Thinking about it, given that Maven actively dislikes tasks not contained within its own project directory, our having dropped xalan-c, and that by today's standards the bandwidth and storage of always getting both is not much of a difference from just fetching the code project, I am becoming more receptive to the idea of making xalan-test a subproject of xalan-java and having the old top-level xalan-test phased out.
I am not convinced I like it, but I'm less convinced that I dislike it. My remaining objections are architectural aesthetics, and I am not sure anyone else cares or should care. As long as we retain the ability to run the test suite against other processors, publishing it inside our processor's box is inelegant but not necessarily unacceptable. The downside is that we would lose the ability to easily collaborate on maintaining this tool with folks developing other processors. But realistically, we haven't been seeing a lot of community activity for a very long time, so... It's half an idea, and a half an hour I may decide I'm being a half wit. But I thought it worth documenting where my head is currently at. Just in case I lose it again. -- /_ Joe Kesselman (he/him/his) -/ _) My Alexa skill for New Music/New Sounds fans: / https://www.amazon.com/dp/B09WJ3H657/ Caveat: Opinionated old geezer with overcompensated writer's block. May be redundant, verbose, prolix, sesquipedalian, didactic, officious, or redundant.
