std.tar: copy another project's full test suite and make them all pass

[andrewrk]

The implementation is the easy part!

Perhaps https://cs.opensource.google/go/go/+/refs/tags/go1.19.5:src/archive/tar/

Perhaps https://github.com/python/cpython/blob/3.11/Lib/test/test_tarfile.py

Make sure the license is MIT-compatible so we can copy the test data.

Depending on how bulky the test data is, maybe this would be better to go into test/standalone rather than lib/std/

I'm marking this as "bug" since it's extremely likely that adding these tests will find bugs.

Bonus: write a handy benchmark script that we can use to find out if any implementation changes to std.tar improved performance.

Please enjoy this inspirational material, while searching for some other project's data to plunder.

[nektro]

The Go team has an extensive set of tar test data and a harness that shows how its used, and is BSD-3 licensed

https://go.dev/src/archive/tar/testdata/

https://go.dev/src/archive/tar/tar_test.go
https://go.dev/src/archive/tar/fuzz_test.go
https://go.dev/src/archive/tar/reader_test.go
https://go.dev/src/archive/tar/writer_test.go

[jhuntwork]

Just to offer some feedback and testing related to archives created with libarchive's bsdtar...

Using this sample zig code:

const std = @import("std");

pub fn main() !void {
    var file = try std.fs.cwd().openFile("test.tar", .{});
    defer file.close();
    const outdir = try std.fs.cwd().openDir(
        "out",
        .{},
    );

    try std.tar.pipeToFileSystem(
        outdir,
        file.reader(),
        .{ .strip_components = 1, .mode_mode = .ignore },
    );
}

If test.tar is created with the tar command from busybox, the archive can be extracted successfully. When created with bsdtar with the same content, I receive the following error:

$ ./zig-out/bin/mere-zig
error: InvalidCharacter
/usr/lib/zig/std/fmt.zig:1930:17: 0x232a77 in charToDigit (mere-zig)
        else => return error.InvalidCharacter,
                ^
/usr/lib/zig/std/fmt.zig:1858:23: 0x2124ea in parseWithSign__anon_4202 (mere-zig)
        const digit = try charToDigit(c, buf_radix);
                      ^
/usr/lib/zig/std/fmt.zig:1757:5: 0x20c3bc in parseInt__anon_3712 (mere-zig)
    return parseWithSign(T, buf, radix, .Pos);
    ^
/usr/lib/zig/std/tar.zig:40:9: 0x20bc80 in fileSize (mere-zig)
        return std.fmt.parseInt(u64, rtrimmed, 8);
        ^
/usr/lib/zig/std/tar.zig:117:27: 0x20d7af in pipeToFileSystem__anon_3316 (mere-zig)
        const file_size = try header.fileSize();
                          ^
/home/jeremy/mere-zig/src/main.zig:11:5: 0x20ec1c in main (mere-zig)
    try std.tar.pipeToFileSystem(
    ^

busybox tar:

$ tar -cf test.tar zig-out
$ file test.tar
test.tar: POSIX tar archive (GNU)

libarchive bsdtar:

$ bsdtar -cf test.tar zig-out
$ file test.tar
test.tar: POSIX tar archive