ruby--ruby/NEWS

# -*- markdown -*- markup: Markdown

# NEWS for Ruby 2.6.0

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. `https://bugs.ruby-lang.org/issues/$FEATURE_OR_BUG_NUMBER`)

## Changes since the 2.5.0 release

### Language changes

* `$SAFE` is a process global state and we can set 0 again.
  [Feature #14250]

* refinements take place at block passing.  [Feature #14223]

* `else` without `rescue` now causes a syntax error.  [EXPERIMENTAL]

* constant names may start with a non-ASCII capital letter.
  [Feature #13770]

* An endless range is introduced.  You can write a range that has no end,
  like `(0..)`.  The following shows typical use cases.  [Feature #12912]

    ```ruby
    ary[1..]                            # identical to ary[1..-1]
    (1..).each {|index| ... }           # infinite loop from index 1
    ary.zip(1..) {|elem, index| ... }   # ary.each.with_index(1) { }
    ```

* Non-`Symbol` key in keyword arguments hash causes an exception.

### Core classes updates (outstanding ones only)

* `Array`

    * Aliased methods:

        * `Array#filter` is a new alias for `Array#select`.
          [Feature #13784]
        * `Array#filter!` is a new alias for `Array#select!`.
          [Feature #13784]

* `Binding`

    * New methods:

        * added `Binding#source_location`.  [Feature #14230]

            This method returns the source location of binding, a 2-element
            array of `__FILE__` and `__LINE__`.  Traditionally, the same
            information could be retrieved by `eval("[__FILE__, __LINE__]",
            binding)`, but we are planning to change this behavior so that
            `Kernel#eval` ignores binding's source location [Bug #4352].
            So, users should use this newly-introduced method instead of
            `Kernel#eval`.

* `Dir`

    * New methods:

        * added `Dir#each_child` and `Dir#children` instance methods.
          [Feature #13969]

* `Enumerable`

    * Aliased methods:

        * `Enumerable#filter` is a new alias for `Enumerable#select`.
          [Feature #13784]

* `Enumerator::Lazy`

    * Aliased methods:

        * `Enumerator::Lazy#filter` is a new alias for
          `Enumerator::Lazy#select`.  [Feature #13784]

* `Exception`

    * `Exception#full_message` takes `:highlight` and `:order` options.
      [Bug #14324]

* `Hash`

    * Aliased methods:

        * `Hash#filter` is a new alias for `Hash#select`.  [Feature #13784]
        * `Hash#filter!` is a new alias for `Hash#select!`.
          [Feature #13784]

* `IO`

    * new mode character `'x'` to open files for exclusive access.
      [Feature #11258]

* `Kernel`

    * `Kernel#then` is a new alias for `Kernel#yield_self`.
      [Feature #14594]

    * `Kernel.#Complex` takes `:exception` option to specify the way of
      error handling.  [Feature #12732]

    * `Kernel.#Float` takes `:exception` option to specify the way of error
      handling.  [Feature #12732]

    * `Kernel.#Integer` takes `:exception` option to specify the way of
      error handling.  [Feature #12732]

    * `Kernel.#Rational` takes `:exception` option to specify the way of
      error handling.  [Feature #12732]

    * `Kernel.#system` takes `:exception` option to raise an exception on
      failure.  [Feature #14386]

    * incompatible changes:
        * `system()` and `exec()` do not close non-standard file descriptors
          (The default of `:close_others` option is changed to `false` by
          default.  but we still set the `FD_CLOEXEC` flag on descriptors we
          create).  [Misc #14907]

* `KeyError`

    * `KeyError#initialize` accepts `:receiver` and `:key` options to set
      receiver and key in Ruby code.  [Feature #14313]

* `Module`

    * `Module#method_defined?`, `Module#private_method_defined?` and
      `Module#protected_method_defined?` now accepts the second parameter as
      optional. If it's `true` (=default), checks ancestor modules/classes,
      or checks only the class itself.  [Feature #14944]

* `NameError`

    * `NameError#initialize` accepts `:receiver` option to set receiver in
      Ruby code.  [Feature #14313]

* `NoMethodError`

    * `NoMethodError#initialize` accepts `:receiver` option to set receiver
      in Ruby code.  [Feature #14313]

* `Proc`

    * `Proc#call` doesn't change `$SAFE` any more.  [Feature #14250]

* `Random`

    * New methods:

        * added `Random.bytes`.  [Feature #4938]

* `Range`

    * `Range#===` now uses `#cover?` instead of `#include?` method.
      [Feature #14575]

* `RubyVM::AST`

    * `RubyVM::AST.parse` parses a given string and returns AST nodes.
      [experimental]
    * `RubyVM::AST.parse_file` parses a given file and returns AST nodes.
      [experimental]

* `String`

    * `String#split` yields each substring to the block if given.
      [Feature #4780]

* `TracePoint`

    * New methods:

        * `TracePoint#parameters` [Feature #14694]

### Stdlib updates (outstanding ones only)

* `ERB`

    * Add `:trim_mode` and `:eoutvar` keyword arguments to `ERB.new`. Now
      non-keyword arguments other than first one are softly deprecated and
      will be removed when Ruby 2.5 becomes EOL.  [Feature #14256]

    * erb command's `-S` option is deprecated, which will be removed in the
      next version.

* `FileUtils`

    * New method:

        * `FileUtils#cp_lr`.  [Feature #4189]

* `Matrix`

    * New method:

        * `Matrix#antisymmetric?`

* `Net`

    * New method:

        * Add `:write_timeout` keyword argument to `Net::BufferedIO.new`.
          [Feature #13396]

        * Add `Net::BufferedIO#write_timeout`,
          `Net::BufferedIO#write_timeout=`, `Net::HTTP#write_timeout`, and
          `Net::HTTP#write_timeout=`.  [Feature #13396]

    * New constant:

        * Add `Net::HTTPClientException` to deprecate
          `Net::HTTPServerException`, whose name is misleading.  [Bug #14688]

* `REXML`

    * Improved some XPath implementations:

        * `concat()` function: Stringify all arguments before concatenating

        * `string()` function: Support context node

        * `string()` function: Support processing instruction node

        * Support `"*:#{ELEMENT_NAME}"` syntax in XPath 2.0

    * Fixed some XPath implementations:

        * `"//#{ELEMENT_NAME}[#{POSITION}]"` case

        * `string()` function: Fix `function(document)` returns nodes that
          are out of root elements.

        * `"/ #{ELEMENT_NAME} "` case

        * `"/ #{ELEMENT_NAME} [ #{PREDICATE} ]"` case

        * `"/ #{AXIS}:: #{ELEMENT_NAME} "` case

        * `"#{N}-#{M}"` case: One ore more white spaces were required before
          `"-"`

        * `"/child::node()"` case

        * `"#{FUNCTION}()/#{PATH}"` case

        * `"@#{ATTRIBUTE}/parent::"` case

        * `"name(#{NODE_SET})"` case

* `RSS`

    * `RSS::Parser.parse`: Accept options as `Hash`. `:validate`,
      `:ignore_unknown_element`, `:parser_class` options are available.

* `Set`

    * Aliased methods:

        * `Set#filter!` is a new alias for `Set#select!`.  [Feature #13784]

* `URI`

    * Add `URI::File` to handle file URI scheme.  [Feature #14035]

### Compatibility issues (excluding feature bug fixes)

* `File`

    * `File.read`, `File.binread`, `File.write`, `File.binwrite`,
      `File.foreach`, and `File.readlines` do not invoke external commands
      even if the path starts with the pipe character `'|'`.
      [Feature #14245]

* `Dir`

    * `Dir.glob` with `'\0'`-separated pattern list will be deprecated, and
      is now warned.  [Feature #14643]

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

### C API updates

### Supported platform changes

### Implementation improvements

* Speedup `Proc#call` because we don't need to care about `$SAFE` any more.
  [Feature #14318]

    With `lc_fizzbuzz` benchmark which uses so many `Proc#call` we can
    measure x1.4 improvements.  [Bug #10212]

* Speedup `block.call` where `block` is passed block parameter.
  [Feature #14330]

    Ruby 2.5 improves block passing performance. [Feature #14045]

    Additionally, Ruby 2.6 improves the performance of passed block calling.

* Introduce an initial implementation of JIT (Just-in-time) compiler.
  [Feature #14235] [experimental]
    * `--jit` option is added to enable JIT. `--jit-verbose=1` is good for
      inspection.  See `ruby --help` for others.
    * This JIT relies on C compiler used to build Ruby, on runtime.  Only
      gcc and clang are supported for the JIT for now, and MinGW support has
      some issues.
    * As of 2.6.0-preview1, we're just preparing infrastructure for JIT and
      very few optimizations are implemented.  So it's not ready for
      benchmarking Ruby's JIT performance yet. It's known that current JIT
      enablement makes Rails application slower for now.

* VM generator script renewal; makes the generated VM more optimized.
  [GH-1779]

* Thread cache enabled for pthreads platforms (for `Thread.new` and
  `Thread.start`).  [Feature #14757]

### Miscellaneous changes

* On macOS, shared libraries no longer include a full version number of ruby
  in their names.  This eliminates the burden of each teeny upgrade on the
  platform that users need to rebuild every extension library.

  * Before:
    * libruby.2.6.0.dylib
    * libruby.2.6.dylib -> libruby.2.6.0.dylib
    * libruby.dylib -> libruby.2.6.0.dylib

  * After:
    * libruby.2.6.dylib
    * libruby.dylib -> libruby.2.6.dylib