Raymond Augé commented on FELIX-5782:

> 1. You now eagerly create ResolutionExceptions during all phases of 
> resolution.  I know as part of a performance improvement Guillaume did was to 
> avoid this.  I would prefer to keep the usage ResolutionError unchanged in 
> the resolver implementation and perhaps make ResolutionError.toException() be 
> abstract such that each ResolutionError implementation can return specific 
> ResolutionException types when we go to throw the exception.

I'll try this.

> 2. I would like to see some javadoc on these specific exception types now 
> since you seem to be introducing new API of sorts here.

Sure thing, I just didn't want to do all that work if there wasn't consensus.

> allow resolver errors to be introspected
> ----------------------------------------
>                 Key: FELIX-5782
>                 URL: https://issues.apache.org/jira/browse/FELIX-5782
>             Project: Felix
>          Issue Type: Improvement
>          Components: Resolver
>            Reporter: Raymond Augé
>            Assignee: Thomas Watson
>            Priority: Minor
> The current model for resolver errors does not provide any means of 
> introspecting deeper knowledge that the resolver gained. The information is 
> internal only, which makes user feedback more difficult to produce than 
> necessary.
> I propose to expose the error types by means of an exported package 
> {{org.apache.felix.resolver.error}}. This will allow interested clients to 
> dig more deeply into the reasons for resolution failure in order to provide 
> better feedback to users.

This message was sent by Atlassian JIRA

Reply via email to