Re: [Bibdesk-develop] BibItem isEqual

[Maxwell, A. R <ada...@pn...>]

On Sep 21, 2005, at 17:38, Michael McCracken wrote:

>
> On Sep 21, 2005, at 5:27 PM, Maxwell, Adam R wrote:
>
>> I discovered an interesting problem while debugging Stefano's  
>> problem generating cite keys:
>>
>> For adding to the set of search results, NSSet compares using - 
>> hash and isEqual:; if all cite keys are identical, it will call  
>> isEqual for every one of them.
>>
>> - (BOOL)isEqual:(BibItem *)aBI{
>>     return [[self  
>> bibTeXStringUnexpandedAndDeTeXifiedWithoutInternalFields]  
>> isEqualToString:[aBI  
>> bibTeXStringUnexpandedAndDeTeXifiedWithoutInternalFields]];
>> }
>>
>> With that comparison, these two items (I think) compared as  
>> equal!  I think the problem is that @other types don't have any  
>> keys defined, so the actual strings being compared were "@other 
>> {FixMe}" and "@other{FixMe}".  What should we do in this case?  My  
>> thought is that if [knownKeys count] == 0, we probably ought to  
>> just include all of the keys.
>
> That sounds like the only way to get a meaningful comparison in  
> that case, so I'd agree with it...

Also, if we're dropping keys, I think we should drop the ones from  
default fields also ([[OFPreferenceWrapper sharedPreferenceWrapper]  
stringArrayForKey:BDSKDefaultFieldsKey]), or else [knownKeys count]  
is never zero and my idea won't work.

>
> I'm a little confused, though - why are we comparing without  
> internal fields?

Well, see our last discussion of this on May 13 2005 for background.   
I guess my thought was what Local-Url &co don't really affect the  
equality of the item; in other words,

@article{myArticle,
author = "me",
title = "foo",
Date-Modified = "some date"
}

isn't different from

@article{myArticle,
author = "me",
title = "foo",
Date-Modified = "some other date"
}

except for the Date-Modified, which is BibDesk metadata.

> It seems like isEqual should compare the entire contents of the  
> item, but I guess if you told me we never actually need to do that,  
> I wouldn't be too shocked.

I won't say that; it depends on the context in which you're using it,  
I guess.

Adam