Auto merge of #28130 - alexcrichton:fix-msvc-static-tls-dtor, r=brson
Running TLS destructors for a MSVC Windows binary requires the linker doesn't
elide the `_tls_used` or `__tls_used` symbols (depending on the architecture).
This is currently achieved via a `#[link_args]` hack but this only works for
dynamically linked binaries because the link arguments aren't propagated to
statically linked binaries.
This commit alters the strategy to instead emit a volatile load from those
symbols so LLVM can't elide it, forcing the reference to the symbol to stay
alive as long as the callback function stays alive (which we've made sure of
with the `#[linkage]` attribute).
Alex Crichton [Mon, 31 Aug 2015 17:28:07 +0000 (10:28 -0700)]
std: Run TLS destructors in a statically linked binary
Running TLS destructors for a MSVC Windows binary requires the linker doesn't
elide the `_tls_used` or `__tls_used` symbols (depending on the architecture).
This is currently achieved via a `#[link_args]` hack but this only works for
dynamically linked binaries because the link arguments aren't propagated to
statically linked binaries.
This commit alters the strategy to instead emit a volatile load from those
symbols so LLVM can't elide it, forcing the reference to the symbol to stay
alive as long as the callback function stays alive (which we've made sure of
with the `#[linkage]` attribute).
Auto merge of #28152 - llogiq:master, r=Manishearth
This changes a lot of `.to_string()` to `.to_owned()`, removes a few redundant closures, and changes some `match`es to `if let`s.
I'm currently in the process of trying out clippy and acting on its suggestions. I started with compiletest, because we use it to test clippy, too. If this finds positive reception, I may continue refactoring other parts of the rust codebase.
Auto merge of #28137 - nrc:remove-non-multi, r=huonw
This is a [breaking-change] for syntax extension authors. The fix is to use MultiModifier or MultiDecorator, which have the same functionality but are more flexible. Users of syntax extensions are unaffected.
I believe everything that doesn't take a constant integer up to SSE4.2
should now be correct (I don't have any reason to believe that those
that do take constant integers are wrong; they're just more complicated
and I just haven't tested them in detail).
(Also, takes out two unused code paths from trans.)
Huon Wilson [Mon, 31 Aug 2015 23:47:58 +0000 (16:47 -0700)]
Fix some typos in SSE-AVX intrinsics.
I believe everything that doesn't take a constant integer up to SSE4.2
should now be correct (I don't have any reason to believe that those
that do take constant integers are wrong; they're just more complicated
and I just haven't tested them in detail).
Nick Cameron [Tue, 1 Sep 2015 01:16:03 +0000 (13:16 +1200)]
Remove the Modifier and Decorator kinds of syntax extensions.
This is a [breaking-change] for syntax extension authors. The fix is to use MultiModifier or MultiDecorator, which have the same functionality but are more flexible. Users of syntax extensions are unaffected.
bors [Mon, 31 Aug 2015 21:34:08 +0000 (21:34 +0000)]
Auto merge of #28079 - huonw:simd, r=alexcrichton
This adds a new Python script (compatible with 2.7 and 3.x) that will consume some JSON files that define a platform's intrinsics. It can output a file that defines the intrinsics in the compiler, or an `extern` block that will import them.
The complexity of the generator is to be DRY: platforms (especially ARM and AArch64) have a lot of repetition with their intrinsics, for different versions with different types, so being able to write it once is nice.
Tobias Bucher [Sat, 29 Aug 2015 16:30:05 +0000 (18:30 +0200)]
Implement `FixedSizeArray` for all fixed size arrays
Do so by using the fact that fixed size arrays (like `[u8; 8]` can be coerced
to slices `&[u8]`, this is expressed through the trait `Unsize<[T]>` that all
fixed size arrays implement.
bors [Mon, 31 Aug 2015 07:24:39 +0000 (07:24 +0000)]
Auto merge of #27980 - tbu-:pr_cloexec_dup, r=alexcrichton
Still needs values of F_DUPFD_CLOEXEC on other OSs.
For Bitrig, NetBSD and OpenBSD the constant was incorrectly in posix01, when
it's actually posix08. In order to maintain backwards-compatiblity, the
constant was only copied, not moved.
bors [Sun, 30 Aug 2015 23:18:34 +0000 (23:18 +0000)]
Auto merge of #27983 - barosl:respect-color, r=nrc
Currently, `early_error` and `early_warn` in `librustc::session` always use `ColorConfig::Auto`. Modify them to follow the color configuration set by the `--color` option.
As colored output is also printed during the early stage, parsing the `--color` option should be done as early as possible. However, there are still some cases when the output needs to be colored before knowing the exact color settings. In these cases, it will be defaulted to `ColorConfig::Auto`, which is the same as before.
Tobias Bucher [Mon, 24 Aug 2015 11:57:11 +0000 (13:57 +0200)]
Atomically set CLOEXEC on duplicated sockets
For Bitrig, NetBSD and OpenBSD the constant was incorrectly in posix01, when
it's actually posix08, so we move it. This is a [breaking-change], but we
already had one in #27930.
Fix NetBSD's F_DUPFD_CLOEXEC constant.
For a similar feature detection, see this musl thread:
http://comments.gmane.org/gmane.linux.lib.musl.general/2963
This assumes that an int literal has type `c_int` for varidic functions.
Huon Wilson [Fri, 28 Aug 2015 23:39:29 +0000 (16:39 -0700)]
Add the platform intrinsic generator script.
This python script will consume an appropriately formatted JSON file and
output either a Rust file for use in librustc_platform_intrinsics, or an
extern block for importing the intrinsics in an external library.
Huon Wilson [Fri, 28 Aug 2015 23:34:07 +0000 (16:34 -0700)]
Add support for aggregates in platform intrinsics.
This adds support for flattened intrinsics, which are called in Rust
with tuples but in LLVM without them (e.g. `foo((a, b))` becomes `foo(a,
b)`). Unflattened ones could be supported, but are not yet.
Huon Wilson [Fri, 28 Aug 2015 23:32:25 +0000 (16:32 -0700)]
Separate integers into signed and unsigned.
This is necessary to reflect the ARM APIs accurately, since some
functions explicitly take an unsigned parameter and a signed one, of the
same integer shape, so the no-duplicates check will fail unless we
distinguish.
bors [Sat, 29 Aug 2015 06:36:51 +0000 (06:36 +0000)]
Auto merge of #28080 - huonw:travis++, r=alexcrichton
It is very difficult to find tidy problems in the midst of the output of
the LLVM/jemalloc/etc. build, and travis is great for the former, so
lets remove that problem.
Huon Wilson [Sat, 29 Aug 2015 05:59:00 +0000 (22:59 -0700)]
Run tidy by itself on travis.
It is very difficult to find tidy problems in the midst of the output of
the LLVM/jemalloc/etc. build, and travis is great for the former, so
lets remove that problem.
bors [Sat, 29 Aug 2015 01:43:16 +0000 (01:43 +0000)]
Auto merge of #28042 - Jake-Shadle:master, r=steveklabnik
Encountered an issue with `pacman` while going through the guide for installing the `mingw` toolchain on Windows with `msys2`, after some googling I found the [solution](https://github.com/Alexpux/MSYS2-packages/issues/163#issuecomment-73555971).
I thought it would be good to update the README so people don't get frustrated. :smiley: