a note regarding net.shibboleth:parent-v3:2

[Cantor, Scott]

On 11/5/13, 1:35 PM, "Tom Zeller" <tzeller at dragonacea.biz> wrote:
>
>Sounds like a good idea. Maybe I should have done that instead, oh well.

Not at the time I suspect, but having seen the chaos now, I think we can
make it better.

And now that I'm looking, it's really just the two you ran into. We
already don't set the OpenSAML version anywhere but idp-parent:

 <properties>
        
<svn.relative.location>java-identity-provider</svn.relative.location>
        <opensaml.groupId>org.opensaml</opensaml.groupId>
        <opensaml.version>3.0-SNAPSHOT</opensaml.version>
    </properties>

So I think it's all good, we just need to move the version for
java-support down into opensaml-parent, and the spring-extensions into the
IdP and anything else using it.

Seems pretty minimal.


>Someone should work through this idea, hopefully it will be trivial. I
>will unless someone else is already there, just let me know to avoid
>duplication of effort.

Just need to verify where every version is being managed from, but it
looks like it's mostly right already.

>>>The difference, to me, is the loss of history on the thing copied from.

I wasn't clear on Ian's terms, but I see by check-in and tag he just means
a tag that's part of the actual artifact, and not a copy that loses
history.

>Which is what I consider to be a loss of history. Not really, not
>technically, but I just wanted the history of trunk to include "Tag as
>version X", so that history is "linear", meaning you only have to look
>in one place, one thread, by which I mean trunk. The "Tag as version
>X" comment should exist both on the thing copied from and to, IMHO,
>for sanity.

I agree.

-- Scott