[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index] [List Home]
Re: [p2-dev] Dependency management synergies

Hi Oleg,

Oleg Gusakov a Ãcrit :
> To introduce myself - I lead the Mercury project
> (http://docs.codehaus.org/display/MAVEN/Mercury,
> http://blogs.sonatype.com/people/author/oleg/), in which we try to
> clearly externalize Maven dependency management into a separate library,
> consumable by any client, not necessarily Maven.
> And the nature of the project calls for generalization - after all we
> all are trying to solve the same problem - calculate dependency graph,
> given a set of requirements and a set of repositories, that can satisfy
> those requirements. It really does not matter whether we calculate java
> or OSGi dependencies because all boils down to creating a system of
> linear inequations and solving it, in both cases with SAT4J.
> Why don't we try to unite our efforts and produce a library that all of
> can use: express our requirements in the unified intermediate language
> and process these statements with this library to produce a solution.
> Which could then be interpreted into either environment.
> The mancoosi project - http://mancoosi.org as one of it's goals - states
> "Develop better algorithms and tools to plan upgrade paths based on
> various information sources about software packages and on optimization
> criteria". Why don't we try similar approach in the java world?
> Please share what you think about it. Pascal ? :)
> On the side issue: P2 uses an indirect way of creating those inequations
> - creates a file and then feeds it to solver. In Mercury we took a
> different approach and bypass file stage, feeding expressions directly
> to the solver, which is more efficient. If you'd like we can work
> together to fix that.

The code of p2 available in the branch "explanation_work" does use
direct calls to SAT4J API. This was needed to gather explanations from
SAT4J when no solution can be found.

All the translation is done through a Facade called "DependencyHelper".

That class allows to express constraints with domain objects without
worrying about the details of the SAT encoding.

You need the code of SAT4J HEAD to get that class (upcoming release 2.1).


             Daniel Le Berre mailto:leberre@xxxxxxxxxxxxxxxxxxx
             MCF,    CRIL-CNRS UMR 8188,    Universite d'Artois