| ... | @@ -1,5 +1,5 @@ | ... | @@ -1,5 +1,5 @@ |
| 1 | // | 1 | // |
| 2 | // Decompressor for ZLIB data streams (RFC1950) | 2 | // Compressor/Decompressor for ZLIB data streams (RFC1950) |
| 3 | | 3 | |
| 4 | const std = @import("std"); | 4 | const std = @import("std"); |
| 5 | const io = std.io; | 5 | const io = std.io; |
| ... | @@ -8,7 +8,19 @@ const testing = std.testing; | ... | @@ -8,7 +8,19 @@ const testing = std.testing; |
| 8 | const mem = std.mem; | 8 | const mem = std.mem; |
| 9 | const deflate = std.compress.deflate; | 9 | const deflate = std.compress.deflate; |
| 10 | | 10 | |
| 11 | pub fn ZlibStream(comptime ReaderType: type) type { | 11 | // Zlib header format as specified in RFC1950 |
| | 12 | const ZLibHeader = packed struct { |
| | 13 | checksum: u5, |
| | 14 | preset_dict: u1, |
| | 15 | compression_level: u2, |
| | 16 | compression_method: u4, |
| | 17 | compression_info: u4, |
| | 18 | |
| | 19 | const DEFLATE = 8; |
| | 20 | const WINDOW_32K = 7; |
| | 21 | }; |
| | 22 | |
| | 23 | pub fn DecompressStream(comptime ReaderType: type) type { |
| 12 | return struct { | 24 | return struct { |
| 13 | const Self = @This(); | 25 | const Self = @This(); |
| 14 | | 26 | |
| ... | @@ -24,26 +36,24 @@ pub fn ZlibStream(comptime ReaderType: type) type { | ... | @@ -24,26 +36,24 @@ pub fn ZlibStream(comptime ReaderType: type) type { |
| 24 | | 36 | |
| 25 | fn init(allocator: mem.Allocator, source: ReaderType) !Self { | 37 | fn init(allocator: mem.Allocator, source: ReaderType) !Self { |
| 26 | // Zlib header format is specified in RFC1950 | 38 | // Zlib header format is specified in RFC1950 |
| 27 | const header = try source.readBytesNoEof(2); | 39 | const header_u16 = try source.readIntBig(u16); |
| 28 | | | |
| 29 | const CM = @truncate(u4, header[0]); | | |
| 30 | const CINFO = @truncate(u4, header[0] >> 4); | | |
| 31 | const FCHECK = @truncate(u5, header[1]); | | |
| 32 | _ = FCHECK; | | |
| 33 | const FDICT = @truncate(u1, header[1] >> 5); | | |
| 34 | | 40 | |
| 35 | if ((@as(u16, header[0]) << 8 | header[1]) % 31 != 0) | 41 | // verify the header checksum |
| | 42 | if (header_u16 % 31 != 0) |
| 36 | return error.BadHeader; | 43 | return error.BadHeader; |
| | 44 | const header = @bitCast(ZLibHeader, header_u16); |
| 37 | | 45 | |
| 38 | // The CM field must be 8 to indicate the use of DEFLATE | 46 | // The CM field must be 8 to indicate the use of DEFLATE |
| 39 | if (CM != 8) return error.InvalidCompression; | 47 | if (header.compression_method != ZLibHeader.DEFLATE) |
| | 48 | return error.InvalidCompression; |
| 40 | // CINFO is the base-2 logarithm of the LZ77 window size, minus 8. | 49 | // CINFO is the base-2 logarithm of the LZ77 window size, minus 8. |
| 41 | // Values above 7 are unspecified and therefore rejected. | 50 | // Values above 7 are unspecified and therefore rejected. |
| 42 | if (CINFO > 7) return error.InvalidWindowSize; | 51 | if (header.compression_info > ZLibHeader.WINDOW_32K) |
| | 52 | return error.InvalidWindowSize; |
| 43 | | 53 | |
| 44 | const dictionary = null; | 54 | const dictionary = null; |
| 45 | // TODO: Support this case | 55 | // TODO: Support this case |
| 46 | if (FDICT != 0) | 56 | if (header.preset_dict != 0) |
| 47 | return error.Unsupported; | 57 | return error.Unsupported; |
| 48 | | 58 | |
| 49 | return Self{ | 59 | return Self{ |
| ... | @@ -84,14 +94,96 @@ pub fn ZlibStream(comptime ReaderType: type) type { | ... | @@ -84,14 +94,96 @@ pub fn ZlibStream(comptime ReaderType: type) type { |
| 84 | }; | 94 | }; |
| 85 | } | 95 | } |
| 86 | | 96 | |
| 87 | pub fn zlibStream(allocator: mem.Allocator, reader: anytype) !ZlibStream(@TypeOf(reader)) { | 97 | pub fn decompressStream(allocator: mem.Allocator, reader: anytype) !DecompressStream(@TypeOf(reader)) { |
| 88 | return ZlibStream(@TypeOf(reader)).init(allocator, reader); | 98 | return DecompressStream(@TypeOf(reader)).init(allocator, reader); |
| | 99 | } |
| | 100 | |
| | 101 | pub const CompressionLevel = enum(u2) { |
| | 102 | no_compression = 0, |
| | 103 | fastest = 1, |
| | 104 | default = 2, |
| | 105 | maximum = 3, |
| | 106 | }; |
| | 107 | |
| | 108 | pub const CompressStreamOptions = struct { |
| | 109 | level: CompressionLevel = .default, |
| | 110 | }; |
| | 111 | |
| | 112 | pub fn CompressStream(comptime WriterType: type) type { |
| | 113 | return struct { |
| | 114 | const Self = @This(); |
| | 115 | |
| | 116 | const Error = WriterType.Error || |
| | 117 | deflate.Compressor(WriterType).Error; |
| | 118 | pub const Writer = io.Writer(*Self, Error, write); |
| | 119 | |
| | 120 | allocator: mem.Allocator, |
| | 121 | deflator: deflate.Compressor(WriterType), |
| | 122 | in_writer: WriterType, |
| | 123 | hasher: std.hash.Adler32, |
| | 124 | |
| | 125 | fn init(allocator: mem.Allocator, dest: WriterType, options: CompressStreamOptions) !Self { |
| | 126 | var header = ZLibHeader{ |
| | 127 | .compression_info = ZLibHeader.WINDOW_32K, |
| | 128 | .compression_method = ZLibHeader.DEFLATE, |
| | 129 | .compression_level = @enumToInt(options.level), |
| | 130 | .preset_dict = 0, |
| | 131 | .checksum = 0, |
| | 132 | }; |
| | 133 | header.checksum = @truncate(u5, 31 - @bitCast(u16, header) % 31); |
| | 134 | |
| | 135 | try dest.writeIntBig(u16, @bitCast(u16, header)); |
| | 136 | |
| | 137 | const compression_level: deflate.Compression = switch (options.level) { |
| | 138 | .no_compression => .no_compression, |
| | 139 | .fastest => .best_speed, |
| | 140 | .default => .default_compression, |
| | 141 | .maximum => .best_compression, |
| | 142 | }; |
| | 143 | |
| | 144 | return Self{ |
| | 145 | .allocator = allocator, |
| | 146 | .deflator = try deflate.compressor(allocator, dest, .{ .level = compression_level }), |
| | 147 | .in_writer = dest, |
| | 148 | .hasher = std.hash.Adler32.init(), |
| | 149 | }; |
| | 150 | } |
| | 151 | |
| | 152 | pub fn write(self: *Self, bytes: []const u8) Error!usize { |
| | 153 | if (bytes.len == 0) { |
| | 154 | return 0; |
| | 155 | } |
| | 156 | |
| | 157 | const w = try self.deflator.write(bytes); |
| | 158 | |
| | 159 | self.hasher.update(bytes[0..w]); |
| | 160 | return w; |
| | 161 | } |
| | 162 | |
| | 163 | pub fn writer(self: *Self) Writer { |
| | 164 | return .{ .context = self }; |
| | 165 | } |
| | 166 | |
| | 167 | pub fn deinit(self: *Self) void { |
| | 168 | self.deflator.deinit(); |
| | 169 | } |
| | 170 | |
| | 171 | pub fn finish(self: *Self) !void { |
| | 172 | const hash = self.hasher.final(); |
| | 173 | try self.deflator.close(); |
| | 174 | try self.in_writer.writeIntBig(u32, hash); |
| | 175 | } |
| | 176 | }; |
| | 177 | } |
| | 178 | |
| | 179 | pub fn compressStream(allocator: mem.Allocator, writer: anytype, options: CompressStreamOptions) !CompressStream(@TypeOf(writer)) { |
| | 180 | return CompressStream(@TypeOf(writer)).init(allocator, writer, options); |
| 89 | } | 181 | } |
| 90 | | 182 | |
| 91 | fn testReader(data: []const u8, expected: []const u8) !void { | 183 | fn testDecompress(data: []const u8, expected: []const u8) !void { |
| 92 | var in_stream = io.fixedBufferStream(data); | 184 | var in_stream = io.fixedBufferStream(data); |
| 93 | | 185 | |
| 94 | var zlib_stream = try zlibStream(testing.allocator, in_stream.reader()); | 186 | var zlib_stream = try decompressStream(testing.allocator, in_stream.reader()); |
| 95 | defer zlib_stream.deinit(); | 187 | defer zlib_stream.deinit(); |
| 96 | | 188 | |
| 97 | // Read and decompress the whole file | 189 | // Read and decompress the whole file |
| ... | @@ -110,24 +202,24 @@ test "compressed data" { | ... | @@ -110,24 +202,24 @@ test "compressed data" { |
| 110 | const rfc1951_txt = @embedFile("testdata/rfc1951.txt"); | 202 | const rfc1951_txt = @embedFile("testdata/rfc1951.txt"); |
| 111 | | 203 | |
| 112 | // Compressed with compression level = 0 | 204 | // Compressed with compression level = 0 |
| 113 | try testReader( | 205 | try testDecompress( |
| 114 | @embedFile("testdata/rfc1951.txt.z.0"), | 206 | @embedFile("testdata/rfc1951.txt.z.0"), |
| 115 | rfc1951_txt, | 207 | rfc1951_txt, |
| 116 | ); | 208 | ); |
| 117 | // Compressed with compression level = 9 | 209 | // Compressed with compression level = 9 |
| 118 | try testReader( | 210 | try testDecompress( |
| 119 | @embedFile("testdata/rfc1951.txt.z.9"), | 211 | @embedFile("testdata/rfc1951.txt.z.9"), |
| 120 | rfc1951_txt, | 212 | rfc1951_txt, |
| 121 | ); | 213 | ); |
| 122 | // Compressed with compression level = 9 and fixed Huffman codes | 214 | // Compressed with compression level = 9 and fixed Huffman codes |
| 123 | try testReader( | 215 | try testDecompress( |
| 124 | @embedFile("testdata/rfc1951.txt.fixed.z.9"), | 216 | @embedFile("testdata/rfc1951.txt.fixed.z.9"), |
| 125 | rfc1951_txt, | 217 | rfc1951_txt, |
| 126 | ); | 218 | ); |
| 127 | } | 219 | } |
| 128 | | 220 | |
| 129 | test "don't read past deflate stream's end" { | 221 | test "don't read past deflate stream's end" { |
| 130 | try testReader(&[_]u8{ | 222 | try testDecompress(&[_]u8{ |
| 131 | 0x08, 0xd7, 0x63, 0xf8, 0xcf, 0xc0, 0xc0, 0x00, 0xc1, 0xff, | 223 | 0x08, 0xd7, 0x63, 0xf8, 0xcf, 0xc0, 0xc0, 0x00, 0xc1, 0xff, |
| 132 | 0xff, 0x43, 0x30, 0x03, 0x03, 0xc3, 0xff, 0xff, 0xff, 0x01, | 224 | 0xff, 0x43, 0x30, 0x03, 0x03, 0xc3, 0xff, 0xff, 0xff, 0x01, |
| 133 | 0x83, 0x95, 0x0b, 0xf5, | 225 | 0x83, 0x95, 0x0b, 0xf5, |
| ... | @@ -142,31 +234,49 @@ test "sanity checks" { | ... | @@ -142,31 +234,49 @@ test "sanity checks" { |
| 142 | // Truncated header | 234 | // Truncated header |
| 143 | try testing.expectError( | 235 | try testing.expectError( |
| 144 | error.EndOfStream, | 236 | error.EndOfStream, |
| 145 | testReader(&[_]u8{0x78}, ""), | 237 | testDecompress(&[_]u8{0x78}, ""), |
| 146 | ); | 238 | ); |
| 147 | // Failed FCHECK check | 239 | // Failed FCHECK check |
| 148 | try testing.expectError( | 240 | try testing.expectError( |
| 149 | error.BadHeader, | 241 | error.BadHeader, |
| 150 | testReader(&[_]u8{ 0x78, 0x9D }, ""), | 242 | testDecompress(&[_]u8{ 0x78, 0x9D }, ""), |
| 151 | ); | 243 | ); |
| 152 | // Wrong CM | 244 | // Wrong CM |
| 153 | try testing.expectError( | 245 | try testing.expectError( |
| 154 | error.InvalidCompression, | 246 | error.InvalidCompression, |
| 155 | testReader(&[_]u8{ 0x79, 0x94 }, ""), | 247 | testDecompress(&[_]u8{ 0x79, 0x94 }, ""), |
| 156 | ); | 248 | ); |
| 157 | // Wrong CINFO | 249 | // Wrong CINFO |
| 158 | try testing.expectError( | 250 | try testing.expectError( |
| 159 | error.InvalidWindowSize, | 251 | error.InvalidWindowSize, |
| 160 | testReader(&[_]u8{ 0x88, 0x98 }, ""), | 252 | testDecompress(&[_]u8{ 0x88, 0x98 }, ""), |
| 161 | ); | 253 | ); |
| 162 | // Wrong checksum | 254 | // Wrong checksum |
| 163 | try testing.expectError( | 255 | try testing.expectError( |
| 164 | error.WrongChecksum, | 256 | error.WrongChecksum, |
| 165 | testReader(&[_]u8{ 0x78, 0xda, 0x03, 0x00, 0x00, 0x00, 0x00, 0x00 }, ""), | 257 | testDecompress(&[_]u8{ 0x78, 0xda, 0x03, 0x00, 0x00, 0x00, 0x00, 0x00 }, ""), |
| 166 | ); | 258 | ); |
| 167 | // Truncated checksum | 259 | // Truncated checksum |
| 168 | try testing.expectError( | 260 | try testing.expectError( |
| 169 | error.EndOfStream, | 261 | error.EndOfStream, |
| 170 | testReader(&[_]u8{ 0x78, 0xda, 0x03, 0x00, 0x00 }, ""), | 262 | testDecompress(&[_]u8{ 0x78, 0xda, 0x03, 0x00, 0x00 }, ""), |
| 171 | ); | 263 | ); |
| 172 | } | 264 | } |
| | 265 | |
| | 266 | test "compress data" { |
| | 267 | const allocator = testing.allocator; |
| | 268 | const rfc1951_txt = @embedFile("testdata/rfc1951.txt"); |
| | 269 | |
| | 270 | for (std.meta.tags(CompressionLevel)) |level| { |
| | 271 | var compressed_data = std.ArrayList(u8).init(allocator); |
| | 272 | defer compressed_data.deinit(); |
| | 273 | |
| | 274 | var compressor = try compressStream(allocator, compressed_data.writer(), .{ .level = level }); |
| | 275 | defer compressor.deinit(); |
| | 276 | |
| | 277 | try compressor.writer().writeAll(rfc1951_txt); |
| | 278 | try compressor.finish(); |
| | 279 | |
| | 280 | try testDecompress(compressed_data.items, rfc1951_txt); |
| | 281 | } |
| | 282 | } |