Remove the parts of atomic::Ordering's intro that wrongly claimed that
SeqCst prevents all reorderings around it.

Closes rust-lang#55196

Given the function

fn foo((_x, _): (LogDrop, LogDrop), (_, _y): (LogDrop, LogDrop)) {}

Prior to 1.12 we dropped both `_x` and `_y` before the rest of their
respective parameters, since then we dropped `_x` and `_y` after. The
original order appears to be the correct order, as the value created
later is dropped first, so we revert to that order and add a test for
it.

This requires adding a new method, `P::filter_map`.

This commit reduces instruction counts for various benchmarks by up to
0.7%.

This makes the error style consistent with the convention in error messages.

The std::io::read main documentation can lead to error because the
buffer is prefilled with 10 zeros that will pad the response.
Using an empty vector is better.

The `read_to_end` documentation is already correct though.

This is my first rust PR, don't hesitate to tell me if I did something
wrong.

…-dereferencing" error

…learer

…type

…ents

…ch_path`.

Instead of maybe storing its own sysroot and maybe deferring to the one
in `Session::opts`, just clone the latter when necessary so one is
always directly available. This removes the need for the getter.

`TryFrom<&[T]> for &[T; $N]` (note *reference* to an array) already exists,
but not needing to dereference makes type inference easier
for example when using `u32::from_be_bytes`.

Also add doc examples doing just that.

It's more idiomatic, makes the code shorter, and will help with the next
commit.

`FileSearch::search()` traverses one or more directories. For each
directory it generates a `Vec<PathBuf>` containing one element per file
in that directory.

In some benchmarks this occurs enough that the allocations done for the
`PathBuf`s are significant, and in practice a small number of
directories are being traversed over and over again. For example, when
compiling the `tokio-webpush-simple` benchmark, two directories are
traversed 58 times each. Each of these directories have more than 100
files.

This commit changes things so that all the `Vec<PathBuf>`s that will be
needed by a `Session` are precomputed when that `Session` is created;
they are stored in `SearchPath`. `FileSearch` gets a reference to the
necessary `SearchPath`s. This reduces instruction counts on several
benchmarks by 1--5%.

The commit also removes the barely-used `visited_dirs` hash in
`for_each_lib_searchPath`. It only detects if `tlib_path` is the same as
one of the previously seen paths, which is unlikely.

Returning an iterator leads to nicer code all around.

Within this `Iterator` implementation, a function `unsafe_get` is
defined which unsafely allows _unchecked_ indexing of any element in a
slice. This should be marked as _unsafe_, but it is not.

To address this issue, I removed that inner function.

Fixes rust-lang#21335.

This way all files needed by packagers now reside in toplevel

Signed-off-by: Marc-Antoine Perennou <[email protected]>

The alignment for the second element of a scalar pair is not the
same as for the first element. Make sure it is computed correctly
based on the element size.

…imNN

bootstrap: clean up a few clippy findings

remove useless format!()s
remove redundant field names in a few struct initializations
pass slice instead of a vector to a function
use is_empty() instead of comparisons to .len()

No functional change intended.

…oerister

Use sort_by_cached_key when the key function is not trivial/free

I'm not 100% sure about `def_path_hash` (everything it does is inlined) but it seems like a good idea at least for the rest, as they are cloning.

add test for issue rust-lang#21335

Running this test with LLVM assertions enabled doesn't seem to trigger an assertion on my Mac.

Fixes rust-lang#21335.

avoid features_untracked

The docs say to not use `features_untracked` when we have a tcx.

@oli-obk any particular reason why the untracked version is used all over const qualification?

…stjepang

atomic::Ordering: Get rid of misleading parts of intro

Remove the parts of atomic::Ordering's intro that wrongly claimed that
SeqCst prevents all reorderings around it.

Closes rust-lang#55196

This is a (minimal) alternative to rust-lang#55233.

I also wonder if it would be worth adding at least some warnings that atomics are often a footgun/hard to use correctly, similarly like `mem::transmute` or other functions have.

…rder, r=cramertj

Drop partially bound function parameters in the expected order

Given the function

```rust
fn foo((_x, _): (LogDrop, LogDrop), (_, _y): (LogDrop, LogDrop)) {}
```

Prior to 1.12.0 we dropped both `_x` and `_y` before the rest of their
respective parameters, since then we dropped `_x` and `_y` after. The
original order appears to be the correct order, as the value created
later is dropped first, so we revert to that order and add a test for
it.

While this is technically a breaking change, I can't work out how
anyone could be relying on this without making their code very
brittle. If this is considered to be too likely to break real world code
then I can revert the change and change the test to check for the
current order.

Reduce the amount of bold text at doc.rlo

Currently, all of the text is either huge or small. IMHO this is hard to read, so I propose that we make the second-level headings smaller, without making them any less prominent.

# Before:

![image](https://user-images.githubusercontent.com/8827840/48737315-50447e80-ec13-11e8-8243-6211dbba1aa1.png)

# After:

![image](https://user-images.githubusercontent.com/8827840/48737258-32771980-ec13-11e8-814e-e7851954c05a.png)

Overhaul `FileSearch` and `SearchPaths`

`FileSearch::search()` traverses one or more directories. For each
directory it generates a `Vec<PathBuf>` containing one element per file
in that directory.

In some benchmarks this occurs enough that the allocations done for the
`PathBuf`s are significant, and in practice a small number of
directories are being traversed over and over again. For example, when
compiling the `tokio-webpush-simple` benchmark, two directories are
traversed 58 times each. Each of these directories have more than 100
files.

We can do all the necessary traversals up front, when `Session` is created,
and get the `Vec<PathBuf>`s then.

This reduces instruction counts on several benchmarks by 1--5%.

r? @alexcrichton

CC @eddyb, @michaelwoerister, @nikomatsakis

…ikomatsakis

Enclose type in backticks for "non-exhaustive patterns" error

This makes the error style consistent with the convention in error messages.

…ake, r=TimNN

Fix small doc mistake on std::io::read::read_to_end

The std::io::read main documentation can lead to error because the buffer is prefilled with 10 zeros that will pad the response.
Using an empty vector is better.

The `read_to_end` documentation is already correct though.

This is my first rust PR, don't hesitate to tell me if I did something wrong.

…tsakis

Update an outdated comment in mir building

r? @eddyb

Assorted tweaks

- preallocate `VecDeque` in `Decodable::decode` (as it is done with other collections which can do it)
- add a FIXME to `String::from_utf16`

r? @RalfJung

…ikomatsakis

Add rustc-guide as a submodule

Adding this as a submodule will allow two things:
- Linking to the guide from doc.rlo
- Doing a link check as part of the rust CI build

Key question: Do we want to wait for the book to be filled out more? e.g. do we ever want to move it out of the nursery?

r? @nikomatsakis

cc @steveklabnik

Make std::os::unix/linux::fs::MetadataExt::a/m/ctime* documentation clearer

I was confused by this API so I clarified what they are doing.

I was wondering if I should try to unify more documentation and examples between `unix` and `linux` (e.g. “of the file” is used in `unix` to refer to the file these metadata is for, “of this file” in `linux`, “of the underlying file” in `std::fs::File`).

…komatsakis

drop glue takes in mutable references, it should reflect that in its type

When drop glue begins, it should retag, like all functions taking references do. But to do that, it needs to take the reference at a proper type: `&mut T`, not `*mut T`.

Failing to retag can mean that the memory the reference points to remains frozen, and `EscapeToRaw` on a frozen location is a NOP, meaning later mutations cause a Stacked Borrows violation.

Cc @nikomatsakis @gankro because Stacked Borrows
Cc @eddyb for the changes to miri argument passing (the intention is to allow passing `*mut [u8]` when `&mut [u8]` is expected and vice versa)

Fix ICE with feature self_struct_ctor

Fix rust-lang#56202.

…ithoutboats

Add TryFrom<&[T]> for [T; $N] where T: Copy

`TryFrom<&[T]> for &[T; $N]` (note *reference* to an array) already exists, but not needing to dereference makes type inference easier for example when using `u32::from_be_bytes`.

Also add doc examples doing just that.

…ikomatsakis

Suggest appropriate place for lifetime when declared after type arguments

…r=wesleywiser

Make JSON output from -Zprofile-json valid

r? @wesleywiser

cc rust-lang/rustc-perf#299

…cramertj

Remove unsafe `unsafe` inner function.

Within this `Iterator` implementation, a function `unsafe_get` is
defined which unsafely allows _unchecked_ indexing of any element in a
slice. This should be marked as _unsafe_, but it is not.

To address this issue, I removed that inner function.

…lexcrichton

Stabilize feature `macro_at_most_once_rep`

a.k.a. `?` Kleene operator 🎉

cc rust-lang#48075

r? @Centril

…aelwoerister

Update outdated code comments in StringReader

For the detection of newlines in the lexer, this is now done in `analyze_source_file.rs`.

…atsakis

rustc-guide has moved to rust-lang/

r? @nikomatsakis

…r=petrochenkov

Reuse the `P` in `InvocationCollector::fold_{,opt_}expr`.

This requires adding a new method, `P::filter_map`.

This commit reduces instruction counts for various benchmarks by up to
0.7%.

…ng-link, r=steveklabnik

Add missing doc link

r? @steveklabnik

move stage0.txt to toplevel directory

This way all files needed by packagers now reside in toplevel

Fix small typo in comment of thread::stack_size

Fix a typo in the documentation of std::ffi

Fix alignment of stores to scalar pair

The alignment for the second element of a scalar pair is not the same as for the first element, make sure it is calculated correctly. This fixes rust-lang#56267.

r? @eddyb