Alex Crichton [Wed, 30 Aug 2017 16:11:14 +0000 (11:11 -0500)]
Rollup merge of #44157 - alexcrichton:no-specializes-cache, r=eddyb
rustc: Remove `specialization_cache` in favor of a query
This commit removes the `specialization_cache` field of `TyCtxt` by moving it to
a dedicated query, which it turned out was already quite easily structured to do
so!
Alex Crichton [Wed, 30 Aug 2017 16:11:12 +0000 (11:11 -0500)]
Rollup merge of #44125 - SergioBenitez:master, r=nrc
Initial diagnostic API for proc-macros.
This commit introduces the ability to create and emit `Diagnostic` structures from proc-macros, allowing for proc-macro authors to emit warning, error, note, and help messages just like the compiler does.
The API is somewhat based on the diagnostic API already present in `rustc` with several changes that improve usability. The entry point into the diagnostic API is a new `Diagnostic` type which is primarily created through new `error`, `warning`, `help`, and `note` methods on `Span`. The `Diagnostic` type records the diagnostic level, message, and optional `Span` for the top-level diagnostic and contains a `Vec` of all of the child diagnostics. Child diagnostics can be added through builder methods on `Diagnostic`.
A typical use of the API may look like:
```rust
let token = parse_token();
let val = parse_val();
val.span
.error(format!("expected A but found {}", val))
.span_note(token.span, "because of this token")
.help("consider using a different token")
.emit();
```
Alex Crichton [Wed, 30 Aug 2017 16:11:10 +0000 (11:11 -0500)]
Rollup merge of #44089 - alexcrichton:trait-proc-macro, r=nrc
rustc: Fix proc_macro expansions on trait methods
This commit fixes procedural macro attributes being attached to trait methods,
ensuring that they get resolved and expanded as other procedural macro
attributes. The bug here was that `current_module` on the resolver was
accidentally set to be a trait when it's otherwise only ever expecting a
`mod`/block module. The actual fix here came from @jseyfried, I'm just helping
to land it in the compiler!
bors [Wed, 30 Aug 2017 11:08:26 +0000 (11:08 +0000)]
Auto merge of #43968 - petrochenkov:span2, r=michaelwoerister
Make fields of `Span` private
I actually tried to intern spans and benchmark the result<sup>*</sup>, and this was a prerequisite.
This kind of encapsulation will be a prerequisite for any other attempt to compress span's representation, so I decided to submit this change alone.
The issue https://github.com/rust-lang/rust/issues/43088 seems relevant, but it looks like `SpanId` won't be able to reuse this interface, unless the tables are global (like interner that I tried) and are not a part of HIR.
r? @michaelwoerister anyway
<sup>*</sup> Interning means 2-3 times more space is required for a single span, but duplicates are free. In practice it turned out that duplicates are not *that* common, so more memory was wasted by interning rather than saved.
bors [Wed, 30 Aug 2017 08:06:42 +0000 (08:06 +0000)]
Auto merge of #43903 - oli-obk:alignto, r=aturon
Add align_offset intrinsic
see https://github.com/rust-lang/rfcs/pull/2043 for details and the plan towards stabilization (reexport in `core::mem` via various convenience functions)
as per @scottmcm 's [comment](https://github.com/rust-lang/rfcs/pull/2043#issuecomment-316818169), this is just the intrinsic (which is obviously unstable).
bors [Wed, 30 Aug 2017 05:30:28 +0000 (05:30 +0000)]
Auto merge of #43880 - arielb1:noninvasive-probe, r=nikomatsakis
Remove the trait selection impl in method::probe
This removes the hacky trait selection reimplementation in `method::probe`, which occasionally comes and causes problems.
There are 2 issues I've found with this approach:
1. The older implementation sometimes had a "guess" type from an impl, which allowed subtyping to work. This is why I needed to make a change in `libtest`: there's an `impl<A> Clone for fn(A)` and we're calling `<for<'a> fn(&'a T) as Clone>::clone`. The older implementation would do a subtyping between the impl type and the trait type, so it would do the check for `<fn(A) as Clone>::clone`, and confirmation would continue with the subtyping. The newer implementation directly passes `<for<'a> fn(&'a T) as Clone>::clone` to selection, which fails. I'm not sure how big of a problem that would be in reality, especially after #43690 would remove the `Clone` problem, but I still want a crater run to avoid breaking the world.
2. The older implementation "looked into" impls to display error messages. I'm not sure that's an advantage - it looked exactly 1 level deep.
Ariel Ben-Yehuda [Tue, 29 Aug 2017 21:41:02 +0000 (21:41 +0000)]
Rollup merge of #44141 - nrc:rls-preview-beta, r=alexcrichton
Rename the rls component to rls-preview on beta/stable
Background is that we will have automatic renaming with the next rustup release. We'll then rename rls to rls-preview. In the meantime, this ensures beta/stable users will always have rls-preview.
Ariel Ben-Yehuda [Tue, 29 Aug 2017 21:40:59 +0000 (21:40 +0000)]
Rollup merge of #44126 - laumann:config-doc-comments, r=nikomatsakis
compiletest: Change Config comments to doc comments
I plan to make the same change in compiletest-rs, to have some documentation in [the docs](https://docs.rs/compiletest_rs/0.2.9/compiletest_rs/common/struct.Config.html).
Ariel Ben-Yehuda [Tue, 29 Aug 2017 21:40:56 +0000 (21:40 +0000)]
Rollup merge of #44076 - RalfJung:src, r=alexcrichton
include Cargo.{toml,lock} in rust-src tarball
The lock file is interesting because e.g. xargo could use it to build libstd against the same dependencies that were used for the main build. More generally speaking, just documenting in this form which exact dependencies should be used IMHO makes lots of sense.
I added the Cargo.toml mostly because having the lock without the toml feels odd. Of course, the toml contains references to paths that don't actually exist in the rust-src tarball. Not sure if that is considered a problem.
Ariel Ben-Yehuda [Tue, 29 Aug 2017 21:40:55 +0000 (21:40 +0000)]
Rollup merge of #43918 - mystor:rustdoc-pound, r=QuietMisdreavus
Don't highlight # which does not start an attribute in rustdoc
Currently when we highlight some macros for rustdoc (e.g. `quote!` from https://github.com/dtolnay/quote), we get really bad syntax highlighting, because we assume that every token between a `#` character and the next `]` in the source must be an attribute.
This patch improves that highlighting behavior to instead only highlight after finding the `[` token after the `#` token.
(NOTE: I've only run this patch against https://github.com/nrc/rustdoc-highlight so if it doesn't build on travis that's why - I don't have a recent rustc build on this laptop)
Ariel Ben-Yehuda [Tue, 29 Aug 2017 21:40:54 +0000 (21:40 +0000)]
Rollup merge of #43778 - topecongiro:handler-reset-err-count, r=arielb1
Add reset_err_count() to errors::Handler
The motivation here is to allow rustfmt to recover from parse errors after failing to parse macros (cc https://github.com/rust-lang-nursery/rustfmt/issues/1742).
r? @nrc
bors [Tue, 29 Aug 2017 17:32:13 +0000 (17:32 +0000)]
Auto merge of #43648 - RalfJung:jemalloc-debug, r=alexcrichton
Fix alloc_jemalloc debug feature
At least, I think that's how it should be. 'debug' is how the feature is called in liballoc_jemalloc/Cargo.toml and libstd/Cargo.toml. I verified this by making the build script panic rather than adding `--enable-debug`, and without this PR, the panic does not occur even when I set `debug-jemalloc = true` in config.toml. With the PR, the panic occurs as expected.
However, I actually have no idea what I am doing here.
Alex Crichton [Tue, 29 Aug 2017 16:25:25 +0000 (09:25 -0700)]
rustc: Remove `specailization_cache` in favor of a query
This commit removes the `specialization_cache` field of `TyCtxt` by moving it to
a dedicated query, which it turned out was already quite easily structured to do
so!
Ariel Ben-Yehuda [Thu, 17 Aug 2017 14:38:16 +0000 (17:38 +0300)]
Track closure signatures & kinds in freshened types
This allows caching closure signatures and kinds in the normal selection
and evaluation caches, and fixes the exponential worst-case in
@remram44's example, which is a part of #43787.
This improvement is complenentary to #43999 - they fix different cases.
bors [Tue, 29 Aug 2017 10:22:14 +0000 (10:22 +0000)]
Auto merge of #44111 - zackmdavis:feature_attr_error_span, r=nikomatsakis
feature error span on attribute for fn_must_use, SIMD/align reprs, macro reëxport
There were several feature-gated attributes for which the feature-not-available
error spans would point to the item annotated with the gated attribute, when it
would make more sense for the span to point to the attribute itself: if the
attribute is removed, the function/struct/_&c._ likely still makes sense and the
program will compile. (Note that we decline to make the analogous change for
the `main`, `start`, and `plugin_registrar` features, for in those cases it
makes sense for the span to implicate the entire function, of which there is
little hope of using without the gated attribute.)
bors [Tue, 29 Aug 2017 00:58:17 +0000 (00:58 +0000)]
Auto merge of #44049 - alexcrichton:nounwind-allocators, r=BurntSushi
std: Mark allocation functions as nounwind
This commit flags all allocation-related functions in liballoc as "this can't
unwind" which should largely resolve the size-related issues found on #42808.
The documentation on the trait was updated with such a restriction (they can't
panic) as well as some other words about the relative instability about
implementing a bullet-proof allocator.
bors [Mon, 28 Aug 2017 20:42:27 +0000 (20:42 +0000)]
Auto merge of #43999 - arielb1:immediate-project, r=nikomatsakis
clear out projection subobligations after they are processed
After a projection was processed, its derived subobligations no longer
need any processing when encountered, and can be removed. This improves
the status of #43787.
This is actually complementary to #43938 - that PR fixes selection
caching (and @remram44's example, which "accidentally" worked because of
the buggy projection caching) while this PR fixes projection caching.
bors [Mon, 28 Aug 2017 16:36:03 +0000 (16:36 +0000)]
Auto merge of #43076 - Zoxc:gen, r=arielb1
Generator support
This adds experimental support for generators intended to land once https://github.com/rust-lang/rfcs/pull/2033 is approved.
This is not yet ready to be merged. Things to do:
- [x] Make closure arguments on generators an error
- [x] Spot FIXMEs
- [x] Pass make tidy
- [x] Write tests
- [x] Document the current syntax and semantics for generators somewhere
- [x] Use proper error message numbers
- [x] ~~Make the implicit argument type default to `()`~~
Alex Crichton [Tue, 22 Aug 2017 21:36:49 +0000 (14:36 -0700)]
std: Mark allocation functions as nounwind
This commit flags all allocation-related functions in liballoc as "this can't
unwind" which should largely resolve the size-related issues found on #42808.
The documentation on the trait was updated with such a restriction (they can't
panic) as well as some other words about the relative instability about
implementing a bullet-proof allocator.
bors [Mon, 28 Aug 2017 10:42:11 +0000 (10:42 +0000)]
Auto merge of #44114 - daboross:patch-1, r=dtolnay
Clarify that VecDeque::swap can panic
The previous documentation mentioned this, but ambiguously used the term "fail".
This clarifies that the function will panic if the index is out of bounds, instead of silently failing and not doing anything.
If there's anything else I can do to improve this PR, I'd be happy to do so! Just saw this when reading through the docs in passing - it was slightly unclear what "fail" meant.
Sergio Benitez [Mon, 28 Aug 2017 09:56:43 +0000 (02:56 -0700)]
Initial diagnostic API for proc-macros.
This commit introduces the ability to create and emit `Diagnostic`
structures from proc-macros, allowing for proc-macro authors to emit
warning, error, note, and help messages just like the compiler does.
Zack M. Davis [Sun, 27 Aug 2017 01:00:33 +0000 (18:00 -0700)]
feature error span on attr. for fn_must_use, SIMD/align, macro reëxport
There were several feature-gated attributes for which the
feature-not-available error spans would point to the item annotated with
the gated attribute, when it would make more sense for the span to point
to the attribute itself: if the attribute is removed, the
function/struct/&c. likely still makes sense and the program will
compile. (Note that we decline to make the analogous change for the
`main`, `start`, and `plugin_registrar` features, for in those cases it
makes sense for the span to implicate the entire function, of which
there is little hope of using without the gated attribute.)
bors [Mon, 28 Aug 2017 06:04:10 +0000 (06:04 +0000)]
Auto merge of #44107 - alexcrichton:no-shell-configure, r=Mark-Simulacrum
rustbuild: Rewrite the configure script in Python
This commit rewrites our ancient `./configure` script from shell into Python.
The impetus for this change is to remove `config.mk` which is just a vestige of
the old makefile build system at this point. Instead all configuration is now
solely done through `config.toml`.
The python script allows us to more flexibly program (aka we can use loops
easily) and create a `config.toml` which is based off `config.toml.example`.
This way we can preserve comments and munge various values as we see fit.
It is intended that the configure script here is a drop-in replacement for the
previous configure script, no functional change is intended. Also note that the
rationale for this is also because our build system requires Python, so having a
python script a bit earlier shouldn't cause too many problems.
Alex Crichton [Sat, 26 Aug 2017 22:01:48 +0000 (15:01 -0700)]
rustbuild: Rewrite the configure script in Python
This commit rewrites our ancient `./configure` script from shell into Python.
The impetus for this change is to remove `config.mk` which is just a vestige of
the old makefile build system at this point. Instead all configuration is now
solely done through `config.toml`.
The python script allows us to more flexibly program (aka we can use loops
easily) and create a `config.toml` which is based off `config.toml.example`.
This way we can preserve comments and munge various values as we see fit.
It is intended that the configure script here is a drop-in replacement for the
previous configure script, no functional change is intended. Also note that the
rationale for this is also because our build system requires Python, so having a
python script a bit earlier shouldn't cause too many problems.
Ariel Ben-Yehuda [Sun, 20 Aug 2017 16:16:36 +0000 (19:16 +0300)]
clear out projection subobligations after they are processed
After a projection was processed, its derived subobligations no longer
need any processing when encountered, and can be removed. This improves
the status of #43787.
This is actually complementary to #43938 - that PR fixes selection
caching (and @remram44's example, which "accidentally" worked because of
the buggy projection caching) while this PR fixes projection caching
bors [Sun, 27 Aug 2017 12:53:48 +0000 (12:53 +0000)]
Auto merge of #44060 - taleks:issue-43205, r=arielb1
Fixes issue #43205: ICE in Rvalue::Len evaluation.
- fixes evaluation of array length for zero-sized type referenced by rvalue operand.
- adds test to verify fix.
*Cause of the issue*.
Zero-sized aggregates are handled as operands, not lvalues. Therefore while visiting `Assign` statement by `LocalAnalyser`, `mark_as_lvalue()` is not called for related `Local`. This behaviour is controlled by `rvalue_creates_operand()` method.
As result it causes error later, when rvalue operand is evaluated in `trans_rvalue_operand()` while handling `Rvalue::Len` case. Array length evaluation invokes `trans_lvalue()` which expects referenced `Local` to be value, not operand.
*How it is fixed*.
In certain cases result of `Rvalue::Len` can be evaluated without calling
`trans_lvalue()`. Method `evaluate_array_len()` is introduced to handle length
evaluation for zero-sized types referenced by Locals.
*Some concerns*.
- `trans_lvalue()` has two other entry points in `rvalue.rs`: it is invoked while handling `Rvalue::Ref` and `Rvalue::Discriminant`. There is a chance those may produce the same issue, but I've failed to write a specific test that leads to this.
- `evaluate_array_len()` performs the same check (matches lvalue and `Local`), which is performed again in `trans_lvalue()`. Without changing `trans_lvalue()` signature to make it aware that caller deals with rvalue, it seems there is no cheap solution to avoid this check.
bors [Sun, 27 Aug 2017 10:02:51 +0000 (10:02 +0000)]
Auto merge of #42588 - ishitatsuyuki:patch-1, r=petrochenkov
Make unused-extern-crate warn-by-default
Apart from enabling the lint, this pull request also removes existing unused crates in the codebase, and fix some amount of false positives on crates with special purposes.
Now that all false positive issues are closed, it should be possible to make it available to wider users.
Quote:
> Now that macro modularization is implemented, this is true today! *https://github.com/rust-lang/rust/issues/30849#issuecomment-286573218*
bors [Sun, 27 Aug 2017 01:41:45 +0000 (01:41 +0000)]
Auto merge of #44102 - Mark-Simulacrum:update-cargo, r=alexcrichton
Update cargo
Should permit https://github.com/rust-lang/rust/pull/41991 to move forward. I think it's best to land this as a separate patch and rebase that, though.
bors [Sat, 26 Aug 2017 23:11:44 +0000 (23:11 +0000)]
Auto merge of #44096 - Dushistov:master, r=japaric
Add test for wrong code generation for HashSet creation on arm cpu
This is test for #42918.
To reproduce bug you need machine with arm cpu and compile with optimization.
I tried with rustc 1.19.0-nightly (3d5b8c626 2017-06-09),
if compile test with -C opt-level=3 for target=arm-linux-androideabi
and run on "Qualcomm MSM 8974 arm cpu" then assert fails,
if compile and run with -C opt-level=2 it gives segmentation fault.
So I add `compile-flags: -O`.
With rustc 1.19.0 (0ade33941 2017-07-17) all works fine.
Closes #42918
bors [Sat, 26 Aug 2017 20:21:28 +0000 (20:21 +0000)]
Auto merge of #44084 - alexcrichton:msvc-ninja, r=Mark-Simulacrum
rustbuild: Automatically enable Ninja on MSVC
Discovered in #43767 it turns out the default MSBuild generator in CMake for
whatever reason isn't supporting many of the configuration options we give to
LLVM. To improve the contributor experience automatically enable Ninja if we
find it to ensure that "flavorful" configurations of LLVM work by default in
more situations.
bors [Sat, 26 Aug 2017 17:48:29 +0000 (17:48 +0000)]
Auto merge of #44082 - pnkfelix:issue-43457, r=eddyb
Fix destruction extent lookup during HIR -> HAIR translation
My method for finding the destruction extent, if any, from cbed41a174aad44e069bec09bf1e502591c132ae (in #39409), was buggy in that it sometimes failed to find an extent that was nonetheless present.