Chris

@szqtc , I'm working on the fix to this bug. I just want to check that the following output looks good to you as well: ``` +58 s => TAI...

Good to know. Concerning the time scale management, prior to version 4, Hifitime would store all input epochs in TAI, so the difference between the 59th second over a leap...

I need to pause this bug fix for a week or two while I tackle urgent work on other repos. Sorry for the delay.

Ha, yeah, you're right that there seems to be a bug for printing years way out in the future here. I will need to fix this ... converting to Gregorian...

> This statement does not sound like a mirror right? Because on the positive direction, centuries = 1 and nanoseconds = 1 IS a greater duration (further from zero). _Greater_...

Let me know if that works. I believe this logic to be correct because it's just an overflowing counter of years.

Yes, it's automatically generated by the Rust documentation, which is above the functions. So you can fix them there and it should propagate to the pyi by running the generate...

@translunar https://github.com/nyx-space/hifitime/pull/437 should address these. However, I'll wait for you to confirm before closing this. I'll release the next version of hifitime soon and embed it in the soon-to-be released...

I'm gonna call this good enough: ``` python -m mypy.stubtest hifitime error: hifitime.hifitime failed to find stubs Stub: MISSING Runtime: Found 1 error (checked 2 modules) ``` It's down from...

What other time types are you referring to? To convert to Python's datetime, you can use to_datetime (now also available with the underscore). If you want the conversion between units,...