Feature #15504: Freeze all Range objects

• Status: Closed
• Priority: Normal
• Assignee: matz (Yukihiro Matsumoto)
• Target version: -

[ko1 (Koichi Sasada)]

Abstract

Range is currently non-frozen. How about freezing all Range objects?

Background

We froze some types of objects: Numerics (r47523) and Symbols [Feature #8906]. I believe that making objects immutable solves some kinds of programming difficulties.

Range is mutable at least when written as Range literal. So we can write the following weird program:

2.times{
  r = (1..3)
  p r.instance_variable_get(:@foo)
  #=> 1st time: nil
  #=> 2nd time: :bar
  r.instance_variable_set(:@foo, :bar)
}

In range.c, there is a comment (thanks znz-san):

static void
range_modify(VALUE range)
{
    rb_check_frozen(range);
    /* Ranges are immutable, so that they should be initialized only once. */
    if (RANGE_EXCL(range) != Qnil) {
	rb_name_err_raise("`initialize' called twice", range, ID2SYM(idInitialize));
    }
}

Patch

Index: range.c
===================================================================
--- range.c	(リビジョン 66699)
+++ range.c	(作業コピー)
@@ -45,6 +45,8 @@
     RANGE_SET_EXCL(range, exclude_end);
     RANGE_SET_BEG(range, beg);
     RANGE_SET_END(range, end);
+
+    rb_obj_freeze(range);
 }
 
 VALUE

Discussion

There are several usages of mutable Range in the tests.

• (1) Taint-flag
• (2) Add singleton methods.
• (3) Subclass with mutable states

Maybe (2) and (3) are crucial.

Thanks,
Koichi

[marcandre (Marc-Andre Lafortune)]

I think that (2) and (3) are indeed capital points. Freezing range litterals (only) might be a better idea? with an approach like frozen string literals?

Not that even frozen ranges aren't completely immutable:

r = ('a'..'z').freeze
r.end.upcase!
r # => "a".."Z" 

[Eregon (Benoit Daloze)]

I think it would make sense to freeze Range literals.

Adding methods to Range might be reasonable, but singleton methods, I would think much less so.

[mame (Yusuke Endoh)]

I guess @Eregon (Benoit Daloze) came from #15950. Will ary[-3..] be as efficient as ary[-3, 3] by freezing and deduping a literal (-3..)? Looks good if we can confirm it by an experiment.

Some thoughts:

• Even if a range is frozen, ("a".."z") should not be deduped because of the reason @marcandre (Marc-Andre Lafortune) said.
• I'm for freezing all Ranges, not only Range literals. I hate the idea of freezing only literals because casually mixing frozen and unfrozen objects leads to hard-to-debug bugs that depend upon data flow.
• It would be the most elegant if the combination of MJIT and escape analysis solves this kind of performance problems.

[ko1 (Koichi Sasada)]

I got an issue on Ractor.

def test
  (1..2)
end

r1 = test

Ractor.new do
  r2 = test
end.take

• compiler cached Range (1..2) because they begin and end are frozen literals. The test method returns a cached same range object.
• it means test returns not-immutable but same object. It violate the Ractor's memory model.

To solve it, there are two options.

(1) avoid cache at compile time.
(2) freeze Range objects which will be cached by th compiler.

For performance reason, I want to choose (2).

After that, could you please discuss all Range objects should be frozen or not.

Thanks,
Koichi

[sawa (Tsuyoshi Sawada)]

• Description updated (diff)
• Subject changed from Freeze all Range object to Freeze all Range objects

[Eregon (Benoit Daloze)]

Right, every object cached at parse time must be deeply immutable, I would think that was an oversight.

[matz (Yukihiro Matsumoto)]

I agree with making ranges frozen. I don't see any particular case that could be broken by frozen ranges.
Since there's possiblity of breakage, I'd like to experiment it.

Matz.

[ko1 (Koichi Sasada)]

Ok, I freeze all Ranges except sub-classes.
https://github.com/ruby/ruby/pull/3583

[ko1 (Koichi Sasada)]

• Status changed from Open to Closed

Applied in changeset git|0096d2b895395df5ab8696d3b6d444dc1b7730b6.

---

freeze all Range objects.

Matz want to try to freeze all Range objects.
[Feature #15504]

[Eregon (Benoit Daloze)]

Great!

Related, should Enumerator::ArithmeticSequence be frozen too?

[ko1 (Koichi Sasada)]

Eregon (Benoit Daloze) wrote in #note-10:

Related, should Enumerator::ArithmeticSequence be frozen too?

new ticket?

[Eregon (Benoit Daloze)]

• Related to Feature #17195: Freeze Enumerator::ArithmeticSequence objects added

[Eregon (Benoit Daloze)]

ko1 (Koichi Sasada) wrote in #note-11:

Eregon (Benoit Daloze) wrote in #note-10:

Related, should Enumerator::ArithmeticSequence be frozen too?

new ticket?

I filed #17195

[AlexWayfer (Alexander Popov)]

I can't now mock Range objects with RSpec (gorilla_patch gem). What should I do? I see no work-around, like +'foo' for strings. Range.new gives frozen objects too.

I found a work-around: (5..7).dup, but it's weird anyway.

[zverok (Victor Shepelev)]

@AlexWayfer (Alexander Popov)

https://github.com/AlexWayfer/gorilla_patch/blob/master/lib/gorilla_patch/cover.rb#L8 -- may be for this particular case it is better to have version guard as an outer check?..

if RUBY_VERSION < '2.6'
  def cover?(value)
    #...
  end
end

...and have the same guard in specs?..

[AlexWayfer (Alexander Popov)]

zverok (Victor Shepelev) wrote in #note-15:

@AlexWayfer (Alexander Popov)

https://github.com/AlexWayfer/gorilla_patch/blob/master/lib/gorilla_patch/cover.rb#L8 -- may be for this particular case it is better to have version guard as an outer check?..

if RUBY_VERSION < '2.6'
  def cover?(value)
    #...
  end
end

...and have the same guard in specs?..

Thank you, I agree, it's better. But… if I want to check was called refined method or core? Right now I'm doing it via have_received once or never (depending on Ruby version), and for this RSpec should change the object (Range), but it's frozen since Ruby 3. Do I have other ways to check which implementation of method was used? value.method(:cover?).source_location returns nil in both cases.

[zverok (Victor Shepelev)]

if I want to check was called refined method or core?

It actually might be a good feature proposal for Ruby. Because currently, you can tell whether the method is defined by this class, by its parent, by included module, by singleton class... via Method#owner. But as far as I can recall, there is no way to ask "whether the method is defined by refinement".

But this whole discussion is unrelated to Range frozenness, honestly :)

[AlexWayfer (Alexander Popov)]

zverok (Victor Shepelev) wrote in #note-17:

if I want to check was called refined method or core?

It actually might be a good feature proposal for Ruby. Because currently, you can tell whether the method is defined by this class, by its parent, by included module, by singleton class... via Method#owner. But as far as I can recall, there is no way to ask "whether the method is defined by refinement".

I'm too lazy for such proposals, especially well formatted, but I can try. UPD: Created https://bugs.ruby-lang.org/issues/17674

zverok (Victor Shepelev) wrote in #note-17:

But this whole discussion is unrelated to Range frozenness, honestly :)

No, I disagree! My issue was raised by resolving this issue, even without changes mention in the article about Ruby 3: https://www.ruby-lang.org/en/news/2020/12/25/ruby-3-0-0-released/

[Eregon (Benoit Daloze)]

@ko1 (Koichi Sasada) Do you know why only Range instances and not Range subclass instances were frozen? (https://bugs.ruby-lang.org/issues/15504#note-8)

This issue title is a bit confusing, also https://bugs.ruby-lang.org/issues/15504#note-9.
Probably https://bugs.ruby-lang.org/issues/15504#note-9 was reverted?

I guess "Freeze all Range objects" means all Range instances, literal or not, but still not subclass instances, that confused me.

[ko1 (Koichi Sasada)]

Eregon (Benoit Daloze) wrote in #note-19:

This issue title is a bit confusing, also https://bugs.ruby-lang.org/issues/15504#note-9.
Probably https://bugs.ruby-lang.org/issues/15504#note-9 was reverted?

I guess "Freeze all Range objects" means all Range instances, literal or not, but still not subclass instances, that confused me.

I see. Not sure but maybe it is only implementation issue.