• R/O
  • HTTP
  • SSH
  • HTTPS

List of commits

Tags
No Tags

Frequently used words (click to add to your profile)

javac++androidlinuxc#windowsobjective-ccocoa誰得qtpythonphprubygameguibathyscaphec計画中(planning stage)翻訳omegatframeworktwitterdomtestvb.netdirectxゲームエンジンbtronarduinopreviewer

rayonパッケージ


RSS
Rev. Time Author
834ae29 upstream upstream/1.3.0 2020-04-07 20:54:23 dyknon

New upstream version 1.3.0

7ffaf34 2019-12-22 00:20:40 bors[bot]


Merge #716

716: Release rayon 1.3.0 / rayon-core 1.7.0 r=cuviper a=cuviper

This also removes `rayon-futures` and `cfg(rayon_unstable)` -- see #715 for their deprecation.

Co-authored-by: Josh Stone <cuviper@gmail.com>

163003e 2019-12-21 23:56:44 Josh Stone

bump the release date

f7d7553 2019-12-20 05:40:12 Josh Stone

Release rayon 1.3.0 / rayon-core 1.7.0

b98bb23 2019-12-20 05:15:06 Josh Stone

Remove unneeded extern crate statements

9b92c7d 2019-12-20 04:52:51 Josh Stone

cargo fmt

e02f439 2019-12-20 04:49:49 Josh Stone

Remove useless 'use crate;' in favor of the crate prelude

a1db0fe 2019-12-20 04:32:36 Josh Stone

Fix clippy::flat_map_identity

4e40256 2019-12-20 04:08:33 Josh Stone

Update ci/compat-Cargo.lock

1f9d117 2019-12-20 03:48:00 Josh Stone

Remove cfg(rayon_unstable)

198ca7f 2019-12-20 03:43:53 Josh Stone

Remove rayon-futures

ab4d244 2019-12-20 02:27:22 bors[bot]


Merge #707

707: Update rand to 0.7, rand_xorshift to 0.2 r=cuviper a=silwol

There are some method calls which require a mutable reference to the
RNG instance. I chose to create the instance and then immediately create
a reference to it which is then used within this scope. If you know a better
solution, feel free to adjust accordingly.

Co-authored-by: Wolfgang Silbermayr <wolfgang@silbermayr.at>

7a396f5 2019-12-20 02:10:05 Wolfgang Silbermayr

Update ci/compat-Cargo.lock for rand 0.7 and rand_xorshift 0.2 updates

b90bfff 2019-12-20 02:07:12 Wolfgang Silbermayr

Update rand to 0.7, rand_xorshift to 0.2

There are some method calls which require a mutable reference to the
RNG instance. I chose to create the instance and then immediately create
a reference to it which is then used within this scope.

0aefd61 2019-12-19 07:33:45 bors[bot]


Merge #711

711: impl IntoParallelIterator for tuples => MultiZip r=nikomatsakis a=cuviper

This is implemented for tuples up to arity 12, much like the standard
library's trait implementations.

- For `(a, b, ...)`, it calls `into_par_iter()` on each member.
- For `&(a, b, ...)`, it calls `par_iter()` on each member.
- For `&mut (a, b, ...)`, it calls `par_iter_mut()` on each member.

The resulting `MultiZip` iterator returns a tuple of the zipped items
from each input iterator. Internally, it is implemented with macros that
forward to a series of regular `zip`s, mapping to a flattened tuple.

Closes #567.

Co-authored-by: Josh Stone <cuviper@gmail.com>

b6449cb 2019-12-19 06:45:50 bors[bot]


Merge #712

712: Rust 2018 r=cuviper a=c410-f3r

Converts code-base to Rust 2018 and bumps minimum version to 1.31.

Brought to you by `rustfix`

Co-authored-by: Caio <c410.f3r@gmail.com>

9fea4bb 2019-12-19 06:06:57 bors[bot]


Merge #714

714: Reduce Windows CI to just stable Rust r=cuviper a=cuviper

CI for #711 ran into a broken beta windows-gnu toolchain, which is frustrating, but in general I already feel that it's overkill to test so much on Windows. The Linux builds on Travis CI already test a breadth of compiler versions, and it does so in parallel, whereas Appveyor runs serially. We can just test Windows stable to be sure the platform works at all, and leave it at that.

Co-authored-by: Josh Stone <cuviper@gmail.com>

b029b4e 2019-12-19 06:02:22 Josh Stone

Reduce Windows CI to just stable Rust

9d552c4 2019-12-11 07:34:34 Caio

Rust 2018

c0dd154 2019-12-06 06:49:06 Josh Stone

Simplify multizip reduction macros

963a4d3 2019-12-06 06:49:06 Josh Stone

Rework the multizip macros for log2 nesting depth

faf347a 2019-12-06 06:49:00 Josh Stone

impl IntoParallelIterator for tuples => MultiZip

This is implemented for tuples up to arity 12, much like the standard
library's trait implementations.

- For `(a, b, ...)`, it calls `into_par_iter()` on each member.
- For `&(a, b, ...)`, it calls `par_iter()` on each member.
- For `&mut (a, b, ...)`, it calls `par_iter_mut()` on each member.

The resulting `MultiZip` iterator returns a tuple of the zipped items
from each input iterator. Internally, it is implemented with macros that
forward to a series of regular `zip`s, mapping to a flattened tuple.

401678e 2019-11-21 08:06:17 bors[bot]


Merge #709

709: Release rayon 1.2.1 / rayon-core 1.6.1 r=cuviper a=cuviper



Co-authored-by: Josh Stone <cuviper@gmail.com>

33e9843 2019-11-21 05:14:36 Josh Stone

Release rayon 1.2.1 / rayon-core 1.6.1

dd874ac 2019-11-21 05:07:22 Josh Stone

Bump crate versions and dependencies

- Bump to rayon 1.2.1 and rayon-core 1.6.1.
- Increase crossbeam-deque to 0.7.2 to avoid duping crossbeam-utils.
- Increase demo's fixedbitset to 0.2.0 and glium to 0.25.1.
- Regenerate compat-Cargo.lock with the most current versions possible.

0c6338d 2019-11-21 05:01:55 Josh Stone

Reduce Option complexity in demo cpu_time

be99e50 2019-11-21 04:45:07 Josh Stone

cargo fmt

9b4d979 2019-11-21 04:44:12 Josh Stone

Avoid mem::uninitialized in the demo cpu_time

For both Unix `getrusage` and Windows `GetProcessTimes`, we can pass
`MaybeUninit` pointers to those FFI calls and `assume_init()` afterward.

5a46643 2019-11-21 04:39:05 Josh Stone

Avoid mem::uninitialized in par_sort_unstable

A couple temporary `[u8; 128]` arrays were uninitialized, and are now
initialized to zeros. This has no measurable affect on the benchmarks
for `par_sort_unstable`.

73b1061 2019-11-08 05:23:16 bors[bot]


Merge #705

705: Update compat-Cargo.lock r=cuviper a=matklad

I've noticed that https://github.com/rayon-rs/rayon/pull/704 uses old compat lockfile, which is strange, given that it changes Cargo.lock.

What happens here is that cargo writes new compat lockfile during CI.

That means that CI might break randomly, if rayon's dependency is published with the same major version, but different MSRV

With this PR, CI will break deterministically if you change Cargo.toml and don't update compat-Cargo.lock.

Not sure if this actually makes sense :)

Co-authored-by: Aleksey Kladov <aleksey.kladov@gmail.com>