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
use std::cell::{Cell, RefCell};
use std::cmp;
use std::fs;
use std::io;
use std::io::prelude::*;
use std::marker;
use std::path::Path;

use crate::entry::{EntryFields, EntryIo};
use crate::error::TarError;
use crate::other;
use crate::pax::pax_extensions_size;
use crate::{Entry, GnuExtSparseHeader, GnuSparseHeader, Header};

/// A top-level representation of an archive file.
///
/// This archive can have an entry added to it and it can be iterated over.
pub struct Archive<R: ?Sized + Read> {
    inner: ArchiveInner<R>,
}

pub struct ArchiveInner<R: ?Sized> {
    pos: Cell<u64>,
    unpack_xattrs: bool,
    preserve_permissions: bool,
    preserve_mtime: bool,
    overwrite: bool,
    ignore_zeros: bool,
    obj: RefCell<R>,
}

/// An iterator over the entries of an archive.
pub struct Entries<'a, R: 'a + Read> {
    fields: EntriesFields<'a>,
    _ignored: marker::PhantomData<&'a Archive<R>>,
}

struct EntriesFields<'a> {
    archive: &'a Archive<dyn Read + 'a>,
    next: u64,
    done: bool,
    raw: bool,
}

impl<R: Read> Archive<R> {
    /// Create a new archive with the underlying object as the reader.
    pub fn new(obj: R) -> Archive<R> {
        Archive {
            inner: ArchiveInner {
                unpack_xattrs: false,
                preserve_permissions: false,
                preserve_mtime: true,
                overwrite: true,
                ignore_zeros: false,
                obj: RefCell::new(obj),
                pos: Cell::new(0),
            },
        }
    }

    /// Unwrap this archive, returning the underlying object.
    pub fn into_inner(self) -> R {
        self.inner.obj.into_inner()
    }

    /// Construct an iterator over the entries in this archive.
    ///
    /// Note that care must be taken to consider each entry within an archive in
    /// sequence. If entries are processed out of sequence (from what the
    /// iterator returns), then the contents read for each entry may be
    /// corrupted.
    pub fn entries(&mut self) -> io::Result<Entries<R>> {
        let me: &mut Archive<dyn Read> = self;
        me._entries().map(|fields| Entries {
            fields: fields,
            _ignored: marker::PhantomData,
        })
    }

    /// Unpacks the contents tarball into the specified `dst`.
    ///
    /// This function will iterate over the entire contents of this tarball,
    /// extracting each file in turn to the location specified by the entry's
    /// path name.
    ///
    /// This operation is relatively sensitive in that it will not write files
    /// outside of the path specified by `dst`. Files in the archive which have
    /// a '..' in their path are skipped during the unpacking process.
    ///
    /// # Examples
    ///
    /// ```no_run
    /// use std::fs::File;
    /// use tar::Archive;
    ///
    /// let mut ar = Archive::new(File::open("foo.tar").unwrap());
    /// ar.unpack("foo").unwrap();
    /// ```
    pub fn unpack<P: AsRef<Path>>(&mut self, dst: P) -> io::Result<()> {
        let me: &mut Archive<dyn Read> = self;
        me._unpack(dst.as_ref())
    }

    /// Indicate whether extended file attributes (xattrs on Unix) are preserved
    /// when unpacking this archive.
    ///
    /// 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.inner.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.inner.preserve_permissions = preserve;
    }

    /// Indicate whether files and symlinks should be overwritten on extraction.
    pub fn set_overwrite(&mut self, overwrite: bool) {
        self.inner.overwrite = overwrite;
    }

    /// 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.inner.preserve_mtime = preserve;
    }

    /// Ignore zeroed headers, which would otherwise indicate to the archive that it has no more
    /// entries.
    ///
    /// This can be used in case multiple tar archives have been concatenated together.
    pub fn set_ignore_zeros(&mut self, ignore_zeros: bool) {
        self.inner.ignore_zeros = ignore_zeros;
    }
}

impl<'a> Archive<dyn Read + 'a> {
    fn _entries(&mut self) -> io::Result<EntriesFields> {
        if self.inner.pos.get() != 0 {
            return Err(other(
                "cannot call entries unless archive is at \
                 position 0",
            ));
        }
        Ok(EntriesFields {
            archive: self,
            done: false,
            next: 0,
            raw: false,
        })
    }

    fn _unpack(&mut self, dst: &Path) -> io::Result<()> {
        if dst.symlink_metadata().is_err() {
            fs::create_dir_all(&dst)
                .map_err(|e| TarError::new(&format!("failed to create `{}`", dst.display()), e))?;
        }

        // Canonicalizing the dst directory will prepend the path with '\\?\'
        // on windows which will allow windows APIs to treat the path as an
        // extended-length path with a 32,767 character limit. Otherwise all
        // unpacked paths over 260 characters will fail on creation with a
        // NotFound exception.
        let dst = &dst.canonicalize().unwrap_or(dst.to_path_buf());

        for entry in self._entries()? {
            let mut file = entry.map_err(|e| TarError::new("failed to iterate over archive", e))?;
            file.unpack_in(dst)?;
        }
        Ok(())
    }

    fn skip(&self, mut amt: u64) -> io::Result<()> {
        let mut buf = [0u8; 4096 * 8];
        while amt > 0 {
            let n = cmp::min(amt, buf.len() as u64);
            let n = (&self.inner).read(&mut buf[..n as usize])?;
            if n == 0 {
                return Err(other("unexpected EOF during skip"));
            }
            amt -= n as u64;
        }
        Ok(())
    }
}

impl<'a, R: Read> Entries<'a, R> {
    /// Indicates whether this iterator will return raw entries or not.
    ///
    /// If the raw list of entries are returned, then no preprocessing happens
    /// on account of this library, for example taking into account GNU long name
    /// or long link archive members. Raw iteration is disabled by default.
    pub fn raw(self, raw: bool) -> Entries<'a, R> {
        Entries {
            fields: EntriesFields {
                raw: raw,
                ..self.fields
            },
            _ignored: marker::PhantomData,
        }
    }
}
impl<'a, R: Read> Iterator for Entries<'a, R> {
    type Item = io::Result<Entry<'a, R>>;

    fn next(&mut self) -> Option<io::Result<Entry<'a, R>>> {
        self.fields
            .next()
            .map(|result| result.map(|e| EntryFields::from(e).into_entry()))
    }
}

impl<'a> EntriesFields<'a> {
    fn next_entry_raw(
        &mut self,
        pax_size: Option<u64>,
    ) -> io::Result<Option<Entry<'a, io::Empty>>> {
        let mut header = Header::new_old();
        let mut header_pos = self.next;
        loop {
            // Seek to the start of the next header in the archive
            let delta = self.next - self.archive.inner.pos.get();
            self.archive.skip(delta)?;

            // EOF is an indicator that we are at the end of the archive.
            if !try_read_all(&mut &self.archive.inner, header.as_mut_bytes())? {
                return Ok(None);
            }

            // If a header is not all zeros, we have another valid header.
            // Otherwise, check if we are ignoring zeros and continue, or break as if this is the
            // end of the archive.
            if !header.as_bytes().iter().all(|i| *i == 0) {
                self.next += 512;
                break;
            }

            if !self.archive.inner.ignore_zeros {
                return Ok(None);
            }
            self.next += 512;
            header_pos = self.next;
        }

        // Make sure the checksum is ok
        let sum = header.as_bytes()[..148]
            .iter()
            .chain(&header.as_bytes()[156..])
            .fold(0, |a, b| a + (*b as u32))
            + 8 * 32;
        let cksum = header.cksum()?;
        if sum != cksum {
            return Err(other("archive header checksum mismatch"));
        }

        let file_pos = self.next;
        let mut size = header.entry_size()?;
        if size == 0 {
            if let Some(pax_size) = pax_size {
                size = pax_size;
            }
        }
        let ret = EntryFields {
            size: size,
            header_pos: header_pos,
            file_pos: file_pos,
            data: vec![EntryIo::Data((&self.archive.inner).take(size))],
            header: header,
            long_pathname: None,
            long_linkname: None,
            pax_extensions: None,
            unpack_xattrs: self.archive.inner.unpack_xattrs,
            preserve_permissions: self.archive.inner.preserve_permissions,
            preserve_mtime: self.archive.inner.preserve_mtime,
            overwrite: self.archive.inner.overwrite,
        };

        // Store where the next entry is, rounding up by 512 bytes (the size of
        // a header);
        let size = (size + 511) & !(512 - 1);
        self.next += size;

        Ok(Some(ret.into_entry()))
    }

    fn next_entry(&mut self) -> io::Result<Option<Entry<'a, io::Empty>>> {
        if self.raw {
            return self.next_entry_raw(None);
        }

        let mut gnu_longname = None;
        let mut gnu_longlink = None;
        let mut pax_extensions = None;
        let mut pax_size = None;
        let mut processed = 0;
        loop {
            processed += 1;
            let entry = match self.next_entry_raw(pax_size)? {
                Some(entry) => entry,
                None if processed > 1 => {
                    return Err(other(
                        "members found describing a future member \
                         but no future member found",
                    ));
                }
                None => return Ok(None),
            };

            let is_recognized_header =
                entry.header().as_gnu().is_some() || entry.header().as_ustar().is_some();

            if is_recognized_header && entry.header().entry_type().is_gnu_longname() {
                if gnu_longname.is_some() {
                    return Err(other(
                        "two long name entries describing \
                         the same member",
                    ));
                }
                gnu_longname = Some(EntryFields::from(entry).read_all()?);
                continue;
            }

            if is_recognized_header && entry.header().entry_type().is_gnu_longlink() {
                if gnu_longlink.is_some() {
                    return Err(other(
                        "two long name entries describing \
                         the same member",
                    ));
                }
                gnu_longlink = Some(EntryFields::from(entry).read_all()?);
                continue;
            }

            if is_recognized_header && entry.header().entry_type().is_pax_local_extensions() {
                if pax_extensions.is_some() {
                    return Err(other(
                        "two pax extensions entries describing \
                         the same member",
                    ));
                }
                pax_extensions = Some(EntryFields::from(entry).read_all()?);
                if let Some(pax_extensions_ref) = &pax_extensions {
                    pax_size = pax_extensions_size(pax_extensions_ref);
                }
                continue;
            }

            let mut fields = EntryFields::from(entry);
            fields.long_pathname = gnu_longname;
            fields.long_linkname = gnu_longlink;
            fields.pax_extensions = pax_extensions;
            self.parse_sparse_header(&mut fields)?;
            return Ok(Some(fields.into_entry()));
        }
    }

    fn parse_sparse_header(&mut self, entry: &mut EntryFields<'a>) -> io::Result<()> {
        if !entry.header.entry_type().is_gnu_sparse() {
            return Ok(());
        }
        let gnu = match entry.header.as_gnu() {
            Some(gnu) => gnu,
            None => return Err(other("sparse entry type listed but not GNU header")),
        };

        // Sparse files are represented internally as a list of blocks that are
        // read. Blocks are either a bunch of 0's or they're data from the
        // underlying archive.
        //
        // Blocks of a sparse file are described by the `GnuSparseHeader`
        // structure, some of which are contained in `GnuHeader` but some of
        // which may also be contained after the first header in further
        // headers.
        //
        // We read off all the blocks here and use the `add_block` function to
        // incrementally add them to the list of I/O block (in `entry.data`).
        // The `add_block` function also validates that each chunk comes after
        // the previous, we don't overrun the end of the file, and each block is
        // aligned to a 512-byte boundary in the archive itself.
        //
        // At the end we verify that the sparse file size (`Header::size`) is
        // the same as the current offset (described by the list of blocks) as
        // well as the amount of data read equals the size of the entry
        // (`Header::entry_size`).
        entry.data.truncate(0);

        let mut cur = 0;
        let mut remaining = entry.size;
        {
            let data = &mut entry.data;
            let reader = &self.archive.inner;
            let size = entry.size;
            let mut add_block = |block: &GnuSparseHeader| -> io::Result<_> {
                if block.is_empty() {
                    return Ok(());
                }
                let off = block.offset()?;
                let len = block.length()?;
                if len != 0 && (size - remaining) % 512 != 0 {
                    return Err(other(
                        "previous block in sparse file was not \
                         aligned to 512-byte boundary",
                    ));
                } else if off < cur {
                    return Err(other(
                        "out of order or overlapping sparse \
                         blocks",
                    ));
                } else if cur < off {
                    let block = io::repeat(0).take(off - cur);
                    data.push(EntryIo::Pad(block));
                }
                cur = off
                    .checked_add(len)
                    .ok_or_else(|| other("more bytes listed in sparse file than u64 can hold"))?;
                remaining = remaining.checked_sub(len).ok_or_else(|| {
                    other(
                        "sparse file consumed more data than the header \
                         listed",
                    )
                })?;
                data.push(EntryIo::Data(reader.take(len)));
                Ok(())
            };
            for block in gnu.sparse.iter() {
                add_block(block)?
            }
            if gnu.is_extended() {
                let mut ext = GnuExtSparseHeader::new();
                ext.isextended[0] = 1;
                while ext.is_extended() {
                    if !try_read_all(&mut &self.archive.inner, ext.as_mut_bytes())? {
                        return Err(other("failed to read extension"));
                    }

                    self.next += 512;
                    for block in ext.sparse.iter() {
                        add_block(block)?;
                    }
                }
            }
        }
        if cur != gnu.real_size()? {
            return Err(other(
                "mismatch in sparse file chunks and \
                 size in header",
            ));
        }
        entry.size = cur;
        if remaining > 0 {
            return Err(other(
                "mismatch in sparse file chunks and \
                 entry size in header",
            ));
        }
        Ok(())
    }
}

impl<'a> Iterator for EntriesFields<'a> {
    type Item = io::Result<Entry<'a, io::Empty>>;

    fn next(&mut self) -> Option<io::Result<Entry<'a, io::Empty>>> {
        if self.done {
            None
        } else {
            match self.next_entry() {
                Ok(Some(e)) => Some(Ok(e)),
                Ok(None) => {
                    self.done = true;
                    None
                }
                Err(e) => {
                    self.done = true;
                    Some(Err(e))
                }
            }
        }
    }
}

impl<'a, R: ?Sized + Read> Read for &'a ArchiveInner<R> {
    fn read(&mut self, into: &mut [u8]) -> io::Result<usize> {
        self.obj.borrow_mut().read(into).map(|i| {
            self.pos.set(self.pos.get() + i as u64);
            i
        })
    }
}

/// Try to fill the buffer from the reader.
///
/// If the reader reaches its end before filling the buffer at all, returns `false`.
/// Otherwise returns `true`.
fn try_read_all<R: Read>(r: &mut R, buf: &mut [u8]) -> io::Result<bool> {
    let mut read = 0;
    while read < buf.len() {
        match r.read(&mut buf[read..])? {
            0 => {
                if read == 0 {
                    return Ok(false);
                }

                return Err(other("failed to read entire block"));
            }
            n => read += n,
        }
    }
    Ok(true)
}