From: James Gray <james@...>
Date: 2008-08-16T06:54:41+09:00
Subject: Re: Wig-Wug Simulator

On Aug 15, 2008, at 4:17 PM, Chris Shea wrote:

> On Aug 15, 2:20 pm, James Gray <ja...@grayproductions.net> wrote:
>> On Aug 15, 2008, at 11:12 AM, Chris Shea wrote:
>>
>>
>>
>>> On Aug 14, 9:19 pm, James Gray <ja...@grayproductions.net> wrote:
>>>> I have started a Wig-Wug (http://therubyist.com/homework/wig-wug/)
>>>> simulator and thrown it on github (http://github.com/JEG2/wig-wug/tree/master
>>>> ).  It probably still has issues, but it may be of interest to  
>>>> others
>>>> playing with the problem.  It includes a map and a digger class I
>>>> wrote that does badly on that map.  Enjoy.
>>
>>>> James Edward Gray II
>>
>>> Sorry for the empty message. Hit send WAY too early.
>>
>>> I mean to send this patch, which lets your game server not crash  
>>> if a
>>> player ends up on the edge of the map:
>>
>>> 91a92
>>>>      return 'E' if x == -1 || y == -1 || x == @width || y ==  
>>>> @height
>>
>> Interesting.  Is this a patch to Map#[]?  I wouldn't think it would  
>> be
>> needed, due to lines 46-50 of game.rb.  Am I missing something?
>>
>> James Edward Gray II
>
> Yes. That was for Map#[]. But taking that out and going back to
> game.rb, I added "pp surrounding" after surrounding gets settled, and
> you can see that there's something up:
>
>  $ ./bin/wig-wug -w 2 -h 2 ~/Desktop/bots/digger.rb     [15:06:44]
>  FO
>  OR
>  Surprise:  Starts at x=1, y=0.
>
>  Turn 1
>  [["E", "E", "E"], ["F", "P", nil], ["O", "R", nil]]
>  Surprise:  (down) Reached the ruby.
>
>  Result:  Surprise won.
>
> I think the issue is that Comparable#between? includes both endpoints,
> so if a bot is on the bottom edge of the map, the Map ends up asking
> for points not on the map. And those will be nil if the player's on
> the right edge, and will raise a NoMethodError for nil[] if it's on
> the bottom edge.
>
> Subtracting one from width and height when they appear as arguments to
> #between? on line 46 of game.rb should actually take care of it better
> than what I suggested earlier.
>
> Or am I missing something now?

I think you are exactly right.  I've committed the fix.  Thanks!

James Edward Gray II