Further clarifying after face-to-face with John: I'm proposing, for all
modules touching a given java package, *union over modules of (<includes/> -
<excludes/>)*

it'd be nice to find an optimized expression that reduced the list away, but
the obvious *union(<includes/>) - intersection(<excludes/>)* doesn't do
that.



On Mon, Apr 12, 2010 at 1:53 PM, Freeland Abbott <[email protected]> wrote:

>
>
> On Mon, Apr 12, 2010 at 12:10 PM, John Tamplin <[email protected]> wrote:
>
>>
>>
> How would you handle the following:
>>
>> A: include foo*.java, exclude b*.java
>> B: include bar*.java, exclude foobaz.java
>>
>> I am assuming your union would be the following:
>>
>>    - all foo*.java
>>    - all bar*.java
>>
>> Is that correct?
>>
>
> Close, but it'd actually include all foo*.java and all b*.java, not just
> bar*.java.  But you probably knew that and just typo'ed....
>
>
> It seems like maybe you would want to merge in the excludes as well so that
>> the most specific rule applies, which would remove foobaz.java from the
>> list.
>>
>
> No, I think that's wrong.  A wants foobaz.java, so if B takes it away, then
> it breaks A, which my app needed.  So either I should remove the inherits of
> A (in which case the problem goes away 'cause I didn't need A after all), or
> I need GWT to see the union of anything passed by *either* A *or* B.  B
> can't yank the rug out from under A.
>
> This also means I don't believe in the "foobaz is toxic to me, keep it away
> at all costs" storyline for <excludes/>.  I instead read <source
> ...><excludes/></source> as "I need ... but don't care about  ...," which
> leaves the option for some other module A to pull in the files that it needs
> but B didn't care about.
>
> Find me a counter-example and I'll change my mind, but I think the "toxic"
> interpretation of <excludes/> introduces a new error case (B removing what A
> wants) without solving an actual problem (if foobaz is toxic, it's toxic to
> everybody, not just B, and A can have no legitimate usecase for it).
>

-- 
http://groups.google.com/group/Google-Web-Toolkit-Contributors

Reply via email to