Maybe we can broaden "relative path" semantics therefore it may not
necessarily be mandatory any longer to introduce "scdlResource".

e.g. the ClassPath is dir1, dir2, jar1, jar2.
Even if a resolved relative path is dir1/com/example/included.scdl,
we can also search for dir2/com/example/included.scdl and
jar2!/com/example/included.scdl.
Similarly evev if a resolved relative path is jar2!/org/tuscant/system.scdl,
we can search for jar1!/org/tuscant/system.scdl and
dir2/org/tuscant/system.scdl

On 9/6/06, Jeremy Boynes <[EMAIL PROTECTED]> wrote:

The spec says that <include> takes the name of the composite to be
included but does not say how it should be located. We currently
support a scdlLocation attribute that takes a URL, with relative URLs
being resolved against the location of the containing scdl.

However, it would be nice to allow users to place SCDL somewhere on
the composite's classpath. I think we should also support a
scdlResource attribute that takes the name of a Java resource that
will be located using the same classpath as we use for Java classes
(i.e. the same classpath used to locate the class in a
<implementation.java> or <interface.java>).

One thing this would let us do is publish scdl fragments in the
kernel for things like the set of loader components (the current
loader.scdl) that can then be reused by the various hosts; this would
get rid of the duplicates and the need to maintain them. This will
help with the webapp host I am working on so I plan to add the
scdlResouce attribute soon.

The same would also apply to things like <implementation.composite>

--
Jeremy

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




--

Yang ZHONG

Reply via email to