Re: trunk r114545: * lisp/progmodes/ruby-mode.el: Fix recently added tests.

Re: trunk r114545: * lisp/progmodes/ruby-mode.el: Fix recently added tests.

[Dmitry Gutov]

Stefan Monnier <monnier@iro.umontreal.ca> writes:
>  a = foo(j, k) -
> -  bar_tee
> +    bar_tee

Hmm, that indentation looks meaningful, and it's even close to
implementing one of the long-standing feature requests
(http://bugs.ruby-lang.org/issues/5357), but I suspect it'll go against
some established codebases.

I think we'll need a toggle here, eventually.

>  foo.
>    bar
>
> +# FIXME: is this really valid Ruby?  Isn't the newline after "foo" treated as
> +# an implicit semi-colon?
>  foo
>    .bar

It is, I guess it's the main exception to the newline termination rule.

This style is actually preferred over keeping the dot on the first line:

https://github.com/bbatsov/ruby-style-guide (search for "chained method
invocation")

Re: trunk r114545: * lisp/progmodes/ruby-mode.el: Fix recently added tests.

[Stefan Monnier]

>> a = foo(j, k) -
>> -  bar_tee
>> +    bar_tee
> Hmm, that indentation looks meaningful, and it's even close to
> implementing one of the long-standing feature requests
> (http://bugs.ruby-lang.org/issues/5357),

Indeed, it's the indentation style I see in most examples.

>> +# FIXME: is this really valid Ruby?  Isn't the newline after "foo" treated as
>> +# an implicit semi-colon?
>> foo
>> .bar
> It is, I guess it's the main exception to the newline termination rule.

Is the newline termination rule documented somewhere?  All I could find
was discussions about the style, but no definitive clear definition of
when a newline is treated as a semi-colon.

Anyway, you can adjust ruby-smie--implicit-semi-p to reflect the "rule".


        Stefan

Re: trunk r114545: * lisp/progmodes/ruby-mode.el: Fix recently added tests.

[Dmitry Gutov]

On 07.10.2013 18:22, Stefan Monnier wrote:
>>> +# FIXME: is this really valid Ruby?  Isn't the newline after "foo" treated as
>>> +# an implicit semi-colon?
>>> foo
>>> .bar
>> It is, I guess it's the main exception to the newline termination rule.
>
> Is the newline termination rule documented somewhere?  All I could find
> was discussions about the style, but no definitive clear definition of
> when a newline is treated as a semi-colon.

This is the best I'm aware of: 
http://whitequark.org/blog/2013/04/01/ruby-hacking-guide-ch-11-finite-state-lexer/ 
(search for "Line-break handling").

I don't see the exception mentioned.

> Anyway, you can adjust ruby-smie--implicit-semi-p to reflect the "rule".

Done, seems to work.