ruby/NEWS.md

Ignoring revisions in .git-blame-ignore-revs. Click here to bypass and see the normal blame view.

472 строки
17 KiB
Markdown
Исходник Обычный вид История

2022-12-26 05:42:36 +03:00
# NEWS for Ruby 3.3.0
2021-12-15 19:24:31 +03:00
This document is a list of user-visible feature changes
2022-12-26 05:42:36 +03:00
since the **3.2.0** release, except for bug fixes.
Note that each entry is kept to a minimum, see links for details.
## Language changes
Evaluate multiple assignment left hand side before right hand side In regular assignment, Ruby evaluates the left hand side before the right hand side. For example: ```ruby foo[0] = bar ``` Calls `foo`, then `bar`, then `[]=` on the result of `foo`. Previously, multiple assignment didn't work this way. If you did: ```ruby abc.def, foo[0] = bar, baz ``` Ruby would previously call `bar`, then `baz`, then `abc`, then `def=` on the result of `abc`, then `foo`, then `[]=` on the result of `foo`. This change makes multiple assignment similar to single assignment, changing the evaluation order of the above multiple assignment code to calling `abc`, then `foo`, then `bar`, then `baz`, then `def=` on the result of `abc`, then `[]=` on the result of `foo`. Implementing this is challenging with the stack-based virtual machine. We need to keep track of all of the left hand side attribute setter receivers and setter arguments, and then keep track of the stack level while handling the assignment processing, so we can issue the appropriate topn instructions to get the receiver. Here's an example of how the multiple assignment is executed, showing the stack and instructions: ``` self # putself abc # send abc, self # putself abc, foo # send abc, foo, 0 # putobject 0 abc, foo, 0, [bar, baz] # evaluate RHS abc, foo, 0, [bar, baz], baz, bar # expandarray abc, foo, 0, [bar, baz], baz, bar, abc # topn 5 abc, foo, 0, [bar, baz], baz, abc, bar # swap abc, foo, 0, [bar, baz], baz, def= # send abc, foo, 0, [bar, baz], baz # pop abc, foo, 0, [bar, baz], baz, foo # topn 3 abc, foo, 0, [bar, baz], baz, foo, 0 # topn 3 abc, foo, 0, [bar, baz], baz, foo, 0, baz # topn 2 abc, foo, 0, [bar, baz], baz, []= # send abc, foo, 0, [bar, baz], baz # pop abc, foo, 0, [bar, baz] # pop [bar, baz], foo, 0, [bar, baz] # setn 3 [bar, baz], foo, 0 # pop [bar, baz], foo # pop [bar, baz] # pop ``` As multiple assignment must deal with splats, post args, and any level of nesting, it gets quite a bit more complex than this in non-trivial cases. To handle this, struct masgn_state is added to keep track of the overall state of the mass assignment, which stores a linked list of struct masgn_attrasgn, one for each assigned attribute. This adds a new optimization that replaces a topn 1/pop instruction combination with a single swap instruction for multiple assignment to non-aref attributes. This new approach isn't compatible with one of the optimizations previously used, in the case where the multiple assignment return value was not needed, there was no lhs splat, and one of the left hand side used an attribute setter. This removes that optimization. Removing the optimization allowed for removing the POP_ELEMENT and adjust_stack functions. This adds a benchmark to measure how much slower multiple assignment is with the correct evaluation order. This benchmark shows: * 4-9% decrease for attribute sets * 14-23% decrease for array member sets * Basically same speed for local variable sets Importantly, it shows no significant difference between the popped (where return value of the multiple assignment is not needed) and !popped (where return value of the multiple assignment is needed) cases for attribute and array member sets. This indicates the previous optimization, which was dropped in the evaluation order fix and only affected the popped case, is not important to performance. Fixes [Bug #4443]
2021-04-21 20:49:19 +03:00
## Command line options
* A new `performance` warning category was introduced.
They are not displayed by default even in verbose mode.
Turn them on with `-W:performance` or `Warning[:performance] = true`. [[Feature #19538]]
* A new `RUBY_CRASH_REPORT` environment variable was introduced to allow
redirecting Ruby crash reports to a file or sub command. See the `BUG REPORT ENVIRONMENT`
section of the ruby manpage for further details. [[Feature #19790]]
## Core classes updates
2021-12-15 19:24:31 +03:00
Note: We're only listing outstanding class updates.
* Array
2023-08-29 07:52:18 +03:00
* Array#pack now raises ArgumentError for unknown directives. [[Bug #19150]]
* Dir
2023-08-29 07:52:18 +03:00
* Dir.for_fd added for returning a Dir object for the directory specified
by the provided directory file descriptor. [[Feature #19347]]
2023-08-29 07:52:18 +03:00
* Dir.fchdir added for changing the directory to the directory specified
by the provided directory file descriptor. [[Feature #19347]]
2023-08-29 07:52:18 +03:00
* Dir#chdir added for changing the directory to the directory specified by
the provided `Dir` object. [[Feature #19347]]
* Encoding
* Encoding#replicate has been removed, it was already deprecated. [[Feature #18949]]
2023-12-06 09:47:33 +03:00
* Fiber
* Introduce Fiber#kill. [[Bug #595]]
2023-12-06 09:47:33 +03:00
```ruby
fiber = Fiber.new do
while true
puts "Yielding..."
Fiber.yield
end
ensure
puts "Exiting..."
end
fiber.resume
# Yielding...
fiber.kill
# Exiting...
```
* MatchData
2023-08-29 07:52:18 +03:00
* MatchData#named_captures now accepts optional `symbolize_names`
keyword. [[Feature #19591]]
* Module
* Module#set_temporary_name added for setting a temporary name for a
module. [[Feature #19521]]
* ObjectSpace::WeakKeyMap
* New core class to build collections with weak references.
The class use equality semantic to lookup keys like a regular hash,
but it doesn't hold strong references on the keys. [[Feature #18498]]
* ObjectSpace::WeakMap
* `ObjectSpace::WeakMap#delete` was added to eagerly clear weak map
entries. [[Feature #19561]]
2023-10-26 10:52:17 +03:00
* Proc
* Now Proc#dup and Proc#clone call `#initialize_dup` and `#initialize_clone`
hooks respectively. [[Feature #19362]]
* Process
* New `Process.warmup` method that notify the Ruby virtual machine that the boot sequence is finished,
and that now is a good time to optimize the application. This is useful
for long-running applications. The actual optimizations performed are entirely
implementation-specific and may change in the future without notice. [[Feature #18885]]
* Process::Status
* Process::Status#& and Process::Status#>> are deprecated. [[Bug #19868]]
* Queue
* Queue#freeze now raises TypeError. [[Bug #17146]]
* Range
* Range#reverse_each can now process beginless ranges with an Integer endpoint. [[Feature #18515]]
* Range#reverse_each now raises TypeError for endless ranges. [[Feature #18551]]
2023-12-06 17:19:38 +03:00
* Range#overlap? added for checking if two ranges overlap. [[Feature #19839]]
* Refinement
* Add Refinement#target as an alternative of Refinement#refined_class.
2023-08-29 07:52:18 +03:00
Refinement#refined_class is deprecated and will be removed in Ruby
3.4. [[Feature #19714]]
* SizedQueue
* SizedQueue#freeze now raises TypeError. [[Bug #17146]]
* String
* String#unpack now raises ArgumentError for unknown directives. [[Bug #19150]]
* String#bytesplice now accepts new arguments index/length or range of the
source string to be copied. [[Feature #19314]]
* TracePoint
* TracePoint supports `rescue` event. When the raised exception was rescued,
the TracePoint will fire the hook. `rescue` event only supports Ruby-level
`rescue`. [[Feature #19572]]
## Stdlib updates
* RubyGems and Bundler warn if users do `require` the following gems without adding them to Gemfile or gemspec.
This is because they will become the bundled gems in the future version of Ruby.
[[Feature #19351]] [[Feature #19776]] [[Feature #19843]]
* abbrev
* base64
* bigdecimal
* csv
* drb
* getoptlong
* mutex_m
* nkf
* observer
* racc
* resolv-replace
* rinda
* syslog
* Socket#recv and Socket#recv_nonblock returns `nil` instead of an empty string on closed
connections. Socket#recvmsg and Socket#recvmsg_nonblock returns `nil` instead of an empty packet on closed
connections. [[Bug #19012]]
* Name resolution such as `Socket.getaddrinfo`, `Socket.getnameinfo`, `Addrinfo.getaddrinfo`, etc.
can now be interrupted. [[Feature #19965]]
* Random::Formatter#alphanumeric is extended to accept optional `chars`
keyword argument. [[Feature #18183]]
2023-10-20 02:25:08 +03:00
The following default gem is added.
* prism 0.19.0
2023-10-20 02:25:08 +03:00
The following default gems are updated.
* RubyGems 3.5.1
* base64 0.2.0
* benchmark 0.3.0
* bigdecimal 3.1.5
* bundler 2.5.1
* cgi 0.4.1
* csv 3.2.8
* date 3.3.4
* delegate 0.3.1
* drb 2.2.0
* english 0.8.0
* erb 4.0.3
* error_highlight 0.6.0
* etc 1.4.3
* fcntl 1.1.0
* fiddle 1.1.2
* fileutils 1.7.2
* find 0.2.0
* getoptlong 0.2.1
* io-console 0.7.0
* io-nonblock 0.3.0
* io-wait 0.3.1
* irb 1.10.1
* json 2.7.1
* logger 1.6.0
* mutex_m 0.2.0
* net-http 0.4.0
* net-protocol 0.2.2
* nkf 0.1.3
* observer 0.1.2
* open-uri 0.4.1
* open3 0.2.0
* openssl 3.2.0
* optparse 0.4.0
* ostruct 0.6.0
* pathname 0.3.0
* pp 0.5.0
* prettyprint 0.2.0
* pstore 0.1.3
* psych 5.1.1.1
* rdoc 6.6.1
* reline 0.4.0
* resolv 0.3.0
* rinda 0.2.0
* securerandom 0.3.0
* set 1.0.4
* shellwords 0.2.0
* singleton 0.2.0
* stringio 3.1.1
* strscan 3.0.8
* syntax_suggest 2.0.0
* tempfile 0.2.1
* time 0.3.0
* timeout 0.4.1
* tmpdir 0.2.0
* tsort 0.2.0
* un 0.3.0
* uri 0.13.0
* weakref 0.1.3
* win32ole 1.8.10
* yaml 0.3.0
* zlib 3.1.0
The following bundled gem is promoted from default gems.
* racc 1.7.3
The following bundled gems are updated.
2023-09-07 09:59:55 +03:00
* minitest 5.20.0
* rake 13.1.0
2023-06-25 10:00:17 +03:00
* test-unit 3.6.1
2023-07-28 09:59:38 +03:00
* rexml 3.2.6
2023-08-12 10:00:20 +03:00
* rss 0.3.0
* net-ftp 0.3.3
* net-imap 0.4.8
2023-09-28 05:47:00 +03:00
* net-smtp 0.4.0
* rbs 3.3.2
* typeprof 0.21.8
* debug 1.9.0
See GitHub releases like [Logger](https://github.com/ruby/logger/releases) or
changelog for details of the default gems or bundled gems.
## Supported platforms
## Compatibility issues
* Subprocess creation/forking via the following file open methods is deprecated. [[Feature #19630]]
* Kernel#open
* URI.open
* IO.binread
* IO.foreach
* IO.readlines
* IO.read
* IO.write
* When given a non-lambda, non-literal block, Kernel#lambda with now raises
ArgumentError instead of returning it unmodified. These usages have been
issuing warnings under the `Warning[:deprecated]` category since Ruby 3.0.0.
[[Feature #19777]]
2023-11-30 09:38:49 +03:00
* The `RUBY_GC_HEAP_INIT_SLOTS` environment variable has been deprecated and
removed. Environment variables `RUBY_GC_HEAP_%d_INIT_SLOTS` should be
used instead. [[Feature #19785]]
* `it` calls without arguments in a block with no ordinary parameters are
deprecated. `it` will be a reference to the first block parameter in Ruby 3.4.
[[Feature #18980]]
## Stdlib compatibility issues
2020-06-18 15:11:19 +03:00
* `racc` is promoted to bundled gems.
* You need to add `racc` to your `Gemfile` if you use `racc` under bundler environment.
* `ext/readline` is retired
2023-12-14 03:54:23 +03:00
* We have `reline` that is pure Ruby implementation compatible with `ext/readline` API.
We rely on `reline` in the future. If you need to use `ext/readline`, you can install
`ext/readline` via rubygems.org with `gem install readline-ext`.
* We no longer need to install libraries like `libreadline` or `libedit`.
2023-05-10 03:56:54 +03:00
## C API updates
2023-12-15 22:39:31 +03:00
* `rb_postponed_job` updates
* New APIs and deprecated APIs (see comments for details)
* added: `rb_postponed_job_preregister()`
* added: `rb_postponed_job_trigger()`
* deprecated: `rb_postponed_job_register()` (and semantic change. see below)
* deprecated: `rb_postponed_job_register_once()`
* The postponed job APIs have been changed to address some rare crashes.
To solve the issue, we introduced new two APIs and deprecated current APIs.
The semantics of these functions have also changed slightly; `rb_postponed_job_register`
now behaves like the `once` variant in that multiple calls with the same
`func` might be coalesced into a single execution of the `func`
[[Feature #20057]]
2023-12-15 22:39:31 +03:00
* Some updates for internal thread event hook APIs
* `rb_internal_thread_event_data_t` with a target Ruby thread (VALUE)
and callback functions (`rb_internal_thread_event_callback`) receive it.
https://github.com/ruby/ruby/pull/8885
* The following functions are introduced to manipulate Ruby thread local data
from internal thread event hook APIs (they are introduced since Ruby 3.2).
https://github.com/ruby/ruby/pull/8936
* `rb_internal_thread_specific_key_create()`
* `rb_internal_thread_specific_get()`
* `rb_internal_thread_specific_set()`
* `rb_profile_thread_frames()` is introduced to get a frames from
a specific thread.
[[Feature #10602]]
* `rb_data_define()` is introduced to define `Data`. [[Feature #19757]]
* `rb_ext_resolve_symbol()` is introduced to search a function from
extension libraries. [[Feature #20005]]
* IO related updates:
* The details of `rb_io_t` will be hidden and deprecated attributes
are added for each members. [[Feature #19057]]
* `rb_io_path(VALUE io)` is introduced to get a path of `io`.
* `rb_io_closed_p(VALUE io)` to get opening or closing of `io`.
* `rb_io_mode(VALUE io)` to get the mode of `io`.
* `rb_io_open_descriptor()` is introduced to make an IO object from a file
descriptor.
## Implementation improvements
* `defined?(@ivar)` is optimized with Object Shapes.
### GC
* Major performance improvements over 3.2
* Young objects referenced by old objects are no longer immediately
promoted to the old generation. This significantly reduces the frequency of
major GC collections. [[Feature #19678]]
* A new `REMEMBERED_WB_UNPROTECTED_OBJECTS_LIMIT_RATIO` tuning variable was
introduced to control the number of unprotected objects cause a major GC
collection to trigger. The default is set to `0.01` (1%). This significantly
reduces the frequency of major GC collection. [[Feature #19571]]
* Write Barriers were implemented for many core types that were missing them,
notably `Time`, `Enumerator`, `MatchData`, `Method`, `File::Stat`, `BigDecimal`
and several others. This significantly reduces minor GC collection time and major
GC collection frequency.
* Most core classes are now using Variable Width Allocation, notably `Hash`, `Time`,
`Thread::Backtrace`, `Thread::Backtrace::Location`, `File::Stat`, `Method`.
This makes these classes faster to allocate and free, use less memory and reduce
heap fragmentation.
### YJIT
* Major performance improvements over 3.2
* Support for splat and rest arguments has been improved.
* Registers are allocated for stack operations of the virtual machine.
* More calls with optional arguments are compiled. Exception handlers are also compiled.
* Unsupported call types and megamorphic call sites no longer exit to the interpreter.
* Basic methods like Rails `#blank?` and
[specialized `#present?`](https://github.com/rails/rails/pull/49909) are inlined.
* `Integer#*`, `Integer#!=`, `String#!=`, `String#getbyte`,
`Kernel#block_given?`, `Kernel#is_a?`, `Kernel#instance_of?`, and `Module#===`
are specially optimized.
* Compilation speed is now slightly faster than 3.2.
* Now more than 3x faster than the interpreter on Optcarrot!
* Significantly improved memory usage over 3.2
* Metadata for compiled code uses a lot less memory.
* `--yjit-call-threshold` is automatically raised from 30 to 120
when the application has more than 40,000 ISEQs.
* `--yjit-cold-threshold` is added to skip compiling cold ISEQs.
* More compact code is generated on Arm64.
* Code GC is now disabled by default
* `--yjit-exec-mem-size` is treated as a hard limit where compilation of new code stops.
* No sudden drops in performance due to code GC.
Better copy-on-write behavior on servers reforking with
[Pitchfork](https://github.com/shopify/pitchfork).
* You can still enable code GC if desired with `--yjit-code-gc`
* Add `RubyVM::YJIT.enable` that can enable YJIT at run-time
* You can start YJIT without modifying command-line arguments or environment variables.
Rails 7.2 will [enable YJIT by default](https://github.com/rails/rails/pull/49947)
using this method.
* This can also be used to enable YJIT only once your application is
done booting. `--yjit-disable` can be used if you want to use other
YJIT options while disabling YJIT at boot.
* Add more profiling capabilities
* `ratio_in_yjit` stat produced by `--yjit-stats` is now available in release builds,
a special stats or dev build is no longer required to access most stats.
* `--yjit-perf` is added to facilitate profiling with Linux perf.
* `--yjit-trace-exits` now supports sampling with `--yjit-trace-exits-sample-rate=N`
* More thorough testing and multiple bug fixes
* `--yjit-stats=quiet` is added to avoid printing stats on exit.
2023-09-20 20:14:55 +03:00
### MJIT
* MJIT is removed.
* `--disable-jit-support` is removed. Consider using `--disable-yjit --disable-rjit` instead.
2023-09-20 20:14:55 +03:00
### RJIT
2023-09-20 20:14:55 +03:00
* Introduced a pure-Ruby JIT compiler RJIT.
* RJIT supports only x86\_64 architecture on Unix platforms.
* Unlike MJIT, it doesn't require a C compiler at runtime.
* RJIT exists only for experimental purposes.
* You should keep using YJIT in production.
2023-12-01 10:45:17 +03:00
### M:N Thread scheduler
2023-11-08 10:16:58 +03:00
* M:N Thread scheduler is introduced. [[Feature #19842]]
* Background: Ruby 1.8 and before, M:1 thread scheduler (M Ruby threads
with 1 native thread. Called as User level threads or Green threads)
is used. Ruby 1.9 and later, 1:1 thread scheduler (1 Ruby thread with
1 native thread). M:1 threads takes lower resources compare with 1:1
threads because it needs only 1 native threads. However it is difficult
to support context switching for all of blocking operation so 1:1
threads are employed from Ruby 1.9. M:N thread scheduler uses N native
threads for M Ruby threads (N is small number in general). It doesn't
need same number of native threads as Ruby threads (similar to the M:1
thread scheduler). Also our M:N threads supports blocking operations
well same as 1:1 threads. See the ticket for more details.
2023-12-01 10:45:17 +03:00
Our M:N thread scheduler refers on the goroutine scheduler in the
Go language.
* In a ractor, only 1 thread can run in a same time because of
implementation. Therefore, applications that use only one Ractor
(most applications) M:N thread scheduler works as M:1 thread scheduler
with further extension from Ruby 1.8.
* M:N thread scheduler can introduce incompatibility for C-extensions,
so it is disabled by default on the main Ractors.
`RUBY_MN_THREADS=1` environment variable will enable it.
On non-main Ractors, M:N thread scheduler is enabled (and can not
disable it now).
* `N` (the number of native threads) can be specified with `RUBY_MAX_CPU`
environment variable. The default is 8.
Note that more than `N` native threads are used to support many kind of
blocking operations.
2023-11-08 10:16:58 +03:00
2023-12-06 09:47:33 +03:00
[Bug #595]: https://bugs.ruby-lang.org/issues/595
2023-12-15 22:39:31 +03:00
[Feature #10602]: https://bugs.ruby-lang.org/issues/10602
[Bug #17146]: https://bugs.ruby-lang.org/issues/17146
[Feature #18183]: https://bugs.ruby-lang.org/issues/18183
[Feature #18498]: https://bugs.ruby-lang.org/issues/18498
[Feature #18515]: https://bugs.ruby-lang.org/issues/18515
[Feature #18551]: https://bugs.ruby-lang.org/issues/18551
[Feature #18885]: https://bugs.ruby-lang.org/issues/18885
[Feature #18949]: https://bugs.ruby-lang.org/issues/18949
2023-12-07 12:44:05 +03:00
[Feature #18980]: https://bugs.ruby-lang.org/issues/18980
2023-09-28 20:32:04 +03:00
[Bug #19012]: https://bugs.ruby-lang.org/issues/19012
2023-12-15 22:39:31 +03:00
[Feature #19057]: https://bugs.ruby-lang.org/issues/19057
2023-04-18 03:28:47 +03:00
[Bug #19150]: https://bugs.ruby-lang.org/issues/19150
2023-01-23 03:34:03 +03:00
[Feature #19314]: https://bugs.ruby-lang.org/issues/19314
[Feature #19347]: https://bugs.ruby-lang.org/issues/19347
[Feature #19351]: https://bugs.ruby-lang.org/issues/19351
2023-10-26 10:52:17 +03:00
[Feature #19362]: https://bugs.ruby-lang.org/issues/19362
2023-07-14 03:01:10 +03:00
[Feature #19521]: https://bugs.ruby-lang.org/issues/19521
2023-04-17 16:34:20 +03:00
[Feature #19538]: https://bugs.ruby-lang.org/issues/19538
2023-11-30 11:39:35 +03:00
[Feature #19561]: https://bugs.ruby-lang.org/issues/19561
2023-12-01 10:39:22 +03:00
[Feature #19571]: https://bugs.ruby-lang.org/issues/19571
[Feature #19572]: https://bugs.ruby-lang.org/issues/19572
2023-05-15 16:35:41 +03:00
[Feature #19591]: https://bugs.ruby-lang.org/issues/19591
[Feature #19630]: https://bugs.ruby-lang.org/issues/19630
2023-12-01 10:39:22 +03:00
[Feature #19678]: https://bugs.ruby-lang.org/issues/19678
2023-08-01 03:27:25 +03:00
[Feature #19714]: https://bugs.ruby-lang.org/issues/19714
2023-12-15 22:39:31 +03:00
[Feature #19757]: https://bugs.ruby-lang.org/issues/19757
[Feature #19776]: https://bugs.ruby-lang.org/issues/19776
[Feature #19777]: https://bugs.ruby-lang.org/issues/19777
[Feature #19785]: https://bugs.ruby-lang.org/issues/19785
2023-11-30 11:39:35 +03:00
[Feature #19790]: https://bugs.ruby-lang.org/issues/19790
2023-12-07 12:44:05 +03:00
[Feature #19839]: https://bugs.ruby-lang.org/issues/19839
2023-11-08 10:16:58 +03:00
[Feature #19842]: https://bugs.ruby-lang.org/issues/19842
[Feature #19843]: https://bugs.ruby-lang.org/issues/19843
[Bug #19868]: https://bugs.ruby-lang.org/issues/19868
[Feature #19965]: https://bugs.ruby-lang.org/issues/19965
2023-12-15 22:39:31 +03:00
[Feature #20005]: https://bugs.ruby-lang.org/issues/20005
[Feature #20057]: https://bugs.ruby-lang.org/issues/20057