- version update to 4.0
  4.0
  ---
    BREAKING & SECURITY: The archive extraction functions now pass 3 security flags (SECURE_NODOTDOT, SECURE_NOABSOLUTEPATHS 
      and SECURE_SYMLINKS) to libarchive by default, unless the current directory is the root.
    BREAKING: The ArchiveEntry properties atime, mtime, ctime, birthtime and size now have the value None instead of 0 when they're unset.
    BREAKING: The ArchiveEntry.pathname property now attempts to decode bytes using UTF-8. This breaks reading archives that
      contain file names which look like UTF-8 but aren't, if such a thing exists. Proper support of encodings will probably be added
      in the next version.
    Multiple entries from the same archive can now be kept in memory, however only the current entry's content can be read.
    The filetype, linkpath, size, mode, redvmajor, redvminor, uid and gid attributes of an archive entry can now be modified.
    The four time properties of archive entries now have setters, so for example entry.set_mtime(0, 0) can be replaced by entry.mtime = 0.
    Archive entries have 4 new properties: perm, rdev, uname and gname.
    When adding files to an archive, a destination path can now be specified.
    The ArchiveRead class now has a bytes_read property.
    The ArchiveWrite class now has a bytes_written property.
    Python 3.6 is no longer tested.
    The documentation has been improved.
  3.2
  ---
    This release fixes the seek callbacks passed to libarchive by the custom_reader and stream_reader function. See #116 for more details.
  3.1
  ---
  This release is mostly meant to fix #113 (a bug introduced in 3.0), but it also contains the following changes:
    the ffi.archive_error() function now returns an exception instead of raising it
    seekable_stream_reader is now an alias of stream_reader
    the misleadingly named ffi.VOID_CB function has been removed
  3.0
  ---
  Changes to the library:

• Files Changed
• Browse Source