Commit Graph

78817 Commits

Author SHA1 Message Date
bors
63cd4a39ea Auto merge of #51171 - faern:const-bswap-ctpop-cttz-ctlz, r=oli-obk
Make some std::intrinsics `const fn`s

Making some rustc intrinsics (`ctpop`, `cttz`, `ctlz` and `bswap`) `const fn`s.

This is a pre-step to being able to make `swap_bytes`, `to_be` and `from_be` constant functions. That in itself could be ergonomic and useful. But even better is that it would allow `Ipv4Addr::new` etc becoming `const fn`s as well. Which might be really useful since I find it quite common to want to define them as constants.

r? @oli-obk
2018-06-01 03:24:54 +00:00
bors
b7a9d4ea74 Auto merge of #50836 - jsgf:arc-downcast, r=SimonSapin
Arc downcast

Implement `downcast` for `Arc<Any + Send + Sync>` as part of #44608, and gated by the same `rc_downcast` feature.

This PR is mostly lightly-edited cut'n'paste.

This has two additional changes:
- The `downcast` implementation needs `Any + Send + Sync` implementations for `is` and `Debug`, and I added `downcast_ref` and `downcast_mut` for completeness/consistency. (Can these be insta-stabilized?)
- At @SimonSapin's suggestion, I converted `Arc` and `Rc` to use `NonNull::cast` to avoid an `unsafe` block in each which tidied things up nicely.
2018-06-01 01:09:25 +00:00
bors
efc508ba4c Auto merge of #51096 - matthewjasper:reverse-normalization-bounds, r=nikomatsakis
Register outlives predicates from queries the right way around.

Closes #49354
The region constraints from queries need to be reversed from sub to outlives.

Note: wf checking reports these errors before NLL, so I'm not sure if there's any case when these predicates need to be created at all.

cc @nikomatsakis
2018-05-31 22:58:37 +00:00
bors
1ffb321477 Auto merge of #51257 - GuillaumeGomez:rollup, r=GuillaumeGomez
Rollup of 7 pull requests

Successful merges:

 - #49546 (Stabilize short error format)
 - #51123 (Update build instructions)
 - #51146 (typeck: Do not pass the field check on field error)
 - #51193 (Fixes some style issues in rustdoc "implementations on Foreign types")
 - #51213 (fs: copy: Use File::set_permissions instead of fs::set_permissions)
 - #51227 (mod.rs isn't beautiful)
 - #51240 (Two minor parsing tweaks)

Failed merges:
2018-05-31 20:44:04 +00:00
Jeremy Fitzhardinge
87c3d7bee5 Fix up Any doc examples
Make the Any+Send+Sync examples use the right trait bounds, and fix a small whitespace issue.
2018-05-31 13:27:08 -07:00
Jeremy Fitzhardinge
0c7bf56d80 Update Arc and Rc to use NonNull::cast to cast the inner pointers
This avoids an `unsafe` block in each case.
2018-05-31 13:27:08 -07:00
Jeremy Fitzhardinge
37f5cf563c Implement downcast for Arc<Any + Send + Sync>
We only need to implement it for `Any + Send + Sync` because in practice
that's the only useful combination for `Arc` and `Any`.

Implementation for #44608 under the `rc_downcast` feature.
2018-05-31 13:27:08 -07:00
Jeremy Fitzhardinge
72433e179d Add implementations for Any + Send + Sync
Implement `is`, `downcast_ref`, `downcast_mut` and `Debug` for
`Any + Send + Sync`.
2018-05-31 13:27:08 -07:00
Guillaume Gomez
7552c2ea64
Rollup merge of #51240 - nnethercote:parse-2, r=nikomatsakis
Two minor parsing tweaks
2018-05-31 22:17:16 +02:00
Guillaume Gomez
33ad846740
Rollup merge of #51227 - uuttff8:master, r=dtolnay
mod.rs isn't beautiful

I hate this spaces.
2018-05-31 22:17:15 +02:00
Guillaume Gomez
af4acbe5e7
Rollup merge of #51213 - nicokoch:copy_permissions, r=cramertj
fs: copy: Use File::set_permissions instead of fs::set_permissions

We already got the open file descriptor at this point.
Don't make the kernel resolve the path again.
2018-05-31 22:17:14 +02:00
Guillaume Gomez
8efecc1102
Rollup merge of #51193 - GuillaumeGomez:fixes-style-issues, r=QuietMisdreavus
Fixes some style issues in rustdoc "implementations on Foreign types"

Fixes #51113.

r? @QuietMisdreavus
2018-05-31 22:17:13 +02:00
Guillaume Gomez
a729d97848
Rollup merge of #51146 - dlrobertson:fix_51102, r=estebank
typeck: Do not pass the field check on field error

If a struct pattern has a field error return an error.

Fixes: #51102
2018-05-31 22:17:12 +02:00
Guillaume Gomez
2f0e3d3046
Rollup merge of #51123 - akoserwal:master, r=Mark-Simulacrum
Update build instructions

It get stuck at the cloning step.
`./x.py build `
Updating only changed submodules
Updating submodule src/llvm
Submodule 'src/llvm' (https://github.com/rust-lang/llvm.git) registered for path 'src/llvm'
Cloning into '/home/username/rust/src/llvm'...
2018-05-31 22:17:10 +02:00
Guillaume Gomez
0345dac0d8
Rollup merge of #49546 - GuillaumeGomez:stabilize-short-error-format, r=oli-obk
Stabilize short error format

r? @oli-obk

Added in https://github.com/rust-lang/rust/pull/44636
2018-05-31 22:17:09 +02:00
Matthew Jasper
b83daea479 Register outlives predicates from queries the right way around. 2018-05-31 20:40:25 +01:00
Abhishek koserwal
bad9fefdae Update build instructions 2018-05-31 23:53:23 +05:30
Guillaume Gomez
426b63f8a3 Make short-error format GNU compatible 2018-05-31 20:09:27 +02:00
Guillaume Gomez
177337d5cc Stabilize short error format 2018-05-31 20:09:27 +02:00
bors
5342d40c1f Auto merge of #51050 - symphorien:fstatat, r=kennytm
std::fs::DirEntry.metadata(): use fstatat instead of lstat when possible

When reading a directory with `read_dir`, querying metadata for a resulting `DirEntry` is done by building the whole path and then `lstat`ing it, which requires the kernel to resolve the whole path. Instead, one
can use the file descriptor to the enumerated directory and use `fstatat`. This make the resolving step
unnecessary.
This PR implements using `fstatat` on linux, android and emscripten.

## Compatibility across targets
`fstatat` is POSIX.
* Linux >= 2.6.19 according to https://linux.die.net/man/2/fstatat
* android according to https://android.googlesource.com/platform/bionic/+/master/libc/libc.map.txt#392
* emscripten according to 7f89560101/system/include/libc/sys/stat.h (L76)

The man page says "A similar system call exists on Solaris." but I haven't found it.

## Compatibility with old platforms
This was introduced with glibc 2.4 according to the man page. The only information I could find about the minimal version of glibc rust must support is this discussion https://internals.rust-lang.org/t/bumping-glibc-requirements-for-the-rust-toolchain/5111/10
The conclusion, if I understand correctly, is that currently rust supports glibc >= 2.3.4 but the "real" requirement is Centos 5 with glibc 2.5. This PR would make the minimal version 2.4, so this should be fine.

## Benefit
I did the following silly benchmark:
```rust
use std::io;
use std::fs;
use std::os::linux::fs::MetadataExt;
use std::time::Instant;

fn main() -> Result<(), io::Error> {
    let mut n = 0;
    let mut size = 0;
    let start = Instant::now();
    for entry in fs::read_dir("/nix/store/.links")? {
        let entry = entry?;
        let stat = entry.metadata()?;
        size += stat.st_size();
        n+=1;
    }
    println!("{} files, size {}, time {:?}", n, size, Instant::now().duration_since(start));
    Ok(())
}
```
On warm cache, with current rust nightly:
```
1014099 files, size 76895290022, time Duration { secs: 2, nanos: 65832118 }
```
(between 2.1 and 2.9 seconds usually)
With this PR:
```
1014099 files, size 76895290022, time Duration { secs: 1, nanos: 581662953 }
```
(1.5 to 1.6 seconds usually).

approximately 40% faster :)

On cold cache there is not much to gain because path lookup (which we spare) would have been a cache hit:
Before
```
1014099 files, size 76895290022, time Duration { secs: 391, nanos: 739874992 }
```
After
```
1014099 files, size 76895290022, time Duration { secs: 388, nanos: 431567396 }
```
## Testing
The tests were run on linux `x86_64`
```
python x.py test src/tools/tidy
./x.py test src/libstd
```
and the above benchmark.
I did not test any other target.
2018-05-31 17:47:12 +00:00
Linus Färnstrand
97a0d46dec Rename num -> bits and num -> out_val 2018-05-31 19:43:56 +02:00
Linus Färnstrand
8939fe60a1 Rename bytes -> bits 2018-05-31 19:21:34 +02:00
Guillaume Girol
8dec03b71a libstd/sys/unix/fs.rs: fix compilation on fuchsia 2018-05-31 19:18:58 +02:00
Linus Färnstrand
0d7d5e994f Rewrite numeric_intrinsic without macro 2018-05-31 19:18:42 +02:00
Linus Färnstrand
12c330c4bf Add test for intrinsics::bswap 2018-05-31 19:18:42 +02:00
Linus Färnstrand
376b6407a8 Make ctpop, cttz, ctlz and bswap const fns 2018-05-31 19:18:42 +02:00
bors
6de4ec679d Auto merge of #51235 - nikomatsakis:issue-51117-borrowck-implicit-deref, r=eddyb
remove notion of Implicit derefs from mem-cat

`PointerKind` is included in `LoanPath` and hence forms part of the equality check; this led to having two unequal paths that both represent `*x`, depending on whether the `*` was inserted automatically or explicitly. Bad mojo.

Fixes #51117

r? @eddyb
2018-05-31 15:24:04 +00:00
Dan Robertson
8f6441513e
typeck: Do not pass the field check on field error
If a struct pattern has a field error return an error.
2018-05-31 14:45:49 +00:00
Niko Matsakis
8a624b737a change PointerKind::Implicit to a note
`PointerKind` is included in `LoanPath` and hence forms part of the
equality check; this led to having two unequal paths that both
represent `*x`, depending on whether the `*` was inserted
automatically or explicitly. Bad mojo. The `note` field, in contrast,
is intended more-or-less primarily for this purpose of adding extra
data.
2018-05-31 10:17:51 -04:00
bors
ae2f299186 Auto merge of #51203 - nnethercote:obligations-2, r=michaelwoerister
Two minor `obligation_forest` tweaks.

Pretty minimal improvements, but improvements nonetheless.
2018-05-31 13:14:18 +00:00
bors
30cae58709 Auto merge of #51182 - eddyb:not-just-visibility, r=nikomatsakis
[MIR] Change "scopes" from "visibility scopes" to "source scopes".

These scopes are already used for source-oriented diagnostics, lint levels and unsafety checks.
This PR generalizes the naming around scopes, making the type `SourceScope`, and flips (across several commits) the relationship/priority between `LocalDecl`'s "visibility" and "syntactic" scopes.

r? @nikomatsakis
2018-05-31 08:55:09 +00:00
bors
a47b2d09d6 Auto merge of #51127 - frewsxcv:frewsxcv-discriminant, r=GuillaumeGomez
Add doc link from discriminant struct to function.

None
2018-05-31 06:38:24 +00:00
Nicholas Nethercote
5adba8e9d9 Avoid an unnecessary match when lexing "<-". 2018-05-31 16:05:39 +10:00
Nicholas Nethercote
3af6291eff Tweak identifer lexing.
By calling `bump()` after getting the first char, to avoid a redundant
`ident_continue()` test on it.
2018-05-31 16:04:33 +10:00
Corey Farwell
1c883d6761 Add doc link from discriminant struct to function. 2018-05-30 23:25:30 -04:00
bors
e38554cd80 Auto merge of #51145 - petrochenkov:npbot, r=alexcrichton
resolve: Make sure indeterminate and inconsistent macro resolutions always generate errors

Addresses the issue described in https://github.com/rust-lang/rust/pull/50911#issuecomment-392560525

I haven't come up with a minimized reproduction yet, but confirmed that `npbot` now generates the correct error with `![feature(use_extern_macros)]`.
2018-05-31 03:18:00 +00:00
bors
c1287c0183 Auto merge of #51220 - nikomatsakis:issue-51008-false-positive-lifetime-must-be-declared, r=cramertj
reset anonymous-lifetime-mode as we enter `()` scopes

Background:

The anonymous lifetime mode is used to prohibit elided lifetimes where
they didn't used to be permitted, and instead require that `'_` be
used. For example:

```rust
impl Trait for Ref<T> { .. }
//             ^^^^^^ ERROR: should be `Ref<'_, T>`
```

When we are parsing the parts of the impl header, we enter into an alternate mode called `CreateParameter`. In this mode, we give an error for things like `Ref<T>`, but for elided lifetimes in a reference type like `&T` we make the elided lifetime into an in-band lifetime:

4f99f37b7e/src/librustc/hir/lowering.rs (L4017-L4035)

This was not intended to change behavior because we only enter into that mode in contexts where elision was not historically permitted. However, the problem is that we fail to reset the mode when we enter into bounds like `Fn(&u32)`, where elision *was* allowed -- the same occurs for fn types like `fn(&u32`). This PR restores the original mode in those contexts.

Fixes #51008

r? @cramertj
2018-05-31 01:06:33 +00:00
Nicholas Nethercote
a5ffcf6dd8 Inline NodeIndex methods.
Because they are small and hot.
2018-05-31 09:22:07 +10:00
Nicholas Nethercote
f46d0213e4 Remove ObligationForest::cache_list.
It's never used in a meaningful way.
2018-05-31 09:22:07 +10:00
bors
e1eed38fed Auto merge of #51138 - spastorino:add-polonius-compare-mode, r=pnkfelix
Add polonius compare mode

**This is now ready to review/merge**
2018-05-30 22:59:21 +00:00
bors
5d0631a643 Auto merge of #51215 - eddyb:visit-for-a-lifetime, r=nikomatsakis
rustc: don't visit lifetime parameters through visit_lifetime.

Ideally we'd also not use the `Lifetime` struct for parameters, but I'll leave that to @varkor (#48149).
Fixes #51185 (discovered while auditing all the `visit_lifetime` implementations).
r? @nikomatsakis
2018-05-30 20:34:23 +00:00
uuttff8
d6b8c67ee6
mod.rs isn't beautiful 2018-05-30 22:24:24 +03:00
Vadim Petrochenkov
345e7c3597 resolve: Make sure indeterminate and inconsistent macro resolutions always generate errors 2018-05-30 22:21:50 +03:00
Niko Matsakis
da69bbce68 also reset anon-param-mode for fn() types 2018-05-30 14:54:51 -04:00
Niko Matsakis
9acb351f57 reset the "anonymous lifetime mode" for parenthesized where clauses
Background:

The anonymous lifetime mode is used to prohibit elided lifetimes where
they didn't used to be permitted, and instead require that `'_` be
used. For example:

```rust
impl Trait for Ref<T> { .. }
//             ^^^^^^ ERROR: should be `Ref<'_, T>`
```

When we are parsing the parts of the impl header, we enter into an
alternate mode called `CreateParameter`. In this mode, we give an
error for things like `Ref<T>`, but for elided lifetimes in a
reference type like `&T` we make the elided lifetime into an in-band
lifetime:

4f99f37b7e/src/librustc/hir/lowering.rs (L4017-L4035)

This was not intended to change behavior because we only enter into
that mode in contexts where elision was not historically
permitted. However, the problem is that we fail to reset the mode when
we enter into bounds like `Fn(&u32)`, where elision *was* allowed --
the same occurs for fn types like `fn(&u32`). This PR restores the
original mode in those contexts.
2018-05-30 14:54:51 -04:00
Guillaume Girol
cb2a0d61ad std::fs::DirEntry.metadata(): use fstatat instead of lstat when possible 2018-05-30 20:52:30 +02:00
Santiago Pastorino
b39a1d6f1a
Run rustfmt 2018-05-30 14:36:23 -03:00
Santiago Pastorino
b970feedc2
Add compare-mode to x.py 2018-05-30 14:35:43 -03:00
Eduard-Mihai Burtescu
06d88cda08 rustc: rename mir::LocalDecl's syntactic_source_info to source_info. 2018-05-30 20:30:10 +03:00
Eduard-Mihai Burtescu
6c53972478 rustc: turn mir::LocalDecl's visibility_source_info into a SourceScope. 2018-05-30 20:30:10 +03:00