kuul. ügyes vagy.

--g
--
Dr. Gulyás, László
igazgató / director
Intelligens Alkalmazások és Webszolgáltatások /
Intelligent Applications and Web Services
AITIA International, Budapest
----- Original Message ----- 
From: "Richard O. Legendi" <[log in to unmask]>
To: <[log in to unmask]>
Sent: Tuesday, December 06, 2011 9:43 AM
Subject: Re: [MASON-INTEREST-L] Patch for minor issues


> Hi Sean,
>
> Thanks for the reply and applying the patch! It's good to hear some of 
> them were useful.
>
> If I find anything in the future, I'll send you a mail about it.
>
> Best,
> Richard
> -- 
> Richard O. Legendi
> Software developer
> Intelligent Applications and Web Services
> AITIA International, Inc.
> http://people.inf.elte.hu/legendi/
>
> On 2011.12.05. 17:36, Sean Luke wrote:
>> On Dec 5, 2011, at 9:26 AM, Richard O. Legendi wrote:
>>
>>> I downloaded the latest source of MASON and made some minor 
>>> modifications on the code (some errors caused by typos, created but not 
>>> thrown exceptions, some NPEs, corrected error messages and alike).
>>
>> Thanks for these, Richard!  There are a few that I disagreed with, but 
>> all in all a solid collection of bugfixes, I've committed nearly all of 
>> them.  See below for responses.  For future bug fixes, feel free to mail 
>> me directly.
>>
>> Sean
>>
>>
>> - this.indexTo = indexTo;
>> Fixed.
>>
>> - if compareTo() returns 0 then equals() should return true as well
>> I'm bleary-eyed this morning (spent all night up with kids) but this is, 
>> I think, another example of surprisingly bad reasoning on Sun's part. 
>> compareTo is largely used for sorting.  equals is largely used for 
>> hashing.  Edges are not unique: they must hash by pointer, but we 
>> *always* want to compare them by value for sorting.
>>
>> - if (obj instanceof Key)
>> Agreed, a good safety fix.
>>
>> - if (min==max)
>> I think changing that from an elseif breaks the logScale check below it 
>> for min>  max.
>>
>> - this.frameRate = frameRate
>> Fixed.
>>
>> - public int compareTo(Object o) // See comment for 
>> sim.field.network.Edge
>> In this case I think I'd prefer that sleeping dogs lie: this is a private 
>> object.
>>
>> - import org.jfree.chart.renderer.xy.XYLineAndShapeRenderer;
>> I did a few tweaks of the imports.  Note that MASON traditionally uses * 
>> rather than enumerating every single import: we code much of this by 
>> hand, not using Eclipse, and enumeration is extraordinarily painful.
>>
>> - super.setName(val);
>> Fixed.
>>
>> - target.printStackTrace(stream);
>> Fixed.
>>
>> - if (lambda<= 0.0) throw new IllegalArgumentException();
>> Fixed.
>>
>> - private static final long serialVersionUID = 1;
>> Fixed in a different way already.
>>
>> - static final long serialVersionUID = 9115981605874680023L;
>> Fixed already.
>>
>> - private static final long serialVersionUID = 1;
>> Actually the issue is that KMeansEngine is not Serializable (it should 
>> have been)
>>
>> - Boolean[] usable;
>> Fixed (I think).  Now creating new Boolean values rather than using the 
>> standard constants.
>>
>> - !scaledFont.getFamily().equals(labelFont.getFamily())
>> Updated for clarity.  But in fact the original works: strings are hashed 
>> constants.
>