|Re: B3 uses the wrong artifact version from the wrong site [message #714063 is a reply to message #713989]
||Tue, 09 August 2011 17:23
|| Thomas Hallgren
Registered: July 2009
The b3 aggregator uses the p2 planner to figure out the best way to create the aggregation. In essence, it performs an
install of all root features that should be mirrored. Hence, you can never "select feature A from site 1". The planner
will find it among all sites in a particular validation set (the concept of validation set was introduced in the 3.7
based release. Prior releases acts as everything lives in one single validation set).
An exclusion rule affects the set of root features that is fed to the planner. That doesn't always help since some other
feature might have a hard requirement on the excluded feature. Unless you also exclude that feature (and any feature
that in turn might include that), you will still see the "excluded" feature in your end result.
You should be able to control which version that gets mirrored if the feature in question is a root feature, and not
included by any other feature that you mirror.
On 2011-08-09 16:23, Carsten Reckord wrote:
> I am trying to build an aggregate from features of multiple sites and am a
> bit confused regarding the result.
> I have two sites containing overlapping sets of features, and if I select a
> feature A from site 1, where it is available in version 1.2.3, what I see in
> my aggregate is actually version 1.3.0 from site 2, where I explicitly added
> an exclusion rule for that feature.
> Is there a reason for this behavior (disregarding from where I include a
> feature and always selecting the highest possible version from all sites)?
> Is there a way to prevent this?
> Best regards,
Powered by FUDForum
. Page generated in 0.01549 seconds