> I've just committed a couple of changes to the sources. First of I
> renamed getTargetUrls() to getBlockUrls() because the word "targets" is
> ambigouse with respect to usage of the term in the configuration targets
> context. Secondly I updated String getRemoteRepositoryUrl() to be
> String[] getRemoteRepositoryUrls() because there can be multiple remote
> fallback repositories declared to the local repository.
Sounds good and it makes sense.
> Concerning the approach on aquisition of block urls (which I copied over
> for remote host urls) .. the current code checks for system properties
> beginning with BLOCK_URLS_BASE (was TARGET_URL_BASE). I'm thinking
> around with an alternative approach of just a single property and
> treating it as a comma delimited enumeration. I think this would be
> easier to override for users at the environment or property level.
Yes you are right about it being much easier with one property. I just did not want
to get into problems with URL escape sequences and the delimiter. I like one property
myself. Is there a chance for a URL to have an escaped comma in it?
> I'm going to tweak some of the property names and setup some default
> values in the merlin.properties file. Finally, two things I think are
Yeah I left it barren for now. I'm going to start looking into writing a
DefaultLoader for the Kernel today that sets up a kernel using a KernelConfig. Next I
think we can retro fit this into the embedded kernel in the kernel-unit subproject as
a test run. Basically this guy replaces the user repo and system repo using maven
properties and paths after defaults are found.
BTW, I think we should retrofit the mechanism to all things that fire up a Kernel and
not just an embedded scenario. You can look at every situation as an embedded
situation even the CLI program.
> still open - (a) we cannot assume that ${user.home} is non-null. The
> are situations where a null value can be returned in a servlet
I never knew that user.home with a servlet env would be null - you never stop
learning. I thought it would be the home of the user who started the JVM.
> environment. Second item - we need to be able to override the location
> of the user repository at the environment level.
Could you elaborate what you mean about the 'environment level'. I'm thinking you
mean the shell env variable for the user repo will be available. Is this what you
meant?
Cheers,
Alex
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
