[#98098] [Ruby master Feature#16824] Follow RubyGems naming conventions for the stdlib — shannonskipper@...

Issue #16824 has been reported by shan (Shannon Skipper).

14 messages 2020/05/01

[#98147] [Ruby master Feature#16832] Use #name rather than #inspect to build "uninitialized constant" error messages — jean.boussier@...

Issue #16832 has been reported by byroot (Jean Boussier).

20 messages 2020/05/06

[#98174] [Ruby master Bug#16837] Can we make Ruby 3.0 as fast as Ruby 2.7 with the new assertions? — takashikkbn@...

Issue #16837 has been reported by k0kubun (Takashi Kokubun).

10 messages 2020/05/07

[#98241] [Ruby master Bug#16845] Building Ruby with old existing system Ruby results in make error with ./tool/file2lastrev.rb — erik@...

Issue #16845 has been reported by ErikSwan (Erik Swan).

7 messages 2020/05/09

[#98256] [Ruby master Feature#16847] Cache instruction sequences by default — jean.boussier@...

Issue #16847 has been reported by byroot (Jean Boussier).

16 messages 2020/05/11

[#98257] [Ruby master Feature#16848] Allow callables in $LOAD_PATH — jean.boussier@...

Issue #16848 has been reported by byroot (Jean Boussier).

27 messages 2020/05/11

[#98318] [Ruby master Bug#16853] calling bla(hash, **kw) with a string-based hash passes the strings into **kw (worked < 2.7) — sylvain.joyeux@...4x.org

Issue #16853 has been reported by sylvain.joyeux (Sylvain Joyeux).

12 messages 2020/05/13

[#98355] [Ruby master Bug#16889] TracePoint.enable { ... } also activates the TracePoint for other threads, even outside the block — eregontp@...

Issue #16889 has been reported by Eregon (Benoit Daloze).

16 messages 2020/05/14

[#98363] [Ruby master Feature#16891] Restore Positional Argument to Keyword Conversion — merch-redmine@...

Issue #16891 has been reported by jeremyevans0 (Jeremy Evans).

23 messages 2020/05/14

[#98371] [Ruby master Feature#16894] Integer division for Ruby 3 — andrew@...

Issue #16894 has been reported by ankane (Andrew Kane).

18 messages 2020/05/15

[#98391] [Ruby master Bug#16896] MakeMakefile methods should be private — eregontp@...

Issue #16896 has been reported by Eregon (Benoit Daloze).

10 messages 2020/05/15

[#98396] [Ruby master Feature#16897] Can a Ruby 3.0 compatible general purpose memoizer be written in such a way that it matches Ruby 2 performance? — sam.saffron@...

Issue #16897 has been reported by sam.saffron (Sam Saffron).

25 messages 2020/05/16

[#98453] [Ruby master Bug#16904] rubygems: psych: superclass mismatch for class Mark (TypeError) — jaruga@...

Issue #16904 has been reported by jaruga (Jun Aruga).

18 messages 2020/05/20

[#98486] [Ruby master Bug#16908] Strange behaviour of Hash#shift when used with `default_proc`. — samuel@...

Issue #16908 has been reported by ioquatix (Samuel Williams).

14 messages 2020/05/23

[#98569] [Ruby master Bug#16921] s390x: ramdom test failures for timeout or segmentation fault — jaruga@...

Issue #16921 has been reported by jaruga (Jun Aruga).

9 messages 2020/05/29

[#98599] [Ruby master Bug#16926] Kernel#require does not load a feature twice when $LOAD_PATH has been modified spec fails only on 2.7 — eregontp@...

Issue #16926 has been reported by Eregon (Benoit Daloze).

12 messages 2020/05/31

[ruby-core:98284] [Ruby master Feature#16832] Use #name rather than #inspect to build "uninitialized constant" error messages

From: jean.boussier@...
Date: 2020-05-12 09:53:26 UTC
List: ruby-core #98284
Issue #16832 has been updated by byroot (Jean Boussier).


> (also #name can be nil, and calling #name on arbitrary objects could easily be worse than #inspect).

So the issue title is no longer relevant. Since then I dug more into this issue, and I think the behavior should be similar to `rb_profile_frame_classpath`, in short try to use `rb_class2name` for T_MODULE and T_CLASS, see this PR for instance: https://github.com/ruby/ruby/pull/3084

I'll try to better explain my reasoning:

  - `#inspect` semantic is to describe an object
  - `uninitialized constant` error message tells you about a missing constant, constants are not objects, they are references to objects.
  - So that message should try to tell you the name of the missing constant (AKA class path) rather than describe the object that was supposed to hold the constant.

The fact is, that message builder basically does `"#{receiver.inspect}::#{missing_constant_name}"`, that joining `::` indicates that it assumes `receiver.inspect` returned a class path.

----------------------------------------
Feature #16832: Use #name rather than #inspect to build "uninitialized constant" error messages
https://bugs.ruby-lang.org/issues/16832#change-85524

* Author: byroot (Jean Boussier)
* Status: Open
* Priority: Normal
----------------------------------------
While debugging a bug in Rails (https://github.com/rails/rails/pull/37632#issuecomment-623387954) I noticed `NameError` calls `inspect` on the `const_get` receiver to build its error message.

The problem is that some libraries such as Active Record have been redefining `inspect` for years to provide human readable information, e.g.: 

```
>> Shipit::Stack.inspect
=> "Shipit::Stack (call 'Shipit::Stack.connection' to establish a connection)"
>> Shipit::Stack.connection; nil
>> Shipit::Stack.inspect
=> "Shipit::Stack(id: integer, environment: string, ...)"
```

Which makes for fairly unhelpful error messages:

```
>> Shipit::Stack.const_get(:Foo)
Traceback (most recent call last):
        2: from (irb):4
        1: from (irb):4:in `const_get'
NameError (uninitialized constant #<Class:0x00007fc8cadf2dd0>::Foo)
```

So perhaps it's Active Record that is at fault here, but from my understanding since the goal is to display the constant path that was requested, `name` is much more likely to return a relevant constant name.

Proposed patch: https://github.com/ruby/ruby/pull/3080



-- 
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>

In This Thread