Tracking issue for extra Duration methods

[clarfon]

Right now, we have from_secs, from_millis, and from_micros, and also as_secs and subsec_nanos.

It makes sense to also add from_nanos, subsec_micros, and subsec_millis.

[clarfon]

@sfackler this should be marked as a tracking issue now

[sfackler]

Done!

[SimonSapin]

Looks good to me to stabilize.

@rfcbot fcp merge

[rfcbot]

Team member @SimonSapin has proposed to merge this. The next step is review by the rest of the tagged teams:

• @Kimundi
• @SimonSapin
• @alexcrichton
• @aturon
• @dtolnay
• @sfackler
• @withoutboats

No concerns currently listed.

Once a majority of reviewers approve (and none object), this will enter its final comment period. If you spot a major issue that hasn't been raised at any point in this process, please speak up!

See this document for info about what commands tagged team members can give me.

[rfcbot]

🔔 This is now entering its final comment period, as per the review above. 🔔

[axos88]

I feel like subsec_* methods have limited use.
I find myself interested in converting the whole duration value to e.g. millis most of the time rather than interested in only the fractional part.
I would propose to add as_millis, as_micros, as_nanos, etc. conforming with the from_* methods.

[birkenfeld]

And similarly, for formatting/display purposes, as_floating_secs etc would be helpful.

[rfcbot]

The final comment period is now complete.

[clarfon]

@birkenfeld personally I'd rather have subsec_float which returns a number between 0 and 1 for maximal precision, and then you can do subsec_float + sec as f64 if you want the total number. but that's more picky imho and I doubt many people want that use case

[xfix]

I know it's late (but it isn't in stable yet), but I wonder if from_nanos shouldn't use u128 instead of u64. Although I suppose it matches rest of methods, so whatever.

[icefoxen]

@xfix I'm not involved in this decision but it appears that 2^64 seconds is about 42x the age of the universe. So, u64 is Probably Enough For Anyone. :-P

[retep998]

Yeah but from_nanos is 2^64 nanoseconds, not seconds, which is only about 585 years. That's really short. A mostly forward compatible change would be to make it take a generic type parameter of either u128 or u64 but that could have subtle inference changes. However this stabilization is for Rust 1.27 which hasn't landed yet so we still have a chance to revert stabilization.