Repository navigation
Replies: 1 comment
|
Thanks, but I'll keep the current behavior. The history plot is a rolling window anchored at now, and the grid lines are the ruler for that: they tell you how far back something happened, in every view, with the same meaning. Calendar alignment only makes sense for the 1-day and 1-week views - there's no clock break to align the 10-minute view to, and months are of unequal length - so it would mean two different grid semantics in the same control. More importantly, the history buffer stores samples at a fixed period, not timestamps. A midnight line would have to be placed by clock arithmetic, so an unset clock, a timezone change or a DST transition would put it somewhere that isn't actually midnight in the recorded data. The current lines are derived from the buffer and are always correct. For calendar-referenced analysis, the data log is the right tool: it has real timestamps and plots nicely in GeigerLog. |
Uh oh!
There was an error while loading. Please reload this page.
Right now, the vertical grid lines in the History view have fixed positions, e.g. in the week view they are 24h, 48h, etc. back from the current moment in time.
It would be more logical to have them at the respective day/time breaks, i.e. the rightmost line in the week-view would not be -24h from now, but at last midnight, the next one would be the midnight before, etc.
All reactions