authorgravatar for igor.anic@gmail.comIgor Anić <igor.anic@gmail.com> 2023-11-29 15:28:38+01:00
committergravatar for andrew@ziglang.orgAndrew Kelley <andrew@ziglang.org> 2024-01-13 19:37:33-07:00
loge1424b84b87903df265cc052f3dac17d1ec1c3be
tree1f19c18bb24edf23fe662706174bf179e57fe554
parent6d5283e83550998953f8784ba2b08a413a41baf2

tar: add parsing size in gnu extended format

Reference: https://www.gnu.org/software/tar/manual/html_node/Extensions.html#Extensions If the leading byte is 0x80 (128), the non-leading bytes of the field are concatenated in big-endian order, with the result being a positive number expressed in binary form.

1 files changed, 56 insertions(+), 19 deletions(-)

lib/std/tar.zig+56-19
......@@ -62,10 +62,10 @@ pub const Options = struct {
6262 };
6363};
6464
65const block_size = 512;
65const BLOCK_SIZE = 512;
6666
6767pub const Header = struct {
68 bytes: *const [block_size]u8,
68 bytes: *const [BLOCK_SIZE]u8,
6969
7070 pub const FileType = enum(u8) {
7171 normal_alias = 0,
......@@ -84,6 +84,19 @@ pub const Header = struct {
8484
8585 pub fn fileSize(header: Header) !u64 {
8686 const raw = header.bytes[124..][0..12];
87 // If the leading byte is 0xff (255), all the bytes of the field
88 // (including the leading byte) are concatenated in big-endian order,
89 // with the result being a negative number expressed in two’s
90 // complement form.
91 if (raw[0] == 0xff) return error.SizeNegative;
92 // If the leading byte is 0x80 (128), the non-leading bytes of the
93 // field are concatenated in big-endian order.
94 if (raw[0] == 0x80) {
95 if (raw[1] + raw[2] + raw[3] != 0) return error.SizeTooBig;
96 return std.mem.readInt(u64, raw[4..12], .big);
97 }
98 // Zero-filled octal number in ASCII. Each numeric field of width w
99 // contains w minus 1 digits, and a null
87100 const ltrimmed = std.mem.trimLeft(u8, raw, "0 ");
88101 const rtrimmed = std.mem.trimRight(u8, ltrimmed, " \x00");
89102 if (rtrimmed.len == 0) return 0;
......@@ -148,7 +161,7 @@ pub const Header = struct {
148161fn BufferedReader(comptime ReaderType: type) type {
149162 return struct {
150163 unbuffered_reader: ReaderType,
151 buffer: [block_size * 8]u8 = undefined,
164 buffer: [BLOCK_SIZE * 8]u8 = undefined,
152165 start: usize = 0,
153166 end: usize = 0,
154167
......@@ -164,14 +177,14 @@ fn BufferedReader(comptime ReaderType: type) type {
164177 }
165178
166179 pub fn readBlock(self: *Self) !?[]const u8 {
167 const block_bytes = try self.readChunk(block_size * 2);
180 const block_bytes = try self.readChunk(BLOCK_SIZE * 2);
168181 switch (block_bytes.len) {
169182 0 => return null,
170 1...(block_size - 1) => return error.UnexpectedEndOfStream,
183 1...(BLOCK_SIZE - 1) => return error.UnexpectedEndOfStream,
171184 else => {},
172185 }
173 self.advance(block_size);
174 return block_bytes[0..block_size];
186 self.advance(BLOCK_SIZE);
187 return block_bytes[0..BLOCK_SIZE];
175188 }
176189
177190 pub fn advance(self: *Self, count: usize) void {
......@@ -258,7 +271,7 @@ fn BufferedReader(comptime ReaderType: type) type {
258271 pub fn sliceReader(self: *Self, size: usize, auto_advance: bool) Self.SliceReader {
259272 return .{
260273 .size = size,
261 .chunk_size = roundedFileSize(size) + block_size,
274 .chunk_size = roundedFileSize(size) + BLOCK_SIZE,
262275 .offset = 0,
263276 .reader = self,
264277 .auto_advance = auto_advance,
......@@ -267,12 +280,12 @@ fn BufferedReader(comptime ReaderType: type) type {
267280 };
268281}
269282
270// file_size rouneded to te block boundary
283// File size rounded to te block boundary.
271284inline fn roundedFileSize(file_size: usize) usize {
272 return std.mem.alignForward(usize, file_size, block_size);
285 return std.mem.alignForward(usize, file_size, BLOCK_SIZE);
273286}
274287
275// number of padding bytes at the last file block
288// Number of padding bytes in the last file block.
276289inline fn filePadding(file_size: usize) usize {
277290 return roundedFileSize(file_size) - file_size;
278291}
......@@ -341,17 +354,18 @@ fn Iterator(comptime ReaderType: type) type {
341354 }
342355 };
343356
344 // Externally, Next iterates through the tar archive as if it is a series of
345 // files. Internally, the tar format often uses fake "files" to add meta
346 // data that describes the next file. These meta data "files" should not
347 // normally be visible to the outside. As such, this loop iterates through
348 // one or more "header files" until it finds a "normal file".
357 // Externally, `next` iterates through the tar archive as if it is a
358 // series of files. Internally, the tar format often uses fake "files"
359 // to add meta data that describes the next file. These meta data
360 // "files" should not normally be visible to the outside. As such, this
361 // loop iterates through one or more "header files" until it finds a
362 // "normal file".
349363 pub fn next(self: *Self) !?File {
350364 var file: File = .{ .reader = &self.reader };
351365 self.attrs.free();
352366
353367 while (try self.reader.readBlock()) |block_bytes| {
354 const block: Header = .{ .bytes = block_bytes[0..block_size] };
368 const block: Header = .{ .bytes = block_bytes[0..BLOCK_SIZE] };
355369 if (block.isZero()) return null;
356370 const file_type = block.fileType();
357371 const file_size = try block.fileSize();
......@@ -572,6 +586,7 @@ const TestCase = struct {
572586 size: usize = 0,
573587 link_name: []const u8 = empty_string,
574588 file_type: Header.FileType = .normal,
589 truncated: bool = false, // when there is no file body, just header, usefull for huge files
575590 };
576591
577592 path: []const u8,
......@@ -794,10 +809,32 @@ test "Go test cases" {
794809 },
795810 },
796811 },
812 .{
813 // Has size in gnu extended format. To represent size bigger than 8 GB.
814 .path = "writer-big.tar",
815 .files = &[_]TestCase.File{
816 .{
817 .name = "tmp/16gig.txt",
818 .size = 16 * 1024 * 1024 * 1024,
819 .truncated = true,
820 },
821 },
822 },
823 .{
824 // Size in gnu extended format, and name in pax attribute.
825 .path = "writer-big-long.tar",
826 .files = &[_]TestCase.File{
827 .{
828 .name = "longname/" ** 15 ++ "16gig.txt",
829 .size = 16 * 1024 * 1024 * 1024,
830 .truncated = true,
831 },
832 },
833 },
797834 };
798835
799836 for (cases) |case| {
800 // if (!std.mem.eql(u8, case.path, "pax.tar")) continue;
837 //if (!std.mem.eql(u8, case.path, "pax-pos-size-file.tar")) continue;
801838
802839 var fs_file = try test_dir.openFile(case.path, .{});
803840 defer fs_file.close();
......@@ -825,7 +862,7 @@ test "Go test cases" {
825862 // std.debug.print("actual chksum: {s}\n", .{std.fmt.fmtSliceHexLower(&actual_chksum)});
826863 try std.testing.expectEqualStrings(expected_chksum, &actual_chksum);
827864 } else {
828 try actual.skip(); // skip file content
865 if (!expected.truncated) try actual.skip(); // skip file content
829866 }
830867 i += 1;
831868 }