Mark Simulacrum [Sun, 1 Apr 2018 16:04:54 +0000 (18:04 +0200)]
Rollup merge of #49451 - QuietMisdreavus:epoch-doctests, r=GuillaumeGomez
rustdoc: add an --edition flag to compile docs/doctests with a certain edition
To correspond with the 2018 edition, this adds a (currently unstable) `--edition` flag to rustdoc that makes it compile crates and doctests with the given edition. Once this lands, Cargo should be updated to pass this flag when the edition configuration option is given.
Auto merge of #49478 - Phlosioneer:fix-windows-file-not-found, r=petrochenkov
Fix escaped backslash in windows file not found message
When a module is declared, but no matching file exists, rustc gives
an error like `help: name the file either foo.rs or foo/mod.rs inside
the directory "src/bar"`. However, at on windows, the backslash was
double-escaped when naming the directory.
It did this because the string was printed in debug mode (`"{:?}"`) to
surround it with quotes. However, it should just be printed like any
other directory in an error message and surrounded by escaped quotes,
rather than relying on the debug print to add quotes (`"\"{}\""`).
I also checked the test suite to see if this output is being correctly tested. It's not - it only tests up to the word "directory". Presumably this is so that the test is not dependent on its exact position in the source tree. I don't know a better way to test this, unless the test suite supports regex?
Auto merge of #49545 - alexcrichton:proc-macro-fixes, r=eddyb
proc_macro: Tweak doc comments and negative literals
This commit tweaks the tokenization of a doc comment to use `#[doc = "..."]`
like `macro_rules!` does (instead of treating it as a `Literal` token).
Additionally it fixes treatment of negative literals in the compiler, for
exapmle `Literal::i32(-1)`. The current fix is a bit of a hack around the
current compiler implementation, providing a fix at the proc-macro layer rather
than the libsyntax layer.
Auto merge of #49530 - petrhosek:empty-extra-flags, r=cramertj
Only include space in RUSTFLAGS extra flags if not empty
When the RUSTFLAGS_STAGE_{1,2} is not set, including a space means
the string will always be non-empty and RUSTFLAGS will be always be
reset which breaks other ways of setting these such as through config
in CARGO_HOME.
Auto merge of #49527 - petrhosek:fast-submodules, r=alexcrichton
Handle fast-submodules option correctly
This option was introduced in 72cb109bec8, but it uses two different
spellings (`fast-submodule` in `bootstrap.py` vs `fast-submodules` in
`config.toml.example`) and isn't handled by Rust bootstrap which means
that any attempt to set this flag fails.
Alex Crichton [Fri, 30 Mar 2018 09:55:54 +0000 (02:55 -0700)]
proc_macro: Tweak doc comments and negative literals
This commit tweaks the tokenization of a doc comment to use `#[doc = "..."]`
like `macro_rules!` does (instead of treating it as a `Literal` token).
Additionally it fixes treatment of negative literals in the compiler, for
exapmle `Literal::i32(-1)`. The current fix is a bit of a hack around the
current compiler implementation, providing a fix at the proc-macro layer rather
than the libsyntax layer.
bors [Sat, 31 Mar 2018 04:09:43 +0000 (04:09 +0000)]
Auto merge of #49472 - nikomatsakis:nll-optimize-constraint-prop-1, r=pnkfelix
optimize NLL constraint propagation a little
Removes a bone-headed hot spot in NLL constant propagation; we were re-allocating the stack vector and hashmap as we repeated the DFS. This change shares those resources across each call.
It also modifies the constraint list to be a linked list; arguably I should revert that, though, as this didn't turn out to be a perf hit and perhaps the old code was clearer? (Still, the new style appeals to me.)
Petr Hosek [Sat, 31 Mar 2018 02:37:08 +0000 (19:37 -0700)]
Only include space in RUSTFLAGS extra flags if not empty
When the RUSTFLAGS_STAGE_{1,2} is not set, including a space means
the string will always be non-empty and RUSTFLAGS will be always be
reset which breaks other ways of setting these such as through config
in CARGO_HOME.
Petr Hosek [Fri, 30 Mar 2018 23:42:57 +0000 (16:42 -0700)]
Handle fast-submodules option correctly
This option was introduced in 72cb109bec8, but it uses two different
spellings (fast-submodule vs fast-submodules) and isn't handled by
Rust bootstrap which means that any attempt to set this flag fails.
This addition landed in https://github.com/rust-lang/rust/pull/49305 and turned out to break crates that had their own copy of `TryFrom` in order to use it on the Stable channel :(
We’ll explore the possibility of the 2018 edition having a different prelude that includes this traits. However per the editions RFC this requires implementing a warning in the 2015 edition for code that *would* break.
bors [Fri, 30 Mar 2018 14:11:35 +0000 (14:11 +0000)]
Auto merge of #49425 - alexcrichton:disallow-inline-always, r=petrochenkov
rustc: Forbid #[inline(always)] with #[target_feature]
Once a target feature is enabled for a function that means that it in general
can't be inlined into other functions which don't have that target feature
enabled. This can cause both safety and LLVM issues if we were to actually
inline it, so `#[inline(always)]` both can't be respected and would be an error
if we did so!
Today LLVM doesn't inline functions with different `#[target_feature]`
annotations, but it turns out that if one is tagged with `#[inline(always)]`
it'll override this and cause scary LLVM error to arise!
This commit fixes this issue by forbidding these two attributes to be used in
conjunction with one another.
kennytm [Thu, 29 Mar 2018 23:31:16 +0000 (01:31 +0200)]
Rollup merge of #49466 - glandium:master, r=rkruppe
Use f{32,64}::to_bits for is_zero test in vec::SpecFromElem
vec::SpecFromElem provides an optimization to use calloc to fill a Vec
when the element given to fill the Vec is represented by 0.
For floats, the test for that currently used is `x == 0. &&
x.is_sign_positive()`. When compiled in a standalone function, rustc
generates the following assembly:
```
xorps xmm1, xmm1
ucomisd xmm0, xmm1
setnp al
sete cl
and cl, al
movq rax, xmm0
test rax, rax
setns al
and al, cl
ret
```
A simpler test telling us whether the value is represented by 0, is
`x.to_bits() == 0`, which rustc compiles to:
```
movq rax, xmm0
test rax, rax
sete al
ret
```
Not that the test is hot in any way, but it also makes it clearer what
the intent in the rust code is.
kennytm [Thu, 29 Mar 2018 23:31:15 +0000 (01:31 +0200)]
Rollup merge of #49464 - ollie27:rustbuild_junction_handle_leak, r=alexcrichton
rustbuild: Don't leak file handles when creating junctions on Windows
This fixes building the compiler docs because stage1-rustc\x86_64-pc-windows-msvc\doc is used twice which doesn't work if we still have a handle from the first time.
bors [Thu, 29 Mar 2018 23:19:52 +0000 (23:19 +0000)]
Auto merge of #49316 - alexcrichton:start-group-end-group, r=michaelwoerister
rustc: Group linked libraries where needed
This commit fixes a longstanding issue with the compiler with circular
dependencies between libcore and libstd. The `core` crate requires at least one
symbol, the ability to unwind. The `std` crate is the crate which actually
defines this symbol, but the `std` crate also depends on the `core` crate.
This circular dependency is in general disallowed in Rust as crates cannot have
cycles amongst them. A special exception is made just for core/std, but this is
also unfortunately incompatible with how GNU linkers work. GNU linkers will
process undefined symbols in a left-to-right fashion, only actually linking an
rlib like libstd if there are any symbols used from it. This strategy is
incompatible with circular dependencies because if we otherwise don't use
symbols from libstd we don't discover that we needed it until we're later
processing libcore's symbols!
To fix this GNU linkers support the `--start-group` and `--end-group` options
which indicate "libraries between these markers may have circular dependencies
amongst them. The linker invocation has been updated to automatically pass these
arguments when we're invoking a GNU linker and automatically calculate where the
arguments need to go (around libstd and libcore)
Alex Crichton [Fri, 23 Mar 2018 21:33:22 +0000 (14:33 -0700)]
rustc: Group linked libraries where needed
This commit fixes a longstanding issue with the compiler with circular
dependencies between libcore and libstd. The `core` crate requires at least one
symbol, the ability to unwind. The `std` crate is the crate which actually
defines this symbol, but the `std` crate also depends on the `core` crate.
This circular dependency is in general disallowed in Rust as crates cannot have
cycles amongst them. A special exception is made just for core/std, but this is
also unfortunately incompatible with how GNU linkers work. GNU linkers will
process undefined symbols in a left-to-right fashion, only actually linking an
rlib like libstd if there are any symbols used from it. This strategy is
incompatible with circular dependencies because if we otherwise don't use
symbols from libstd we don't discover that we needed it until we're later
processing libcore's symbols!
To fix this GNU linkers support the `--start-group` and `--end-group` options
which indicate "libraries between these markers may have circular dependencies
amongst them. The linker invocation has been updated to automatically pass these
arguments when we're invoking a GNU linker and automatically calculate where the
arguments need to go (around libstd and libcore)
bors [Thu, 29 Mar 2018 11:34:17 +0000 (11:34 +0000)]
Auto merge of #49163 - SimonSapin:range-bounds, r=alexcrichton
Rename RangeArgument to RangeBounds, move it and Bound to libcore
As proposed in the tracking issue: https://github.com/rust-lang/rust/issues/30877
Changes to *stable* items:
* `core::ops::Bound` / `std::ops::Bound` is new
* `std::collections::Bound` is a deprecated reexport of it (does this actually cause a warning?)
Changes to *unstable* items
* `alloc::Bound` is gone
* `alloc::range::RangeArgument` is moved to `core::ops::RangeBounds` / `std::ops::RangeBounds`
* `alloc::range` is gone
* `std::collections::range::RangeArgument` is deprecated reexport, to be removed later
* `std::collections::range` is deprecated, to be removed later
* `impl RangeBounds<T> for Range{,From,To,Inclusive,ToInclusive}<&T>` are added
The idea of replacing this trait with a type to be used with `Into<_>` is left for future consideration / work.
Phlosioneer [Thu, 29 Mar 2018 11:03:24 +0000 (07:03 -0400)]
Fix escaped backslash in windows file not found message
When a module is declared, but no matching file exists, rustc gives
an error like 'help: name the file either foo.rs or foo/mod.rs inside
the directory "src/bar"'. However, at on windows, the backslash was
double-escaped when naming the directory.
It did this because the string was printed in debug mode ( "{:?}" ) to
surround it with quotes. However, it should just be printed like any
other directory in an error message and surrounded by escaped quotes,
rather than relying on the debug print to add quotes ( "\"{}\"" ).
bors [Thu, 29 Mar 2018 05:44:40 +0000 (05:44 +0000)]
Auto merge of #49313 - sgrif:sg-revert-stuff, r=nikomatsakis
Remove universes from `ty::ParamEnv`
This change was never meant to land. #48407 takes an alternate approach. However, that PR is now blocked on some issues with canonicalization, and rebasing these reverts gets harder each time, so let's just get this bit out of the way now.
Mike Hommey [Thu, 29 Mar 2018 00:34:39 +0000 (09:34 +0900)]
Use f{32,64}::to_bits for is_zero test in vec::SpecFromElem
vec::SpecFromElem provides an optimization to use calloc to fill a Vec
when the element given to fill the Vec is represented by 0.
For floats, the test for that currently used is `x == 0. &&
x.is_sign_positive()`. When compiled in a standalone function, rustc
generates the following assembly:
```
xorps xmm1, xmm1
ucomisd xmm0, xmm1
setnp al
sete cl
and cl, al
movq rax, xmm0
test rax, rax
setns al
and al, cl
ret
```
A simpler test telling us whether the value is represented by 0, is
`x.to_bits() == 0`, which rustc compiles to:
```
movq rax, xmm0
test rax, rax
sete al
ret
```
Not that the test is hot in any way, but it also makes it clearer what
the intent in the rust code is.
Oliver Middleton [Wed, 28 Mar 2018 21:43:23 +0000 (22:43 +0100)]
rustbuild: Don't leak file handles when creating junctions on Windows
This fixes building the compiler docs because stage1-rustc\x86_64-pc-windows-msvc\doc is used twice which
doesn't work if we still have a handle from the first time.
kennytm [Wed, 28 Mar 2018 15:55:13 +0000 (17:55 +0200)]
Rollup merge of #49428 - cuviper:llvm6-target_feature, r=alexcrichton
Enable target_feature on any LLVM 6+
In `LLVMRustHasFeature()`, rather than using `MCInfo->getFeatureTable()`
that is specific to Rust's LLVM fork, we can use this in LLVM 6:
/// Check whether the subtarget features are enabled/disabled as per
/// the provided string, ignoring all other features.
bool checkFeatures(StringRef FS) const;
Now rustc using external LLVM can also have `target_feature`.
kennytm [Wed, 28 Mar 2018 15:55:08 +0000 (17:55 +0200)]
Rollup merge of #49364 - wesleywiser:incr_handle_load_failure, r=michaelwoerister
[incremental] Don't panic if decoding the cache fails
If the cached data can't be loaded from disk, just issue a warning to
the user so they know why compilation is taking longer than usual but
don't fail the entire compilation since we can recover by ignorning the
on disk cache.
In the same way, if the disk cache can't be deserialized (because it has
been corrupted for some reason), report the issue as a warning and
continue without failing the compilation. `Decodable::decode()` tends to
panic with various errors like "entered unreachable code" or "index out
of range" if the input data is corrupted. Work around this by catching
panics from the `decode()` calls and continuing without the cached data.
bors [Wed, 28 Mar 2018 15:31:29 +0000 (15:31 +0000)]
Auto merge of #49394 - cramertj:stable-match-bindings, r=nikomatsakis
Stabilize match_default_bindings
This includes a submodule update to rustfmt
in order to allow a stable feature declaration.
r? @nikomatsakis
cc https://github.com/rust-lang/rust/issues/42640
Many of the tests this PR touches are merely testing the current lack of desired future behavior around https://github.com/rust-lang/rust/issues/44849 and https://github.com/rust-lang/rust/issues/44848 (cc @tschottdorf). I noticed the bullets for those items were checked on the tracking issue-- I've unchecked them, as they don't appear to have been completed and I don't see any comments indicating that we don't want to pursue them further. Still, I think it's fine to stabilize the current behavior, as I think expanding it in the future should be backwards-compatible.