From: "Wolfgang Nádasi-Donner" <ed.odanow@...>
Date: 2007-11-25T07:54:56+09:00
Subject: Re: retry semantics changed

I want to inform you about two interesting reactions
to my questions in the German Ruby forum.

I presented the code...

class Otto
   def a
     puts "In 'a'"
     self
   end
   def b
     puts "In 'b'"
     self
   end
   def c
     puts "In 'c'"
     yield
   end
end

p = 2
(puts "before 'Otto'"; Otto)
   .new
   .a
   .b
   .c{retry if (p-=1) > 0}

...which produces...

before 'Otto'
In 'a'
In 'b'
In 'c'
before 'Otto'
In 'a'
In 'b'
In 'c'

The real interesting reactions were:

1) Nobody was confused or surprized about the new syntax,
    which allows these method call chains written with the
    dot at the beginning of a line (tere were a lot of
    discussion in a different thread).

2) The real surprize was, that not only method "c" is
    evaluated again, but the rest of the chain before, too.
    It was recognized as very surprizing, that a retry in
    relation with method "c" effects all already finished
    calls and calculations.

I think both intermediate results are of interest.

Wolfgang N��dasi-Donner