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

[Cantor, Scott]

On 11/5/13, 9:12 AM, "Ian Young" <ian at iay.org.uk> wrote:
>
>The whole idea makes me uncomfortable enough to suggest that we should
>resolve it by *not* having such circular dependency management. That
>would mean taking spring-extensions and java-support out of the parent
>project and moving them down into the parent POMs for the individual
>projects.
>
>Is this worth discussing on the next dev call?

Yeah, and FWIW I think that's a simple enough fix to make. It also will
decouple our parent POM from being tied to specific versions of our own
utilities, thus reducing the number of times we would need to change it. I
think we should stick to managing only external dependency versions and
not our own.

It's really up to a given IdP POM what version of OpenSAML to pull in,
IMHO.

>I am not, personally, very fond of the tag-from-workspace way of doing
>things, although that's what we document as our procedures. I think
>that's mainly a svn-specific convention and I come from the "check it in
>and then tag it" generation (although that's obviously how Git works,
>too, so everything old is new again).

Not sure I see the difference, though with svn you can do a direct cp of a
particular revision to become the tag.

>I'm a bit surprised that you seem to be losing history, though.

It's that the cp doesn't create a history on the original trunk/branch.
The tag knows its history, but there's no record of the tag in the history
of the revisions the tag was created from.

So if you look at a given source file's history from a branch, for
example, there's no log of which revisions became which tags.

-- Scott