[10:49:28 CST(-0600)] <drewwills> Are you there EricDalquist?
[10:50:01 CST(-0600)] <drewwills> I'm kicking the tires of pull #113 https://github.com/Jasig/uPortal/pull/113/files (bundled calendar portlet)
[10:50:43 CST(-0600)] <drewwills> The code changes (imho) look pretty elegant, and the whole thing works as expected
[10:51:14 CST(-0600)] <drewwills> the one issue I have is the number of extra build cycles in portlets, and the noise and time it adds to the build
[10:52:07 CST(-0600)] <drewwills> because both 'db-hibernate-portlets' and 'portlet-data-import' targets call $mvn package on up-portlets-overlay
[10:52:17 CST(-0600)] <EricDalquist> uhg
[10:52:25 CST(-0600)] <EricDalquist> yeah that would be good to minimize
[10:52:40 CST(-0600)] <drewwills> so an $ant initportal ends up packaging them 3 times
[10:52:43 CST(-0600)] <EricDalquist> also I've merged the work that james did into <plugin>
[10:52:43 CST(-0600)] <EricDalquist> <groupId>org.jasig.portal.maven</groupId>
[10:52:43 CST(-0600)] <EricDalquist> <artifactId>uportal-maven-plugin</artifactId>
[10:52:43 CST(-0600)] <EricDalquist> <version>1.0.0</version>
[10:52:43 CST(-0600)] <EricDalquist> <!-- extensions must be set to true to pick up the custom phases -->
[10:52:43 CST(-0600)] <EricDalquist> <extensions>true</extensions>
[10:52:43 CST(-0600)] <EricDalquist> </plugin>
[10:52:47 CST(-0600)] <EricDalquist> yeah
[10:52:51 CST(-0600)] <EricDalquist> thats not great
[10:53:51 CST(-0600)] <drewwills> we could (1) skip rnning the package goal in these new tasks, or (2) somehow detect we're running within an initportal (or initdb) and skip it because we've just run it
[10:56:03 CST(-0600)] <drewwills> yep – we can use parameters in the <antcall> to shut off the package if we're coming from an aggregating task (like initportal): http://ant.apache.org/manual/index.html
[10:56:32 CST(-0600)] <EricDalquist> sounds good
[11:39:07 CST(-0600)] <drewwills> Got it worked out EricDalquist and it has the desired effect...
[11:39:14 CST(-0600)] <EricDalquist> cool
[11:39:21 CST(-0600)] <drewwills> but looking at the build logs, i find 1 thing surprising
[11:39:39 CST(-0600)] <drewwills> the portlet imports seem to be happening BEFORE the portal imports
[11:40:23 CST(-0600)] <drewwills> snippet: https://gist.github.com/drewwills/5058581
[11:40:40 CST(-0600)] <drewwills> portlet imports kick off after portlet-data-import:
[11:41:33 CST(-0600)] <drewwills> and after them, there's a up-shell: sectiuon, that starts the portal data
[11:41:51 CST(-0600)] <EricDalquist> yeah
[11:41:57 CST(-0600)] <EricDalquist> that is because the uportal data import is done via ant
[11:42:10 CST(-0600)] <EricDalquist> you'd have to switch the uportal data import over to use maven as well
[11:42:15 CST(-0600)] <EricDalquist> if you want it to run first
[11:42:17 CST(-0600)] <drewwills> complete log: https://gist.github.com/drewwills/5058590
[11:42:32 CST(-0600)] <drewwills> at this point i don't believe it matters
[11:43:23 CST(-0600)] <EricDalquist> not really
[11:43:35 CST(-0600)] <EricDalquist> and for 4.1 we can look at actually switching it over to use maven for data import
[11:43:42 CST(-0600)] <drewwills> sounds good
[11:44:14 CST(-0600)] <drewwills> this is really slick... i'm very pleased to see it emerge