desc 1.4.3

• $set() and desc_set() now can omit checks if check = FALSE
  is set.

desc 1.4.2

• The description$write() method, and thus all desc_*() functions work
  correctly now on R 4.3.x for packages that declare a non-UTF-8 encoding.

desc 1.4.1

• The $coerce_authors_at_r() method now does a much better job at setting
  the authors' roles (#114, @dpprdan).

v1.4.0

• DESCRIPTION objects created with the !new command now omit LazyData: true
  to match new CRAN checks (#105, @malcolmbarrett)

• description$write() now writes out the file in the correct encoding
  (#109).

• Authors@R fields are now formatted differently when normalizing a
  DESCRIPTION file (#78).

• New description$get_list(), description$set_list() and
  corresponding desc_get_list() and desc_set_list() values to query
  and create comma separated fields (#86).

Breaking change

• desc_get_field() gains a boolean squish_ws parameter to normalize
  whitespace within the retrieved value. It defaults to the value of trim_ws
  (TRUE by default). Example with desc's current DESCRIPTION:

  Old behaviour:

  > desc::desc_get_field("Description")
  [1] "... DESCRIPTION files.\n    It is intended for packages ..."

  New behaviour:

  > desc::desc_get_field("Description")
  [1] "... DESCRIPTION files. It is intended for packages ..."

  If you want the old behaviour, just set squish_ws = FALSE.

v1.3.0

• Adding authors with long names or other fields (comment, typically)
  works well now (#91).

• get_deps() now removes unneeded whitespace from version requirements
  (#84).

• normalize() now does not drop Authors@R on non-UTF-8 systems
  when it has non-ASCII characters (#80).

• has_dep() now works well with dependencies listed multiple times
  (#97, @richfitz).

• Add coerce_authors_at_r() method to convert Author to
  Authors@R (#44, @muschellij2).

• add_author() and similar functions now allow a character vector of
  multiple roles (@niceume, #89).

• desc_set_deps() now inserts new packages in (case-insensitive)
  alphabetical order, if the existing packages are already in alphabetical
  order.

• New add_author_gh() method and desc_add_author_gh() function to add
  an author using the information available from GitHub V3 API. This method
  and function depend on gh and are limited when the GitHub user full
  name is incomplete or not well parsed by as.person() and when their
  email address isn't available (@maelle, #71).

• When using desc_normalize() the package dependencies are now
  alphabetically sorted (#66, @llrs).

• New add_orcid() method and desc_add_orcid() functions make it
  possible to add ORCID IDs to authors directly instead of via the
  comment argument (@maelle, #70).

• All functions and methods managing authors (add_me, add_author(),
  del_author(), add_role(), del_role(), change_maintainer(),
  search_for_author(), add_me(), etc.) gain an orcid argument
  (@maelle, #70).

• In person() within the Authors@R field, comment can now be a
  named character vector (@maelle, #69; @gvegayon, #65).

• When using desc(text=) parameter, set
  textConnection(encoding = "bytes") to handle cases when the input
  text is in a different marked encoding than the default encoding, such
  as UTF-8 input on Windows.