bors [Wed, 10 May 2017 08:54:50 +0000 (08:54 +0000)]
Auto merge of #41764 - scottmcm:faster-reverse, r=brson
Make [u8]::reverse() 5x faster
Since LLVM doesn't vectorize the loop for us, do unaligned reads of a larger type and use LLVM's bswap intrinsic to do the reversing of the actual bytes. cfg!-restricted to x86 and x86_64, as I assume it wouldn't help on things like ARMv5.
Also makes [u16]::reverse() a more modest 1.5x faster by loading/storing u32 and swapping the u16s with ROT16.
Thank you ptr::*_unaligned for making this easy :)
Benchmark results (from my i5-2500K):
```text
# Before
test slice::reverse_u8 ... bench: 273,836 ns/iter (+/- 15,592) = 3829 MB/s
test slice::reverse_u16 ... bench: 139,793 ns/iter (+/- 17,748) = 7500 MB/s
test slice::reverse_u32 ... bench: 74,997 ns/iter (+/- 5,130) = 13981 MB/s
test slice::reverse_u64 ... bench: 47,452 ns/iter (+/- 2,213) = 22097 MB/s
# After
test slice::reverse_u8 ... bench: 52,170 ns/iter (+/- 3,962) = 20099 MB/s
test slice::reverse_u16 ... bench: 93,330 ns/iter (+/- 4,412) = 11235 MB/s
test slice::reverse_u32 ... bench: 74,731 ns/iter (+/- 1,425) = 14031 MB/s
test slice::reverse_u64 ... bench: 47,556 ns/iter (+/- 3,025) = 22049 MB/s
```
If you're curious about the assembly, instead of doing this
```
movzx eax, byte ptr [rdi]
movzx ecx, byte ptr [rsi]
mov byte ptr [rdi], cl
mov byte ptr [rsi], al
```
it does this
```
mov rax, qword ptr [rdx]
mov rbx, qword ptr [r11 + rcx - 8]
bswap rbx
mov qword ptr [rdx], rbx
bswap rax
mov qword ptr [r11 + rcx - 8], rax
```
bors [Wed, 10 May 2017 03:27:36 +0000 (03:27 +0000)]
Auto merge of #41659 - bluss:clone-split-whitespace, r=aturon
impl Clone for .split_whitespace()
Use custom closure structs for the predicates so that the iterator's
clone can simply be derived. This should also reduce virtual call
overhead by not using function pointers.
bors [Tue, 9 May 2017 19:31:35 +0000 (19:31 +0000)]
Auto merge of #41830 - jonhoo:update-cargo, r=alexcrichton
Bump cargo for rust-lang/cargo#4000
rust-lang/cargo#4000 recently landed, which fixes warnings about using `-Z` when `CARGO_INCREMENTAL` is set while running stable/beta builds. As #41751 has now landed, these warnings will turn to errors in the next release, so getting the cargo fix in place is necessary unless we want people confused about why they can no longer compile anything on stable/beta.
bors [Tue, 9 May 2017 16:31:07 +0000 (16:31 +0000)]
Auto merge of #41814 - gamazeps:thread-struct-doc, r=steveklabnik
[Doc] improve `thread::Thread` and `thread::Builder` documentations
Part of #29378
- Adds information about the stack_size when using `Builder`. This might be considered too low level, but I assume that if someone wants to create their own builder instead of using `thread::spawn` they may be interested in that info.
- Updates the `thread::Thread` structure doc, mostly by explaining how to get one, the previous example was removed because it was not related to `thread::Thread`, but rather to `thread::Builder::name`.
Not much is present there, mostly because this API is not often used (the only method that seems useful is `unpark`, which is documented in #41809).
bors [Tue, 9 May 2017 11:55:37 +0000 (11:55 +0000)]
Auto merge of #41709 - michaelwoerister:close-metadata-ich-holes, r=nikomatsakis
incr.comp.: Hash more pieces of crate metadata to detect changes there.
This PR adds incr. comp. hashes for non-`Entry` pieces of data in crate metadata.
The first part of it I like: `EntryBuilder` is refactored into the more generally applicable `IsolatedEncoder` which provides means of encoding something into metadata while also feeding the encoded data into an incr. comp. hash. We already did this for `Entry`, now we are doing it for various other pieces of data too, like the set of exported symbols and so on. The hashes generated there are persisted together with the per-`Entry` hashes and are also used for dep-graph dirtying the same way.
The second part of the PR I'm not entirely happy with: In order to make sure that we don't forget registering a read to the new `DepNodes` introduced here, I added the `Tracked<T>` struct. This struct wraps a value and requires a `DepNode` when accessing the wrapped value. This makes it harder to overlook adding read edges in the right places and works just fine.
However, crate metadata is already used in places where there is no `tcx` yet or even in places where no `cnum` has been assigned -- this makes it harder to apply this feature consistently or implement it ergonomically. The result is not too bad but there's a bit more code churn and a bit more opportunity to get something wrong than I would have liked. On the other hand, wrapping things in `Tracked<T>` already has revealed some bugs, so there's definitely some value in it.
This is still a work in progress:
- [x] I need to write some test cases.
- [x] Accessing the CodeMap should really be dependency tracked too, especially with the new path-remapping feature.
Corey Farwell [Tue, 9 May 2017 02:34:52 +0000 (22:34 -0400)]
Rollup merge of #41838 - z1mvader:fix_fn_args_coerce_closure, r=nikomatsakis
Fixed argument inference for closures when coercing into 'fn'
This fixes https://github.com/rust-lang/rust/issues/41755. The tests `compile-fail/closure-no-fn.rs` and `compile-fail/issue-40000.rs` were modified. A new test `run-pass/closure_to_fn_coercion-expected-types.rs` was added
1) changed "long way into" to "long way toward"
2) changed "developer lives" to "developers' lives"
3) removed the "either... or..." format from second paragraph because there are more than 2 options
4) Minor revisions to paragraphs 3-6 to make them more consistent in format and to fix minor grammar issues.
Jing Zhao [Mon, 8 May 2017 20:33:54 +0000 (13:33 -0700)]
Grammar fixes to rustc-ux-guidelines.md
1) changed "long way into" to "long way toward"
2) changed "developer lives" to "developers' lives"
3) removed the "either... or..." format from second paragraph because there are more than 2 options
4) Minor revisions to paragraphs 3-6 to make them more consistent in format and to fix minor grammar issues.
bors [Mon, 8 May 2017 12:00:22 +0000 (12:00 +0000)]
Auto merge of #41745 - oli-obk:diagnostics, r=jonathandturner
Remove need for &format!(...) or &&"" dances in `span_label` calls
These were always a thorn in my eye. Note that this will monomorphize to two impls, one for `String` and one for `&str`. But I think that cost is worth the ergonomics at the call sites that can be seen throughout this PR.
bors [Mon, 8 May 2017 05:29:24 +0000 (05:29 +0000)]
Auto merge of #41818 - michaelwu:hvx-v60, r=nagisa
Add support for Hexagon v60 HVX intrinsics
HVX is a SIMD coprocessor available on newer hexagon cores. It can be configured for 512 or 1024 bit registers, and some instructions use pairs of registers. It only does integer operations, but it probably has every integer operation you'd want for 8/16/32 bit elements.
There are a lot of intrinsics. The generator outputs 582 of them. I probably got some wrong. I did some scripting to make sure that every llvm intrinsic name exists, but intrinsic names provided for programs have only been compared by eye to Qualcomm's own names. 64/128 is also appended to the names to select between 512/1024 bit. The C intrinsics don't do this, but they only expose one set, selected at compile time.
The json specifying the intrinsics required a bit of duplication since I didn't see an easy way to specify combinations of signed/unsigned types (eg. u(8-16) and s(16-32)). I also didn't see an easy way to specify variants of instructions like saturating or rounding.
Basic multiplication and load/store tested on the hexagon simulator.
bors [Sun, 7 May 2017 16:20:15 +0000 (16:20 +0000)]
Auto merge of #41791 - Mark-Simulacrum:doc-guidelines, r=frewsxcv
Minor cleanup of UX guidelines.
I think this fixes https://github.com/rust-lang/rust/issues/34808. It covers the [long error code explanations normalization] by linking to the RFC, and cleaning up the list where long diagnostics are defined. While the [error message overhaul] isn't covered directly, I'm not really sure that more than the [existing section] on the error/warning/help messages is warranted; the overhaul linked didn't really specify any new guidelines, primarily just changing the output format.
bors [Sun, 7 May 2017 13:57:36 +0000 (13:57 +0000)]
Auto merge of #40857 - estebank:recursive, r=arielb1
Point at fields that make the type recursive
On recursive types of infinite size, point at all the fields that make
the type recursive.
```rust
struct Foo {
bar: Bar,
}
struct Bar {
foo: Foo,
}
```
outputs
```
error[E0072]: recursive type `Foo` has infinite size
--> file.rs:1:1
1 | struct Foo {
| ^^^^^^^^^^ recursive type has infinite size
2 | bar: Bar,
| -------- recursive here
|
= help: insert indirection (e.g., a `Box`, `Rc`, or `&`) at some point to make `Foo` representable
error[E0072]: recursive type `Bar` has infinite size
--> file.rs:5:1
|
5 | struct Bar {
| ^^^^^^^^^^ recursive type has infinite size
6 | foo: Foo,
| -------- recursive here
|
= help: insert indirection (e.g., a `Box`, `Rc`, or `&`) at some point to make `Bar` representable
```
bors [Sun, 7 May 2017 10:52:26 +0000 (10:52 +0000)]
Auto merge of #41785 - Mark-Simulacrum:issue-41783, r=GuillaumeGomez
Allow # to appear in rustdoc code output.
"##" at the start of a trimmed rustdoc line is now cut to "#" and then
shown. If the user wanted to show "##", they can type "###".
I'm somewhat concerned about the potential implications for users, since this does make a potentially backwards-incompatible change. Previously, `##` had no special handling, and now we do change it. However, I'm not really sure what we can do here to improve this, and I can't think of any cases where `##` would likely be correct in a code block, though of course I could be wrong.
bors [Sat, 6 May 2017 22:14:43 +0000 (22:14 +0000)]
Auto merge of #41787 - jsheard:ulongptr, r=alexcrichton
Fix definitions of ULONG_PTR
The Windows type `ULONG_PTR` is supposed to be equivalent to `usize`, but several parts of the codebase currently define it as `u64`. Evidently this hasn't broken anything yet but it might cause annoying 32-bit-specific breakage in future.
See https://msdn.microsoft.com/en-gb/library/windows/desktop/aa383751(v=vs.85).aspx
bors [Sat, 6 May 2017 19:45:10 +0000 (19:45 +0000)]
Auto merge of #41786 - acdenisSK:an_to_a, r=frewsxcv
Fix "an" usage
Since the pr i reviewed on got merged way before the author had a chance to quickly change it, i just did it myself. (Or well, someone else asked me to, if you want me to be honest)
bors [Sat, 6 May 2017 02:01:00 +0000 (02:01 +0000)]
Auto merge of #41768 - rap2hpoutre:patch-4, r=frewsxcv
Add an example to std::thread::Result type
This PR is a part of https://github.com/rust-lang/rust/issues/29378. I submit this PR with the help (mentoring) of @steveklabnik. I'm still not sure my request is good enough but I don't want to spoil the issue with too much questions so I continue here. r? @steveklabnik
Corey Farwell [Fri, 5 May 2017 21:35:29 +0000 (17:35 -0400)]
Rollup merge of #41734 - nikomatsakis:incr-comp-refactor-variance, r=pnkfelix
Refactor variance and remove last `[pub]` map
This PR refactors variance to work in a more red-green friendly way. Because red-green doesn't exist yet, it has to be a bit hacky. The basic idea is this:
- We compute a big map with the variance for all items in the crate; when you request variances for a particular item, we read it from the crate
- We now hard-code that traits are invariant (which they are, for deep reasons, not gonna' change)
- When building constraints, we compute the transitive closure of all things within the crate that depend on what using `TransitiveRelation`
- this lets us gin up the correct dependencies when requesting variance of a single item
Ah damn, just remembered, one TODO:
- [x] Update the variance README -- ah, I guess the README updates I did are sufficient
Corey Farwell [Fri, 5 May 2017 21:35:28 +0000 (17:35 -0400)]
Rollup merge of #41722 - F001:warnTilde, r=petrochenkov
Suggest `!` for bitwise negation when encountering a `~`
Fix #41679
Here is a program
```rust
fn main() {
let x = ~1;
}
```
It's output:
```
error: `~` can not be used as an unary operator
--> /home/fcc/temp/test.rs:4:13
|
4 | let x = ~1;
| ^^
|
= help: use `!` instead of `~` if you meant to bitwise negation
```
Corey Farwell [Fri, 5 May 2017 21:35:26 +0000 (17:35 -0400)]
Rollup merge of #41582 - jonhoo:reread-nameservers-on-lookup-fail, r=alexcrichton
Reload nameserver information on lookup failure
As discussed in #41570, UNIX systems often cache the contents of `/etc/resolv.conf`, which can cause lookup failures to persist even after a network connection becomes available. This patch modifies lookup_host to force a reload of the nameserver entries following a lookup failure. This is in line with what many C programs already do (see #41570 for details). On systems with nscd, this should not be necessary, but not all systems run nscd.
Corey Farwell [Fri, 5 May 2017 21:35:25 +0000 (17:35 -0400)]
Rollup merge of #41512 - alexcrichton:fix-windows-tls-deadlock, r=BurntSushi
std: Avoid locks during TLS destruction on Windows
Gecko recently had a bug reported [1] with a deadlock in the Rust TLS
implementation for Windows. TLS destructors are implemented in a sort of ad-hoc
fashion on Windows as it doesn't natively support destructors for TLS keys. To
work around this the runtime manages a list of TLS destructors and registers a
hook to get run whenever a thread exits. When a thread exits it takes a look at
the list and runs all destructors.
Unfortunately it turns out that there's a lock which is held when our "at thread
exit" callback is run. The callback then attempts to acquire a lock protecting
the list of TLS destructors. Elsewhere in the codebase while we hold a lock over
the TLS destructors we try to acquire the same lock held first before our
special callback is run. And as a result, deadlock!
This commit sidesteps the issue with a few small refactorings:
* Removed support for destroying a TLS key on Windows. We don't actually ever
exercise this as a public-facing API, and it's only used during `lazy_init`
during racy situations. To handle that we just synchronize `lazy_init`
globally on Windows so we never have to call `destroy`.
* With no need to support removal the global synchronized `Vec` was tranformed
to a lock-free linked list. With the removal of locks this means that
iteration no long requires a lock and as such we won't run into the deadlock
problem mentioned above.
Note that it's still a general problem that you have to be extra super careful
in TLS destructors. For example no code which runs a TLS destructor on Windows
can call back into the Windows API to do a dynamic library lookup. Unfortunately
I don't know of a great way around that, but this at least fixes the immediate
problem that Gecko was seeing which is that with "well behaved" destructors the
system would still deadlock!
Corey Farwell [Fri, 5 May 2017 21:35:24 +0000 (17:35 -0400)]
Rollup merge of #41064 - Gankro:ptr-redux, r=alexcrichton
refactor NonZero, Shared, and Unique APIs
Major difference is that I removed Deref impls, as apparently LLVM has
trouble maintaining metadata with a `&ptr -> &ptr` API. This was cited
as a blocker for ever stabilizing this API. It wasn't that ergonomic
anyway.
* Added `get` to NonZero to replace Deref impl
* Added `ptr` getter to Shared/Unique to replace Deref impl
* Added Unique's `get` and `get_mut` conveniences to Shared
* Deprecated `as_mut_ptr` on Shared in favour of `ptr`
Note that Shared used to primarily expose only `*const` but there isn't
a good justification for that, so I made it `*mut`.
bors [Fri, 5 May 2017 20:44:15 +0000 (20:44 +0000)]
Auto merge of #41769 - alexcrichton:fix-doc-test, r=aturon
std: Prevent deadlocks in doctests on Windows
Windows historically has problems with threads panicking and the main thread
exiting at the same time, typically causing deadlocks. In the past (#25824)
we've joined on threads but this just prevents running the test for now to avoid
tampering with the example.
Alex Crichton [Fri, 5 May 2017 14:02:48 +0000 (07:02 -0700)]
std: Prevent deadlocks in doctests on Windows
Windows historically has problems with threads panicking and the main thread
exiting at the same time, typically causing deadlocks. In the past (#25824)
we've joined on threads but this just prevents running the test for now to avoid
tampering with the example.
Alex Crichton [Mon, 24 Apr 2017 18:34:16 +0000 (11:34 -0700)]
std: Avoid locks during TLS destruction on Windows
Gecko recently had a bug reported [1] with a deadlock in the Rust TLS
implementation for Windows. TLS destructors are implemented in a sort of ad-hoc
fashion on Windows as it doesn't natively support destructors for TLS keys. To
work around this the runtime manages a list of TLS destructors and registers a
hook to get run whenever a thread exits. When a thread exits it takes a look at
the list and runs all destructors.
Unfortunately it turns out that there's a lock which is held when our "at thread
exit" callback is run. The callback then attempts to acquire a lock protecting
the list of TLS destructors. Elsewhere in the codebase while we hold a lock over
the TLS destructors we try to acquire the same lock held first before our
special callback is run. And as a result, deadlock!
This commit sidesteps the issue with a few small refactorings:
* Removed support for destroying a TLS key on Windows. We don't actually ever
exercise this as a public-facing API, and it's only used during `lazy_init`
during racy situations. To handle that we just synchronize `lazy_init`
globally on Windows so we never have to call `destroy`.
* With no need to support removal the global synchronized `Vec` was tranformed
to a lock-free linked list. With the removal of locks this means that
iteration no long requires a lock and as such we won't run into the deadlock
problem mentioned above.
Note that it's still a general problem that you have to be extra super careful
in TLS destructors. For example no code which runs a TLS destructor on Windows
can call back into the Windows API to do a dynamic library lookup. Unfortunately
I don't know of a great way around that, but this at least fixes the immediate
problem that Gecko was seeing which is that with "well behaved" destructors the
system would still deadlock!
Jon Gjengset [Thu, 27 Apr 2017 16:58:52 +0000 (12:58 -0400)]
Reload nameserver information on lookup failure
As discussed in #41570, UNIX systems often cache the contents of
/etc/resolv.conf, which can cause lookup failures to persist even after
a network connection becomes available. This patch modifies lookup_host
to force a reload of the nameserver entries following a lookup failure.
This is in line with what many C programs already do (see #41570 for
details). On systems with nscd, this should not be necessary, but not
all systems run nscd.
Introduces an std linkage dependency on libresolv on macOS/iOS (which
also makes it necessary to update run-make/tools.mk).
Scott McMurray [Fri, 5 May 2017 03:28:34 +0000 (20:28 -0700)]
Make [u8]::reverse() 5x faster
Since LLVM doesn't vectorize the loop for us, do unaligned reads
of a larger type and use LLVM's bswap intrinsic to do the
reversing of the actual bytes. cfg!-restricted to x86 and
x86_64, as I assume it wouldn't help on things like ARMv5.
Also makes [u16]::reverse() a more modest 1.5x faster by
loading/storing u32 and swapping the u16s with ROT16.
Thank you ptr::*_unaligned for making this easy :)
Esteban Küber [Mon, 27 Mar 2017 16:35:27 +0000 (09:35 -0700)]
Point at fields that make the type recursive
On recursive types of infinite size, point at all the fields that make
the type recursive.
```rust
struct Foo {
bar: Bar,
}
struct Bar {
foo: Foo,
}
```
outputs
```
error[E0072]: recursive type `Foo` has infinite size
--> file.rs:1:1
1 | struct Foo {
| _^ starting here...
2 | | bar: Bar,
| | -------- recursive here
3 | | }
| |_^ ...ending here: recursive type has infinite size
|
= help: insert indirection (e.g., a `Box`, `Rc`, or `&`) at some point to make `Foo` representable
error[E0072]: recursive type `Bar` has infinite size
--> file.rs:5:1
|
5 | struct Bar {
| _^ starting here...
6 | | foo: Foo,
| | -------- recursive here
7 | | }
| |_^ ...ending here: recursive type has infinite size
|
= help: insert indirection (e.g., a `Box`, `Rc`, or `&`) at some point to make `Bar` representable
```