[Mondrian] Re: Conversion from Member[] to List<Member>

Luis F. Canals luis.canals at stratebi.com
Tue Dec 11 04:26:07 EST 2007


Luis F. Canals wrote:
> I'm with you: changing to high cardinality IS destabilizing the code, 
> and thanks to current test (thanks God for holy Test Driven 
> Delopment!) I can advance step by step...
>
> The idea is to mark a dimension as "highCardinality" and to be managed 
> in a little different way, avoiding to iterate over all members for 
> intermediate operations.
>
> Best regards!
>
> - luis
>
> Julian Hyde wrote:
>>> Luis wrote:
>>>
>>> as we are working on high cardinality over the changes we've sent 
>>> you, instead of generating a patch firstly for the change from 
>>> Membe[] to List<Member> and then for the high cardinality support, I 
>>> propose you to wait for the next improvement and then make all 
>>> change once (including your comments to improve the code).
>>>     
>>
>> My concern is that the high cardinality change will destabilize the 
>> code. If
>> there is any chance that the the change is destabilizing, it would be 
>> better
>> to make the API change as a separate edit.
>>
>> It's your call. If you think that we can make both changes without
>> destabilizing the code, let's do it. But as Colin Powell said, "You 
>> break
>> it, you own it." :)
>>
>> Would it help if I created a perforce account for your committer? 
>> Then at
>> least you could apply the changes to the head revision. I am willing to
>> merge your changes with the head revision, but it won't be trivial 
>> because a
>> couple of large changes have gone in since 2.4.
>>
>> Julian
>>
>>   
>




More information about the Mondrian mailing list