> Just a heads up about why I'm suddenly pushing so hard on getting > everything converted to JDOM 1.0. > > First off, since we're installing extensions into ant/lib, there is a high > likelihood of breakage unless all components in that library use JDOM 1.0. > (For the record, I've tried twice before to avoid installing hackystat > components into ant/lib and just use the taskdef classpath stuff, but it > doesn't work due to classloader issues: > <http://ant.apache.org/faq.html#delegating-classloader>
I've always used the command 'ant -lib lib\ant' instead of actually installing copying stuff into the ant/lib directory. Since Ant 1.6, the -lib option allows you to specify a directory (or more using 'ant -lib dir1 -lib dir2') to merge with your local Ant installation. If I recall correctly, there are a few projects in Hackystat that don't play nice with this idea, but it's helped keep my Ant directory cleaner for the most part. > Second, I am in the process of teaching a new cohort of advanced software > engineering students to use Hackystat. They have just installed Eclipse, > and I am planning on having them install the Jupiter sensor and do a review > starting on Monday. On Wednesday, I plan to walk through the StackyHack > system (chapter four of the user guide). Thus, there is some urgency to > get everything upgraded and consistent since there are a couple dozen users > being introduced this week. > > So. In the best of all possible worlds, I would like to get JBlanket and > Jupiter upgraded to JDOM 1.0 by early tomorrow morning, so that I can make > a new build and install it on the public server by noon, so that by 1:30pm > I can teach my class, have them update their sensors, and have everything > work just fine. I've updated the only line that seems to have broken with the 1.0 API. There's one test failure, but I've checked it against the current CVS and it looks like that test fails there too. So the newest CVS should be as good as the old, and converted to the JDom 1.0 API. Use it at your peril! :-) I've got company from out of state, so I don't have time to make a more thorough analysis. Sorry! > I'll find out tomorrow if this corresponds to the world that I actually > live in. :-) Good luck, Tim
