Bug #9448: Assignment with overridden []= returns return value of method rather than new value

• Status: Closed
• Assignee: -
• Target version: 2.2.0
• ruby -v: ruby 2.2.0dev
• Backport: 1.9.3: DONTNEED, 2.0.0: DONTNEED, 2.1: DONTNEED

[Sirupsen (Simon Eskildsen)]

When calling []= the assignment will return the return value of the []= method, rather than the value of the right-hand side, which is the behaviour in 2.1.0-p0, 2.0.0-p353 and 1.9.3-p484.

Example code:

class Animal                                                                                                                  
  def []=(key, new_value)                                                                                                      
    "walrus"                                                                                                                 
  end                                                                                                                          
end                                                                                                                            
animal = Animal.new                                                                                                               
puts animal['current'] = "unicorn" 

In trunk, this returns 'walrus'. In other releases (tested in the ones mentioned above), this returns 'unicorn'.

[nobu (Nobuyoshi Nakada)]

• Status changed from Open to Rejected

It's a spec.
An assignment returns the assigned value, not the value returned by the method.

[normalperson (Eric Wong)]

[e-mail] wrote:

Status changed from Open to Rejected

It's a spec.
An assignment returns the assigned value, not the value returned by the method.

I think you misread, he is saying assignment returns the value returned
by the method in trunk. It looks like trunk is violating spec.
I think it is r44551, opt_aset_str causing the bug. Fix coming.

[nobu (Nobuyoshi Nakada)]

• Status changed from Rejected to Open

Indeed, thank you.

[normalperson (Eric Wong)]

[e-mail] wrote:

I think you misread, he is saying assignment returns the value returned
by the method in trunk. It looks like trunk is violating spec.
I think it is r44551, opt_aset_str causing the bug. Fix coming.

I fail at fixing this so far. Maybe I need food or sleep, first :x

Work-in-progress here:
http://bogomips.org/ruby.git/patch/?id=1095d48fe6
(also git://bogomips.org/ruby.git bug9448-wip)

[nobu (Nobuyoshi Nakada)]

• Status changed from Open to Closed
• % Done changed from 0 to 100

Applied in changeset r44705.

---

compile.c: assignment result of aset_with

• compile.c (iseq_compile_each): result of assignment should be
  its rhs instead of returned value from a method.
  [ruby-core:60071] [Bug #9448]

[nobu (Nobuyoshi Nakada)]

• Backport changed from 1.9.3: UNKNOWN, 2.0.0: UNKNOWN, 2.1: UNKNOWN to 1.9.3: DONTNEED, 2.0.0: DONTNEED, 2.1: DONTNEED

[normalperson (Eric Wong)]

[e-mail] wrote:

Applied in changeset r44705.

Thank you! Sorry for the breakage.
I need to study/understand compile.c better :x

[Sirupsen (Simon Eskildsen)]

Thanks a lot for the quick fix!