Rename plus and minus to add and sub

[christian-kolb]

Change Log

Changed

• Changed all methods from `plus` to `add` and `minus` to `sub`.

---

Description

Resolves #281

[christian-kolb]

@norberttech When running the tests locally with a PHP version that has the date.timezone UTC configured (I think that's the cause), the tests are randomly failing.

For example:

There was 1 failure:

1) Aeon\Calendar\Tests\Unit\Gregorian\DateTimeTest::test_creating_datetime_from_string with data set #25 ('2020-10-25 02:30:00+02:00', '2020-10-25 02:30:00 Europe/Warsaw', 'Y-m-d H:i:sP')
Failed asserting that two strings are identical.
--- Expected
+++ Actual
@@ @@
-'2020-10-25 02:30:00+02:00'
+'2020-10-25 02:30:00+01:00'

There was 1 failure:

1) Aeon\Calendar\Tests\Unit\Gregorian\DateTimeTest::test_creating_datetime_from_string with data set #25 ('2020-10-25 02:30:00+02:00', '2020-10-25 02:30:00 Europe/Warsaw', 'Y-m-d H:i:sP')
Failed asserting that two strings are identical.
--- Expected
+++ Actual
@@ @@
-'2020-10-25 02:30:00+02:00'
+'2020-10-25 02:30:00+01:00'

I don't think that it's related to the PR but something that should be looked into 🙂

[norberttech]

Thanks, this looks great!
Before I merge this one, let me first try to figure out what's wrong with those failing tests since local timezone setting should not affect this test 🤔

[christian-kolb]

Thanks, this looks great! Before I merge this one, let me first try to figure out what's wrong with those failing tests since local timezone setting should not affect this test 🤔

As they're also failing on the CI, I guess it's more because of a summer time / winter time switch where the tests aren't taking this into account. It might be better to create the dates in the timezone UTC. Otherwise the tests need to account for that change explicitly 🙂

[norberttech]

Unfortunately, I'm afraid it's another BC break in PHP, unrelated to the UTC in php.ini settings, check here: #283

It would not be the first one, https://github.com/aeon-php/rate-limiter is broken due to a bug in how PHP started to handle microseconds in DateInterval, it was already fixed but it looks like it will be released in 8.1.10

[christian-kolb]

So the current 8.1 release has an issue with handling DateIntervals? Good to know.
Only the mutation tests are still failing.

[norberttech]

Yeah, in 8.1 definitely rate limiter is affected but if you don't need bigger precision than 1s you should be fine.
Hmm those mutation tests are failing because you added new methods and old ones are now not covered which means MSI score is not 100% anymore.

Could you maybe try to use @infection-ignore-all above those all methods to see if this will make it happy? Another way around would be to keep in tests old methods as they internally use new ones and switch while migrating to 2.0

[christian-kolb]

I've updated it, but don't know what information I can get out of the report. Can you help me here?

[norberttech]

Yeah, I see it didn't help.
In general, if you never used infection I recommend you to go through their readme but in this case, you can also simply reduce the MSI score in the infection.json file to 99 so I can merge it and tackle those mutants later as they are harmless in this case.

[christian-kolb]

I would like to dive deeper into the topic first. So for now I reduced the score to 99. Thanks for your help 🙂

[norberttech]

Thanks @christian-kolb this was truly needed contribution!
I will tag the new release once I resolve this MSI issue

[christian-kolb]

Awesome, thanks for your help 🙂

[norberttech]

@christian-kolb https://github.com/aeon-php/calendar/releases done 👍

[christian-kolb]

@norberttech That was quick, thanks a lot for your help on this 🙂