Herwin

There is a bit of inconsistent behaviour in this pull request. A snippet of the old/unchanged timeout setter: ```ruby attr_reader :timeout def timeout=(new_timeout) @timeout = new_timeout @http.read_timeout = @timeout @http.open_timeout...

That solution is not 100% backwards compatible, there is still a chance that some code changes the open/read timeout directly on the http object (even though the docs say not...

The "Core classes updates" has two sections named "Fiber". The second one is in between "Symbol" and "Thread". The last item of the second section is "Thread#join invokes the scheduler...

```ruby puts 'hello world!' puts 'hello world!' puts 'hello world!' ``` After implementing the deduplication for strings, I tried this simple and repetitive program. Without sections, the generated bytecode looks...

That should do the trick. The `rodata` section now contains strings and everything symbol-like (things like names of methods and variables). All the deduplication logic is shared based on string...

So, about this 8 years old change... Long story short: the remote repository of this pull request is a company account, Scharrels does not work here anymore, and we've refactored...

@ntikhoa You might be looking at the wrong data. Since your output uses single quotes, this is not the original JSON input, JSON requires double quote. My guess is that...

This was indeed Ruby 2.7, I forgot to mention that. Mocha 1.x on Ruby 2.7 has a few other quirks, the following lines are equivalent: ```ruby object.expects(:expected_method).with(:param1, :param2, foo: 1)...

> Did you find that surprising too or was it the behaviour you expected? If it was, then can you explain why? That is what I tried to explain with...

I guess that would make a backwards incompatible change, which should not happen in a minor release (according to semantic versioning). Also: I'm not sure how well this change would...