Commit f85c01d4c7520d2242626f4c0684ab97e47af373 | zig

author	FnControlOption <fncontroloption@noreply.codeberg.org> 2023-01-21 15:47:31-08:00
committer	Andrew Kelley <andrew@ziglang.org> 2023-01-22 00:33:29-05:00
log	f85c01d4c7520d2242626f4c0684ab97e47af373
tree	9385b5c065a3e90449dc1e02883d579799dca1fe
parent	562d52e23dce703b0eb45a212522588274afbd38

Implement gzip header CRC check.

From RFC 1952:

> If FHCRC is set, a CRC16 for the gzip header is present,
> immediately before the compressed data. The CRC16 consists
> of the two least significant bytes of the CRC32 for all
> bytes of the gzip header up to and not including the CRC16.

2 files changed, 54 insertions(+), 8 deletions(-)

lib
- std
  - compress.zig
  - compress
    gzip.zig

lib/std/compress.zig+30

...	...	@@ -4,6 +4,36 @@ pub const deflate = @import("compress/deflate.zig");
4	4	pub const gzip = @import("compress/gzip.zig");
5	5	pub const zlib = @import("compress/zlib.zig");
6	6
	7	pub fn HashedReader(
	8	comptime ReaderType: anytype,
	9	comptime HasherType: anytype,
	10	) type {
	11	return struct {
	12	child_reader: ReaderType,
	13	hasher: HasherType,
	14
	15	pub const Error = ReaderType.Error;
	16	pub const Reader = std.io.Reader(*@This(), Error, read);
	17
	18	pub fn read(self: *@This(), buf: []u8) Error!usize {
	19	const amt = try self.child_reader.read(buf);
	20	self.hasher.update(buf);
	21	return amt;
	22	}
	23
	24	pub fn reader(self: *@This()) Reader {
	25	return .{ .context = self };
	26	}
	27	};
	28	}
	29
	30	pub fn hashedReader(
	31	reader: anytype,
	32	hasher: anytype,
	33	) HashedReader(@TypeOf(reader), @TypeOf(hasher)) {
	34	return .{ .child_reader = reader, .hasher = hasher };
	35	}
	36
7	37	test {
8	38	_ = deflate;
9	39	_ = gzip;

lib/std/compress/gzip.zig+24-8

...	...	@@ -44,8 +44,11 @@ pub fn GzipStream(comptime ReaderType: type) type {
44	44	},
45	45
46	46	fn init(allocator: mem.Allocator, source: ReaderType) !Self {
	47	var hasher = std.compress.hashedReader(source, std.hash.Crc32.init());
	48	const hashed_reader = hasher.reader();
	49
47	50	// gzip header format is specified in RFC1952
48		const header = try source.readBytesNoEof(10);
	51	const header = try hashed_reader.readBytesNoEof(10);
49	52
50	53	// Check the ID1/ID2 fields
51	54	if (header[0] != 0x1f or header[1] != 0x8b)
...	...	@@ -66,31 +69,31 @@ pub fn GzipStream(comptime ReaderType: type) type {
66	69	_ = XFL;
67	70
68	71	const extra = if (FLG & FEXTRA != 0) blk: {
69		const len = try source.readIntLittle(u16);
	72	const len = try hashed_reader.readIntLittle(u16);
70	73	const tmp_buf = try allocator.alloc(u8, len);
71	74	errdefer allocator.free(tmp_buf);
72	75
73		try source.readNoEof(tmp_buf);
	76	try hashed_reader.readNoEof(tmp_buf);
74	77	break :blk tmp_buf;
75	78	} else null;
76	79	errdefer if (extra) \|p\| allocator.free(p);
77	80
78	81	const filename = if (FLG & FNAME != 0)
79		try source.readUntilDelimiterAlloc(allocator, 0, max_string_len)
	82	try hashed_reader.readUntilDelimiterAlloc(allocator, 0, max_string_len)
80	83	else
81	84	null;
82	85	errdefer if (filename) \|p\| allocator.free(p);
83	86
84	87	const comment = if (FLG & FCOMMENT != 0)
85		try source.readUntilDelimiterAlloc(allocator, 0, max_string_len)
	88	try hashed_reader.readUntilDelimiterAlloc(allocator, 0, max_string_len)
86	89	else
87	90	null;
88	91	errdefer if (comment) \|p\| allocator.free(p);
89	92
90	93	if (FLG & FHCRC != 0) {
91		// TODO: Evaluate and check the header checksum. The stdlib has
92		// no CRC16 yet :(
93		_ = try source.readIntLittle(u16);
	94	const hash = try source.readIntLittle(u16);
	95	if (hash != @truncate(u16, hasher.hasher.final()))
	96	return error.WrongChecksum;
94	97	}
95	98
96	99	return Self{
...	...	@@ -230,3 +233,16 @@ test "sanity checks" {
230	233	}, ""),
231	234	);
232	235	}
	236
	237	test "header checksum" {
	238	try testReader(&[_]u8{
	239	// GZIP header
	240	0x1f, 0x8b, 0x08, 0x12, 0x00, 0x09, 0x6e, 0x88, 0x00, 0xff, 0x48, 0x65, 0x6c, 0x6c, 0x6f, 0x00,
	241
	242	// header.FHCRC (should cover entire header)
	243	0x99, 0xd6,
	244
	245	// GZIP data
	246	0x01, 0x00, 0x00, 0xff, 0xff, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
	247	}, "");
	248	}