Re: [osgi-users] Possible mistake in Feature spec?

[Mark Raynsford <org.eclipse@xxxxxxxx>]

On 2022-05-25T17:01:50 +0100
David Bosschaert <david.bosschaert@xxxxxxxxx> wrote:

> Hi Mark,
> 
> Features can be complete which means that they provide all their
> dependencies, or they can be non-complete which means that they have
> requirements and imports that are not met by the feature itself. In the
> latter case the feature depends on another feature to provide these
> capabilities.
> 
> So I think the sentence is fine. The dependency is not a 'require-bundle'
> type dependency where another feature is required by id or name, its a
> declarative requirement-based dependency based on what the bundles, and
> possibly the feature itself declare as their requirements and which are not
> addressed by the features constituents.

Hello!

I've got to admit... I'm not completely clear on this. If I'm
understanding correctly, a Feature may be "incomplete" in that it may
have requirements that aren't met by the set of bundles listed in the
Feature itself. 

I'm not seeing anywhere in the spec that allows the Feature itself to
declare extra requirements or capabilities, so I'm interpreting this to
mean that there may be, for example, a feature F that contains a set of
bundles R where some bundle B in R may have requirements that aren't
satisfied by any of the other bundles in R. The implication would be
then that there might be some other feature G somewhere else that would
fill in the remaining requirements of F. I'm not sure how anyone would
know how to find G.

If that is the right interpretation, then I think the sentence a "a
feature may depend on another feature" is a pretty loose interpretation.

Is there some way a Feature itself can define extra requirements or
capabilities that I'm not seeing?

-- 
Mark Raynsford | https://www.io7m.com