abhinav-phi commented on PR #44212:
URL: https://github.com/apache/superset/pull/44212#issuecomment-5685903891

   @rusackas good catch — you're right, they have exactly the same problem, and 
there's no reason to scope the year to day/week. My first pass was just 
matching the shape of the report (a daily-granularity scatter), not a 
deliberate design choice.
   
   I've extended `%Y` to the `hour`, `minute` and `second` tiers too:
   
   | tier | before | after |
   | --- | --- | --- |
   | second | `Thu Mar 20, 03:45:12 PM` | `Thu Mar 20 2016, 03:45:12 PM` |
   | minute | `Thu Mar 20, 03:45 PM` | `Thu Mar 20 2016, 03:45 PM` |
   | hour | `Thu Mar 20, 03 PM` | `Thu Mar 20 2016, 03 PM` |
   | day / week | `Thu Mar 20` | `Thu Mar 20 2016` |
   
   The year goes with the date part (`%a %b %d %Y, …`) so the date renders the 
same as the day/week tier, and the same-date-different-year case (`2014-03-20` 
vs `2026-03-20`) now produces different strings. Tests cover every tier in the 
default locale and a non-default one.
   
   Two things I left alone on purpose — happy to change either if you disagree:
   
   - `millisecond` (`.%L`, i.e. `.345`): it's the bare fractional-second 
fragment the multi-formatter emits for sub-second resolution, not a full 
timestamp, so a year there would read oddly. Making ms-precision tooltips 
self-describing is a separate fix, since it needs a real date+time prefix 
rather than `.345`.
   - `%d` stays zero-padded in the sub-day tiers (`Mar 03`) while day/week use 
`%-e` (`Mar 3`). Pre-existing, kept as-is to keep this diff focused.
   
   Unrelated to the review: `master` currently resolves to two Alembic heads 
(#43939 and #43490), which made every `flask db upgrade head` job fail on this 
PR's merge commit (playwright, cypress, docker-build (dev) and their 
`*-required` gates). I rebased onto current `master` and added the no-op merge 
revision joining both heads — the added file is byte-identical to the one in 
#44288, so whichever PR lands first the other still merges cleanly without 
creating a third head. `git merge-tree` of this branch against `master` now 
resolves to exactly one head (`e2f3a1b9c640`). If someone can approve the 
workflow runs for the new commit, everything should come back green.


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: [email protected]

For queries about this service, please contact Infrastructure at:
[email protected]


---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]

Reply via email to