[NonSwag]

Supersedes #14086

Based on main...feature/clock-api

Notes

• I removed the methods from the Server class from the original branch, because clock managers are now per-world
  When time.affects-all-worlds is enabled, changes apply to the shared global clocks.
• I did not add a clock version for World#getTime/setTime (time of day) as those assume a 24000 tick day, which not every clock has. (should we deprecate those?)
• The WorldCreator clock and timeline overrides from the original branch are not included.

Behavior changes

• /time set <ticks> no longer changes the clokc when its event is cacneled.
• sleeping now fires ClockTimeSkipEvent instead of TimeSkipEvent when time.affects-all-worlds is enabled.
• I added equals and hashCode overrides to the vanilla Timeline so that the registry builder round-trip test can compare rebuilt timelines.

TimeSkipEvent extends ClockTimeSKipEvent but has its own HandlerList, so ClockTimeSkipEvent listeners don't receive per-world skips. That was already the case after #13814. Should this be documented or changed?

[kennytv]

Doesn't need the experimental annotations I'd say

[Machine-Maker]

no API records, we've been bitten by this before. Make it an interface and add static methods to create one.

[Doc94]

Why Range and not NonNegative/Positive?

[NonSwag]

Why Range and not NonNegative/Positive?

Pretty much just personal preference. Most projects I contribute to, have jetbrains' annotations, so kinda my go to.

[Doc94]

Pretty much just personal preference. Most projects I contribute to, have jetbrains' annotations, so kinda my go to.

okay i suppose is not a problem but in general we try to use that for that static ranges and the Paper Util for range check later (in implementation with https://github.com/PaperMC/Paper/blob/main/paper-api/src/main/java/io/papermc/paper/util/BoundChecker.java)