It's not exactly a side effect, but I see why its a bug. The collector
can't select the repository as it doesn't know where it will be present.

The current logic is based on the use of snapshots - it considered that
releases would only be in one so the first found is used, whereas for
snapshots all must be checked and the newest used. I believe it works
for the latter.

It seems to be falling down only for version ranges, correct? I had
thought these should also operate as a transformation but that doesn't
sit with the whole resolution technique.

The actual version selection should be assigning a single repository to
it where the correct version was found.

This area of the code needs a lot more tests, in the order of what the
collector itself has. I suspect doing so will probably yield a lot of
clues for a better design also.

- Brett

Garrett Conaty wrote:
> Thought it was appropriate to bring my comments out of the jira issue
> for discussion on the list.
> 
> Basically, I want to work on a more elegant fix for this issue but am
> requesting some community guidance.  Specifically for the issue:
> 
> 
> 1) Should the DefaultArtifactResolver really use
> artifact.getRepository() exclusively if it's not null?  Perhaps the
> Artifact really ought to contain a list of repositories that are
> acceptable (from the transformation phase) from which to try.  This may
> be a good enhancment.
> 
> but the more pressing issue is
> 2) Shouldn't the DefaultArtifactCollector actually do the repository
> selection, not have it be a side effect of getting the metadata.
> 
> Thanks,
> Garrett
> 
> 
> 
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: [EMAIL PROTECTED]
> For additional commands, e-mail: [EMAIL PROTECTED]
> 

---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]

Reply via email to