Commit Graph

79643 Commits

Author SHA1 Message Date
Mike Hommey
3373204ac4 Replace impl GlobalAlloc for Global with a set of free functions 2018-06-11 13:47:23 -07:00
Mike Hommey
f6ab74b8e7 Remove alloc::Opaque and use *mut u8 as pointer type for GlobalAlloc 2018-06-11 13:47:23 -07:00
Guillaume Gomez
dadfa13f47 Fix extern prelude failure in rustdoc 2018-06-11 22:24:35 +02:00
bors
1d4dbf488a Auto merge of #51442 - tinaun:more-future-impls, r=cramertj
[futures] add a few blanket impls to std

these were defined in the futures crate, but with the core definitions moving to std these would need to move too.
2018-06-11 20:14:39 +00:00
PSeitz
6fd9ede0cd
Make span_fatal and parse_block public
span_fatal and parse_block  were made private in #51265. These methods are used in stainless.

Related #51498 #51504
2018-06-11 21:19:12 +02:00
Aaron Power
33ceb03efd Removed incorrect release note. fixes #51091 2018-06-11 20:11:28 +01:00
Josh Stone
f81e34b825 Regenerate character tables for Unicode 11 2018-06-11 10:54:30 -07:00
Oliver Schneider
c53b76324c Update tests 2018-06-11 19:54:22 +02:00
jeb
7e56261fcb Make parse_seq_to_end and parse_path public 2018-06-11 10:40:58 -06:00
bors
0b7c9e756e Auto merge of #51490 - Havvy:diagnostic-list, r=GuillaumeGomez
Long diagnostic for E0541

r? @GuillaumeGomez
2018-06-11 15:28:44 +00:00
Seiichi Uchida
ed74b0b016 Make parse_ident public 2018-06-11 23:04:11 +09:00
bors
0d763178e5 Auto merge of #51478 - crlf0710:patch-2, r=QuietMisdreavus
Add #[doc(inline)] in std::task

Add #[doc(inline)] in `std::task` to make the nightly docs seem right.
2018-06-11 13:24:00 +00:00
Havvy
cf3fcab973 More E0541 long diagnostic wording tweaks 2018-06-11 03:51:04 -07:00
bors
13f8d073fe Auto merge of #51480 - dtolnay:lifetime, r=kennytm
Enable fall through past $:lifetime matcher

```rust
macro_rules! is_lifetime {
    ($lifetime:lifetime) => { true };
    ($other:tt) => { false };
}

fn main() {
    println!("{}", is_lifetime!('lifetime));
    println!("{}", is_lifetime!(@));
}
```

Before this fix, the `is_lifetime!` invocation would fail to compile with:

```
error: expected a lifetime, found `@`
 --> src/main.rs:8:33
  |
8 |     println!("{}", is_lifetime!(@));
  |                                 ^
```

Fixes #50903.
Fixes #51477.

r? @kennytm
2018-06-11 10:43:20 +00:00
Havvy
0c91bd200e Make tidy happy 2018-06-11 03:04:08 -07:00
Havvy
2d7b046d71 Fix typos in previous commit 2018-06-11 02:44:55 -07:00
Havvy
f7df1f38dd Long diagnostic for E0541 2018-06-11 01:52:12 -07:00
bors
18a00bd985 Auto merge of #51467 - toidiu:ak-51409, r=nikomatsakis
create separate dep-nodes for predicates_of and explicit_predicates_of

Fix for https://github.com/rust-lang/rust/issues/51409

- added incremental compilation test for infer_outlives_requirements
- created separate dep-node for explicit_predicates_of
2018-06-11 08:37:01 +00:00
bors
a32e979ba5 Auto merge of #51423 - RalfJung:from_raw_parts, r=sfackler
Improve docs for slice::from_raw_parts

Explain why the pointer must be non-null and aligned.

Cc @hsivonen
2018-06-11 03:48:23 +00:00
bors
7dc61f3022 Auto merge of #51461 - estebank:e0423-doc, r=petrochenkov
Update E0423 description

E0423 doesn't apply only to structs, update the error index description
to make this clear.

Fix #42077.
2018-06-11 00:32:22 +00:00
bors
b202bebdee Auto merge of #51475 - GuillaumeGomez:fix-error-codes, r=Manishearth
Fix error codes
2018-06-10 22:30:14 +00:00
David Tolnay
987020846c
Enable fall through past $:lifetime matcher 2018-06-10 14:39:16 -07:00
bors
a805a2a5eb Auto merge of #50205 - topecongiro:include-parens-to-type-parameter, r=petrochenkov
Include parens to type parameter

The motivation of this PR is to fix a bug in rustfmt (cc https://github.com/rust-lang-nursery/rustfmt/issues/2630).
2018-06-10 20:09:22 +00:00
CrLF0710
02c96d4733
Add #[doc(inline)] in std::task
Add #[doc(inline)] in `std::task` to make the doc seem right.
2018-06-11 03:09:30 +08:00
bors
cabb679bf1 Auto merge of #51474 - llogiq:from-docs, r=TimNN
add some docs to `From` conversions

This adds a helpful document to the bool → int* conversions as well as to the lossless integer conversions.

One of #51430 down, some more to go.
2018-06-10 18:00:12 +00:00
bors
c5a129e809 Auto merge of #51200 - tmccombs:stable-iter-repeat-with, r=Centril,kennytm
Stabilize iterator_repeat_with

Fixes #48169
2018-06-10 15:48:14 +00:00
bors
900037e3c3 Auto merge of #51079 - GuillaumeGomez:stabilize-entry-or-default, r=SimonSapin
Stabilize entry-or-default

Fixes #44324.

cc @SimonSapin
2018-06-10 12:46:14 +00:00
Guillaume Gomez
f2349d5ec6 Fix error codes 2018-06-10 14:04:48 +02:00
Andre Bogus
553a44a5cc add some docs to conversions 2018-06-10 13:16:34 +02:00
Guillaume Gomez
861c7cb9fd Stabilize entry-or-default 2018-06-10 12:59:12 +02:00
bors
684b8d3d7b Auto merge of #51196 - Havvy:remove-keywords, r=petrochenkov
Implement RFC 2421, 'Keyword unreservations (pure, sizeof, alignof, offsetof)

On my local machine I was getting failures in `test/ui-fulldeps`, but I was also getting them again after reverting the change, so I'm hoping that this works.

I removed the test because that's what the other PR that unreserved a keyword did and I feel that it doesn't make sense/keep value to keep a test for something removed years ago.
2018-06-10 08:26:10 +00:00
toidiu
ea2012b2c7 create separate dep-nodes for predicates_of and explicit_predicates_of 2018-06-10 02:33:55 -04:00
bors
b4d836c98b Auto merge of #51320 - tmccombs:step-by, r=SimonSapin
Stabilize Iterator::step_by

Fixes #27741
2018-06-10 01:45:42 +00:00
Seiichi Uchida
dc87d0d1fb Update tests 2018-06-10 10:37:38 +09:00
Seiichi Uchida
699be41809 Simplify an error handling in the parser 2018-06-10 10:37:38 +09:00
Seiichi Uchida
78a19d9b88 Include parens to type parameter 2018-06-10 10:37:38 +09:00
bors
ae0659c97d Auto merge of #51265 - Mark-Simulacrum:cleanup-syntax-parse, r=petrochenkov
crate-ify and delete unused code from syntax::parse

This is intended primarily to ensure the compiler catches dead code for us in
more cases.
2018-06-09 23:32:01 +00:00
Mark Simulacrum
60058e5dbe Crate-ify and delete unused code in syntax::parse 2018-06-09 16:57:19 -06:00
Havvy
b09bc88e9f Refactor: Rename ExistentialPredicate::cmp to ExistentialPredicate::stable_cmp
See https://github.com/rust-lang/rust/pull/51276#discussion_r193549404 for
rationale.
2018-06-09 15:09:10 -07:00
Esteban Küber
a6782d9e4e Update E0423 description
E0423 doesn't apply only to structs, update the error index description
to make this clear.
2018-06-09 14:27:42 -07:00
bors
2a0062974a Auto merge of #51459 - kennytm:dist-at-stage-0, r=Mark-Simulacrum
Miscellaneous changes to rustbuild and CI.

1. Don't build LLVM when running rust-installer.
2. If toolstate is unchanged, don't push a commit to the toolstate repo.
3. Allow `./x.py build src/librustc_codegen_llvm`
4. Added log to track #50887.
2018-06-09 21:22:36 +00:00
Niko Matsakis
2e25bed9b1 remove some #[inline(never)] 2018-06-09 17:20:15 -04:00
Havvy
679a5219e7 Test keyword unreservations 2018-06-09 14:15:57 -07:00
Havvy
ebc315bbbb Implement RFC 2421, 'Keyword unreservations (pure, sizeof, alignof, offsetof)' 2018-06-09 14:15:57 -07:00
Niko Matsakis
9240025afa introduce any_unifications flag
Resetting unifications is expensive. Only do if there is something to
reset.
2018-06-09 16:54:20 -04:00
kennytm
ab5e3e66b5
Added comment to explain why only RustIstaller has llvm_tools = false. 2018-06-10 04:50:42 +08:00
kennytm
4a0e92eee5
Log the clang_rt.asan-dynamic-i386.vers on failure to track #50887 2018-06-10 04:31:50 +08:00
kennytm
b00ea6c2da
Allowing building the codegen backend specifically.
Use `./x.py build src/librustc_codegen_llvm` to build the codegen backend.
2018-06-10 04:31:49 +08:00
kennytm
d2b5b7603b
Do not push a commit if the toolstate is unchanged.
This should greatly reduce the commits on the rust-toolstate repository.

`publish_toolstate.py` defaults to keep the old status if a new one is not
found, so nothing needs to be changed to that file.
2018-06-10 04:31:49 +08:00
kennytm
14d50bfecb
Allow some tools to be run without first building LLVM.
Conservatively only disable LLVM for rust-installer. This should shave
5 minutes from the x86_64-gnu-distcheck job by not building LLVM twice.
2018-06-10 04:31:49 +08:00