Feature #17393: `Ractor::Moved#inspect`

• Status: Assigned
• Assignee: ko1 (Koichi Sasada)
• Target version: -

[marcandre (Marc-Andre Lafortune)]

It could be helpful to define Ractor::Moved#inspect and output the source location of when the data was moved. If preferred, it could raise an error with this information:

x = []
Ractor.new{ receive }.send(x, move: true)
p x # => "Data was moved in `example.rb:4`"
# or
p x # => "Data was moved in `example.rb:4`" (Ractor::MovedError)

Also @zverok (Victor Shepelev) and myself were wondering if there was a technical reason to freeze Ractor::Moved? If not, is it only to "force" people to use refinements (which are allowed on frozen classes)? It's already known that it is in general a bad idea to modify builtin classes, so it's not clear to me that freezing that class is best.

[marcandre (Marc-Andre Lafortune)]

Maybe refining the message from method_missing is the best and simplest. (giving info that object was moved, and where).
I would also unfreeze the class unless there's a good reason not to.

[marcandre (Marc-Andre Lafortune)]

• Assignee changed from matz (Yukihiro Matsumoto) to ko1 (Koichi Sasada)

[marcandre (Marc-Andre Lafortune)]

Discussed with ko1, who said he is afraid that unfreezing the class might create issues, as implementation is quite special. Possible implementation of #inspect for Ruby 3.1

Two things to keep in mind:

• usecase for unfreezing the class would be to implement #inspect in pure Ruby for 3.0, for example.
• it is possible to refine frozen classes, including Ractor::Moved, so freezing that class may not have the desired effect?

[ko1 (Koichi Sasada)]

I don't think providing #inspect is good idea because any methods should respond to the moved object.
Otherwise there is possibility to ignore the accessing moved objects.

I prefer to show the moved location in the error message.

x = []
Ractor.new{ receive }.send(x, move: true)
x.size

...:in `method_missing': can not send any methods to a moved object (moved at foo.rb:123) (Ractor::MovedError)

[marcandre (Marc-Andre Lafortune)]

Yes, that looks good, that is what I was proposing in note 1 too 👍

[ko1 (Koichi Sasada)]

Try it on 3.1, not 3.0.

[hsbt (Hiroshi SHIBATA)]

• Status changed from Open to Assigned