mirror/zig
1
0
Fork 0
mirror of https://github.com/ziglang/zig.git synced 2026-01-20 22:35:24 +00:00
Code Issues Packages Projects Releases Wiki Activity

std.tar: don't return in iterator init

Browse Source 
Don't assert min buffer size on iterator init. That was changing public
interface. This way we don't break that interface.
This commit is contained in:
Igor Anić 2024-03-02 10:59:25 +01:00
parent 10add7c677
commit 5ccbb196ad
2 changed files with 369 additions and 363 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

34
lib/std/tar.zig
View File

@ -115,9 +115,10 @@ pub const Header = struct {
/// Includes prefix concatenated, if any.
/// TODO: check against "../" and other nefarious things
pub fn fullName(header: Header, buffer: *[MAX_NAME_SIZE]u8) ![]const u8 {
pub fn fullName(header: Header, buffer: []u8) ![]const u8 {
const n = name(header);
const p = prefix(header);
if (buffer.len < n.len + p.len + 1) return error.TarInsufficientBuffer;
if (!is_ustar(header) or p.len == 0) {
@memcpy(buffer[0..n.len], n);
return buffer[0..n.len];
@ -130,11 +131,12 @@ pub const Header = struct {
/// When kind is symbolic_link linked-to name (target_path) is specified in
/// the linkname field.
pub fn linkName(header: Header, buffer: *[LINK_NAME_SIZE]u8) []const u8 {
pub fn linkName(header: Header, buffer: []u8) ![]const u8 {
const link_name = header.str(157, 100);
if (link_name.len == 0) {
return buffer[0..0];
}
if (buffer.len < link_name.len) return error.TarInsufficientBuffer;
const buf = buffer[0..link_name.len];
@memcpy(buf, link_name);
return buf;
@ -248,13 +250,7 @@ pub const IteratorOptions = struct {
/// Iterates over files in tar archive.
/// `next` returns each file in `reader` tar archive.
/// Provided buffers should be at least 256 bytes for file_name and 100 bytes
/// for link_name.
pub fn iterator(reader: anytype, options: IteratorOptions) !Iterator(@TypeOf(reader)) {
if (options.file_name_buffer.len < Header.MAX_NAME_SIZE or
options.link_name_buffer.len < Header.LINK_NAME_SIZE)
return error.TarInsufficientBuffer;
pub fn iterator(reader: anytype, options: IteratorOptions) Iterator(@TypeOf(reader)) {
return .{
.reader = reader,
.diagnostics = options.diagnostics,
@ -372,10 +368,10 @@ fn Iterator(comptime ReaderType: type) type {
self.file.size = size;
}
if (self.file.link_name.len == 0) {
self.file.link_name = header.linkName(self.link_name_buffer[0..Header.LINK_NAME_SIZE]);
self.file.link_name = try header.linkName(self.link_name_buffer);
}
if (self.file.name.len == 0) {
self.file.name = try header.fullName(self.file_name_buffer[0..Header.MAX_NAME_SIZE]);
self.file.name = try header.fullName(self.file_name_buffer);
}
self.padding = blockPadding(self.file.size);
@ -565,7 +561,7 @@ pub fn pipeToFileSystem(dir: std.fs.Dir, reader: anytype, options: Options) !voi
var file_name_buffer: [std.fs.MAX_PATH_BYTES]u8 = undefined;
var link_name_buffer: [std.fs.MAX_PATH_BYTES]u8 = undefined;
var iter = try iterator(reader, .{
var iter = iterator(reader, .{
.file_name_buffer = &file_name_buffer,
.link_name_buffer = &link_name_buffer,
.diagnostics = options.diagnostics,
@ -871,17 +867,3 @@ test "create file and symlink" {
_ = try createDirAndSymlink(root.dir, "../../../g/h/i/file4", "j/k/l/symlink3");
_ = try createDirAndFile(root.dir, "g/h/i/file4");
}
test "insufficient buffer for iterator" {
var file_name_buffer: [10]u8 = undefined;
var link_name_buffer: [10]u8 = undefined;
var fsb = std.io.fixedBufferStream("");
try std.testing.expectError(
error.TarInsufficientBuffer,
iterator(fsb.reader(), .{
.file_name_buffer = &file_name_buffer,
.link_name_buffer = &link_name_buffer,
}),
);
}

698
lib/std/tar/test.zig
View File

@ -2,333 +2,352 @@ const std = @import("std");
const tar = @import("../tar.zig");
const testing = std.testing;
const Case = struct {
const File = struct {
name: []const u8,
size: u64 = 0,
mode: u32 = 0,
link_name: []const u8 = &[0]u8{},
kind: tar.Header.Kind = .normal,
truncated: bool = false, // when there is no file body, just header, usefull for huge files
};
data: []const u8, // testdata file content
files: []const File = &[_]@This().File{}, // expected files to found in archive
chksums: []const []const u8 = &[_][]const u8{}, // chksums of each file content
err: ?anyerror = null, // parsing should fail with this error
};
const cases = [_]Case{
.{
.data = @embedFile("testdata/gnu.tar"),
.files = &[_]Case.File{
.{
.name = "small.txt",
.size = 5,
.mode = 0o640,
},
.{
.name = "small2.txt",
.size = 11,
.mode = 0o640,
},
},
.chksums = &[_][]const u8{
"e38b27eaccb4391bdec553a7f3ae6b2f",
"c65bd2e50a56a2138bf1716f2fd56fe9",
},
},
.{
.data = @embedFile("testdata/sparse-formats.tar"),
.err = error.TarUnsupportedHeader,
},
.{
.data = @embedFile("testdata/star.tar"),
.files = &[_]Case.File{
.{
.name = "small.txt",
.size = 5,
.mode = 0o640,
},
.{
.name = "small2.txt",
.size = 11,
.mode = 0o640,
},
},
.chksums = &[_][]const u8{
"e38b27eaccb4391bdec553a7f3ae6b2f",
"c65bd2e50a56a2138bf1716f2fd56fe9",
},
},
.{
.data = @embedFile("testdata/v7.tar"),
.files = &[_]Case.File{
.{
.name = "small.txt",
.size = 5,
.mode = 0o444,
},
.{
.name = "small2.txt",
.size = 11,
.mode = 0o444,
},
},
.chksums = &[_][]const u8{
"e38b27eaccb4391bdec553a7f3ae6b2f",
"c65bd2e50a56a2138bf1716f2fd56fe9",
},
},
.{
.data = @embedFile("testdata/pax.tar"),
.files = &[_]Case.File{
.{
.name = "a/123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100",
.size = 7,
.mode = 0o664,
},
.{
.name = "a/b",
.size = 0,
.kind = .symbolic_link,
.mode = 0o777,
.link_name = "123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100",
},
},
.chksums = &[_][]const u8{
"3c382e8f5b6631aa2db52643912ffd4a",
},
},
.{
// pax attribute don't end with \n
.data = @embedFile("testdata/pax-bad-hdr-file.tar"),
.err = error.PaxInvalidAttributeEnd,
},
.{
// size is in pax attribute
.data = @embedFile("testdata/pax-pos-size-file.tar"),
.files = &[_]Case.File{
.{
.name = "foo",
.size = 999,
.kind = .normal,
.mode = 0o640,
},
},
.chksums = &[_][]const u8{
"0afb597b283fe61b5d4879669a350556",
},
},
.{
// has pax records which we are not interested in
.data = @embedFile("testdata/pax-records.tar"),
.files = &[_]Case.File{
.{
.name = "file",
},
},
},
.{
// has global records which we are ignoring
.data = @embedFile("testdata/pax-global-records.tar"),
.files = &[_]Case.File{
.{
.name = "file1",
},
.{
.name = "file2",
},
.{
.name = "file3",
},
.{
.name = "file4",
},
},
},
.{
.data = @embedFile("testdata/nil-uid.tar"),
.files = &[_]Case.File{
.{
.name = "P1050238.JPG.log",
.size = 14,
.kind = .normal,
.mode = 0o664,
},
},
.chksums = &[_][]const u8{
"08d504674115e77a67244beac19668f5",
},
},
.{
// has xattrs and pax records which we are ignoring
.data = @embedFile("testdata/xattrs.tar"),
.files = &[_]Case.File{
.{
.name = "small.txt",
.size = 5,
.kind = .normal,
.mode = 0o644,
},
.{
.name = "small2.txt",
.size = 11,
.kind = .normal,
.mode = 0o644,
},
},
.chksums = &[_][]const u8{
"e38b27eaccb4391bdec553a7f3ae6b2f",
"c65bd2e50a56a2138bf1716f2fd56fe9",
},
},
.{
.data = @embedFile("testdata/gnu-multi-hdrs.tar"),
.files = &[_]Case.File{
.{
.name = "GNU2/GNU2/long-path-name",
.link_name = "GNU4/GNU4/long-linkpath-name",
.kind = .symbolic_link,
},
},
},
.{
// has gnu type D (directory) and S (sparse) blocks
.data = @embedFile("testdata/gnu-incremental.tar"),
.err = error.TarUnsupportedHeader,
},
.{
// should use values only from last pax header
.data = @embedFile("testdata/pax-multi-hdrs.tar"),
.files = &[_]Case.File{
.{
.name = "bar",
.link_name = "PAX4/PAX4/long-linkpath-name",
.kind = .symbolic_link,
},
},
},
.{
.data = @embedFile("testdata/gnu-long-nul.tar"),
.files = &[_]Case.File{
.{
.name = "0123456789",
.mode = 0o644,
},
},
},
.{
.data = @embedFile("testdata/gnu-utf8.tar"),
.files = &[_]Case.File{
.{
.name = "☺☻☹☺☻☹☺☻☹☺☻☹☺☻☹☺☻☹☺☻☹☺☻☹☺☻☹☺☻☹☺☻☹☺☻☹☺☻☹☺☻☹☺☻☹☺☻☹☺☻☹☺☻☹",
.mode = 0o644,
},
},
},
.{
.data = @embedFile("testdata/gnu-not-utf8.tar"),
.files = &[_]Case.File{
.{
.name = "hi\x80\x81\x82\x83bye",
.mode = 0o644,
},
},
},
.{
// null in pax key
.data = @embedFile("testdata/pax-nul-xattrs.tar"),
.err = error.PaxNullInKeyword,
},
.{
.data = @embedFile("testdata/pax-nul-path.tar"),
.err = error.PaxNullInValue,
},
.{
.data = @embedFile("testdata/neg-size.tar"),
.err = error.TarHeader,
},
.{
.data = @embedFile("testdata/issue10968.tar"),
.err = error.TarHeader,
},
.{
.data = @embedFile("testdata/issue11169.tar"),
.err = error.TarHeader,
},
.{
.data = @embedFile("testdata/issue12435.tar"),
.err = error.TarHeaderChksum,
},
.{
// has magic with space at end instead of null
.data = @embedFile("testdata/invalid-go17.tar"),
.files = &[_]Case.File{
.{
.name = "aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa/foo",
},
},
},
.{
.data = @embedFile("testdata/ustar-file-devs.tar"),
.files = &[_]Case.File{
.{
.name = "file",
.mode = 0o644,
},
},
},
.{
.data = @embedFile("testdata/trailing-slash.tar"),
.files = &[_]Case.File{
.{
.name = "123456789/" ** 30,
.kind = .directory,
},
},
},
.{
// Has size in gnu extended format. To represent size bigger than 8 GB.
.data = @embedFile("testdata/writer-big.tar"),
.files = &[_]Case.File{
.{
.name = "tmp/16gig.txt",
.size = 16 * 1024 * 1024 * 1024,
.truncated = true,
.mode = 0o640,
},
},
},
.{
// Size in gnu extended format, and name in pax attribute.
.data = @embedFile("testdata/writer-big-long.tar"),
.files = &[_]Case.File{
.{
.name = "longname/" ** 15 ++ "16gig.txt",
.size = 16 * 1024 * 1024 * 1024,
.mode = 0o644,
.truncated = true,
},
},
},
.{
.data = @embedFile("testdata/fuzz1.tar"),
.err = error.TarInsufficientBuffer,
},
.{
.data = @embedFile("testdata/fuzz2.tar"),
.err = error.PaxSizeAttrOverflow,
},
};
// used in test to calculate file chksum
const Md5Writer = struct {
h: std.crypto.hash.Md5 = std.crypto.hash.Md5.init(.{}),
pub fn writeAll(self: *Md5Writer, buf: []const u8) !void {
self.h.update(buf);
}
pub fn writeByte(self: *Md5Writer, byte: u8) !void {
self.h.update(&[_]u8{byte});
}
pub fn chksum(self: *Md5Writer) [32]u8 {
var s = [_]u8{0} ** 16;
self.h.final(&s);
return std.fmt.bytesToHex(s, .lower);
}
};
test "run test cases" {
const Case = struct {
const File = struct {
name: []const u8,
size: u64 = 0,
mode: u32 = 0,
link_name: []const u8 = &[0]u8{},
kind: tar.Header.Kind = .normal,
truncated: bool = false, // when there is no file body, just header, usefull for huge files
};
data: []const u8, // testdata file content
files: []const File = &[_]@This().File{}, // expected files to found in archive
chksums: []const []const u8 = &[_][]const u8{}, // chksums of each file content
err: ?anyerror = null, // parsing should fail with this error
};
const cases = [_]Case{
.{
.data = @embedFile("testdata/gnu.tar"),
.files = &[_]Case.File{
.{
.name = "small.txt",
.size = 5,
.mode = 0o640,
},
.{
.name = "small2.txt",
.size = 11,
.mode = 0o640,
},
},
.chksums = &[_][]const u8{
"e38b27eaccb4391bdec553a7f3ae6b2f",
"c65bd2e50a56a2138bf1716f2fd56fe9",
},
},
.{
.data = @embedFile("testdata/sparse-formats.tar"),
.err = error.TarUnsupportedHeader,
},
.{
.data = @embedFile("testdata/star.tar"),
.files = &[_]Case.File{
.{
.name = "small.txt",
.size = 5,
.mode = 0o640,
},
.{
.name = "small2.txt",
.size = 11,
.mode = 0o640,
},
},
.chksums = &[_][]const u8{
"e38b27eaccb4391bdec553a7f3ae6b2f",
"c65bd2e50a56a2138bf1716f2fd56fe9",
},
},
.{
.data = @embedFile("testdata/v7.tar"),
.files = &[_]Case.File{
.{
.name = "small.txt",
.size = 5,
.mode = 0o444,
},
.{
.name = "small2.txt",
.size = 11,
.mode = 0o444,
},
},
.chksums = &[_][]const u8{
"e38b27eaccb4391bdec553a7f3ae6b2f",
"c65bd2e50a56a2138bf1716f2fd56fe9",
},
},
.{
.data = @embedFile("testdata/pax.tar"),
.files = &[_]Case.File{
.{
.name = "a/123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100",
.size = 7,
.mode = 0o664,
},
.{
.name = "a/b",
.size = 0,
.kind = .symbolic_link,
.mode = 0o777,
.link_name = "123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100",
},
},
.chksums = &[_][]const u8{
"3c382e8f5b6631aa2db52643912ffd4a",
},
},
.{
// pax attribute don't end with \n
.data = @embedFile("testdata/pax-bad-hdr-file.tar"),
.err = error.PaxInvalidAttributeEnd,
},
.{
// size is in pax attribute
.data = @embedFile("testdata/pax-pos-size-file.tar"),
.files = &[_]Case.File{
.{
.name = "foo",
.size = 999,
.kind = .normal,
.mode = 0o640,
},
},
.chksums = &[_][]const u8{
"0afb597b283fe61b5d4879669a350556",
},
},
.{
// has pax records which we are not interested in
.data = @embedFile("testdata/pax-records.tar"),
.files = &[_]Case.File{
.{
.name = "file",
},
},
},
.{
// has global records which we are ignoring
.data = @embedFile("testdata/pax-global-records.tar"),
.files = &[_]Case.File{
.{
.name = "file1",
},
.{
.name = "file2",
},
.{
.name = "file3",
},
.{
.name = "file4",
},
},
},
.{
.data = @embedFile("testdata/nil-uid.tar"),
.files = &[_]Case.File{
.{
.name = "P1050238.JPG.log",
.size = 14,
.kind = .normal,
.mode = 0o664,
},
},
.chksums = &[_][]const u8{
"08d504674115e77a67244beac19668f5",
},
},
.{
// has xattrs and pax records which we are ignoring
.data = @embedFile("testdata/xattrs.tar"),
.files = &[_]Case.File{
.{
.name = "small.txt",
.size = 5,
.kind = .normal,
.mode = 0o644,
},
.{
.name = "small2.txt",
.size = 11,
.kind = .normal,
.mode = 0o644,
},
},
.chksums = &[_][]const u8{
"e38b27eaccb4391bdec553a7f3ae6b2f",
"c65bd2e50a56a2138bf1716f2fd56fe9",
},
},
.{
.data = @embedFile("testdata/gnu-multi-hdrs.tar"),
.files = &[_]Case.File{
.{
.name = "GNU2/GNU2/long-path-name",
.link_name = "GNU4/GNU4/long-linkpath-name",
.kind = .symbolic_link,
},
},
},
.{
// has gnu type D (directory) and S (sparse) blocks
.data = @embedFile("testdata/gnu-incremental.tar"),
.err = error.TarUnsupportedHeader,
},
.{
// should use values only from last pax header
.data = @embedFile("testdata/pax-multi-hdrs.tar"),
.files = &[_]Case.File{
.{
.name = "bar",
.link_name = "PAX4/PAX4/long-linkpath-name",
.kind = .symbolic_link,
},
},
},
.{
.data = @embedFile("testdata/gnu-long-nul.tar"),
.files = &[_]Case.File{
.{
.name = "0123456789",
.mode = 0o644,
},
},
},
.{
.data = @embedFile("testdata/gnu-utf8.tar"),
.files = &[_]Case.File{
.{
.name = "☺☻☹☺☻☹☺☻☹☺☻☹☺☻☹☺☻☹☺☻☹☺☻☹☺☻☹☺☻☹☺☻☹☺☻☹☺☻☹☺☻☹☺☻☹☺☻☹☺☻☹☺☻☹",
.mode = 0o644,
},
},
},
.{
.data = @embedFile("testdata/gnu-not-utf8.tar"),
.files = &[_]Case.File{
.{
.name = "hi\x80\x81\x82\x83bye",
.mode = 0o644,
},
},
},
.{
// null in pax key
.data = @embedFile("testdata/pax-nul-xattrs.tar"),
.err = error.PaxNullInKeyword,
},
.{
.data = @embedFile("testdata/pax-nul-path.tar"),
.err = error.PaxNullInValue,
},
.{
.data = @embedFile("testdata/neg-size.tar"),
.err = error.TarHeader,
},
.{
.data = @embedFile("testdata/issue10968.tar"),
.err = error.TarHeader,
},
.{
.data = @embedFile("testdata/issue11169.tar"),
.err = error.TarHeader,
},
.{
.data = @embedFile("testdata/issue12435.tar"),
.err = error.TarHeaderChksum,
},
.{
// has magic with space at end instead of null
.data = @embedFile("testdata/invalid-go17.tar"),
.files = &[_]Case.File{
.{
.name = "aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa/foo",
},
},
},
.{
.data = @embedFile("testdata/ustar-file-devs.tar"),
.files = &[_]Case.File{
.{
.name = "file",
.mode = 0o644,
},
},
},
.{
.data = @embedFile("testdata/trailing-slash.tar"),
.files = &[_]Case.File{
.{
.name = "123456789/" ** 30,
.kind = .directory,
},
},
},
.{
// Has size in gnu extended format. To represent size bigger than 8 GB.
.data = @embedFile("testdata/writer-big.tar"),
.files = &[_]Case.File{
.{
.name = "tmp/16gig.txt",
.size = 16 * 1024 * 1024 * 1024,
.truncated = true,
.mode = 0o640,
},
},
},
.{
// Size in gnu extended format, and name in pax attribute.
.data = @embedFile("testdata/writer-big-long.tar"),
.files = &[_]Case.File{
.{
.name = "longname/" ** 15 ++ "16gig.txt",
.size = 16 * 1024 * 1024 * 1024,
.mode = 0o644,
.truncated = true,
},
},
},
.{
.data = @embedFile("testdata/fuzz1.tar"),
.err = error.TarInsufficientBuffer,
},
.{
.data = @embedFile("testdata/fuzz2.tar"),
.err = error.PaxSizeAttrOverflow,
},
};
var file_name_buffer: [std.fs.MAX_PATH_BYTES]u8 = undefined;
var link_name_buffer: [std.fs.MAX_PATH_BYTES]u8 = undefined;
for (cases) |case| {
var fsb = std.io.fixedBufferStream(case.data);
var iter = try tar.iterator(fsb.reader(), .{
var iter = tar.iterator(fsb.reader(), .{
.file_name_buffer = &file_name_buffer,
.link_name_buffer = &link_name_buffer,
});
@ -359,6 +378,10 @@ test "run test cases" {
}
try testing.expectEqual(case.files.len, i);
}
}
test "pax/gnu long names with small buffer" {
// should fail with insufficient buffer error
var min_file_name_buffer: [tar.Header.MAX_NAME_SIZE]u8 = undefined;
var min_link_name_buffer: [tar.Header.LINK_NAME_SIZE]u8 = undefined;
@ -366,7 +389,7 @@ test "run test cases" {
for (long_name_cases) |case| {
var fsb = std.io.fixedBufferStream(case.data);
var iter = try tar.iterator(fsb.reader(), .{
var iter = tar.iterator(fsb.reader(), .{
.file_name_buffer = &min_file_name_buffer,
.link_name_buffer = &min_link_name_buffer,
});
@ -382,24 +405,25 @@ test "run test cases" {
}
}
// used in test to calculate file chksum
const Md5Writer = struct {
h: std.crypto.hash.Md5 = std.crypto.hash.Md5.init(.{}),
test "insufficient buffer in Header name filed" {
var min_file_name_buffer: [9]u8 = undefined;
var min_link_name_buffer: [tar.Header.LINK_NAME_SIZE]u8 = undefined;
pub fn writeAll(self: *Md5Writer, buf: []const u8) !void {
self.h.update(buf);
}
var fsb = std.io.fixedBufferStream(cases[0].data);
var iter = tar.iterator(fsb.reader(), .{
.file_name_buffer = &min_file_name_buffer,
.link_name_buffer = &min_link_name_buffer,
});
pub fn writeByte(self: *Md5Writer, byte: u8) !void {
self.h.update(&[_]u8{byte});
}
var iter_err: ?anyerror = null;
while (iter.next() catch |err| brk: {
iter_err = err;
break :brk null;
}) |_| {}
pub fn chksum(self: *Md5Writer) [32]u8 {
var s = [_]u8{0} ** 16;
self.h.final(&s);
return std.fmt.bytesToHex(s, .lower);
}
};
try testing.expect(iter_err != null);
try testing.expectEqual(error.TarInsufficientBuffer, iter_err.?);
}
test "should not overwrite existing file" {
// Starting from this folder structure: