Issue #15875 has been updated by matsuda (Akira Matsuda).


### Side Note #1

`const_get()` could get this constant for all released versions, but has been intentionally changed in 2.7. Correct?

```
% ruby -e "module A; end; B = 1; p Object.const_get('A::B')"

[1.9.3...2.6.3]
1

[2.7.0-dev (2019-05-25 trunk 559dca509d)]
Traceback (most recent call last):
	1: from -e:1:in `<main>'
-e:1:in `const_get': uninitialized constant A::B (NameError)

```


### Side Note #2

Current stable jruby seems to be following the 2.6 behavior.

```
% ruby -e "module A; end; B = 1; p Object.const_defined?('A::B')"

[jruby-9.2.7.0]
true
```


----------------------------------------
Bug #15875: const_defined? behavior inconsistency in 2.6.x series
https://bugs.ruby-lang.org/issues/15875#change-78211

* Author: matsuda (Akira Matsuda)
* Status: Open
* Priority: Normal
* Assignee: 
* Target version: 
* ruby -v: 2.6.1, 2.6.2, 2.6.3
* Backport: 2.4: UNKNOWN, 2.5: UNKNOWN, 2.6: UNKNOWN
----------------------------------------
`Module#const_defined?` returns true for unreachable nested fully qualified module name.
Here are the results of `const_defined?` for each ruby version on my machine.

```
% ruby -e "module A; end; B = 1; p Object.const_defined?('A::B')"

[1.9.3-p551, 2.0.0-p648]
-e:1:in `const_defined?': wrong constant name A::B (NameError)
	from -e:1:in `<main>'

[2.1.10, 2.2.10, 2.3.8, 2.4.6, 2.5.5, 2.6.0]
false

[2.6.1, 2.6.2, 2.6.3]
true

[2.7.0-dev (2019-05-25 trunk 559dca509d)]
false
```

For (2.6.1...2.6.3), it fails to reference the constant even though `const_defined?` returns true. So IMO this should be `const_defined?`'s bug.

```
% ruby -e "module A; end; B = 1; p Object.const_defined?('A::B') && A::B"
-e:1:in `<main>': uninitialized constant A::B (NameError)
```




-- 
https://bugs.ruby-lang.org/

Unsubscribe: <mailto:ruby-core-request / ruby-lang.org?subject=unsubscribe>
<http://lists.ruby-lang.org/cgi-bin/mailman/options/ruby-core>