From: James Tucker <jftucker@...>
Date: 2008-01-01T01:41:34+09:00
Subject: Re: Detecting Arrow Keys


On 31 Dec 2007, at 08:34, tho_mica_l wrote:
>
> Isn't readline already distributed with ruby by default? Why write a
> ruby version of a standard ruby library?

Well, for one, it's only the binding that's in the stdlib. The one- 
click-installer ships with a copy, which I think came from GnuWin32.  
The implementations of readline available on windows are broken, though.

Simple replication:

In a command prompt <<EOEXAMPLE
mode con: cols=81
irb
["a", "b", "c", "d", "e", "f", "g", "h", "i", "j", "k", "l", "m", "n",  
"o", "p", "q", "r", "s", "t", "u", "v", "w", "x", "y", "z"]
EOEXAMPLE

This will not display correctly, starting from around about character  
70 in width, IIRC

It can take two or three newlines to clear the display corruption, and  
when trying to go up past the wide line in the history will also cause  
the same issues, making irb quite unusable.

The more core issue, is that there isn't a way to do a "getc" that  
actually doesn't violate POLS. I don't know about others here, but  
personally, I would (if I didn't know better) expect "get character"  
to just get a single character.

N.B. We're not talking about the argv get[cs] POLS violation, there  
are two other problems here, STDIN buffering until "\n" (all  
platforms) and readline being broken (Windows).