I've got a script to track this at https://github.com/bgrins/xul-analysis. As best as I can tell, today we have 1406 test xul files and 129 non-test files.
We've built out initial top-level HTML window support (which will result in eventually shipping browser.xhtml in bug 1533881). Separately, we can start migrating the existing XUL test documents that don't test actual XUL stuff, and do some analysis on non-test documents to see what our options are there.
Some things to think about while moving test documents:
- Handle file moves when tracking intermittents (test.xul->test.html should continue to link up known intermittents with test.xul in treeherder with test.html).
- Clean up our boilerplate for mochitests (right now there's a lot of duplicate bug #s, unnecessary attributes, etc).
Some things to think about while moving non-test documents:
- Make it easy to handle XULStore migration for persisted values.