Summary report shows wrong year

[okms]

when viewing the summary report, the wrong year is shown below each week (2020). It appears however that the hours shown are from 2021, which is correct

[damsleth]

This issue does not occur after week 1 2022. Probably related to the dayjs .isoWeek()-bug

[olemp]

This issue does not occur after week 1 2022. Probably related to the dayjs .isoWeek()-bug

Can't reproduce now. What do I need to do to reproduce @okms @damsleth?

[olemp]

This issue does not occur after week 1 2022. Probably related to the dayjs .isoWeek()-bug

Related to #1028?

[damsleth]

This issue does not occur after week 1 2022. Probably related to the dayjs .isoWeek()-bug

Related to #1028?

Most likely, yes.
Rolling our own isoWeek algorithm would probably solve both issues

[okms]

This will most probably happen again 1.jan 2026

[olemp]

This issue does not occur after week 1 2022. Probably related to the dayjs .isoWeek()-bug

Related to #1028?

Most likely, yes. Rolling our own isoWeek algorithm would probably solve both issues

Maybe ChatGPT can make a isoWeek algorithm @okms? ;)

[damsleth]

👻 It's back, week 01-2023

The data returned from graphql contains hours for
weeks 46 to 52, month 12, year 2022 (correctly), as well as
week 52, month 1, year 2022 (incorrect, should be 2023)

Whereas the table seems to be trying to present weeks 46 to 52 2021, as well as week 52 month 1 year 2022.
The effect is that only the 52/01/2022 week is visible in the table, since it's both returned by graphQL and represented in the table.

[damsleth]

The plot thickens. The issue does no longer occur as of 03.01.2023 09:00

The issue still remains though, that the data for 52/01/2022 (i.e. the part of week 52 of 2021 that falls in 2022), are returned instead of 52/01/2023 (the part of week 52 of 2022 that falls in 2023)

[damsleth]

We should have a way of simulating the current date to be able to debug this properly

[okms]

Step 1 in moving forward with this issue is to make test cases where we can specify what date we are testing for

[okms]

Lets re-implement this function from scratch and not wait for dayjs to fix this