From: "John Lam (CLR)" <jflam@...>
Date: 2007-07-07T16:00:35+09:00
Subject: Re: Array#values_at bug?

There's also similar behavior in other range-like entities in strings:

"foo"[3,0] == ""

But

"foo"[4,0] == nil

Do folks depend on this kind of behavior? I'll need to add a special case to my string implementation to replicate this behavior in IronRuby. Or should I do the correct thing and make "foo"[3,0] == nil?

Thanks,
-John


-----Original Message-----
From: yuanyi zhang [mailto:zhangyuanyi@gmail.com]
Sent: Tuesday, July 03, 2007 6:07 PM
To: ruby-core@ruby-lang.org
Subject: Re: Array#values_at bug?

2007/7/3, John Lam (CLR) <jflam@microsoft.com>:
>
>
>
>
> a = [1,2,3,4]
>
>
>
> a[2..4].should == [3,4]
>
>
>
> but
>
>
>
> a.values_at(2..4).should == [3,4,nil]
>
>
>
> The latter behavior does not feel consistent at all, especially since
>
>
>
> a.values_at(2..5).should == [3,4,nil]
>
>
>
> IronRuby right now just calls slice() in a loop to implement values_at. Is
> this what MRI does?
>
>

Array is a built-in class in MRI, and is implemented by C. You can
find the source in array.c, it's called rb_values_at.

The reason why a.values_at(2..5) == a.values_at(2..4) is because
rb_range_beg_len consider range(2..4) and range(2..5) has same len.
You can find rb_range_beg_len in Range.c:

if (end > len) end = len;

This sentence guarantee Range end won't be greater then Array length,
But I think it should let end = len-1, then a.values_at(2..4).should
== [3,4]

Hope this help.

BR

Yuanyi Zhang