Merge pull request #137 from nvinson/symlink-mode-support

Author: Changaco

README.rst

@@ -114,6 +114,14 @@ and the optional third argument is the compression format (called “filter” i
 libarchive). The acceptable values are listed in ``libarchive.ffi.WRITE_FORMATS``
 and ``libarchive.ffi.WRITE_FILTERS``.
 
+Symbolic links
+~~~~~~~~~~~~~~
+
+By default, libarchive preserves symbolic links. If you want it to resolve the
+links and archive the files they point to instead, pass ``symlink_mode='logical'``
+when calling the ``add_files`` method. If you do that, an ``ArchiveError``
+exception will be raised when a symbolic link points to a nonexistent file.
+
 File metadata codecs
 --------------------
 

libarchive/ffi.py

@@ -286,6 +286,9 @@ def get_write_filter_function(filter_name):
 ffi('read_disk_open', [c_archive_p, c_char_p], c_int, check_int)
 ffi('read_disk_open_w', [c_archive_p, c_wchar_p], c_int, check_int)
 ffi('read_disk_descend', [c_archive_p], c_int, check_int)
+ffi('read_disk_set_symlink_hybrid', [c_archive_p], c_int, check_int)
+ffi('read_disk_set_symlink_logical', [c_archive_p], c_int, check_int)
+ffi('read_disk_set_symlink_physical', [c_archive_p], c_int, check_int)
 
 # archive_read_data
 

libarchive/write.py

@@ -46,7 +46,7 @@ def add_entries(self, entries):
 
     def add_files(
         self, *paths, flags=0, lookup=False, pathname=None, recursive=True,
-        **attributes
+        symlink_mode=None, **attributes
     ):
         """Read files through the OS and add them to the archive.
 
@@ -63,6 +63,9 @@ def add_files(
             recursive (bool):
                 when False, if a path in `paths` is a directory,
                 only the directory itself is added.
+            symlink_mode (Literal['hybrid', 'logical', 'physical'] | None):
+                how symbolic links should be handled; see `man archive_read_disk`
+                for meanings
             attributes (dict): passed to `ArchiveEntry.modify()`
 
         Raises:
@@ -75,10 +78,23 @@ def add_files(
         if block_size <= 0:
             block_size = 10240  # pragma: no cover
 
+        set_symlink_mode = None
+        if symlink_mode:
+            try:
+                set_symlink_mode = getattr(
+                    ffi, f'read_disk_set_symlink_{symlink_mode}'
+                )
+            except AttributeError:
+                raise ValueError(
+                    f"symlink_mode value {symlink_mode!r} is invalid"
+                ) from None
+
         entry = ArchiveEntry(header_codec=self.header_codec)
         entry_p = entry._entry_p
         for path in paths:
             with new_archive_read_disk(path, flags, lookup) as read_p:
+                if set_symlink_mode:
+                    set_symlink_mode(read_p)
                 while 1:
                     r = read_next_header2(read_p, entry_p)
                     if r == ARCHIVE_EOF:

tests/test_rwx.py

@@ -2,6 +2,7 @@
 
 import io
 import json
+import os
 
 import libarchive
 from libarchive.entry import format_time
@@ -181,3 +182,32 @@ def write_callback(data):
                 )
             assert archive_entry.uid == 1000
             assert archive_entry.gid == 1000
+
+
+def test_symlinks(tmpdir):
+    os.chdir(tmpdir)
+    with open('empty', 'w'):
+        pass
+    with open('unreadable', 'w') as f:
+        f.write('secret')
+    os.chmod('unreadable', 0)
+
+    os.symlink('empty', 'symlink-to-empty')
+    os.symlink('unreadable', 'symlink-to-unreadable')
+
+    with libarchive.file_writer('archive.tar', 'gnutar') as archive:
+        archive.add_files('symlink-to-empty', symlink_mode='hybrid')
+        with pytest.raises(libarchive.ArchiveError):
+            archive.add_files('symlink-to-unreadable', symlink_mode='logical')
+        archive.add_files('symlink-to-unreadable', symlink_mode='physical')
+
+    with libarchive.file_reader('archive.tar') as archive:
+        entries = iter(archive)
+        e1 = next(entries)
+        assert e1.pathname == 'symlink-to-empty'
+        assert e1.isreg
+        assert e1.size == 0
+        e2 = next(entries)
+        assert e2.pathname == 'symlink-to-unreadable'
+        assert e2.issym
+        assert e2.linkpath == 'unreadable'