Add option to control the order of the days in the log output

[joelostblom]

I think it is quite common to sort with the newest entry at the bottom (e.g with ls and this is also how the times are ordering within each day in the log output), so I suggest adding an option to enable this behavior.

[jmaupetit]

Having a global configuration for this is a good idea, but I also would like to be able to override it while calling the log command. So, adding an extra option to the aforementioned, would be lovely.

[joelostblom]

Will add a flag with the same name --log_latest_day_on_top!

[jmaupetit]

Why not --reverse?

[joelostblom]

That's what I had initially, but then I wanted to be more specific to indicate that it is only reversing the order of the days, not the time entries within the days, so that I would leave it open for someone to add an option to reverse times if desired. This might never be the case though, and I am happy to rename both the option and the flag reverse-log if you prefer that.

[jmaupetit]

--reverse should be fine, thanks!

[jmaupetit]

LGTM @joelostblom 🙏

Do you think you can add a test for this?

[joelostblom]

I can try, but might need some assistance to make it satisfactory. Before giving it a go, I just want to clarify what what you would like this test to contain. Do you want it to be something simple like the current CLI log tests, just adding the --reverse flag and check that no errors are raised? Or do you want there to be a new log section in test_watson.py which checks the actual ordering of the days when using --reverse/--no-reverse?

[jmaupetit]

Testing Watson's API is easier to implement, so adding a new log section in test_watson.py would be perfect.

[joelostblom]

@jmaupetit Let me know what you think of the tests I just added. Since log is defined entirely in cli.py, I could not figure out a way to test it in watson.py. I tried adding it in with the other cli-tests but could not figure out a good way to add data and test it, so I added the tests to test_utils.py. The disadvantage is that we're really only testing the inner most function and things could of course go wrong outside, but I wasn't able to implement anything better. Let me know if you have suggestions. The failing flake8test is unrelated and fails on master as well.

[joelostblom]

@jmaupetit Just a reminder to have a look at this when you get a chance.

[jmaupetit]

[jmaupetit]

Thanks @joelostblom ! If you handle the rebase, I'll merge this asap 🙏

[joelostblom]

Haha thanks! =) I just rebased and squashed a few of the commits.

[joelostblom]

Realized that the changelog entry was in the wrong place since the 1.9 tag was added recently. Ready for merge now 👍