[ 
https://issues.apache.org/jira/browse/MAHOUT-768?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13070230#comment-13070230
 ] 

Sean Owen commented on MAHOUT-768:
----------------------------------

There's a third option, perhaps: put collections back into math. I suggest it 
since 'collections' is rather stranded as a separate release. The idea was to 
make it more reusable apart from Mahout, but it isn't for reasons like this.

I'd prefer not to move all this out of math as it makes some sense there, I 
think that was the intent, and that's where people are looking for it. Really I 
mean I prefer the simplest solution.

> Duplicated DoubleFunction in mahout and mahout-collections (mahout.math 
> package).
> ---------------------------------------------------------------------------------
>
>                 Key: MAHOUT-768
>                 URL: https://issues.apache.org/jira/browse/MAHOUT-768
>             Project: Mahout
>          Issue Type: Improvement
>          Components: collections, Math
>    Affects Versions: 0.5, collections-1.0
>            Reporter: Dawid Weiss
>            Priority: Minor
>             Fix For: 0.6
>
>
> DoubleFunction is duplicated in Mahout Math and Collections. There are also 
> other things inside m.math.* package that are kept there to avoid circular 
> dependencies... Simply removing DoubleFunction from collections is not going 
> to work because it is needed for compilation (again, circular dependency 
> between collections and math). I see two solutions:
> 1) extract the common definitions inside math.function.* into a separate 
> module. This is a clean solution, but obviously scatters the code even 
> further.
> 2) create a compilation-time, optional dependency on mahout 0.5 in 
> collections, remove the entire mahout.math.* subpackage from collections and 
> live with this. I don't know how Maven handles circular dependencies of this 
> type:
> collections [trunk] -> [optional, required at build time] mahout.math 0.5
> mahout.math [trunk] -> collections [trunk]
> it seems tricky and error-prone, but should also work.
> Looking for other ideas of cleaning this mess up, of course.
> Dawid

--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira

        

Reply via email to