From: Caleb Clausen <caleb@...>
Date: 2009-07-22T08:26:35+09:00
Subject: [ruby-core:24489] Re: [Bug #1801] parse error on variable/method collision

Yukihiro Matsumoto wrote:
> Ah, sorry for my weak memory.  I remember the conversation now.
> Now it's a issue of trade off, between compatibility and
> intuitiveness.

While you're trying to decide if this ought to be changed or not, let me 
ask about a similar case. I don't know ruby's parser that well, and I 
can't decide if this is a different manifestation of the same bug (if 
the first is determined to be a bug, that is), or just a highly similar 
bug, maybe you or someone else here can enlighten me:

   p = p m %(2)

I think the %(2) should be treated as a string, since it is in all other 
cases I know of (unless m is an lvar, which obviously it isn't here). 
However, it gets parsed as an operator and a parenthesized numeric literal.

In this variation of the expression, %(2) is a string, as I expect:

   p m %(2)

Is this the same bug (if it is to be a bug)? Should I file a new report?

This expression works as I expected for versions 1.4.0 thru 1.7.1. All 
versions after 1.7.1 don't work the way I expect. In versions before 
1.4.0 it causes a syntax error. Jruby has identical behavior as recent 
mri rubies. In rubinius, it causes a stack dump.

Something interesting happened around about 1.7.1.

(ruby-versions.net sure is useful. I've used it 3 times today.)

I showed you this expression before as well; you said then that this one 
is a bug too.

I recall that when I showed these examples to you and Charles Nutter, 
both of you then went on to come up with a series of more elaborate 
cases which exhibit.... interesting behavior. I couldn't follow what 
either of you were doing after the first couple of passes, but I wish I 
had those more elaborate examples so I could add them to my test suite.