Hello,
Going through section 3.6.4 of the OSGi specification (R4.1), about
package constraints, I noticed something in the example depicted in
Figure 3.17.
According to the description before that (p. 44), bundle A imports a
package q, and exports a package p which uses packages q and r.
Package r is not exported or imported, but it is drawn as an exported
package in Figure 3.17. This could mean several things, which are not
clear from the surrounding text.
1. Including r in both the text and figure is a mistake, and should
have been left out altogether.
2. r is really an exported package (although it is not used in the
example), but has been left out in the textual representation.
3. r is actually a private package, which is used by q. This would
imply that no bundle can be wired to p when it imports (or embeds) r.
In short, I believe that either information is missing that states
what should be done with uses-constraints targeting packages which
are not imported or exported, or the example contains an error.
In either case, it would be useful to state either that uses-
constraints can only target imported or exported packages, or what a
uses-constraint targeting a non-imported or -exported package means.
Thanks in advance,
Angelo van der Sijpt
_______________________________________________
OSGi Developer Mail List
[email protected]
http://www2.osgi.org/mailman/listinfo/osgi-dev