std: Redesign Duration, implementing RFC 1040 #24920

alexcrichton · 2015-04-29T00:41:31Z

This commit is an implementation of RFC 1040 which is a redesign of the
currently-unstable Duration type. The API of the type has been scaled back to
be more conservative and it also no longer supports negative durations.

The inner duration module of the time module has now been hidden (as
Duration is reexported) and the feature name for this type has changed from
std_misc to duration. All APIs accepting durations have also been audited to
take a more flavorful feature name instead of std_misc.

Closes #24874

rust-highfive · 2015-04-29T00:41:37Z

r? @pcwalton

(rust_highfive has picked a reviewer for you, use r? to override)

alexcrichton · 2015-04-29T00:42:19Z

r? @aturon

cc @wycats

The only "surprise" I had during the implementation was the naming of the methods here and there. Abbreviating "seconds" as "secs" didn't feel super rustic to me, but typing millis/nanos felt much better than milliseconds/nanoseconds. Overall I didn't feel too strongly either way!

huonw · 2015-04-29T03:39:23Z

src/libstd/time/duration.rs

+        let seconds = self.secs;
+        let millis = self.nanos / NANOS_PER_MILLI;
+        let nanos = self.nanos % NANOS_PER_MILLI;
+        match (seconds, millis, nanos) {


What's the motivation for this sort of 'complicated' formatting rather than something like, say, 1234.00000567s?

I'm not particularly wed to this format, but it's what was specified in the RFC (as a starting point). I also had some reservations and wouldn't mind changing it to be only numeric as well.

Hm, I also have mixed feelings about this actually looking at this code (I should have raised this with the RFC). I think that automatically figuring out whether to use seconds, millis, or nanos is a good idea, but I don't understand the motivation for XX seconds and XX nanoseconds.

bors · 2015-04-29T04:15:34Z

☔ The latest upstream changes (presumably #24888) made this pull request unmergeable. Please resolve the merge conflicts.

bors · 2015-04-30T02:09:49Z

☔ The latest upstream changes (presumably #24967) made this pull request unmergeable. Please resolve the merge conflicts.

aturon · 2015-05-04T16:44:15Z

src/libstd/sys/unix/condvar.rs

@@ -57,25 +57,20 @@ impl Condvar {
    // https://github.com/llvm-mirror/libcxx/blob/release_35/src/condition_variable.cpp#L46
    // https://github.com/llvm-mirror/libcxx/blob/release_35/include/__mutex_base#L367
    pub unsafe fn wait_timeout(&self, mutex: &Mutex, dur: Duration) -> bool {
-        if dur <= Duration::zero() {


This changes the behavior for a zero duration; intended?

aturon · 2015-05-04T17:08:53Z

@alexcrichton Ok, I've done a preliminary pass. Looks nice a simple. I raised a few concerns (most of which, to be honest, I should have raised on the RFC).

alexcrichton · 2015-05-07T17:34:33Z

Ok I have updated some feedback and discussion with @aturon:

Platforms now round up the duration instead of rounding down if they do not support nanosecond precision.
The Display implementation has become a little less wordy
Duration::nanos was renamed to Duration::extra_nanos. An alternative would be one method accessor returning (u64, u32), but it's unclear what the name for this accessor would be.
Arithmetic on Duration is now always checked.

re-r? @aturon

nagisa · 2015-05-09T21:46:56Z

src/libstd/sys/unix/thread.rs

-            return Thread::yield_now()
-        }
-        let seconds = dur.num_seconds();
-        let ns = dur - Duration::seconds(seconds);
        let mut ts = libc::timespec {


There should exist From<Duration> for libc::timespec in liblibc instead.

Unfortunately this can't be provided in liblibc because Duration doesn't exist by that point and due to the in-tree liblibc being different from the out-of-tree liblibc it unfortunately also wouldn't work to provide this in the standard library as well.

alexcrichton · 2015-05-11T17:16:58Z

Thanks for taking a look @nagisa!

aturon · 2015-05-13T16:57:38Z

@bors: r+

bors · 2015-05-13T16:57:38Z

📌 Commit 086fddc has been approved by aturon

aturon · 2015-05-13T16:58:07Z

Thanks @alexcrichton!

bors · 2015-05-13T17:29:49Z

⌛ Testing commit 086fddc with merge 9eb94a8...

bors · 2015-05-13T18:23:58Z

💔 Test failed - auto-mac-64-nopt-t

Manishearth · 2015-05-13T18:26:17Z

failures:

---- sync::condvar::tests::wait_timeout_with stdout ----
    thread 'sync::condvar::tests::wait_timeout_with' panicked at 'overflow when subtracting durations', /Users/rustbuild/src/rust-buildbot/slave/auto-mac-64-nopt-t/build/src/libcore/option.rs:330



failures:
    sync::condvar::tests::wait_timeout_with

alexcrichton · 2015-05-13T23:11:35Z

@bors: r=aturon 38cfc7a

bors · 2015-05-13T23:46:38Z

⌛ Testing commit 38cfc7a with merge 9b165c0...

bors · 2015-05-14T00:42:32Z

💔 Test failed - auto-win-64-opt

This commit is an implementation of [RFC 1040][rfc] which is a redesign of the currently-unstable `Duration` type. The API of the type has been scaled back to be more conservative and it also no longer supports negative durations. [rfc]: https://github.com/rust-lang/rfcs/blob/master/text/1040-duration-reform.md The inner `duration` module of the `time` module has now been hidden (as `Duration` is reexported) and the feature name for this type has changed from `std_misc` to `duration`. All APIs accepting durations have also been audited to take a more flavorful feature name instead of `std_misc`. Closes rust-lang#24874

alexcrichton · 2015-05-14T00:53:04Z

@bors: r=aturon 556e76b

bors · 2015-05-14T02:32:00Z

⌛ Testing commit 556e76b with merge 0755c09...

bors · 2015-05-14T03:36:36Z

💔 Test failed - auto-linux-32-nopt-t

alexcrichton · 2015-05-14T16:53:52Z

@bors: retry

On Wed, May 13, 2015 at 8:36 PM, bors [email protected] wrote:

[image: 💔] Test failed - auto-linux-32-nopt-t
http://buildbot.rust-lang.org/builders/auto-linux-32-nopt-t/builds/4942

—
Reply to this email directly or view it on GitHub
#24920 (comment).

bors · 2015-05-14T18:18:39Z

⌛ Testing commit 556e76b with merge dd4dad8...

This commit is an implementation of [RFC 1040][rfc] which is a redesign of the currently-unstable `Duration` type. The API of the type has been scaled back to be more conservative and it also no longer supports negative durations. [rfc]: https://github.com/rust-lang/rfcs/blob/master/text/1040-duration-reform.md The inner `duration` module of the `time` module has now been hidden (as `Duration` is reexported) and the feature name for this type has changed from `std_misc` to `duration`. All APIs accepting durations have also been audited to take a more flavorful feature name instead of `std_misc`. Closes #24874

bors · 2015-05-14T20:03:38Z

☀️ Test successful - auto-linux-32-nopt-t, auto-linux-32-opt, auto-linux-64-nopt-t, auto-linux-64-opt, auto-linux-64-x-android-t, auto-mac-32-opt, auto-mac-64-nopt-t, auto-mac-64-opt, auto-win-32-nopt-t, auto-win-32-opt, auto-win-64-nopt-t, auto-win-64-opt

rust-highfive assigned pcwalton Apr 29, 2015

rust-highfive assigned aturon and unassigned pcwalton Apr 29, 2015

alexcrichton force-pushed the duration branch 4 times, most recently from c0d3aa8 to 260ce79 Compare April 29, 2015 00:53

huonw reviewed Apr 29, 2015
View reviewed changes

alexcrichton force-pushed the duration branch from 260ce79 to c30db14 Compare April 29, 2015 15:03

alexcrichton force-pushed the duration branch from c30db14 to d0990f7 Compare April 30, 2015 02:15

aturon reviewed May 4, 2015
View reviewed changes

alexcrichton force-pushed the duration branch from d0990f7 to b59f968 Compare May 7, 2015 17:32

alexcrichton force-pushed the duration branch 2 times, most recently from 2fe4199 to f143b01 Compare May 7, 2015 21:57

nagisa reviewed May 9, 2015
View reviewed changes

alexcrichton force-pushed the duration branch from 1b17654 to 086fddc Compare May 11, 2015 17:16

alexcrichton force-pushed the duration branch 2 times, most recently from cd06e02 to 38cfc7a Compare May 13, 2015 23:11

alexcrichton force-pushed the duration branch from 38cfc7a to 556e76b Compare May 14, 2015 00:52

bors merged commit 556e76b into rust-lang:master May 14, 2015

zmbush added a commit to zmbush/megaprompt that referenced this pull request May 18, 2015

Fix for rust Duration reform (rust-lang/rust#24920)

8f84c64

jooert mentioned this pull request May 27, 2015

Yet another attempt to refactor libstd's Duration #18416

Closed

alexcrichton deleted the duration branch August 17, 2015 19:49

djc mentioned this pull request Sep 7, 2023

Feedback on swapping out the chrono::Duration type chronotope/chrono#1268

Closed

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

std: Redesign Duration, implementing RFC 1040 #24920

std: Redesign Duration, implementing RFC 1040 #24920

alexcrichton commented Apr 29, 2015

rust-highfive commented Apr 29, 2015

alexcrichton commented Apr 29, 2015

huonw Apr 29, 2015

alexcrichton Apr 29, 2015

aturon May 4, 2015

bors commented Apr 29, 2015

bors commented Apr 30, 2015

aturon May 4, 2015

aturon commented May 4, 2015

alexcrichton commented May 7, 2015

nagisa May 9, 2015

alexcrichton May 11, 2015

alexcrichton commented May 11, 2015

aturon commented May 13, 2015

bors commented May 13, 2015

aturon commented May 13, 2015

bors commented May 13, 2015

bors commented May 13, 2015

Manishearth commented May 13, 2015

alexcrichton commented May 13, 2015

bors commented May 13, 2015

bors commented May 14, 2015

alexcrichton commented May 14, 2015

bors commented May 14, 2015

bors commented May 14, 2015

alexcrichton commented May 14, 2015

bors commented May 14, 2015

bors commented May 14, 2015

std: Redesign Duration, implementing RFC 1040 #24920

std: Redesign Duration, implementing RFC 1040 #24920

Conversation

alexcrichton commented Apr 29, 2015

rust-highfive commented Apr 29, 2015

alexcrichton commented Apr 29, 2015

huonw Apr 29, 2015

Choose a reason for hiding this comment

alexcrichton Apr 29, 2015

Choose a reason for hiding this comment

aturon May 4, 2015

Choose a reason for hiding this comment

bors commented Apr 29, 2015

bors commented Apr 30, 2015

aturon May 4, 2015

Choose a reason for hiding this comment

aturon commented May 4, 2015

alexcrichton commented May 7, 2015

nagisa May 9, 2015

Choose a reason for hiding this comment

alexcrichton May 11, 2015

Choose a reason for hiding this comment

alexcrichton commented May 11, 2015

aturon commented May 13, 2015

bors commented May 13, 2015

aturon commented May 13, 2015

bors commented May 13, 2015

bors commented May 13, 2015

Manishearth commented May 13, 2015

alexcrichton commented May 13, 2015

bors commented May 13, 2015

bors commented May 14, 2015

alexcrichton commented May 14, 2015

bors commented May 14, 2015

bors commented May 14, 2015

alexcrichton commented May 14, 2015

bors commented May 14, 2015

bors commented May 14, 2015