ruby/NEWS

# -*- rdoc -*-

= NEWS for Ruby 2.4.0

This document is a list of user visible feature changes made between
releases except for bug fixes.

Note that each entry is kept so brief that no reason behind or
reference information is supplied with.  For a full list of changes
with all sufficient information, see the ChangeLog file or Redmine
(e.g. <tt>https://bugs.ruby-lang.org/issues/$FEATURE_OR_BUG_NUMBER</tt>)

== Changes since the 2.3.0 release

=== Language changes

* Multiple assignment in conditional expression is now allowed.
  [Feature #10617]

=== Core classes updates (outstanding ones only)

* Array

  * Array#max and Array#min.  [Feature #12172]
    This may cause a tiny incompatibility: if you redefine
    Enumerable#max and call max to an Array, your redefinition will be
    now ignored.  You should also redefine Array#max.

  * Array#sum  [Feature #12217]
    This is different from Enumerable#sum in that Array#sum doesn't depend on
    the definition of each method.

* Dir

  * Dir.empty?.  [Feature #10121]

* Enumerable

  * Enumerable#sum  [Feature #12217]

* File

  * File.empty?.  [Feature #9969]

* Float

  * Float#ceil, Float#floor, and Float#truncate now take an optional
    digits, as well as Float#round.  [Feature #12245]

* Integer

  * Integer#ceil, Integer#floor, and Integer#truncate now take an optional
    digits, as well as Integer#round.  [Feature #12245]

  * Fixnum and Bignum are unified into Integer  [Feature #12005]

  * Integer#digits for extracting columns of place-value notation [Feature #12447]

* Regexp

  * Regexp#match? [Feature #8110]
    This returns bool and doesn't save backref.

* String

  * String#upcase, String#downcase, String#capitalize, String#swapcase and
    their bang variants work for all of Unicode, and are no longer limited
    to ASCII. Variations are available with options. See the documentation
    of String#downcase for details. [Feature #10085]

  * String.new(capacity: size) [Feature #12024]

* Symbol

  * Symbol#match now returns MatchData.  [Bug #11991]

  * Symbol#upcase, Symbol#downcase, Symbol#capitalize, and Symbol#swapcase now
    work for all of Unicode. See the documentation of String#downcase
    for details. [Feature #10085]

* MatchData

  * MatchData#named_captures [Feature #11999]
  * MatchData#values_at supports named captures [Feature #9179]

* Thread

  * Thread#report_on_exception and Thread.report_on_exception
    [Feature #6647]

=== Stdlib updates (outstanding ones only)

* CSV

  * Add a liberal_parsing option. [Feature #11839]

* Logger

	* Allow specifying logger parameters in constructor such
	  as level, progname, datetime_format, formatter. [Feature #12224]
	* Add shift_period_suffix option. [Feature #10772]

* optparse

  * Add an into option. [Feature #11191]

=== Compatibility issues (excluding feature bug fixes)

* Array#sum and Enumerable#sum are implemented.  [Feature #12217]
  Ruby itself has no compatibility problem because Ruby didn't have sum method
  for arrays before Ruby 2.4.
  However many third party gems, activesupport, facets, simple_stats, etc,
  defines sum method.  These implementations are mostly compatible but
  there are subtle differences.
  Ruby's sum method should be mostly compatible but it is impossible to
  be perfectly compatible with all of them.

* Fixnum and Bignum are unified into Integer  [Feature #12005]
  Fixnum class and Bignum class is removed.
  Integer class is changed from abstract class to concrete class.
  For example, 0 is an instance of Integer: 0.class returns Integer.
  The constants Fixnum and Bignum is bound to Integer.
  So obj.kind_of?(Fixnum) works as obj.kind_of?(Integer).
  At C-level, Fixnum object and Bignum object should be distinguished by
  FIXNUM_P(obj) and RB_TYPE_P(obj, T_BIGNUM).
  ruby-2.4.0-preview1 (temporally) removes rb_cFixnum and rb_cBignum
  to check the effect of incompatibility.
  (rb_cFixnum and rb_cBignum may be defined as rb_cInteger later if
  compilation failure of extension library is too big problem.)

* String/Symbol#upcase/downcase/swapcase/capitalize(!) now work for all of
  Unicode, not only for ASCII. [Feature #10085]
  No change is needed if the data is in ASCII anyway or if the limitation
  to ASCII was only tolerated while waiting for a more extensive implementation.
  A change (using the :ascii option) is needed in cases where Unicode data
  is processed, but the operation has to be limited to ASCII only.
  A good example of this are internationalized domain names.

=== Stdlib compatibility issues (excluding feature bug fixes)

* Time

  * Time#to_time now preserves timezone.  [Bug #12271]

* DateTime

  * DateTime#to_time now preserves timezone.  [Bug #12189]

* thread

  * the extension library is removed.  Till 2.0 it was a pure ruby script
    "thread.rb", which has precedence over "thread.so", and has been provided
    in $LOADED_FEATURES since 2.1.

=== C API updates

* ruby_show_version() will no longer exits the process, if
  RUBY_SHOW_COPYRIGHT_TO_DIE is set to 0.  This will be the default in
  the future.

=== Supported platform changes

* FreeBSD < 4 is no longer supported

=== Implementation improvements

* In some condition, `[x, y].max` and `[x, y].min` are optimized
  so that a temporal array is not created.  The concrete condition is
  an implementation detail: currently, the array literal must have no
  splat, must have at least one expression but literal, the length must
  be <= 0x100, and Array#max and min must not be redefined.  It will work
  in most casual and real-life use case where it is written with intent
  to `Math.max(x, y)`.

* Thread deadlock detection now shows their backtrace and dependency. [Feature #8214]