ant-ivy-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Jim Adams <Jim.Ad...@sas.com>
Subject Re: Bug in latest-compatible conflict manager ?
Date Thu, 25 Jun 2009 15:28:39 GMT
Isn't that the exact opposite of what that conflict manager is suposed  
to be handling?

Sent from my iPhone

On Jun 25, 2009, at 11:13 AM, "Jarosław Wypychowski" <J.Wypychowski@icm.edu.p 
l> wrote:

> Hello,
>
> I tried to prepare a conflict manager/latest strategy/whatever to
> support a use case in which I want to have latest.integration  
> overriding
> any other revision number.
> During that struggle I tried to understand what exactly does
> LatestCompatibleConflictManager do. In the IVY-2.0.0 (but no changes  
> in
> that code in 2.1.0-RC1 nor svn) in LatestCompatibleConflictManagerTest
> are 2 tests for circular dependencies:
>
>    public void testCompatibilityResolveCircularDependency2() throws
> Exception {
>        fixture
>            .addMD("#A;1->#C;2")
>            .addMD("#C;1->#B;1")
>            .addMD("#C;2->#B;2")
>            .addMD("#C;3->#B;3")
>            .addMD("#B;1->#C;latest.integration") // circular  
> dependency
>            .addMD("#B;2->#C;latest.integration") // circular  
> dependency
>            .addMD("#B;3->#C;latest.integration") // circular  
> dependency
>            .init();
>        resolveAndAssert("#A;1", "#B;2, #C;2");
>    }
>
>    public void testCompatibilityResolveCircularDependency3() throws
> Exception {
>        // same as 2, but A depends on B
>        fixture
>            .addMD("#A;1->#B;2")
>            .addMD("#C;1->#B;1")
>            .addMD("#C;2->#B;2")
>            .addMD("#C;3->#B;3")
>            .addMD("#B;1->#C;latest.integration") // circular  
> dependency
>            .addMD("#B;2->#C;latest.integration") // circular  
> dependency
>            .addMD("#B;3->#C;latest.integration") // circular  
> dependency
>            .init();
>        resolveAndAssert("#A;1", "#B;2, #C;2");
>    }
>
> The first one is intuitive and #C;2 is what I expect to see. But the
> second one is magic.
> We have the following graph:
>
> #A;1->#B;2->C#latest.integration
>
> C#latest.integration doesn't seam to have any dependencies so the rest
> is a garbage IMHO (unless I'm missing something). Could someone please
> enlighten me why the expected result is #C;2 in this test ?
> And secondly - how to achieve the main goal - resolving to
> latest.integration always when it is listed anywhere in the
> dependencies ?
>
> Best Regards,
>
> --jw
>
> -- 
> Jaroslaw Wypychowski, ICM UW, email: jarwyp@icm.edu.pl
>
> KPK Kan. 97 § 2. Małoletni, przed ukończeniem siódmego roku życi 
> a,
> nazywa się dzieckiem i uważany jest za nie posiadającego używania
> rozumu. Po skończonym siódmym roku życia domniemywa się, że posia 
> da
> używanie rozumu.
Mime
View raw message