diff options
| author | Craig Jennings <c@cjennings.net> | 2026-04-22 06:24:57 -0500 |
|---|---|---|
| committer | Craig Jennings <c@cjennings.net> | 2026-04-22 06:24:57 -0500 |
| commit | 86a06eb53c7a07b4b5c7ca889edad09042384cb4 (patch) | |
| tree | 7ff2b6795875233f7afcf79f9066b88588188565 /chime.el | |
| parent | 77b5e8c80b56ffcccb2a44807e15d933f22edb05 (diff) | |
| download | chime-86a06eb53c7a07b4b5c7ca889edad09042384cb4.tar.gz chime-86a06eb53c7a07b4b5c7ca889edad09042384cb4.zip | |
test: pin group-events-by-day clock and add midnight edge cases
The Normal and Boundary tests built events as fixed-minute offsets from real current-time. On late-night runs the +60 minute event in single-day crossed midnight and the test failed because events 1-2 landed "Today" while event 3 landed "Tomorrow". The intent of these Normal tests is routine grouping, not midnight behavior, so they should be deterministic.
Pin the clock with with-test-time using the established let-bind pattern. with-test-time re-evaluates BASE-TIME inside the mocked current-time lambda, so passing test-time-today-at directly would recurse, because the helper itself calls current-time.
Add two new Boundary tests that explicitly exercise the day-crossing logic at 23:50. One event at +5 minutes stays in Today and one at +20 minutes crosses into Tomorrow. These are the first tests in the file that actually target midnight behavior.
Drop stale "REFACTORED: Uses dynamic timestamps" annotations from docstrings since the behavior is documented at the file level.
Diffstat (limited to 'chime.el')
0 files changed, 0 insertions, 0 deletions
