zip 3.0.0

• zip(), zipr(), zip_append(), zipr_append(), zip_process(), and
  unzip() / unzip_process() now support password-protected archives
  using WinZIP AES-256, and other encryption schemes (#38).

• zip_list() now reports an encryption column indicating the encryption
  scheme used for each entry (#38).

• unzip() is now vectorized. If all arguments apart from zipfile,
  exdir and password are the default, then it unprocesses all files
  concurrently, using a thread pool. The size of the thread pool can be
  set with the zip_threads option or the ZIP_THREADS environment
  variable (#147).

• zip_list() and unzip() now work directly on http:// and https://
  URLs. They use HTTP range requests to download only the central directory
  and the requested entries, so listing or extracting a few files from a
  large remote archive does not downloads the whole file. If the server
  does not support range requests, they fall back to downloading the entire
  archive (with a warning). This requires the curl package (#39).

• zip_list() and unzip() now report the Unix permission bits stored in an
  archive on Windows as well. Previously they always reported 700/600 on
  Windows, regardless of the permissions recorded in the ZIP file.

• zip_list() and unzip() now report type == "directory" for directory
  entries whose Unix mode bits lack S_IFDIR but that are marked as
  directories by a trailing slash or the DOS directory attribute (e.g.
  archives created by zip() itself). Previously these were reported as
  "file".

• zip() and unzip() now show a progress bar when the cli package is
  installed. For zip(), progress is byte-level, so large single files are
  tracked smoothly. For unzip(), progress advances once per extracted
  entry. Progress bars are (for now) opt-in via the ZIP_PROGRESS=true
  environment variable or the zip.progress option (#48).

• unzip() now returns a data frame (invisibly) with one row per extracted
  entry, containing the same columns as zip_list() (filename,
  compressed_size, uncompressed_size, timestamp, permissions,
  crc32, offset, type) plus a path column with the absolute path
  to each extracted file on disk (#35).

• zip_list() and unzip() now have an encoding argument for ZIP files
  with non-UTF-8, non-CP437 filenames (e.g. CP932/Shift-JIS on Japanese
  Windows). When encoding is set, filenames without the UTF-8 flag are
  decoded from the specified code page instead of CP437 (#101).

• zip_append() and zipr_append() now replace existing entries when
  appending a file whose archive path already exists in the zip file,
  instead of creating duplicate entries (#111).

• unzip() and zip_list() now correctly handle ZIP files with
  non-UTF-8 filenames (e.g. filenames encoded in IBM CP437, as created
  by many Windows tools). The filenames are converted to UTF-8 using the
  CP437 character map when the UTF-8 flag is not set in the ZIP entry
  (#103).

• New keys argument to zip(), zipr(), zip_append(), and
  zipr_append(). It allows specifying custom paths for entries inside
  the archive, independently of their paths on disk (#50).

• unzip_process() now probes falls back to using an R subprocess when
  the bundled cmdunzip executable cannot be started (#135).

• Updated embedded miniz to version 3.1.1 (#122).