| ... | @@ -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 ZlibStreamReader(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 ZlibStreamReader(comptime ReaderType: type) type { | ... | @@ -24,26 +36,24 @@ pub fn ZlibStreamReader(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 | | 40 | |
| 29 | const CM = @truncate(u4, header[0]); | 41 | // verify the header checksum |
| 30 | const CINFO = @truncate(u4, header[0] >> 4); | 42 | if (header_u16 % 31 != 0) |
| 31 | const FCHECK = @truncate(u5, header[1]); | | |
| 32 | _ = FCHECK; | | |
| 33 | const FDICT = @truncate(u1, header[1] >> 5); | | |
| 34 | | | |
| 35 | if ((@as(u16, header[0]) << 8 | header[1]) % 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{ |
| ... | @@ -84,8 +94,8 @@ pub fn ZlibStreamReader(comptime ReaderType: type) type { | ... | @@ -84,8 +94,8 @@ pub fn ZlibStreamReader(comptime ReaderType: type) type { |
| 84 | }; | 94 | }; |
| 85 | } | 95 | } |
| 86 | | 96 | |
| 87 | pub fn zlibStreamReader(allocator: mem.Allocator, reader: anytype) !ZlibStreamReader(@TypeOf(reader)) { | 97 | pub fn decompressStream(allocator: mem.Allocator, reader: anytype) !DecompressStream(@TypeOf(reader)) { |
| 88 | return ZlibStreamReader(@TypeOf(reader)).init(allocator, reader); | 98 | return DecompressStream(@TypeOf(reader)).init(allocator, reader); |
| 89 | } | 99 | } |
| 90 | | 100 | |
| 91 | pub const CompressionLevel = enum(u2) { | 101 | pub const CompressionLevel = enum(u2) { |
| ... | @@ -95,11 +105,11 @@ pub const CompressionLevel = enum(u2) { | ... | @@ -95,11 +105,11 @@ pub const CompressionLevel = enum(u2) { |
| 95 | maximum = 3, | 105 | maximum = 3, |
| 96 | }; | 106 | }; |
| 97 | | 107 | |
| 98 | pub const CompressionOptions = struct { | 108 | pub const CompressStreamOptions = struct { |
| 99 | level: CompressionLevel = .default, | 109 | level: CompressionLevel = .default, |
| 100 | }; | 110 | }; |
| 101 | | 111 | |
| 102 | pub fn ZlibStreamWriter(comptime WriterType: type) type { | 112 | pub fn CompressStream(comptime WriterType: type) type { |
| 103 | return struct { | 113 | return struct { |
| 104 | const Self = @This(); | 114 | const Self = @This(); |
| 105 | | 115 | |
| ... | @@ -112,17 +122,17 @@ pub fn ZlibStreamWriter(comptime WriterType: type) type { | ... | @@ -112,17 +122,17 @@ pub fn ZlibStreamWriter(comptime WriterType: type) type { |
| 112 | in_writer: WriterType, | 122 | in_writer: WriterType, |
| 113 | hasher: std.hash.Adler32, | 123 | hasher: std.hash.Adler32, |
| 114 | | 124 | |
| 115 | fn init(allocator: mem.Allocator, dest: WriterType, options: CompressionOptions) !Self { | 125 | fn init(allocator: mem.Allocator, dest: WriterType, options: CompressStreamOptions) !Self { |
| 116 | // Zlib header format is specified in RFC1950 | 126 | var header = ZLibHeader{ |
| 117 | const CM: u4 = 8; // DEFLATE | 127 | .compression_info = ZLibHeader.WINDOW_32K, |
| 118 | const CINFO: u4 = 7; // 32K window | 128 | .compression_method = ZLibHeader.DEFLATE, |
| 119 | const CMF: u8 = (@as(u8, CINFO) << 4) | CM; | 129 | .compression_level = @enumToInt(options.level), |
| | 130 | .preset_dict = 0, |
| | 131 | .checksum = 0, |
| | 132 | }; |
| | 133 | header.checksum = @truncate(u5, 31 - @bitCast(u16, header) % 31); |
| 120 | | 134 | |
| 121 | const FLEVEL: u2 = @enumToInt(options.level); | 135 | try dest.writeIntBig(u16, @bitCast(u16, header)); |
| 122 | const FDICT: u1 = 0; // No preset dictionary support | | |
| 123 | const FLG_temp = (@as(u8, FLEVEL) << 6) | (@as(u8, FDICT) << 5); | | |
| 124 | const FCHECK: u5 = 31 - ((@as(u16, CMF) * 256 + FLG_temp) % 31); | | |
| 125 | const FLG = FLG_temp | FCHECK; | | |
| 126 | | 136 | |
| 127 | const compression_level: deflate.Compression = switch (options.level) { | 137 | const compression_level: deflate.Compression = switch (options.level) { |
| 128 | .no_compression => .no_compression, | 138 | .no_compression => .no_compression, |
| ... | @@ -131,8 +141,6 @@ pub fn ZlibStreamWriter(comptime WriterType: type) type { | ... | @@ -131,8 +141,6 @@ pub fn ZlibStreamWriter(comptime WriterType: type) type { |
| 131 | .maximum => .best_compression, | 141 | .maximum => .best_compression, |
| 132 | }; | 142 | }; |
| 133 | | 143 | |
| 134 | try dest.writeAll(&.{ CMF, FLG }); | | |
| 135 | | | |
| 136 | return Self{ | 144 | return Self{ |
| 137 | .allocator = allocator, | 145 | .allocator = allocator, |
| 138 | .deflator = try deflate.compressor(allocator, dest, .{ .level = compression_level }), | 146 | .deflator = try deflate.compressor(allocator, dest, .{ .level = compression_level }), |
| ... | @@ -160,7 +168,7 @@ pub fn ZlibStreamWriter(comptime WriterType: type) type { | ... | @@ -160,7 +168,7 @@ pub fn ZlibStreamWriter(comptime WriterType: type) type { |
| 160 | self.deflator.deinit(); | 168 | self.deflator.deinit(); |
| 161 | } | 169 | } |
| 162 | | 170 | |
| 163 | pub fn close(self: *Self) !void { | 171 | pub fn finish(self: *Self) !void { |
| 164 | const hash = self.hasher.final(); | 172 | const hash = self.hasher.final(); |
| 165 | try self.deflator.close(); | 173 | try self.deflator.close(); |
| 166 | try self.in_writer.writeIntBig(u32, hash); | 174 | try self.in_writer.writeIntBig(u32, hash); |
| ... | @@ -168,15 +176,14 @@ pub fn ZlibStreamWriter(comptime WriterType: type) type { | ... | @@ -168,15 +176,14 @@ pub fn ZlibStreamWriter(comptime WriterType: type) type { |
| 168 | }; | 176 | }; |
| 169 | } | 177 | } |
| 170 | | 178 | |
| 171 | pub fn zlibStreamWriter(allocator: mem.Allocator, writer: anytype, options: CompressionOptions) !ZlibStreamWriter(@TypeOf(writer)) { | 179 | pub fn compressStream(allocator: mem.Allocator, writer: anytype, options: CompressStreamOptions) !CompressStream(@TypeOf(writer)) { |
| 172 | return ZlibStreamWriter(@TypeOf(writer)).init(allocator, writer, options); | 180 | return CompressStream(@TypeOf(writer)).init(allocator, writer, options); |
| 173 | } | 181 | } |
| 174 | | 182 | |
| 175 | | 183 | fn testDecompress(data: []const u8, expected: []const u8) !void { |
| 176 | fn testReader(data: []const u8, expected: []const u8) !void { | | |
| 177 | var in_stream = io.fixedBufferStream(data); | 184 | var in_stream = io.fixedBufferStream(data); |
| 178 | | 185 | |
| 179 | var zlib_stream = try zlibStreamReader(testing.allocator, in_stream.reader()); | 186 | var zlib_stream = try decompressStream(testing.allocator, in_stream.reader()); |
| 180 | defer zlib_stream.deinit(); | 187 | defer zlib_stream.deinit(); |
| 181 | | 188 | |
| 182 | // Read and decompress the whole file | 189 | // Read and decompress the whole file |
| ... | @@ -195,24 +202,24 @@ test "compressed data" { | ... | @@ -195,24 +202,24 @@ test "compressed data" { |
| 195 | const rfc1951_txt = @embedFile("testdata/rfc1951.txt"); | 202 | const rfc1951_txt = @embedFile("testdata/rfc1951.txt"); |
| 196 | | 203 | |
| 197 | // Compressed with compression level = 0 | 204 | // Compressed with compression level = 0 |
| 198 | try testReader( | 205 | try testDecompress( |
| 199 | @embedFile("testdata/rfc1951.txt.z.0"), | 206 | @embedFile("testdata/rfc1951.txt.z.0"), |
| 200 | rfc1951_txt, | 207 | rfc1951_txt, |
| 201 | ); | 208 | ); |
| 202 | // Compressed with compression level = 9 | 209 | // Compressed with compression level = 9 |
| 203 | try testReader( | 210 | try testDecompress( |
| 204 | @embedFile("testdata/rfc1951.txt.z.9"), | 211 | @embedFile("testdata/rfc1951.txt.z.9"), |
| 205 | rfc1951_txt, | 212 | rfc1951_txt, |
| 206 | ); | 213 | ); |
| 207 | // Compressed with compression level = 9 and fixed Huffman codes | 214 | // Compressed with compression level = 9 and fixed Huffman codes |
| 208 | try testReader( | 215 | try testDecompress( |
| 209 | @embedFile("testdata/rfc1951.txt.fixed.z.9"), | 216 | @embedFile("testdata/rfc1951.txt.fixed.z.9"), |
| 210 | rfc1951_txt, | 217 | rfc1951_txt, |
| 211 | ); | 218 | ); |
| 212 | } | 219 | } |
| 213 | | 220 | |
| 214 | test "don't read past deflate stream's end" { | 221 | test "don't read past deflate stream's end" { |
| 215 | try testReader(&[_]u8{ | 222 | try testDecompress(&[_]u8{ |
| 216 | 0x08, 0xd7, 0x63, 0xf8, 0xcf, 0xc0, 0xc0, 0x00, 0xc1, 0xff, | 223 | 0x08, 0xd7, 0x63, 0xf8, 0xcf, 0xc0, 0xc0, 0x00, 0xc1, 0xff, |
| 217 | 0xff, 0x43, 0x30, 0x03, 0x03, 0xc3, 0xff, 0xff, 0xff, 0x01, | 224 | 0xff, 0x43, 0x30, 0x03, 0x03, 0xc3, 0xff, 0xff, 0xff, 0x01, |
| 218 | 0x83, 0x95, 0x0b, 0xf5, | 225 | 0x83, 0x95, 0x0b, 0xf5, |
| ... | @@ -227,32 +234,32 @@ test "sanity checks" { | ... | @@ -227,32 +234,32 @@ test "sanity checks" { |
| 227 | // Truncated header | 234 | // Truncated header |
| 228 | try testing.expectError( | 235 | try testing.expectError( |
| 229 | error.EndOfStream, | 236 | error.EndOfStream, |
| 230 | testReader(&[_]u8{0x78}, ""), | 237 | testDecompress(&[_]u8{0x78}, ""), |
| 231 | ); | 238 | ); |
| 232 | // Failed FCHECK check | 239 | // Failed FCHECK check |
| 233 | try testing.expectError( | 240 | try testing.expectError( |
| 234 | error.BadHeader, | 241 | error.BadHeader, |
| 235 | testReader(&[_]u8{ 0x78, 0x9D }, ""), | 242 | testDecompress(&[_]u8{ 0x78, 0x9D }, ""), |
| 236 | ); | 243 | ); |
| 237 | // Wrong CM | 244 | // Wrong CM |
| 238 | try testing.expectError( | 245 | try testing.expectError( |
| 239 | error.InvalidCompression, | 246 | error.InvalidCompression, |
| 240 | testReader(&[_]u8{ 0x79, 0x94 }, ""), | 247 | testDecompress(&[_]u8{ 0x79, 0x94 }, ""), |
| 241 | ); | 248 | ); |
| 242 | // Wrong CINFO | 249 | // Wrong CINFO |
| 243 | try testing.expectError( | 250 | try testing.expectError( |
| 244 | error.InvalidWindowSize, | 251 | error.InvalidWindowSize, |
| 245 | testReader(&[_]u8{ 0x88, 0x98 }, ""), | 252 | testDecompress(&[_]u8{ 0x88, 0x98 }, ""), |
| 246 | ); | 253 | ); |
| 247 | // Wrong checksum | 254 | // Wrong checksum |
| 248 | try testing.expectError( | 255 | try testing.expectError( |
| 249 | error.WrongChecksum, | 256 | error.WrongChecksum, |
| 250 | testReader(&[_]u8{ 0x78, 0xda, 0x03, 0x00, 0x00, 0x00, 0x00, 0x00 }, ""), | 257 | testDecompress(&[_]u8{ 0x78, 0xda, 0x03, 0x00, 0x00, 0x00, 0x00, 0x00 }, ""), |
| 251 | ); | 258 | ); |
| 252 | // Truncated checksum | 259 | // Truncated checksum |
| 253 | try testing.expectError( | 260 | try testing.expectError( |
| 254 | error.EndOfStream, | 261 | error.EndOfStream, |
| 255 | testReader(&[_]u8{ 0x78, 0xda, 0x03, 0x00, 0x00 }, ""), | 262 | testDecompress(&[_]u8{ 0x78, 0xda, 0x03, 0x00, 0x00 }, ""), |
| 256 | ); | 263 | ); |
| 257 | } | 264 | } |
| 258 | | 265 | |
| ... | @@ -260,14 +267,16 @@ test "compress data" { | ... | @@ -260,14 +267,16 @@ test "compress data" { |
| 260 | const allocator = testing.allocator; | 267 | const allocator = testing.allocator; |
| 261 | const rfc1951_txt = @embedFile("testdata/rfc1951.txt"); | 268 | const rfc1951_txt = @embedFile("testdata/rfc1951.txt"); |
| 262 | | 269 | |
| 263 | var compressed_data = std.ArrayList(u8).init(allocator); | 270 | for (std.meta.tags(CompressionLevel)) |level| { |
| 264 | defer compressed_data.deinit(); | 271 | var compressed_data = std.ArrayList(u8).init(allocator); |
| | 272 | defer compressed_data.deinit(); |
| 265 | | 273 | |
| 266 | var compressor = try zlibStreamWriter(allocator, compressed_data.writer(), .{}); | 274 | var compressor = try compressStream(allocator, compressed_data.writer(), .{ .level = level }); |
| 267 | defer compressor.deinit(); | 275 | defer compressor.deinit(); |
| 268 | | 276 | |
| 269 | try compressor.writer().writeAll(rfc1951_txt); | 277 | try compressor.writer().writeAll(rfc1951_txt); |
| 270 | try compressor.close(); | 278 | try compressor.finish(); |
| 271 | | 279 | |
| 272 | try testReader(compressed_data.items, rfc1951_txt); | 280 | try testDecompress(compressed_data.items, rfc1951_txt); |
| | 281 | } |
| 273 | } | 282 | } |