|
readFn: fn (self: *Self, buffer: []u8) Error!usize,
|
|
|
|
/// Replaces `buffer` contents by reading from the stream until it is finished.
|
|
/// If `buffer.len()` would exceed `max_size`, `error.StreamTooLong` is returned and
|
|
/// the contents read from the stream are lost.
|
|
pub fn readAllBuffer(self: *Self, buffer: *Buffer, max_size: usize) !void {
|
|
try buffer.resize(0);
|
|
|
|
var actual_buf_len: usize = 0;
|
|
while (true) {
|
|
const dest_slice = buffer.toSlice()[actual_buf_len..];
|
|
const bytes_read = try self.readFull(dest_slice);
|
|
actual_buf_len += bytes_read;
|
|
|
|
if (bytes_read != dest_slice.len) {
|
|
buffer.shrink(actual_buf_len);
|
|
return;
|
|
}
|
|
|
|
const new_buf_size = math.min(max_size, actual_buf_len + os.page_size);
|
|
if (new_buf_size == actual_buf_len) return error.StreamTooLong;
|
|
try buffer.resize(new_buf_size);
|
|
}
|
|
}
|
|
|
|
/// Allocates enough memory to hold all the contents of the stream. If the allocated
|
|
/// memory would be greater than `max_size`, returns `error.StreamTooLong`.
|
|
/// Caller owns returned memory.
|
|
/// If this function returns an error, the contents from the stream read so far are lost.
|
|
pub fn readAllAlloc(self: *Self, allocator: *mem.Allocator, max_size: usize) ![]u8 {
|
|
var buf = Buffer.initNull(allocator);
|
|
defer buf.deinit();
|
|
|
|
try self.readAllBuffer(&buf, max_size);
|
|
return buf.toOwnedSlice();
|
|
}
|
|
|
|
/// Replaces `buffer` contents by reading from the stream until `delimiter` is found.
|
|
/// Does not include the delimiter in the result.
|
|
/// If `buffer.len()` would exceed `max_size`, `error.StreamTooLong` is returned and the contents
|
|
/// read from the stream so far are lost.
|
|
pub fn readUntilDelimiterBuffer(self: *Self, buffer: *Buffer, delimiter: u8, max_size: usize) !void {
|
|
try buffer.resize(0);
|
|
|
|
while (true) {
|
|
var byte: u8 = try self.readByte();
|
|
|
|
if (byte == delimiter) {
|
|
return;
|
|
}
|
|
|
|
if (buffer.len() == max_size) {
|
|
return error.StreamTooLong;
|
|
}
|
|
|
|
try buffer.appendByte(byte);
|
|
}
|
|
}
|
|
|
|
/// Allocates enough memory to read until `delimiter`. If the allocated
|
|
/// memory would be greater than `max_size`, returns `error.StreamTooLong`.
|
|
/// Caller owns returned memory.
|
|
/// If this function returns an error, the contents from the stream read so far are lost.
|
|
pub fn readUntilDelimiterAlloc(self: *Self, allocator: *mem.Allocator, delimiter: u8, max_size: usize) ![]u8 {
|
|
var buf = Buffer.initNull(allocator);
|
|
defer buf.deinit();
|
|
|
|
try self.readUntilDelimiterBuffer(&buf, delimiter, max_size);
|
|
return buf.toOwnedSlice();
|
|
}
|
|
|
|
/// Returns the number of bytes read. It may be less than buffer.len.
|
|
/// If the number of bytes read is 0, it means end of stream.
|
|
/// End of stream is not an error condition.
|
|
pub fn read(self: *Self, buffer: []u8) Error!usize {
|
|
return self.readFn(self, buffer);
|
|
}
|
|
|
|
/// Returns the number of bytes read. If the number read is smaller than buf.len, it
|
|
/// means the stream reached the end. Reaching the end of a stream is not an error
|
|
/// condition.
|
|
pub fn readFull(self: *Self, buffer: []u8) Error!usize {
|
|
var index: usize = 0;
|
|
while (index != buffer.len) {
|
|
const amt = try self.read(buffer[index..]);
|
|
if (amt == 0) return index;
|
|
index += amt;
|
|
}
|
|
return index;
|
|
}
|
|
|
|
/// Same as `readFull` but end of stream returns `error.EndOfStream`.
|
|
pub fn readNoEof(self: *Self, buf: []u8) !void {
|
|
const amt_read = try self.read(buf);
|
|
if (amt_read < buf.len) return error.EndOfStream;
|
|
}
|
|
|
|
/// Reads 1 byte from the stream or returns `error.EndOfStream`.
|
|
pub fn readByte(self: *Self) !u8 {
|
|
var result: [1]u8 = undefined;
|
|
try self.readNoEof(result[0..]);
|
|
return result[0];
|
|
}
|
|
|
|
/// Same as `readByte` except the returned byte is signed.
|
|
pub fn readByteSigned(self: *Self) !i8 {
|
|
return @bitCast(i8, try self.readByte());
|
|
}
|
|
|
|
/// Reads a native-endian integer
|
|
pub fn readIntNative(self: *Self, comptime T: type) !T {
|
|
var bytes: [@sizeOf(T)]u8 = undefined;
|
|
try self.readNoEof(bytes[0..]);
|
|
return mem.readIntNative(T, &bytes);
|
|
}
|
|
|
|
/// Reads a foreign-endian integer
|
|
pub fn readIntForeign(self: *Self, comptime T: type) !T {
|
|
var bytes: [@sizeOf(T)]u8 = undefined;
|
|
try self.readNoEof(bytes[0..]);
|
|
return mem.readIntForeign(T, &bytes);
|
|
}
|
|
|
|
pub fn readIntLittle(self: *Self, comptime T: type) !T {
|
|
var bytes: [@sizeOf(T)]u8 = undefined;
|
|
try self.readNoEof(bytes[0..]);
|
|
return mem.readIntLittle(T, &bytes);
|
|
}
|
|
|
|
pub fn readIntBig(self: *Self, comptime T: type) !T {
|
|
var bytes: [@sizeOf(T)]u8 = undefined;
|
|
try self.readNoEof(bytes[0..]);
|
|
return mem.readIntBig(T, &bytes);
|
|
}
|
|
|
|
pub fn readInt(self: *Self, comptime T: type, endian: builtin.Endian) !T {
|
|
var bytes: [@sizeOf(T)]u8 = undefined;
|
|
try self.readNoEof(bytes[0..]);
|
|
return mem.readInt(T, &bytes, endian);
|
|
}
|
|
|
|
pub fn readVarInt(self: *Self, comptime ReturnType: type, endian: builtin.Endian, size: usize) !ReturnType {
|
|
assert(size <= @sizeOf(ReturnType));
|
|
var bytes_buf: [@sizeOf(ReturnType)]u8 = undefined;
|
|
const bytes = bytes_buf[0..size];
|
|
try self.readNoEof(bytes);
|
|
return mem.readVarInt(ReturnType, bytes, endian);
|
|
}
|
|
|
|
pub fn skipBytes(self: *Self, num_bytes: usize) !void {
|
|
var i: usize = 0;
|
|
while (i < num_bytes) : (i += 1) {
|
|
_ = try self.readByte();
|
|
}
|
|
}
|
|
|
|
pub fn readStruct(self: *Self, comptime T: type) !T {
|
|
// Only extern and packed structs have defined in-memory layout.
|
|
comptime assert(@typeInfo(T).Struct.layout != builtin.TypeInfo.ContainerLayout.Auto);
|
|
var res: [1]T = undefined;
|
|
try self.readNoEof(@sliceToBytes(res[0..]));
|
|
return res[0];
|
|
}
|
|
};
|
|
}
|
|
|
|
pub fn OutStream(comptime WriteError: type) type {
|
|
return struct {
|
|
const Self = @This();
|
|
pub const Error = WriteError;
|
|
|
|
writeFn: fn (self: *Self, bytes: []const u8) Error!void,
|
|
|
|
pub fn print(self: *Self, comptime format: []const u8, args: ...) Error!void {
|
|
return std.fmt.format(self, Error, self.writeFn, format, args);
|
|
}
|
|
|
|
pub fn write(self: *Self, bytes: []const u8) Error!void {
|
|
return self.writeFn(self, bytes);
|
|
}
|
|
|
|
pub fn writeByte(self: *Self, byte: u8) Error!void {
|
|
const slice = (*const [1]u8)(&byte)[0..];
|
|
return self.writeFn(self, slice);
|
|
}
|
|
|
|
pub fn writeByteNTimes(self: *Self, byte: u8, n: usize) Error!void {
|
|
const slice = (*const [1]u8)(&byte)[0..];
|
|
var i: usize = 0;
|
|
while (i < n) : (i += 1) {
|
|
try self.writeFn(self, slice);
|
|
}
|
|
}
|
|
|
|
/// Write a native-endian integer.
|
|
pub fn writeIntNative(self: *Self, comptime T: type, value: T) Error!void {
|
|
var bytes: [@sizeOf(T)]u8 = undefined;
|
|
mem.writeIntNative(T, &bytes, value);
|
|
return self.writeFn(self, bytes);
|
|
}
|
|
|
|
/// Write a foreign-endian integer.
|
|
pub fn writeIntForeign(self: *Self, comptime T: type, value: T) Error!void {
|
|
var bytes: [@sizeOf(T)]u8 = undefined;
|
|
mem.writeIntForeign(T, &bytes, value);
|
|
return self.writeFn(self, bytes);
|
|
}
|
|
|
|
pub fn writeIntLittle(self: *Self, comptime T: type, value: T) Error!void {
|
|
var bytes: [@sizeOf(T)]u8 = undefined;
|
|
mem.writeIntLittle(T, &bytes, value);
|
|
return self.writeFn(self, bytes);
|
|
}
|
|
|
|
pub fn writeIntBig(self: *Self, comptime T: type, value: T) Error!void {
|
|
var bytes: [@sizeOf(T)]u8 = undefined;
|
|
mem.writeIntBig(T, &bytes, value);
|
|
return self.writeFn(self, bytes);
|
|
}
|
|
|
|
pub fn writeInt(self: *Self, comptime T: type, value: T, endian: builtin.Endian) Error!void {
|
|
var bytes: [@sizeOf(T)]u8 = undefined;
|
|
mem.writeInt(T, &bytes, value, endian);
|
|
return self.writeFn(self, bytes);
|
|
}
|
|
};
|
|
}
|
|
|
|
pub fn writeFile(path: []const u8, data: []const u8) !void {
|
|
var file = try File.openWrite(path);
|
|
defer file.close();
|
|
try file.write(data);
|
|
}
|
|
|
|
/// On success, caller owns returned buffer.
|
|
pub fn readFileAlloc(allocator: *mem.Allocator, path: []const u8) ![]u8 {
|
|
return readFileAllocAligned(allocator, path, @alignOf(u8));
|
|
}
|
|
|
|
/// On success, caller owns returned buffer.
|
|
pub fn readFileAllocAligned(allocator: *mem.Allocator, path: []const u8, comptime A: u29) ![]align(A) u8 {
|
|
var file = try File.openRead(path);
|
|
defer file.close();
|
|
|
|
const size = try file.getEndPos();
|
|
const buf = try allocator.alignedAlloc(u8, A, size);
|
|
errdefer allocator.free(buf);
|
|
|
|
var adapter = file.inStream();
|
|
try adapter.stream.readNoEof(buf[0..size]);
|
|
return buf;
|
|
}
|
|
|
|
pub fn BufferedInStream(comptime Error: type) type {
|
|
return BufferedInStreamCustom(os.page_size, Error);
|
|
}
|
|
|
|
pub fn BufferedInStreamCustom(comptime buffer_size: usize, comptime Error: type) type {
|
|
return struct {
|
|
const Self = @This();
|
|
const Stream = InStream(Error);
|
|
|
|
pub stream: Stream,
|
|
|
|
unbuffered_in_stream: *Stream,
|
|
|
|
buffer: [buffer_size]u8,
|
|
start_index: usize,
|
|
end_index: usize,
|
|
|
|
pub fn init(unbuffered_in_stream: *Stream) Self {
|
|
return Self{
|
|
.unbuffered_in_stream = unbuffered_in_stream,
|
|
.buffer = undefined,
|
|
|
|
// Initialize these two fields to buffer_size so that
|
|
// in `readFn` we treat the state as being able to read
|
|
// more from the unbuffered stream. If we set them to 0
|
|
// and 0, the code would think we already hit EOF.
|
|
.start_index = buffer_size,
|
|
.end_index = buffer_size,
|
|
|
|
.stream = Stream{ .readFn = readFn },
|
|
};
|
|
}
|
|
|
|
fn readFn(in_stream: *Stream, dest: []u8) !usize {
|
|
const self = @fieldParentPtr(Self, "stream", in_stream);
|
|
|
|
var dest_index: usize = 0;
|
|
while (true) {
|
|
const dest_space = dest.len - dest_index;
|
|
if (dest_space == 0) {
|
|
return dest_index;
|
|
}
|
|
const amt_buffered = self.end_index - self.start_index;
|
|
if (amt_buffered == 0) {
|
|
assert(self.end_index <= buffer_size);
|
|
// Make sure the last read actually gave us some data
|
|
if (self.end_index == 0) {
|
|
// reading from the unbuffered stream returned nothing
|
|
// so we have nothing left to read.
|
|
return dest_index;
|
|
}
|
|
// we can read more data from the unbuffered stream
|
|
if (dest_space < buffer_size) {
|
|
self.start_index = 0;
|
|
self.end_index = try self.unbuffered_in_stream.read(self.buffer[0..]);
|
|
} else {
|
|
// asking for so much data that buffering is actually less efficient.
|
|
// forward the request directly to the unbuffered stream
|
|
const amt_read = try self.unbuffered_in_stream.read(dest[dest_index..]);
|
|
return dest_index + amt_read;
|
|
}
|
|
}
|
|
|
|
const copy_amount = math.min(dest_space, amt_buffered);
|
|
const copy_end_index = self.start_index + copy_amount;
|
|
mem.copy(u8, dest[dest_index..], self.buffer[self.start_index..copy_end_index]);
|
|
self.start_index = copy_end_index;
|
|
dest_index += copy_amount;
|
|
}
|
|
}
|
|
};
|
|
}
|
|
|
|
test "io.BufferedInStream" {
|
|
const OneByteReadInStream = struct {
|
|
const Error = error{NoError};
|
|
const Stream = InStream(Error);
|
|
|
|
stream: Stream,
|
|
str: []const u8,
|
|
curr: usize,
|
|
|
|
fn init(str: []const u8) @This() {
|
|
return @This(){
|
|
.stream = Stream{ .readFn = readFn },
|
|
.str = str,
|
|
.curr = 0,
|
|
};
|
|
}
|
|
|
|
fn readFn(in_stream: *Stream, dest: []u8) Error!usize {
|
|
const self = @fieldParentPtr(@This(), "stream", in_stream);
|
|
if (self.str.len <= self.curr or dest.len == 0)
|
|
return 0;
|
|
|
|
dest[0] = self.str[self.curr];
|
|
self.curr += 1;
|
|
return 1;
|
|
}
|
|
};
|
|
|
|
var buf: [100]u8 = undefined;
|
|
const allocator = &std.heap.FixedBufferAllocator.init(buf[0..]).allocator;
|
|
|
|
const str = "This is a test";
|
|
var one_byte_stream = OneByteReadInStream.init(str);
|
|
var buf_in_stream = BufferedInStream(OneByteReadInStream.Error).init(&one_byte_stream.stream);
|
|
const stream = &buf_in_stream.stream;
|
|
|
|
const res = try stream.readAllAlloc(allocator, str.len + 1);
|
|
testing.expectEqualSlices(u8, str, res);
|
|
}
|
|
|
|
/// Creates a stream which supports 'un-reading' data, so that it can be read again.
|
|
/// This makes look-ahead style parsing much easier.
|
|
pub fn PeekStream(comptime buffer_size: usize, comptime InStreamError: type) type {
|
|
return struct {
|
|
const Self = @This();
|
|
pub const Error = InStreamError;
|
|
pub const Stream = InStream(Error);
|
|
|
|
pub stream: Stream,
|
|
base: *Stream,
|
|
|
|
// Right now the look-ahead space is statically allocated, but a version with dynamic allocation
|
|
// is not too difficult to derive from this.
|
|
buffer: [buffer_size]u8,
|
|
index: usize,
|
|
at_end: bool,
|
|
|
|
pub fn init(base: *Stream) Self {
|
|
return Self{
|
|
.base = base,
|
|
.buffer = undefined,
|
|
.index = 0,
|
|
.at_end = false,
|
|
.stream = Stream{ .readFn = readFn },
|
|
};
|
|
}
|
|
|
|
pub fn putBackByte(self: *Self, byte: u8) void {
|
|
self.buffer[self.index] = byte;
|
|
self.index += 1;
|
|
}
|
|
|
|
pub fn putBack(self: *Self, bytes: []const u8) void {
|
|
var pos = bytes.len;
|
|
while (pos != 0) {
|
|
pos -= 1;
|
|
self.putBackByte(bytes[pos]);
|
|
}
|
|
}
|
|
|
|
fn readFn(in_stream: *Stream, dest: []u8) Error!usize {
|
|
const self = @fieldParentPtr(Self, "stream", in_stream);
|
|
|
|
// copy over anything putBack()'d
|
|
var pos: usize = 0;
|
|
while (pos < dest.len and self.index != 0) {
|
|
dest[pos] = self.buffer[self.index - 1];
|
|
self.index -= 1;
|
|
pos += 1;
|
|
}
|
|
|
|
if (pos == dest.len or self.at_end) {
|
|
return pos;
|
|
}
|
|
|
|
// ask the backing stream for more
|
|
const left = dest.len - pos;
|
|
const read = try self.base.read(dest[pos..]);
|
|
assert(read <= left);
|
|
|
|
self.at_end = (read < left);
|
|
return pos + read;
|
|
}
|
|
};
|
|
}
|
|
|
|
pub const SliceInStream = struct {
|
|
const Self = @This();
|
|
pub const Error = error{};
|
|
pub const Stream = InStream(Error);
|
|
|
|
pub stream: Stream,
|
|
|
|
pos: usize,
|
|
slice: []const u8,
|
|
|
|
pub fn init(slice: []const u8) Self {
|
|
return Self{
|
|
.slice = slice,
|
|
.pos = 0,
|
|
.stream = Stream{ .readFn = readFn },
|
|
};
|
|
}
|
|
|
|
fn readFn(in_stream: *Stream, dest: []u8) Error!usize {
|
|
const self = @fieldParentPtr(Self, "stream", in_stream);
|
|
const size = math.min(dest.len, self.slice.len - self.pos);
|
|
const end = self.pos + size;
|
|
|
|
mem.copy(u8, dest[0..size], self.slice[self.pos..end]);
|
|
self.pos = end;
|
|
|
|
return size;
|
|
}
|
|
};
|
|
|
|
/// Creates a stream which allows for reading bit fields from another stream
|
|
pub fn BitInStream(endian: builtin.Endian, comptime Error: type) type {
|
|
return struct {
|
|
const Self = @This();
|
|
|
|
in_stream: *Stream,
|
|
bit_buffer: u7,
|
|
bit_count: u3,
|
|
stream: Stream,
|
|
|
|
pub const Stream = InStream(Error);
|
|
const u8_bit_count = comptime meta.bitCount(u8);
|
|
const u7_bit_count = comptime meta.bitCount(u7);
|
|
const u4_bit_count = comptime meta.bitCount(u4);
|
|
|
|
pub fn init(in_stream: *Stream) Self {
|
|
return Self{
|
|
.in_stream = in_stream,
|
|
.bit_buffer = 0,
|
|
.bit_count = 0,
|
|
.stream = Stream{ .readFn = read },
|
|
};
|
|
}
|
|
|
|
/// Reads `bits` bits from the stream and returns a specified unsigned int type
|
|
/// containing them in the least significant end, returning an error if the
|
|
/// specified number of bits could not be read.
|
|
pub fn readBitsNoEof(self: *Self, comptime U: type, bits: usize) !U {
|
|
var n: usize = undefined;
|
|
const result = try self.readBits(U, bits, &n);
|
|
if (n < bits) return error.EndOfStream;
|
|
return result;
|
|
}
|
|
|
|
/// Reads `bits` bits from the stream and returns a specified unsigned int type
|
|
/// containing them in the least significant end. The number of bits successfully
|
|
/// read is placed in `out_bits`, as reaching the end of the stream is not an error.
|
|
pub fn readBits(self: *Self, comptime U: type, bits: usize, out_bits: *usize) Error!U {
|
|
comptime assert(trait.isUnsignedInt(U));
|
|
|
|
//by extending the buffer to a minimum of u8 we can cover a number of edge cases
|
|
// related to shifting and casting.
|
|
const u_bit_count = comptime meta.bitCount(U);
|
|
const buf_bit_count = bc: {
|
|
assert(u_bit_count >= bits);
|
|
break :bc if (u_bit_count <= u8_bit_count) u8_bit_count else u_bit_count;
|
|
};
|
|
const Buf = @IntType(false, buf_bit_count);
|
|
const BufShift = math.Log2Int(Buf);
|
|
|
|
out_bits.* = usize(0);
|
|
if (U == u0 or bits == 0) return 0;
|
|
var out_buffer = Buf(0);
|
|
|
|
if (self.bit_count > 0) {
|
|
const n = if (self.bit_count >= bits) @intCast(u3, bits) else self.bit_count;
|
|
const shift = u7_bit_count - n;
|
|
switch (endian) {
|
|
builtin.Endian.Big => {
|
|
out_buffer = Buf(self.bit_buffer >> shift);
|
|
self.bit_buffer <<= n;
|
|
},
|
|
builtin.Endian.Little => {
|
|
const value = (self.bit_buffer << shift) >> shift;
|
|
out_buffer = Buf(value);
|
|
self.bit_buffer >>= n;
|
|
},
|
|
}
|
|
self.bit_count -= n;
|
|
out_bits.* = n;
|
|
}
|
|
//at this point we know bit_buffer is empty
|
|
|
|
//copy bytes until we have enough bits, then leave the rest in bit_buffer
|
|
while (out_bits.* < bits) {
|
|
const n = bits - out_bits.*;
|
|
const next_byte = self.in_stream.readByte() catch |err| {
|
|
if (err == error.EndOfStream) {
|
|
return @intCast(U, out_buffer);
|
|
}
|
|
//@BUG: See #1810. Not sure if the bug is that I have to do this for some
|
|
// streams, or that I don't for streams with emtpy errorsets.
|
|
return @errSetCast(Error, err);
|
|
};
|
|
|
|
switch (endian) {
|
|
builtin.Endian.Big => {
|
|
if (n >= u8_bit_count) {
|
|
out_buffer <<= @intCast(u3, u8_bit_count - 1);
|
|
out_buffer <<= 1;
|
|
out_buffer |= Buf(next_byte);
|
|
out_bits.* += u8_bit_count;
|
|
continue;
|
|
}
|
|
|
|
const shift = @intCast(u3, u8_bit_count - n);
|
|
out_buffer <<= @intCast(BufShift, n);
|
|
out_buffer |= Buf(next_byte >> shift);
|
|
out_bits.* += n;
|
|
self.bit_buffer = @truncate(u7, next_byte << @intCast(u3, n - 1));
|
|
self.bit_count = shift;
|
|
},
|
|
builtin.Endian.Little => {
|
|
if (n >= u8_bit_count) {
|
|
out_buffer |= Buf(next_byte) << @intCast(BufShift, out_bits.*);
|
|
out_bits.* += u8_bit_count;
|
|
continue;
|
|
}
|
|
|
|
const shift = @intCast(u3, u8_bit_count - n);
|
|
const value = (next_byte << shift) >> shift;
|
|
out_buffer |= Buf(value) << @intCast(BufShift, out_bits.*);
|
|
out_bits.* += n;
|
|
self.bit_buffer = @truncate(u7, next_byte >> @intCast(u3, n));
|
|
self.bit_count = shift;
|
|
},
|
|
}
|
|
}
|
|
|
|
return @intCast(U, out_buffer);
|
|
}
|
|
|
|
pub fn alignToByte(self: *Self) void {
|
|
self.bit_buffer = 0;
|
|
self.bit_count = 0;
|
|
}
|
|
|
|
pub fn read(self_stream: *Stream, buffer: []u8) Error!usize {
|
|
var self = @fieldParentPtr(Self, "stream", self_stream);
|
|
|
|
var out_bits: usize = undefined;
|
|
var out_bits_total = usize(0);
|
|
//@NOTE: I'm not sure this is a good idea, maybe alignToByte should be forced
|
|
if (self.bit_count > 0) {
|
|
for (buffer) |*b, i| {
|
|
b.* = try self.readBits(u8, u8_bit_count, &out_bits);
|
|
out_bits_total += out_bits;
|
|
}
|
|
const incomplete_byte = @boolToInt(out_bits_total % u8_bit_count > 0);
|
|
return (out_bits_total / u8_bit_count) + incomplete_byte;
|
|
}
|
|
|
|
return self.in_stream.read(buffer);
|
|
}
|
|
};
|
|
}
|
|
|
|
/// This is a simple OutStream that writes to a slice, and returns an error
|
|
/// when it runs out of space.
|
|
pub const SliceOutStream = struct {
|
|
pub const Error = error{OutOfSpace};
|
|
pub const Stream = OutStream(Error);
|
|
|
|
pub stream: Stream,
|
|
|
|
pub pos: usize,
|
|
slice: []u8,
|
|
|
|
pub fn init(slice: []u8) SliceOutStream {
|
|
return SliceOutStream{
|
|
.slice = slice,
|
|
.pos = 0,
|
|
.stream = Stream{ .writeFn = writeFn },
|
|
};
|
|
}
|
|
|
|
pub fn getWritten(self: *const SliceOutStream) []const u8 {
|
|
return self.slice[0..self.pos];
|
|
}
|
|
|
|
pub fn reset(self: *SliceOutStream) void {
|
|
self.pos = 0;
|
|
}
|
|
|
|
fn writeFn(out_stream: *Stream, bytes: []const u8) Error!void {
|
|
const self = @fieldParentPtr(SliceOutStream, "stream", out_stream);
|
|
|
|
assert(self.pos <= self.slice.len);
|
|
|
|
const n = if (self.pos + bytes.len <= self.slice.len)
|
|
bytes.len
|
|
else
|
|
self.slice.len - self.pos;
|
|
|
|
std.mem.copy(u8, self.slice[self.pos .. self.pos + n], bytes[0..n]);
|
|
self.pos += n;
|
|
|
|
if (n < bytes.len) {
|
|
return Error.OutOfSpace;
|
|
}
|
|
}
|
|
};
|
|
|
|
test "io.SliceOutStream" {
|
|
var buf: [255]u8 = undefined;
|
|
var slice_stream = SliceOutStream.init(buf[0..]);
|
|
const stream = &slice_stream.stream;
|
|
|
|
try stream.print("{}{}!", "Hello", "World");
|
|
testing.expectEqualSlices(u8, "HelloWorld!", slice_stream.getWritten());
|
|
}
|
|
|
|
var null_out_stream_state = NullOutStream.init();
|
|
pub const null_out_stream = &null_out_stream_state.stream;
|
|
|
|
/// An OutStream that doesn't write to anything.
|
|
pub const NullOutStream = struct {
|
|
pub const Error = error{};
|
|
pub const Stream = OutStream(Error);
|
|
|
|
pub stream: Stream,
|
|
|
|
pub fn init() NullOutStream {
|
|
return NullOutStream{
|
|
.stream = Stream{ .writeFn = writeFn },
|
|
};
|
|
}
|
|
|
|
fn writeFn(out_stream: *Stream, bytes: []const u8) Error!void {}
|
|
};
|
|
|
|
test "io.NullOutStream" {
|
|
var null_stream = NullOutStream.init();
|
|
const stream = &null_stream.stream;
|
|
stream.write("yay" ** 10000) catch unreachable;
|
|
}
|
|
|
|
/// An OutStream that counts how many bytes has been written to it.
|
|
pub fn CountingOutStream(comptime OutStreamError: type) type {
|
|
return struct {
|
|
const Self = @This();
|
|
pub const Stream = OutStream(Error);
|
|
pub const Error = OutStreamError;
|
|
|
|
pub stream: Stream,
|
|
pub bytes_written: usize,
|
|
child_stream: *Stream,
|
|
|
|
pub fn init(child_stream: *Stream) Self {
|
|
return Self{
|
|
.stream = Stream{ .writeFn = writeFn },
|
|
.bytes_written = 0,
|
|
.child_stream = child_stream,
|
|
};
|
|
}
|
|
|
|
fn writeFn(out_stream: *Stream, bytes: []const u8) OutStreamError!void {
|
|
const self = @fieldParentPtr(Self, "stream", out_stream);
|
|
try self.child_stream.write(bytes);
|
|
self.bytes_written += bytes.len;
|
|
}
|
|
};
|
|
}
|
|
|
|
test "io.CountingOutStream" {
|
|
var null_stream = NullOutStream.init();
|
|
var counting_stream = CountingOutStream(NullOutStream.Error).init(&null_stream.stream);
|
|
const stream = &counting_stream.stream;
|
|
|
|
const bytes = "yay" ** 10000;
|
|
stream.write(bytes) catch unreachable;
|
|
testing.expect(counting_stream.bytes_written == bytes.len);
|
|
}
|
|
|
|
pub fn BufferedOutStream(comptime Error: type) type {
|
|
return BufferedOutStreamCustom(os.page_size, Error);
|
|
}
|
|
|
|
pub fn BufferedOutStreamCustom(comptime buffer_size: usize, comptime OutStreamError: type) type {
|
|
return struct {
|
|
const Self = @This();
|
|
pub const Stream = OutStream(Error);
|
|
pub const Error = OutStreamError;
|
|
|
|
pub stream: Stream,
|
|
|
|
unbuffered_out_stream: *Stream,
|
|
|
|
buffer: [buffer_size]u8,
|
|
index: usize,
|
|
|
|
pub fn init(unbuffered_out_stream: *Stream) Self {
|
|
return Self{
|
|
.unbuffered_out_stream = unbuffered_out_stream,
|
|
.buffer = undefined,
|
|
.index = 0,
|
|
.stream = Stream{ .writeFn = writeFn },
|
|
};
|
|
}
|
|
|
|
pub fn flush(self: *Self) !void {
|
|
try self.unbuffered_out_stream.write(self.buffer[0..self.index]);
|
|
self.index = 0;
|
|
}
|
|
|
|
fn writeFn(out_stream: *Stream, bytes: []const u8) !void {
|
|
const self = @fieldParentPtr(Self, "stream", out_stream);
|
|
|
|
if (bytes.len >= self.buffer.len) {
|
|
try self.flush();
|
|
return self.unbuffered_out_stream.write(bytes);
|
|
}
|
|
var src_index: usize = 0;
|
|
|
|
while (src_index < bytes.len) {
|
|
const dest_space_left = self.buffer.len - self.index;
|
|
const copy_amt = math.min(dest_space_left, bytes.len - src_index);
|
|
mem.copy(u8, self.buffer[self.index..], bytes[src_index .. src_index + copy_amt]);
|
|
self.index += copy_amt;
|
|
assert(self.index <= self.buffer.len);
|
|
if (self.index == self.buffer.len) {
|
|
try self.flush();
|
|
}
|
|
src_index += copy_amt;
|
|
}
|
|
}
|
|
};
|
|
}
|
|
|
|
/// Implementation of OutStream trait for Buffer
|
|
pub const BufferOutStream = struct {
|
|
buffer: *Buffer,
|
|
stream: Stream,
|
|
|
|
pub const Error = error{OutOfMemory};
|
|
pub const Stream = OutStream(Error);
|
|
|
|
pub fn init(buffer: *Buffer) BufferOutStream {
|
|
return BufferOutStream{
|
|
.buffer = buffer,
|
|
.stream = Stream{ .writeFn = writeFn },
|
|
};
|
|
}
|
|
|
|
fn writeFn(out_stream: *Stream, bytes: []const u8) !void {
|
|
const self = @fieldParentPtr(BufferOutStream, "stream", out_stream);
|
|
return self.buffer.append(bytes);
|
|
}
|
|
};
|
|
|
|
/// Creates a stream which allows for writing bit fields to another stream
|
|
pub fn BitOutStream(endian: builtin.Endian, comptime Error: type) type {
|
|
return struct {
|
|
const Self = @This();
|
|
|
|
out_stream: *Stream,
|
|
bit_buffer: u8,
|
|
bit_count: u4,
|
|
stream: Stream,
|
|
|
|
pub const Stream = OutStream(Error);
|
|
const u8_bit_count = comptime meta.bitCount(u8);
|
|
const u4_bit_count = comptime meta.bitCount(u4);
|
|
|
|
pub fn init(out_stream: *Stream) Self {
|
|
return Self{
|
|
.out_stream = out_stream,
|
|
.bit_buffer = 0,
|
|
.bit_count = 0,
|
|
.stream = Stream{ .writeFn = write },
|
|
};
|
|
}
|
|
|
|
/// Write the specified number of bits to the stream from the least significant bits of
|
|
/// the specified unsigned int value. Bits will only be written to the stream when there
|
|
/// are enough to fill a byte.
|
|
pub fn writeBits(self: *Self, value: var, bits: usize) Error!void {
|
|
if (bits == 0) return;
|
|
|
|
const U = @typeOf(value);
|
|
comptime assert(trait.isUnsignedInt(U));
|
|
|
|
//by extending the buffer to a minimum of u8 we can cover a number of edge cases
|
|
// related to shifting and casting.
|
|
const u_bit_count = comptime meta.bitCount(U);
|
|
const buf_bit_count = bc: {
|
|
assert(u_bit_count >= bits);
|
|
break :bc if (u_bit_count <= u8_bit_count) u8_bit_count else u_bit_count;
|
|
};
|
|
const Buf = @IntType(false, buf_bit_count);
|
|
const BufShift = math.Log2Int(Buf);
|
|
|
|
const buf_value = @intCast(Buf, value);
|
|
|
|
const high_byte_shift = @intCast(BufShift, buf_bit_count - u8_bit_count);
|
|
var in_buffer = switch (endian) {
|
|
builtin.Endian.Big => buf_value << @intCast(BufShift, buf_bit_count - bits),
|
|
builtin.Endian.Little => buf_value,
|
|
};
|
|
var in_bits = bits;
|
|
|
|
if (self.bit_count > 0) {
|
|
const bits_remaining = u8_bit_count - self.bit_count;
|
|
const n = @intCast(u3, if (bits_remaining > bits) bits else bits_remaining);
|
|
switch (endian) {
|
|
builtin.Endian.Big => {
|
|
const shift = @intCast(BufShift, high_byte_shift + self.bit_count);
|
|
const v = @intCast(u8, in_buffer >> shift);
|
|
self.bit_buffer |= v;
|
|
in_buffer <<= n;
|
|
},
|
|
builtin.Endian.Little => {
|
|
const v = @truncate(u8, in_buffer) << @intCast(u3, self.bit_count);
|
|
self.bit_buffer |= v;
|
|
in_buffer >>= n;
|
|
},
|
|
}
|
|
self.bit_count += n;
|
|
in_bits -= n;
|
|
|
|
//if we didn't fill the buffer, it's because bits < bits_remaining;
|
|
if (self.bit_count != u8_bit_count) return;
|
|
try self.out_stream.writeByte(self.bit_buffer);
|
|
self.bit_buffer = 0;
|
|
self.bit_count = 0;
|
|
}
|
|
//at this point we know bit_buffer is empty
|
|
|
|
//copy bytes until we can't fill one anymore, then leave the rest in bit_buffer
|
|
while (in_bits >= u8_bit_count) {
|
|
switch (endian) {
|
|
builtin.Endian.Big => {
|
|
const v = @intCast(u8, in_buffer >> high_byte_shift);
|
|
try self.out_stream.writeByte(v);
|
|
in_buffer <<= @intCast(u3, u8_bit_count - 1);
|
|
in_buffer <<= 1;
|
|
},
|
|
builtin.Endian.Little => {
|
|
const v = @truncate(u8, in_buffer);
|
|
try self.out_stream.writeByte(v);
|
|
in_buffer >>= @intCast(u3, u8_bit_count - 1);
|
|
in_buffer >>= 1;
|
|
},
|
|
}
|
|
in_bits -= u8_bit_count;
|
|
}
|
|
|
|
if (in_bits > 0) {
|
|
self.bit_count = @intCast(u4, in_bits);
|
|
self.bit_buffer = switch (endian) {
|
|
builtin.Endian.Big => @truncate(u8, in_buffer >> high_byte_shift),
|
|
builtin.Endian.Little => @truncate(u8, in_buffer),
|
|
};
|
|
}
|
|
}
|
|
|
|
/// Flush any remaining bits to the stream.
|
|
pub fn flushBits(self: *Self) Error!void {
|
|
if (self.bit_count == 0) return;
|
|
try self.out_stream.writeByte(self.bit_buffer);
|
|
self.bit_buffer = 0;
|
|
self.bit_count = 0;
|
|
}
|
|
|
|
pub fn write(self_stream: *Stream, buffer: []const u8) Error!void {
|
|
var self = @fieldParentPtr(Self, "stream", self_stream);
|
|
|
|
//@NOTE: I'm not sure this is a good idea, maybe flushBits should be forced
|
|
if (self.bit_count > 0) {
|
|
for (buffer) |b, i|
|
|
try self.writeBits(b, u8_bit_count);
|
|
return;
|
|
}
|
|
|
|
return self.out_stream.write(buffer);
|
|
}
|
|
};
|
|
}
|
|
|
|
pub const BufferedAtomicFile = struct {
|
|
atomic_file: os.AtomicFile,
|
|
file_stream: os.File.OutStream,
|
|
buffered_stream: BufferedOutStream(os.File.WriteError),
|
|
allocator: *mem.Allocator,
|
|
|
|
pub fn create(allocator: *mem.Allocator, dest_path: []const u8) !*BufferedAtomicFile {
|
|
// TODO with well defined copy elision we don't need this allocation
|
|
var self = try allocator.create(BufferedAtomicFile);
|
|
self.* = BufferedAtomicFile{
|
|
.atomic_file = undefined,
|
|
.file_stream = undefined,
|
|
.buffered_stream = undefined,
|
|
.allocator = allocator,
|
|
};
|
|
errdefer allocator.destroy(self);
|
|
|
|
self.atomic_file = try os.AtomicFile.init(dest_path, os.File.default_mode);
|
|
errdefer self.atomic_file.deinit();
|
|
|
|
self.file_stream = self.atomic_file.file.outStream();
|
|
self.buffered_stream = BufferedOutStream(os.File.WriteError).init(&self.file_stream.stream);
|
|
return self;
|
|
}
|
|
|
|
/// always call destroy, even after successful finish()
|
|
pub fn destroy(self: *BufferedAtomicFile) void {
|
|
self.atomic_file.deinit();
|
|
self.allocator.destroy(self);
|
|
}
|
|
|
|
pub fn finish(self: *BufferedAtomicFile) !void {
|
|
try self.buffered_stream.flush();
|
|
try self.atomic_file.finish();
|
|
}
|
|
|
|
pub fn stream(self: *BufferedAtomicFile) *OutStream(os.File.WriteError) {
|
|
return &self.buffered_stream.stream;
|
|
}
|
|
};
|
|
|
|
pub fn readLine(buf: *std.Buffer) ![]u8 {
|
|
var stdin = try getStdIn();
|
|
var stdin_stream = stdin.inStream();
|
|
return readLineFrom(&stdin_stream.stream, buf);
|
|
}
|
|
|
|
/// Reads all characters until the next newline into buf, and returns
|
|
/// a slice of the characters read (excluding the newline character(s)).
|
|
pub fn readLineFrom(stream: var, buf: *std.Buffer) ![]u8 {
|
|
const start = buf.len();
|
|
while (true) {
|
|
const byte = try stream.readByte();
|
|
switch (byte) {
|
|
'\r' => {
|
|
// trash the following \n
|
|
_ = try stream.readByte();
|
|
return buf.toSlice()[start..];
|
|
},
|
|
'\n' => return buf.toSlice()[start..],
|
|
else => try buf.appendByte(byte),
|
|
}
|
|
}
|
|
}
|
|
|
|
test "io.readLineFrom" {
|
|
var bytes: [128]u8 = undefined;
|
|
const allocator = &std.heap.FixedBufferAllocator.init(bytes[0..]).allocator;
|
|
|
|
var buf = try std.Buffer.initSize(allocator, 0);
|
|
var mem_stream = SliceInStream.init(
|
|
\\Line 1
|
|
\\Line 22
|
|
\\Line 333
|
|
);
|
|
const stream = &mem_stream.stream;
|
|
|
|
testing.expectEqualSlices(u8, "Line 1", try readLineFrom(stream, &buf));
|
|
testing.expectEqualSlices(u8, "Line 22", try readLineFrom(stream, &buf));
|
|
testing.expectError(error.EndOfStream, readLineFrom(stream, &buf));
|
|
testing.expectEqualSlices(u8, "Line 1Line 22Line 333", buf.toSlice());
|
|
}
|
|
|
|
pub fn readLineSlice(slice: []u8) ![]u8 {
|
|
var stdin = try getStdIn();
|
|
var stdin_stream = stdin.inStream();
|
|
return readLineSliceFrom(&stdin_stream.stream, slice);
|
|
}
|
|
|
|
/// Reads all characters until the next newline into slice, and returns
|
|
/// a slice of the characters read (excluding the newline character(s)).
|
|
pub fn readLineSliceFrom(stream: var, slice: []u8) ![]u8 {
|
|
// We cannot use Buffer.fromOwnedSlice, as it wants to append a null byte
|
|
// after taking ownership, which would always require an allocation.
|
|
var buf = std.Buffer{ .list = std.ArrayList(u8).fromOwnedSlice(debug.failing_allocator, slice) };
|
|
try buf.resize(0);
|
|
return try readLineFrom(stream, &buf);
|
|
}
|
|
|
|
test "io.readLineSliceFrom" {
|
|
var buf: [7]u8 = undefined;
|
|
var mem_stream = SliceInStream.init(
|
|
\\Line 1
|
|
\\Line 22
|
|
\\Line 333
|
|
);
|
|
const stream = &mem_stream.stream;
|
|
|
|
testing.expectEqualSlices(u8, "Line 1", try readLineSliceFrom(stream, buf[0..]));
|
|
testing.expectError(error.OutOfMemory, readLineSliceFrom(stream, buf[0..]));
|
|
}
|
|
|
|
pub const Packing = enum {
|
|
Byte, /// Pack data to byte alignment
|
|
Bit, /// Pack data to bit alignment
|
|
};
|
|
|
|
/// Creates a deserializer that deserializes types from any stream.
|
|
/// If `is_packed` is true, the data stream is treated as bit-packed,
|
|
/// otherwise data is expected to be packed to the smallest byte.
|
|
/// Types may implement a custom deserialization routine with a
|
|
/// function named `deserialize` in the form of:
|
|
/// pub fn deserialize(self: *Self, deserializer: var) !void
|
|
/// which will be called when the deserializer is used to deserialize
|
|
/// that type. It will pass a pointer to the type instance to deserialize
|
|
/// into and a pointer to the deserializer struct.
|
|
pub fn Deserializer(comptime endian: builtin.Endian, comptime packing: Packing, comptime Error: type) type {
|
|
return struct {
|
|
const Self = @This();
|
|
|
|
in_stream: if (packing == .Bit) BitInStream(endian, Stream.Error) else *Stream,
|
|
|
|
pub const Stream = InStream(Error);
|
|
|
|
pub fn init(in_stream: *Stream) Self {
|
|
return Self{ .in_stream = switch (packing) {
|
|
.Bit => BitInStream(endian, Stream.Error).init(in_stream),
|
|
.Byte => in_stream,
|
|
} };
|
|
}
|
|
|
|
pub fn alignToByte(self: *Self) void {
|
|
if (!is_packed) return;
|
|
self.in_stream.alignToByte();
|
|
}
|
|
|
|
//@BUG: inferred error issue. See: #1386
|
|
fn deserializeInt(self: *Self, comptime T: type) (Error || error{EndOfStream})!T {
|
|
comptime assert(trait.is(builtin.TypeId.Int)(T) or trait.is(builtin.TypeId.Float)(T));
|
|
|
|
const u8_bit_count = 8;
|
|
const t_bit_count = comptime meta.bitCount(T);
|
|
|
|
const U = @IntType(false, t_bit_count);
|
|
const Log2U = math.Log2Int(U);
|
|
const int_size = (U.bit_count + 7) / 8;
|
|
|
|
if (packing == .Bit) {
|
|
const result = try self.in_stream.readBitsNoEof(U, t_bit_count);
|
|
return @bitCast(T, result);
|
|
}
|
|
|
|
var buffer: [int_size]u8 = undefined;
|
|
const read_size = try self.in_stream.read(buffer[0..]);
|
|
if (read_size < int_size) return error.EndOfStream;
|
|
|
|
if (int_size == 1) {
|
|
if (t_bit_count == 8) return @bitCast(T, buffer[0]);
|
|
const PossiblySignedByte = @IntType(T.is_signed, 8);
|
|
return @truncate(T, @bitCast(PossiblySignedByte, buffer[0]));
|
|
}
|
|
|
|
var result = U(0);
|
|
for (buffer) |byte, i| {
|
|
switch (endian) {
|
|
builtin.Endian.Big => {
|
|
result = (result << u8_bit_count) | byte;
|
|
},
|
|
builtin.Endian.Little => {
|
|
result |= U(byte) << @intCast(Log2U, u8_bit_count * i);
|
|
},
|
|
}
|
|
}
|
|
|
|
return @bitCast(T, result);
|
|
}
|
|
|
|
//@TODO: Replace this with @unionInit or whatever when it is added
|
|
// see: #1315
|
|
fn setTag(ptr: var, tag: var) void {
|
|
const T = @typeOf(ptr);
|
|
comptime assert(trait.isPtrTo(builtin.TypeId.Union)(T));
|
|
const U = meta.Child(T);
|
|
|
|
const info = @typeInfo(U).Union;
|
|
if (info.tag_type) |TagType| {
|
|
comptime assert(TagType == @typeOf(tag));
|
|
|
|
var ptr_tag = ptr: {
|
|
if (@alignOf(TagType) >= @alignOf(U)) break :ptr @ptrCast(*TagType, ptr);
|
|
const offset = comptime max: {
|
|
var max_field_size: comptime_int = 0;
|
|
for (info.fields) |field_info| {
|
|
const field_size = @sizeOf(field_info.field_type);
|
|
max_field_size = math.max(max_field_size, field_size);
|
|
}
|
|
break :max math.max(max_field_size, @alignOf(U));
|
|
};
|
|
break :ptr @intToPtr(*TagType, @ptrToInt(ptr) + offset);
|
|
};
|
|
ptr_tag.* = tag;
|
|
}
|
|
}
|
|
|
|
/// Deserializes and returns data of the specified type from the stream
|
|
pub fn deserialize(self: *Self, comptime T: type) !T {
|
|
var value: T = undefined;
|
|
try self.deserializeInto(&value);
|
|
return value;
|
|
}
|
|
|
|
/// Deserializes data into the type pointed to by `ptr`
|
|
pub fn deserializeInto(self: *Self, ptr: var) !void {
|
|
const T = @typeOf(ptr);
|
|
comptime assert(trait.is(builtin.TypeId.Pointer)(T));
|
|
|
|
if (comptime trait.isSlice(T) or comptime trait.isPtrTo(builtin.TypeId.Array)(T)) {
|
|
for (ptr) |*v|
|
|
try self.deserializeInto(v);
|
|
return;
|
|
}
|
|
|
|
comptime assert(trait.isSingleItemPtr(T));
|
|
|
|
const C = comptime meta.Child(T);
|
|
const child_type_id = @typeId(C);
|
|
|
|
//custom deserializer: fn(self: *Self, deserializer: var) !void
|
|
if (comptime trait.hasFn("deserialize")(C)) return C.deserialize(ptr, self);
|
|
|
|
if (comptime trait.isPacked(C) and packing != .Bit) {
|
|
var packed_deserializer = Deserializer(endian, .Bit, Error).init(self.in_stream);
|
|
return packed_deserializer.deserializeInto(ptr);
|
|
}
|
|
|
|
switch (child_type_id) {
|
|
builtin.TypeId.Void => return,
|
|
builtin.TypeId.Bool => ptr.* = (try self.deserializeInt(u1)) > 0,
|
|
builtin.TypeId.Float, builtin.TypeId.Int => ptr.* = try self.deserializeInt(C),
|
|
builtin.TypeId.Struct => {
|
|
const info = @typeInfo(C).Struct;
|
|
|
|
inline for (info.fields) |*field_info| {
|
|
const name = field_info.name;
|
|
const FieldType = field_info.field_type;
|
|
|
|
if (FieldType == void or FieldType == u0) continue;
|
|
|
|
//it doesn't make any sense to read pointers
|
|
if (comptime trait.is(builtin.TypeId.Pointer)(FieldType)) {
|
|
@compileError("Will not " ++ "read field " ++ name ++ " of struct " ++
|
|
@typeName(C) ++ " because it " ++ "is of pointer-type " ++
|
|
@typeName(FieldType) ++ ".");
|
|
}
|
|
|
|
try self.deserializeInto(&@field(ptr, name));
|
|
}
|
|
},
|
|
builtin.TypeId.Union => {
|
|
const info = @typeInfo(C).Union;
|
|
if (info.tag_type) |TagType| {
|
|
//we avoid duplicate iteration over the enum tags
|
|
// by getting the int directly and casting it without
|
|
// safety. If it is bad, it will be caught anyway.
|
|
const TagInt = @TagType(TagType);
|
|
const tag = try self.deserializeInt(TagInt);
|
|
|
|
{
|
|
@setRuntimeSafety(false);
|
|
//See: #1315
|
|
setTag(ptr, @intToEnum(TagType, tag));
|
|
}
|
|
|
|
inline for (info.fields) |field_info| {
|
|
if (field_info.enum_field.?.value == tag) {
|
|
const name = field_info.name;
|
|
const FieldType = field_info.field_type;
|
|
@field(ptr, name) = FieldType(undefined);
|
|
try self.deserializeInto(&@field(ptr, name));
|
|
return;
|
|
}
|
|
}
|
|
//This is reachable if the enum data is bad
|
|
return error.InvalidEnumTag;
|
|
}
|
|
@compileError("Cannot meaningfully deserialize " ++ @typeName(C) ++
|
|
" because it is an untagged union. Use a custom deserialize().");
|
|
},
|
|
builtin.TypeId.Optional => {
|
|
const OC = comptime meta.Child(C);
|
|
const exists = (try self.deserializeInt(u1)) > 0;
|
|
if (!exists) {
|
|
ptr.* = null;
|
|
return;
|
|
}
|
|
|
|
ptr.* = OC(undefined); //make it non-null so the following .? is guaranteed safe
|
|
const val_ptr = &ptr.*.?;
|
|
try self.deserializeInto(val_ptr);
|
|
},
|
|
builtin.TypeId.Enum => {
|
|
var value = try self.deserializeInt(@TagType(C));
|
|
ptr.* = try meta.intToEnum(C, value);
|
|
},
|
|
else => {
|
|
@compileError("Cannot deserialize " ++ @tagName(child_type_id) ++ " types (unimplemented).");
|
|
},
|
|
}
|
|
}
|
|
};
|
|
}
|
|
|
|
/// Creates a serializer that serializes types to any stream.
|
|
/// If `is_packed` is true, the data will be bit-packed into the stream.
|
|
/// Note that the you must call `serializer.flush()` when you are done
|
|
/// writing bit-packed data in order ensure any unwritten bits are committed.
|
|
/// If `is_packed` is false, data is packed to the smallest byte. In the case
|
|
/// of packed structs, the struct will written bit-packed and with the specified
|
|
/// endianess, after which data will resume being written at the next byte boundary.
|
|
/// Types may implement a custom serialization routine with a
|
|
/// function named `serialize` in the form of:
|
|
/// pub fn serialize(self: Self, serializer: var) !void
|
|
/// which will be called when the serializer is used to serialize that type. It will
|
|
/// pass a const pointer to the type instance to be serialized and a pointer
|
|
/// to the serializer struct.
|
|
pub fn Serializer(comptime endian: builtin.Endian, comptime packing: Packing, comptime Error: type) type {
|
|
return struct {
|
|
const Self = @This();
|
|
|
|
out_stream: if (packing == .Bit) BitOutStream(endian, Stream.Error) else *Stream,
|
|
|
|
pub const Stream = OutStream(Error);
|
|
|
|
pub fn init(out_stream: *Stream) Self {
|
|
return Self{ .out_stream = switch (packing) {
|
|
.Bit => BitOutStream(endian, Stream.Error).init(out_stream),
|
|
.Byte => out_stream,
|
|
} };
|
|
}
|
|
|
|
/// Flushes any unwritten bits to the stream
|
|
pub fn flush(self: *Self) Error!void {
|
|
if (packing == .Bit) return self.out_stream.flushBits();
|
|
}
|
|
|
|
fn serializeInt(self: *Self, value: var) Error!void {
|
|
const T = @typeOf(value);
|
|
comptime assert(trait.is(builtin.TypeId.Int)(T) or trait.is(builtin.TypeId.Float)(T));
|
|
|
|
const t_bit_count = comptime meta.bitCount(T);
|
|
const u8_bit_count = comptime meta.bitCount(u8);
|
|
|
|
const U = @IntType(false, t_bit_count);
|
|
const Log2U = math.Log2Int(U);
|
|
const int_size = (U.bit_count + 7) / 8;
|
|
|
|
const u_value = @bitCast(U, value);
|
|
|
|
if (packing == .Bit) return self.out_stream.writeBits(u_value, t_bit_count);
|
|
|
|
var buffer: [int_size]u8 = undefined;
|
|
if (int_size == 1) buffer[0] = u_value;
|
|
|
|
for (buffer) |*byte, i| {
|
|
const idx = switch (endian) {
|
|
.Big => int_size - i - 1,
|
|
.Little => i,
|
|
};
|
|
const shift = @intCast(Log2U, idx * u8_bit_count);
|
|
const v = u_value >> shift;
|
|
byte.* = if (t_bit_count < u8_bit_count) v else @truncate(u8, v);
|
|
}
|
|
|
|
try self.out_stream.write(buffer);
|
|
}
|
|
|
|
/// Serializes the passed value into the stream
|
|
pub fn serialize(self: *Self, value: var) Error!void {
|
|
const T = comptime @typeOf(value);
|
|
|
|
if (comptime trait.isIndexable(T)) {
|
|
for (value) |v|
|
|
try self.serialize(v);
|
|
return;
|
|
}
|
|
|
|
//custom serializer: fn(self: Self, serializer: var) !void
|
|
if (comptime trait.hasFn("serialize")(T)) return T.serialize(value, self);
|
|
|
|
if (comptime trait.isPacked(T) and packing != .Bit) {
|
|
var packed_serializer = Serializer(endian, .Bit, Error).init(self.out_stream);
|
|
try packed_serializer.serialize(value);
|
|
try packed_serializer.flush();
|
|
return;
|
|
}
|
|
|
|
switch (@typeId(T)) {
|
|
builtin.TypeId.Void => return,
|
|
builtin.TypeId.Bool => try self.serializeInt(u1(@boolToInt(value))),
|
|
builtin.TypeId.Float, builtin.TypeId.Int => try self.serializeInt(value),
|
|
builtin.TypeId.Struct => {
|
|
const info = @typeInfo(T);
|
|
|
|
inline for (info.Struct.fields) |*field_info| {
|
|
const name = field_info.name;
|
|
const FieldType = field_info.field_type;
|
|
|
|
if (FieldType == void or FieldType == u0) continue;
|
|
|
|
//It doesn't make sense to write pointers
|
|
if (comptime trait.is(builtin.TypeId.Pointer)(FieldType)) {
|
|
@compileError("Will not " ++ "serialize field " ++ name ++
|
|
" of struct " ++ @typeName(T) ++ " because it " ++
|
|
"is of pointer-type " ++ @typeName(FieldType) ++ ".");
|
|
}
|
|
try self.serialize(@field(value, name));
|
|
}
|
|
},
|
|
builtin.TypeId.Union => {
|
|
const info = @typeInfo(T).Union;
|
|
if (info.tag_type) |TagType| {
|
|
const active_tag = meta.activeTag(value);
|
|
try self.serialize(active_tag);
|
|
//This inline loop is necessary because active_tag is a runtime
|
|
// value, but @field requires a comptime value. Our alternative
|
|
// is to check each field for a match
|
|
inline for (info.fields) |field_info| {
|
|
if (field_info.enum_field.?.value == @enumToInt(active_tag)) {
|
|
const name = field_info.name;
|
|
const FieldType = field_info.field_type;
|
|
try self.serialize(@field(value, name));
|
|
return;
|
|
}
|
|
}
|
|
unreachable;
|
|
}
|
|
@compileError("Cannot meaningfully serialize " ++ @typeName(T) ++
|
|
" because it is an untagged union. Use a custom serialize().");
|
|
},
|
|
builtin.TypeId.Optional => {
|
|
if (value == null) {
|
|
try self.serializeInt(u1(@boolToInt(false)));
|
|
return;
|
|
}
|
|
try self.serializeInt(u1(@boolToInt(true)));
|
|
|
|
const OC = comptime meta.Child(T);
|
|
const val_ptr = &value.?;
|
|
try self.serialize(val_ptr.*);
|
|
},
|
|
builtin.TypeId.Enum => {
|
|
try self.serializeInt(@enumToInt(value));
|
|
},
|
|
else => @compileError("Cannot serialize " ++ @tagName(@typeId(T)) ++ " types (unimplemented)."),
|
|
}
|
|
}
|
|
};
|
|
}
|
|
|
|
test "import io tests" {
|
|
comptime {
|
|
_ = @import("io_test.zig");
|
|
}
|
|
}
|
|
|