100% reproducible for me in a vivid chroot. One interesting clue, test-utils loads the same com.canonical.indicator.datetime schema, and /that/ test is passing. They use two different code paths to get there though. I feel like there's something simple here that we're both missing, so I'll stare at the screen a little more. :)
Thanks for explaining the plan with the relocatable schema, that part LGTM.
100% reproducible for me in a vivid chroot. One interesting clue, test-utils loads the same com.canonical. indicator. datetime schema, and /that/ test is passing. They use two different code paths to get there though. I feel like there's something simple here that we're both missing, so I'll stare at the screen a little more. :)
Thanks for explaining the plan with the relocatable schema, that part LGTM.