Fix last_day_of_month for timestamp with time zone #2852
Conversation
| millis = unpackChronology(timestampWithTimeZone).monthOfYear().roundCeiling(millis + 1); | ||
| ISOChronology isoChronology = unpackChronology(timestampWithTimeZone); | ||
| millis = isoChronology.monthOfYear().roundCeiling(millis + 1); | ||
| // millis is currently midnight in timezone of the original value |
There was a problem hiding this comment.
This comment makes me think about the case when there is no midnight and the date/time is near the end of the month.
For example, America/Bahia_Banderas and the local date/time 1970-01-01 00:00:00 ±1ms.
Can you add this to your tests? (along with other zones and date/times you have)
There was a problem hiding this comment.
Added tests for timezone America/Bahia_Banderas and 2019-01-01 00:00:00.000 ±1ms.
Not sure how to test for day without midnight. Could you provide a concrete example for this?
1970-01-01 00:00:00 is epoch and difficult to test.
|
The error seems unrelated. How can I trigger a re-test? |
|
@oneonestar there is a button on the actions page to rerun the jobs, but I'm not sure if you can see it (github permission are weird). You can always re-push to trigger a rebuild. I triggered a rebuild. |
|
@oneonestar Github actions keep failing during the git check out. I'm not sure if there is something wrong with actions right now, but can you try re pushing your code? |
|
Still there are some errors that looks like unrelated. |
|
I tried to rerun those tests, and they fail during check out again. I can't see the original failures anymore. Can you re push? I'll just manually verify next time. |
| millis = isoChronology.monthOfYear().roundCeiling(millis + 1); | ||
| // millis is currently midnight in timezone of the original value | ||
| // convert to UTC | ||
| millis = millis + isoChronology.getZone().getOffset(millis) - MILLISECONDS_IN_DAY; |
|
Merged as 13e53d7, thanks! (i slightly changed comments) |
Fix #2851