1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703 704 705 706 707 708 709 710 711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746 747 748 749 750 751 752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777 778 779 780 781 782 783 784 785 786 787 788 789 790 791 792 793 794 795 796 797 798 799 800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821 822 823 824 825 826 827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843 844 845 846 847 848 849 850 851 852 853 854 855 856 857 858 859 860 861 862 863 864 865 866 867 868 869 870 871 872 873 874 875 876 877 878 879 880 881 882 883 884 885 886 887 888 889 890 891 892 893 894 895 896 897 898 899 900 901 902 903 904 905 906 907 908 909 910 911 912 913 914 915 916 917 918 919 920 921 922 923 924 925 926 927 928 929 930 931 932 933 934 935 936 937 938 939 940 941 942 943 944 945 946 947 948 949 950 951 952 953 954 955 956 957 958 959 960 961 962 963 964 965 966 967 968 969 970 971 972 973 974 975 976 977 978 979 980 981 982 983 984 985 986 987 988 989 990 991 992 993 994 995 996 997 998 999 1000 1001 1002 1003 1004 1005 1006 1007
use std::borrow::Cow;
use std::cmp;
use std::fs;
use std::fs::OpenOptions;
use std::io::prelude::*;
use std::io::{self, Error, ErrorKind, SeekFrom};
use std::marker;
use std::path::{Component, Path, PathBuf};
use filetime::{self, FileTime};
use crate::archive::ArchiveInner;
use crate::error::TarError;
use crate::header::bytes2path;
use crate::other;
use crate::{Archive, Header, PaxExtensions};
/// A read-only view into an entry of an archive.
///
/// This structure is a window into a portion of a borrowed archive which can
/// be inspected. It acts as a file handle by implementing the Reader trait. An
/// entry cannot be rewritten once inserted into an archive.
pub struct Entry<'a, R: 'a + Read> {
fields: EntryFields<'a>,
_ignored: marker::PhantomData<&'a Archive<R>>,
}
// private implementation detail of `Entry`, but concrete (no type parameters)
// and also all-public to be constructed from other modules.
pub struct EntryFields<'a> {
pub long_pathname: Option<Vec<u8>>,
pub long_linkname: Option<Vec<u8>>,
pub pax_extensions: Option<Vec<u8>>,
pub mask: u32,
pub header: Header,
pub size: u64,
pub header_pos: u64,
pub file_pos: u64,
pub data: Vec<EntryIo<'a>>,
pub unpack_xattrs: bool,
pub preserve_permissions: bool,
pub preserve_ownerships: bool,
pub preserve_mtime: bool,
pub overwrite: bool,
}
pub enum EntryIo<'a> {
Pad(io::Take<io::Repeat>),
Data(io::Take<&'a ArchiveInner<dyn Read + 'a>>),
}
/// When unpacking items the unpacked thing is returned to allow custom
/// additional handling by users. Today the File is returned, in future
/// the enum may be extended with kinds for links, directories etc.
#[derive(Debug)]
pub enum Unpacked {
/// A file was unpacked.
File(std::fs::File),
/// A directory, hardlink, symlink, or other node was unpacked.
#[doc(hidden)]
__Nonexhaustive,
}
impl<'a, R: Read> Entry<'a, R> {
/// Returns the path name for this entry.
///
/// This method may fail if the pathname is not valid Unicode and this is
/// called on a Windows platform.
///
/// Note that this function will convert any `\` characters to directory
/// separators, and it will not always return the same value as
/// `self.header().path()` as some archive formats have support for longer
/// path names described in separate entries.
///
/// It is recommended to use this method instead of inspecting the `header`
/// directly to ensure that various archive formats are handled correctly.
pub fn path(&self) -> io::Result<Cow<Path>> {
self.fields.path()
}
/// Returns the raw bytes listed for this entry.
///
/// Note that this function will convert any `\` characters to directory
/// separators, and it will not always return the same value as
/// `self.header().path_bytes()` as some archive formats have support for
/// longer path names described in separate entries.
pub fn path_bytes(&self) -> Cow<[u8]> {
self.fields.path_bytes()
}
/// Returns the link name for this entry, if any is found.
///
/// This method may fail if the pathname is not valid Unicode and this is
/// called on a Windows platform. `Ok(None)` being returned, however,
/// indicates that the link name was not present.
///
/// Note that this function will convert any `\` characters to directory
/// separators, and it will not always return the same value as
/// `self.header().link_name()` as some archive formats have support for
/// longer path names described in separate entries.
///
/// It is recommended to use this method instead of inspecting the `header`
/// directly to ensure that various archive formats are handled correctly.
pub fn link_name(&self) -> io::Result<Option<Cow<Path>>> {
self.fields.link_name()
}
/// Returns the link name for this entry, in bytes, if listed.
///
/// Note that this will not always return the same value as
/// `self.header().link_name_bytes()` as some archive formats have support for
/// longer path names described in separate entries.
pub fn link_name_bytes(&self) -> Option<Cow<[u8]>> {
self.fields.link_name_bytes()
}
/// Returns an iterator over the pax extensions contained in this entry.
///
/// Pax extensions are a form of archive where extra metadata is stored in
/// key/value pairs in entries before the entry they're intended to
/// describe. For example this can be used to describe long file name or
/// other metadata like atime/ctime/mtime in more precision.
///
/// The returned iterator will yield key/value pairs for each extension.
///
/// `None` will be returned if this entry does not indicate that it itself
/// contains extensions, or if there were no previous extensions describing
/// it.
///
/// Note that global pax extensions are intended to be applied to all
/// archive entries.
///
/// Also note that this function will read the entire entry if the entry
/// itself is a list of extensions.
pub fn pax_extensions(&mut self) -> io::Result<Option<PaxExtensions>> {
self.fields.pax_extensions()
}
/// Returns access to the header of this entry in the archive.
///
/// This provides access to the metadata for this entry in the archive.
pub fn header(&self) -> &Header {
&self.fields.header
}
/// Returns access to the size of this entry in the archive.
///
/// In the event the size is stored in a pax extension, that size value
/// will be referenced. Otherwise, the entry size will be stored in the header.
pub fn size(&self) -> u64 {
self.fields.size
}
/// Returns the starting position, in bytes, of the header of this entry in
/// the archive.
///
/// The header is always a contiguous section of 512 bytes, so if the
/// underlying reader implements `Seek`, then the slice from `header_pos` to
/// `header_pos + 512` contains the raw header bytes.
pub fn raw_header_position(&self) -> u64 {
self.fields.header_pos
}
/// Returns the starting position, in bytes, of the file of this entry in
/// the archive.
///
/// If the file of this entry is continuous (e.g. not a sparse file), and
/// if the underlying reader implements `Seek`, then the slice from
/// `file_pos` to `file_pos + entry_size` contains the raw file bytes.
pub fn raw_file_position(&self) -> u64 {
self.fields.file_pos
}
/// Writes this file to the specified location.
///
/// This function will write the entire contents of this file into the
/// location specified by `dst`. Metadata will also be propagated to the
/// path `dst`.
///
/// This function will create a file at the path `dst`, and it is required
/// that the intermediate directories are created. Any existing file at the
/// location `dst` will be overwritten.
///
/// > **Note**: This function does not have as many sanity checks as
/// > `Archive::unpack` or `Entry::unpack_in`. As a result if you're
/// > thinking of unpacking untrusted tarballs you may want to review the
/// > implementations of the previous two functions and perhaps implement
/// > similar logic yourself.
///
/// # Examples
///
/// ```no_run
/// use std::fs::File;
/// use binstall_tar::Archive;
///
/// let mut ar = Archive::new(File::open("foo.tar").unwrap());
///
/// for (i, file) in ar.entries().unwrap().enumerate() {
/// let mut file = file.unwrap();
/// file.unpack(format!("file-{}", i)).unwrap();
/// }
/// ```
pub fn unpack<P: AsRef<Path>>(&mut self, dst: P) -> io::Result<Unpacked> {
self.fields.unpack(None, dst.as_ref())
}
/// Extracts this file under the specified path, avoiding security issues.
///
/// This function will write the entire contents of this file into the
/// location obtained by appending the path of this file in the archive to
/// `dst`, creating any intermediate directories if needed. Metadata will
/// also be propagated to the path `dst`. Any existing file at the location
/// `dst` will be overwritten.
///
/// This function carefully avoids writing outside of `dst`. If the file has
/// a '..' in its path, this function will skip it and return false.
///
/// # Examples
///
/// ```no_run
/// use std::fs::File;
/// use binstall_tar::Archive;
///
/// let mut ar = Archive::new(File::open("foo.tar").unwrap());
///
/// for (i, file) in ar.entries().unwrap().enumerate() {
/// let mut file = file.unwrap();
/// file.unpack_in("target").unwrap();
/// }
/// ```
pub fn unpack_in<P: AsRef<Path>>(&mut self, dst: P) -> io::Result<bool> {
self.fields.unpack_in(dst.as_ref())
}
/// Set the mask of the permission bits when unpacking this entry.
///
/// The mask will be inverted when applying against a mode, similar to how
/// `umask` works on Unix. In logical notation it looks like:
///
/// ```text
/// new_mode = old_mode & (~mask)
/// ```
///
/// The mask is 0 by default and is currently only implemented on Unix.
pub fn set_mask(&mut self, mask: u32) {
self.fields.mask = mask;
}
/// Indicate whether extended file attributes (xattrs on Unix) are preserved
/// when unpacking this entry.
///
/// This flag is disabled by default and is currently only implemented on
/// Unix using xattr support. This may eventually be implemented for
/// Windows, however, if other archive implementations are found which do
/// this as well.
pub fn set_unpack_xattrs(&mut self, unpack_xattrs: bool) {
self.fields.unpack_xattrs = unpack_xattrs;
}
/// Indicate whether extended permissions (like suid on Unix) are preserved
/// when unpacking this entry.
///
/// This flag is disabled by default and is currently only implemented on
/// Unix.
pub fn set_preserve_permissions(&mut self, preserve: bool) {
self.fields.preserve_permissions = preserve;
}
/// Indicate whether access time information is preserved when unpacking
/// this entry.
///
/// This flag is enabled by default.
pub fn set_preserve_mtime(&mut self, preserve: bool) {
self.fields.preserve_mtime = preserve;
}
}
impl<'a, R: Read> Read for Entry<'a, R> {
fn read(&mut self, into: &mut [u8]) -> io::Result<usize> {
self.fields.read(into)
}
}
impl<'a> EntryFields<'a> {
pub fn from<R: Read>(entry: Entry<R>) -> EntryFields {
entry.fields
}
pub fn into_entry<R: Read>(self) -> Entry<'a, R> {
Entry {
fields: self,
_ignored: marker::PhantomData,
}
}
pub fn read_all(&mut self) -> io::Result<Vec<u8>> {
// Preallocate some data but don't let ourselves get too crazy now.
let cap = cmp::min(self.size, 128 * 1024);
let mut v = Vec::with_capacity(cap as usize);
self.read_to_end(&mut v).map(|_| v)
}
pub fn is_pax_sparse(&mut self) -> bool {
if let Some(ref pax) = self.pax_extensions {
let mut extensions = PaxExtensions::new(pax).filter_map(|f| f.ok());
return extensions
.find(|f| f.key_bytes() == b"GNU.sparse.major" && f.value_bytes() == b"1")
.is_some()
&& extensions
.find(|f| f.key_bytes() == b"GNU.sparse.minor" && f.value_bytes() == b"0")
.is_some();
}
false
}
pub fn pax_sparse_name(&mut self) -> Option<Vec<u8>> {
if let Some(ref pax) = self.pax_extensions {
return PaxExtensions::new(pax)
.filter_map(|f| f.ok())
.find(|f| f.key_bytes() == b"GNU.sparse.name")
.map(|f| f.value_bytes().to_vec());
}
None
}
pub fn pax_sparse_realsize(&mut self) -> io::Result<u64> {
if let Some(ref pax) = self.pax_extensions {
let pax = PaxExtensions::new(pax)
.filter_map(|f| f.ok())
.find(|f| f.key_bytes() == b"GNU.sparse.realsize")
.map(|f| f.value_bytes());
if let Some(field) = pax {
let str =
std::str::from_utf8(&field).map_err(|_| other("failed to read string"))?;
return str
.parse::<u64>()
.map_err(|_| other("failed to parse the real size"));
}
}
Err(other("PAX extension GNU.sparse.realsize not found"))
}
fn path(&self) -> io::Result<Cow<Path>> {
bytes2path(self.path_bytes())
}
fn path_bytes(&self) -> Cow<[u8]> {
match self.long_pathname {
Some(ref bytes) => {
if let Some(&0) = bytes.last() {
Cow::Borrowed(&bytes[..bytes.len() - 1])
} else {
Cow::Borrowed(bytes)
}
}
None => {
if let Some(ref pax) = self.pax_extensions {
let pax = PaxExtensions::new(pax)
.filter_map(|f| f.ok())
.find(|f| f.key_bytes() == b"path")
.map(|f| f.value_bytes());
if let Some(field) = pax {
return Cow::Borrowed(field);
}
}
self.header.path_bytes()
}
}
}
/// Gets the path in a "lossy" way, used for error reporting ONLY.
fn path_lossy(&self) -> String {
String::from_utf8_lossy(&self.path_bytes()).to_string()
}
fn link_name(&self) -> io::Result<Option<Cow<Path>>> {
match self.link_name_bytes() {
Some(bytes) => bytes2path(bytes).map(Some),
None => Ok(None),
}
}
fn link_name_bytes(&self) -> Option<Cow<[u8]>> {
match self.long_linkname {
Some(ref bytes) => {
if let Some(&0) = bytes.last() {
Some(Cow::Borrowed(&bytes[..bytes.len() - 1]))
} else {
Some(Cow::Borrowed(bytes))
}
}
None => {
if let Some(ref pax) = self.pax_extensions {
let pax = PaxExtensions::new(pax)
.filter_map(|f| f.ok())
.find(|f| f.key_bytes() == b"linkpath")
.map(|f| f.value_bytes());
if let Some(field) = pax {
return Some(Cow::Borrowed(field));
}
}
self.header.link_name_bytes()
}
}
}
fn pax_extensions(&mut self) -> io::Result<Option<PaxExtensions>> {
if self.pax_extensions.is_none() {
if !self.header.entry_type().is_pax_global_extensions()
&& !self.header.entry_type().is_pax_local_extensions()
{
return Ok(None);
}
self.pax_extensions = Some(self.read_all()?);
}
Ok(Some(PaxExtensions::new(
self.pax_extensions.as_ref().unwrap(),
)))
}
fn unpack_in(&mut self, dst: &Path) -> io::Result<bool> {
// Notes regarding bsdtar 2.8.3 / libarchive 2.8.3:
// * Leading '/'s are trimmed. For example, `///test` is treated as
// `test`.
// * If the filename contains '..', then the file is skipped when
// extracting the tarball.
// * '//' within a filename is effectively skipped. An error is
// logged, but otherwise the effect is as if any two or more
// adjacent '/'s within the filename were consolidated into one
// '/'.
//
// Most of this is handled by the `path` module of the standard
// library, but we specially handle a few cases here as well.
let mut file_dst = dst.to_path_buf();
{
let path = self.path().map_err(|e| {
TarError::new(
format!("invalid path in entry header: {}", self.path_lossy()),
e,
)
})?;
for part in path.components() {
match part {
// Leading '/' characters, root paths, and '.'
// components are just ignored and treated as "empty
// components"
Component::Prefix(..) | Component::RootDir | Component::CurDir => continue,
// If any part of the filename is '..', then skip over
// unpacking the file to prevent directory traversal
// security issues. See, e.g.: CVE-2001-1267,
// CVE-2002-0399, CVE-2005-1918, CVE-2007-4131
Component::ParentDir => return Ok(false),
Component::Normal(part) => file_dst.push(part),
}
}
}
// Skip cases where only slashes or '.' parts were seen, because
// this is effectively an empty filename.
if *dst == *file_dst {
return Ok(true);
}
// Skip entries without a parent (i.e. outside of FS root)
let parent = match file_dst.parent() {
Some(p) => p,
None => return Ok(false),
};
self.ensure_dir_created(&dst, parent)
.map_err(|e| TarError::new(format!("failed to create `{}`", parent.display()), e))?;
let canon_target = self.validate_inside_dst(&dst, parent)?;
self.unpack(Some(&canon_target), &file_dst)
.map_err(|e| TarError::new(format!("failed to unpack `{}`", file_dst.display()), e))?;
Ok(true)
}
/// Unpack as destination directory `dst`.
fn unpack_dir(&mut self, dst: &Path) -> io::Result<()> {
// If the directory already exists just let it slide
fs::create_dir(dst).or_else(|err| {
if err.kind() == ErrorKind::AlreadyExists {
let prev = fs::metadata(dst);
if prev.map(|m| m.is_dir()).unwrap_or(false) {
return Ok(());
}
}
Err(Error::new(
err.kind(),
format!("{} when creating dir {}", err, dst.display()),
))
})
}
/// Returns access to the header of this entry in the archive.
fn unpack(&mut self, target_base: Option<&Path>, dst: &Path) -> io::Result<Unpacked> {
fn set_perms_ownerships(
dst: &Path,
f: Option<&mut std::fs::File>,
header: &Header,
mask: u32,
perms: bool,
ownerships: bool,
) -> io::Result<()> {
// ownerships need to be set first to avoid stripping SUID bits in the permissions ...
if ownerships {
set_ownerships(dst, &f, header.uid()?, header.gid()?)?;
}
// ... then set permissions, SUID bits set here is kept
if let Ok(mode) = header.mode() {
set_perms(dst, f, mode, mask, perms)?;
}
Ok(())
}
fn get_mtime(header: &Header) -> Option<FileTime> {
header.mtime().ok().map(|mtime| {
// For some more information on this see the comments in
// `Header::fill_platform_from`, but the general idea is that
// we're trying to avoid 0-mtime files coming out of archives
// since some tools don't ingest them well. Perhaps one day
// when Cargo stops working with 0-mtime archives we can remove
// this.
let mtime = if mtime == 0 { 1 } else { mtime };
FileTime::from_unix_time(mtime as i64, 0)
})
}
let kind = self.header.entry_type();
if kind.is_dir() {
self.unpack_dir(dst)?;
set_perms_ownerships(
dst,
None,
&self.header,
self.mask,
self.preserve_permissions,
self.preserve_ownerships,
)?;
return Ok(Unpacked::__Nonexhaustive);
} else if kind.is_hard_link() || kind.is_symlink() {
let src = match self.link_name()? {
Some(name) => name,
None => {
return Err(other(&format!(
"hard link listed for {} but no link name found",
String::from_utf8_lossy(self.header.as_bytes())
)));
}
};
if src.iter().count() == 0 {
return Err(other(&format!(
"symlink destination for {} is empty",
String::from_utf8_lossy(self.header.as_bytes())
)));
}
if kind.is_hard_link() {
let link_src = match target_base {
// If we're unpacking within a directory then ensure that
// the destination of this hard link is both present and
// inside our own directory. This is needed because we want
// to make sure to not overwrite anything outside the root.
//
// Note that this logic is only needed for hard links
// currently. With symlinks the `validate_inside_dst` which
// happens before this method as part of `unpack_in` will
// use canonicalization to ensure this guarantee. For hard
// links though they're canonicalized to their existing path
// so we need to validate at this time.
Some(ref p) => {
let link_src = p.join(src);
self.validate_inside_dst(p, &link_src)?;
link_src
}
None => src.into_owned(),
};
fs::hard_link(&link_src, dst).map_err(|err| {
Error::new(
err.kind(),
format!(
"{} when hard linking {} to {}",
err,
link_src.display(),
dst.display()
),
)
})?;
} else {
symlink(&src, dst)
.or_else(|err_io| {
if err_io.kind() == io::ErrorKind::AlreadyExists && self.overwrite {
// remove dest and try once more
std::fs::remove_file(dst).and_then(|()| symlink(&src, dst))
} else {
Err(err_io)
}
})
.map_err(|err| {
Error::new(
err.kind(),
format!(
"{} when symlinking {} to {}",
err,
src.display(),
dst.display()
),
)
})?;
if self.preserve_mtime {
if let Some(mtime) = get_mtime(&self.header) {
filetime::set_symlink_file_times(dst, mtime, mtime).map_err(|e| {
TarError::new(format!("failed to set mtime for `{}`", dst.display()), e)
})?;
}
}
}
return Ok(Unpacked::__Nonexhaustive);
#[cfg(target_arch = "wasm32")]
#[allow(unused_variables)]
fn symlink(src: &Path, dst: &Path) -> io::Result<()> {
Err(io::Error::new(io::ErrorKind::Other, "Not implemented"))
}
#[cfg(windows)]
fn symlink(src: &Path, dst: &Path) -> io::Result<()> {
::std::os::windows::fs::symlink_file(src, dst)
}
#[cfg(unix)]
fn symlink(src: &Path, dst: &Path) -> io::Result<()> {
::std::os::unix::fs::symlink(src, dst)
}
} else if kind.is_pax_global_extensions()
|| kind.is_pax_local_extensions()
|| kind.is_gnu_longname()
|| kind.is_gnu_longlink()
{
return Ok(Unpacked::__Nonexhaustive);
};
// Old BSD-tar compatibility.
// Names that have a trailing slash should be treated as a directory.
// Only applies to old headers.
if self.header.as_ustar().is_none() && self.path_bytes().ends_with(b"/") {
self.unpack_dir(dst)?;
set_perms_ownerships(
dst,
None,
&self.header,
self.mask,
self.preserve_permissions,
self.preserve_ownerships,
)?;
return Ok(Unpacked::__Nonexhaustive);
}
// Note the lack of `else` clause above. According to the FreeBSD
// documentation:
//
// > A POSIX-compliant implementation must treat any unrecognized
// > typeflag value as a regular file.
//
// As a result if we don't recognize the kind we just write out the file
// as we would normally.
// Ensure we write a new file rather than overwriting in-place which
// is attackable; if an existing file is found unlink it.
fn open(dst: &Path) -> io::Result<std::fs::File> {
OpenOptions::new().write(true).create_new(true).open(dst)
}
let mut f = (|| -> io::Result<std::fs::File> {
let mut f = open(dst).or_else(|err| {
if err.kind() != ErrorKind::AlreadyExists {
Err(err)
} else if self.overwrite {
match fs::remove_file(dst) {
Ok(()) => open(dst),
Err(ref e) if e.kind() == io::ErrorKind::NotFound => open(dst),
Err(e) => Err(e),
}
} else {
Err(err)
}
})?;
for io in self.data.drain(..) {
match io {
EntryIo::Data(mut d) => {
let expected = d.limit();
if io::copy(&mut d, &mut f)? != expected {
return Err(other("failed to write entire file"));
}
}
EntryIo::Pad(d) => {
// TODO: checked cast to i64
let to = SeekFrom::Current(d.limit() as i64);
let size = f.seek(to)?;
f.set_len(size)?;
}
}
}
Ok(f)
})()
.map_err(|e| {
let header = self.header.path_bytes();
TarError::new(
format!(
"failed to unpack `{}` into `{}`",
String::from_utf8_lossy(&header),
dst.display()
),
e,
)
})?;
if self.preserve_mtime {
if let Some(mtime) = get_mtime(&self.header) {
filetime::set_file_handle_times(&f, Some(mtime), Some(mtime)).map_err(|e| {
TarError::new(format!("failed to set mtime for `{}`", dst.display()), e)
})?;
}
}
set_perms_ownerships(
dst,
Some(&mut f),
&self.header,
self.mask,
self.preserve_permissions,
self.preserve_ownerships,
)?;
if self.unpack_xattrs {
set_xattrs(self, dst)?;
}
return Ok(Unpacked::File(f));
fn set_ownerships(
dst: &Path,
f: &Option<&mut std::fs::File>,
uid: u64,
gid: u64,
) -> Result<(), TarError> {
_set_ownerships(dst, f, uid, gid).map_err(|e| {
TarError::new(
format!(
"failed to set ownerships to uid={:?}, gid={:?} \
for `{}`",
uid,
gid,
dst.display()
),
e,
)
})
}
#[cfg(unix)]
fn _set_ownerships(
dst: &Path,
f: &Option<&mut std::fs::File>,
uid: u64,
gid: u64,
) -> io::Result<()> {
use std::os::unix::prelude::*;
let uid: libc::uid_t = uid.try_into().map_err(|_| {
io::Error::new(io::ErrorKind::Other, format!("UID {} is too large!", uid))
})?;
let gid: libc::gid_t = gid.try_into().map_err(|_| {
io::Error::new(io::ErrorKind::Other, format!("GID {} is too large!", gid))
})?;
match f {
Some(f) => unsafe {
let fd = f.as_raw_fd();
if libc::fchown(fd, uid, gid) != 0 {
Err(io::Error::last_os_error())
} else {
Ok(())
}
},
None => unsafe {
let path = std::ffi::CString::new(dst.as_os_str().as_bytes()).map_err(|e| {
io::Error::new(
io::ErrorKind::Other,
format!("path contains null character: {:?}", e),
)
})?;
if libc::lchown(path.as_ptr(), uid, gid) != 0 {
Err(io::Error::last_os_error())
} else {
Ok(())
}
},
}
}
// Windows does not support posix numeric ownership IDs
#[cfg(any(windows, target_arch = "wasm32"))]
fn _set_ownerships(
_: &Path,
_: &Option<&mut std::fs::File>,
_: u64,
_: u64,
) -> io::Result<()> {
Ok(())
}
fn set_perms(
dst: &Path,
f: Option<&mut std::fs::File>,
mode: u32,
mask: u32,
preserve: bool,
) -> Result<(), TarError> {
_set_perms(dst, f, mode, mask, preserve).map_err(|e| {
TarError::new(
format!(
"failed to set permissions to {:o} \
for `{}`",
mode,
dst.display()
),
e,
)
})
}
#[cfg(unix)]
fn _set_perms(
dst: &Path,
f: Option<&mut std::fs::File>,
mode: u32,
mask: u32,
preserve: bool,
) -> io::Result<()> {
use std::os::unix::prelude::*;
let mode = if preserve { mode } else { mode & 0o777 };
let mode = mode & !mask;
let perm = fs::Permissions::from_mode(mode as _);
match f {
Some(f) => f.set_permissions(perm),
None => fs::set_permissions(dst, perm),
}
}
#[cfg(windows)]
fn _set_perms(
dst: &Path,
f: Option<&mut std::fs::File>,
mode: u32,
_mask: u32,
_preserve: bool,
) -> io::Result<()> {
if mode & 0o200 == 0o200 {
return Ok(());
}
match f {
Some(f) => {
let mut perm = f.metadata()?.permissions();
perm.set_readonly(true);
f.set_permissions(perm)
}
None => {
let mut perm = fs::metadata(dst)?.permissions();
perm.set_readonly(true);
fs::set_permissions(dst, perm)
}
}
}
#[cfg(target_arch = "wasm32")]
#[allow(unused_variables)]
fn _set_perms(
dst: &Path,
f: Option<&mut std::fs::File>,
mode: u32,
mask: u32,
_preserve: bool,
) -> io::Result<()> {
Err(io::Error::new(io::ErrorKind::Other, "Not implemented"))
}
#[cfg(all(unix, feature = "xattr"))]
fn set_xattrs(me: &mut EntryFields, dst: &Path) -> io::Result<()> {
use std::ffi::OsStr;
use std::os::unix::prelude::*;
let exts = match me.pax_extensions() {
Ok(Some(e)) => e,
_ => return Ok(()),
};
let exts = exts
.filter_map(|e| e.ok())
.filter_map(|e| {
let key = e.key_bytes();
let prefix = b"SCHILY.xattr.";
key.strip_prefix(prefix).map(|rest| (rest, e))
})
.map(|(key, e)| (OsStr::from_bytes(key), e.value_bytes()));
for (key, value) in exts {
xattr::set(dst, key, value).map_err(|e| {
TarError::new(
format!(
"failed to set extended \
attributes to {}. \
Xattrs: key={:?}, value={:?}.",
dst.display(),
key,
String::from_utf8_lossy(value)
),
e,
)
})?;
}
Ok(())
}
// Windows does not completely support posix xattrs
// https://en.wikipedia.org/wiki/Extended_file_attributes#Windows_NT
#[cfg(any(windows, not(feature = "xattr"), target_arch = "wasm32"))]
fn set_xattrs(_: &mut EntryFields, _: &Path) -> io::Result<()> {
Ok(())
}
}
fn ensure_dir_created(&self, dst: &Path, dir: &Path) -> io::Result<()> {
let mut ancestor = dir;
let mut dirs_to_create = Vec::new();
while ancestor.symlink_metadata().is_err() {
dirs_to_create.push(ancestor);
if let Some(parent) = ancestor.parent() {
ancestor = parent;
} else {
break;
}
}
for ancestor in dirs_to_create.into_iter().rev() {
if let Some(parent) = ancestor.parent() {
self.validate_inside_dst(dst, parent)?;
}
fs::create_dir_all(ancestor)?;
}
Ok(())
}
fn validate_inside_dst(&self, dst: &Path, file_dst: &Path) -> io::Result<PathBuf> {
// Abort if target (canonical) parent is outside of `dst`
let canon_parent = file_dst.canonicalize().map_err(|err| {
Error::new(
err.kind(),
format!("{} while canonicalizing {}", err, file_dst.display()),
)
})?;
let canon_target = dst.canonicalize().map_err(|err| {
Error::new(
err.kind(),
format!("{} while canonicalizing {}", err, dst.display()),
)
})?;
if !canon_parent.starts_with(&canon_target) {
let err = TarError::new(
format!(
"trying to unpack outside of destination path: {}",
canon_target.display()
),
// TODO: use ErrorKind::InvalidInput here? (minor breaking change)
Error::new(ErrorKind::Other, "Invalid argument"),
);
return Err(err.into());
}
Ok(canon_target)
}
}
impl<'a> Read for EntryFields<'a> {
fn read(&mut self, into: &mut [u8]) -> io::Result<usize> {
loop {
match self.data.get_mut(0).map(|io| io.read(into)) {
Some(Ok(0)) => {
self.data.remove(0);
}
Some(r) => return r,
None => return Ok(0),
}
}
}
}
impl<'a> Read for EntryIo<'a> {
fn read(&mut self, into: &mut [u8]) -> io::Result<usize> {
match *self {
EntryIo::Pad(ref mut io) => io.read(into),
EntryIo::Data(ref mut io) => io.read(into),
}
}
}