Yuki Okushi [Fri, 7 May 2021 06:20:26 +0000 (15:20 +0900)]
Rollup merge of #84897 - richkadel:cover-closure-macros, r=tmandry
Coverage instruments closure bodies in macros (not the macro body)
Fixes: #84884
This solution might be considered a compromise, but I think it is the
better choice.
The results in the `closure.rs` test correctly resolve all test cases
broken as described in #84884.
One test pattern (in both `closure_macro.rs` and
`closure_macro_async.rs`) was also affected, and removes coverage
statistics for the lines inside the closure, because the closure
includes a macro. (The coverage remains at the callsite of the macro, so
we lose some detail, but there isn't a perfect choice with macros.
Often macro implementations are split across the macro and the callsite,
and there doesn't appear to be a single "right choice" for which body
should be covered. For the current implementation, we can't do both.
The callsite is most likely to be the preferred site for coverage.
Yuki Okushi [Fri, 7 May 2021 06:20:25 +0000 (15:20 +0900)]
Rollup merge of #84875 - richkadel:no-coverage-dont-check-unused, r=tmandry
Removes unneeded check of `#[no_coverage]` in mapgen
There is an anticipated feature request to support a compiler flag that
only adds coverage for specific files (or perhaps mods). As I thought
about where that change would need to be supported, I realized that
checking the attribute in mapgen (for unused functions) was unnecessary.
The unused functions are only synthesized if they have MIR coverage, and
functions with the `no_coverage` attribute will not have been
instrumented with MIR coverage statements in the first place.
New tests confirm this.
Also, while adding tests, I updated resolved comments and FIXMEs in
other tests, and expanded comments and tests on one remaining issue that
is still not resolved.
bors [Fri, 7 May 2021 01:16:08 +0000 (01:16 +0000)]
Auto merge of #85014 - Dylan-DPC:rollup-jzpbkdu, r=Dylan-DPC
Rollup of 11 pull requests
Successful merges:
- #84409 (Ensure TLS destructors run before thread joins in SGX)
- #84500 (Add --run flag to compiletest)
- #84728 (Add test for suggestion to borrow unsized function parameters)
- #84734 (Add `needs-unwind` and beginning of support for testing `panic=abort` std to compiletest)
- #84755 (Allow using `core::` in intra-doc links within core itself)
- #84871 (Disallows `#![feature(no_coverage)]` on stable and beta (using standard crate-level gating))
- #84872 (Wire up tidy dependency checks for cg_clif)
- #84896 (Handle incorrect placement of parentheses in trait bounds more gracefully)
- #84905 (CTFE engine: rename copy → copy_intrinsic, move to intrinsics.rs)
- #84953 (Remove unneeded call to with_default_session_globals in rustdoc highlight)
- #84987 (small nits)
Dylan DPC [Thu, 6 May 2021 22:38:43 +0000 (00:38 +0200)]
Rollup merge of #84905 - RalfJung:copy, r=oli-obk
CTFE engine: rename copy → copy_intrinsic, move to intrinsics.rs
The `copy` name is confusing for this function because we also have `copy_op` which is pretty different. I hope `copy_intrinsic` is clearer. Also `step.rs` should really just contain the main loop and opcode dispatch, so move this helper function to a more appropriate place.
Dylan DPC [Thu, 6 May 2021 22:38:40 +0000 (00:38 +0200)]
Rollup merge of #84871 - richkadel:no-coverage-unstable-only, r=nagisa
Disallows `#![feature(no_coverage)]` on stable and beta (using standard crate-level gating)
Fixes: #84836
Removes the function-level feature gating solution originally implemented, and solves the same problem using `allow_internal_unstable`, so normal crate-level feature gating mechanism can still be used (which disallows the feature on stable and beta).
I tested this, building the compiler with and without `CFG_DISABLE_UNSTABLE_FEATURES=1`
With unstable features disabled, I get the expected result as shown here:
```shell
$ ./build/x86_64-unknown-linux-gnu/stage1/bin/rustc src/test/run-make-fulldeps/coverage/no_cov_crate.rs
error[E0554]: `#![feature]` may not be used on the dev release channel
--> src/test/run-make-fulldeps/coverage/no_cov_crate.rs:2:1
|
2 | #![feature(no_coverage)]
| ^^^^^^^^^^^^^^^^^^^^^^^^
error: aborting due to previous error
For more information about this error, try `rustc --explain E0554`.
```
Dylan DPC [Thu, 6 May 2021 22:38:38 +0000 (00:38 +0200)]
Rollup merge of #84755 - jyn514:core-links, r=kennytm
Allow using `core::` in intra-doc links within core itself
I came up with this idea ages ago, but rustdoc used to ICE on it. Now it doesn't.
Helps with https://github.com/rust-lang/rust/issues/73445. Doesn't fix it completely since `extern crate self as std;` in std still gives strange errors.
Dylan DPC [Thu, 6 May 2021 22:38:37 +0000 (00:38 +0200)]
Rollup merge of #84734 - tmandry:compiletest-needs-unwind, r=Mark-Simulacrum
Add `needs-unwind` and beginning of support for testing `panic=abort` std to compiletest
For the Fuchsia platform we build libstd with `panic=abort` and would like a way to run tests with that enabled. This adds low-level support for this directly to compiletest.
In the future I'd like to add high-level support in rustbuild, e.g. having target-specific flags that allow configuring a panic strategy. (Side note: It would be nice if we could also build multiple configurations for the same target, but I'm getting ahead of myself.)
This plus #84500 have everything that's needed to get ui tests passing on fuchsia targets.
Part of #84766. Note that this change only includes the header on tests which need an unwinder to _build_, not those which need it to _run_.
Dylan DPC [Thu, 6 May 2021 22:38:36 +0000 (00:38 +0200)]
Rollup merge of #84728 - camelid:sized-param-sugg-test, r=Mark-Simulacrum
Add test for suggestion to borrow unsized function parameters
Closes #82820.
This is a regression test for #82820.
This test case is included in more general tests, but I think the error
regressed because there were a bunch of other diagnostic changes in the
test that obscured this regression.
Hopefully, having a test specific to the suggestion, and running rustfix
for the test, will prevent this error from regressing in the future.
Rich Kadel [Mon, 3 May 2021 18:23:40 +0000 (11:23 -0700)]
Removes unneeded check of `#[no_coverage]` in mapgen
And adds tests to validate it still works.
There is an anticipated feature request to support a compiler flag that
only adds coverage for specific files (or perhaps mods). As I thought
about where that change would need to be supported, I realized that
checking the attribute in mapgen (for unused functions) was unnecessary.
The unused functions are only synthesized if they have MIR coverage, and
functions with the `no_coverage` attribute will not have been
instrumented with MIR coverage statements in the first place.
New tests confirm this.
Also, while adding tests, I updated resolved comments and FIXMEs in
other tests.
Add test for suggestion to borrow unsized function parameters
This is a regression test for #82820.
This test case is included in more general tests, but I think the error
regressed because there were a bunch of other diagnostic changes in the
test that obscured this regression.
Hopefully, having a test specific to the suggestion, and running rustfix
for the test, will prevent this error from regressing in the future.
Rich Kadel [Tue, 4 May 2021 06:21:24 +0000 (23:21 -0700)]
Coverage instruments closure bodies in macros (not the macro body)
Fixes: #84884
This solution might be considered a compromise, but I think it is the
better choice.
The results in the `closure.rs` test correctly resolve all test cases
broken as described in #84884.
One test pattern (in both `closure_macro.rs` and
`closure_macro_async.rs`) was also affected, and removes coverage
statistics for the lines inside the closure, because the closure
includes a macro. (The coverage remains at the callsite of the macro, so
we lose some detail, but there isn't a perfect choice with macros.
Often macro implementations are split across the macro and the callsite,
and there doesn't appear to be a single "right choice" for which body
should be covered. For the current implementation, we can't do both.
The callsite is most likely to be the preferred site for coverage.
I applied this fix to all `MacroKinds`, not just `Bang`.
I'm trying to resolve an issue of lost coverage in a
`MacroKind::Attr`-based, function-scoped macro. Instead of only
searching for a body_span that is "not a function-like macro" (that is,
MacroKind::Bang), I'm expanding this to all `MacroKind`s. Maybe I should
expand this to `ExpnKind::Desugaring` and `ExpnKind::AstPass` (or
subsets, depending on their sub-kinds) as well, but I'm not sure that's
a good idea.
I'd like to add a test of the `Attr` macro on functions, but I need time
to figure out how to constract a good, simple example without external
crate dependencies. For the moment, all tests still work as expected (no
change), this new commit shouldn't have a negative affect, and more
importantly, I believe it will have a positive effect. I will try to
confirm this.
bors [Thu, 6 May 2021 17:42:32 +0000 (17:42 +0000)]
Auto merge of #84559 - jackh726:issue-84398, r=nikomatsakis
Deduplicate ParamCandidates with the same value except for bound vars
Fixes #84398
This is kind of a hack. I wonder if we can get other types of candidates that are the same except for bound vars. This won't be a problem with Chalk, since we don't really need to know that there are two different "candidates" if they both give the same final substitution.
Dylan DPC [Thu, 6 May 2021 11:31:01 +0000 (13:31 +0200)]
Rollup merge of #84950 - pietroalbini:revert-3478f83c0, r=Mark-Simulacrum
Revert PR 83866
https://github.com/rust-lang/rust/pull/83866 caused multiple test failures when Rust's channel is changed from nightly to anything else. The PR will have to be landed again after the test suite is fixed.
The two kinds of test failures were:
* Rustdoc tests failed because the links in the generated HTML didn't point at nightly anymore, see https://github.com/rust-lang/rust/pull/84909#issuecomment-832028701
* Rustdoc UI tests failed because the links included in error messages didn't point at nightly anymore, see https://github.com/rust-lang/rust/pull/84909#issuecomment-832517722
Introduces feature flags for the general syntax (`native_link_modifiers`) and each modifier (`native_link_modifiers_{as_needed,bundle,verbatim,whole_archive}`).
For certain sorts of systems, programming, it's deemed essential that
all allocation failures be explicitly handled where they occur. For
example, see Linus Torvald's opinion in [1]. Merely not calling global
panic handlers, or always `try_reserving` first (for vectors), is not
deemed good enough, because the mere presence of the global OOM handlers
is burdens static analysis.
One option for these projects to use rust would just be to skip `alloc`,
rolling their own allocation abstractions. But this would, in my
opinion be a real shame. `alloc` has a few `try_*` methods already, and
we could easily have more. Features like custom allocator support also
demonstrate and existing to support diverse use-cases with the same
abstractions.
A natural way to add such a feature flag would a Cargo feature, but
there are currently uncertainties around how std library crate's Cargo
features may or not be stable, so to avoid any risk of stabilizing by
mistake we are going with a more low-level "raw cfg" token, which
cannot be interacted with via Cargo alone.
Note also that since there is no notion of "default cfg tokens" outside
of Cargo features, we have to invert the condition from
`global_oom_handling` to to `not(no_global_oom_handling)`. This breaks
the monotonicity that would be important for a Cargo feature (i.e.
turning on more features should never break compatibility), but it
doesn't matter for raw cfg tokens which are not intended to be
"constraint solved" by Cargo or anything else.
To support this use-case we create a new feature, "global-oom-handling",
on by default, and put the global OOM handler infra and everything else
it that depends on it behind it. By default, nothing is changed, but
users concerned about global handling can make sure it is disabled, and
be confident that all OOM handling is local and explicit.
For this first iteration, non-flat collections are outright disabled.
`Vec` and `String` don't yet have `try_*` allocation methods, but are
kept anyways since they can be oom-safely created "from parts", and we
hope to add those `try_` methods in the future.
Luqman Aden [Thu, 25 Mar 2021 04:45:09 +0000 (21:45 -0700)]
Implement RFC 2951: Native link modifiers
This commit implements both the native linking modifiers infrastructure
as well as an initial attempt at the individual modifiers from the RFC.
It also introduces a feature flag for the general syntax along with
individual feature flags for each modifier.
The8472 [Wed, 5 May 2021 20:52:58 +0000 (22:52 +0200)]
lazify backtrace formatting for delayed diagnostics
This defers backtrace formatting to the point where we
actually want to flush delayed diagnostics. If they are discarded
before that point then we can avoid invoking the backtrace formatting
machinery which will parse debug info and symbol tables.
for debuginfo=2 this leads to a 20% walltime reduction of the UI testsuite
For certain sorts of systems, programming, it's deemed essential that
all allocation failures be explicitly handled where they occur. For
example, see Linus Torvald's opinion in [1]. Merely not calling global
panic handlers, or always `try_reserving` first (for vectors), is not
deemed good enough, because the mere presence of the global OOM handlers
is burdens static analysis.
One option for these projects to use rust would just be to skip `alloc`,
rolling their own allocation abstractions. But this would, in my
opinion be a real shame. `alloc` has a few `try_*` methods already, and
we could easily have more. Features like custom allocator support also
demonstrate and existing to support diverse use-cases with the same
abstractions.
A natural way to add such a feature flag would a Cargo feature, but
there are currently uncertainties around how std library crate's Cargo
features may or not be stable, so to avoid any risk of stabilizing by
mistake we are going with a more low-level "raw cfg" token, which
cannot be interacted with via Cargo alone.
Note also that since there is no notion of "default cfg tokens" outside
of Cargo features, we have to invert the condition from
`global_oom_handling` to to `not(no_global_oom_handling)`. This breaks
the monotonicity that would be important for a Cargo feature (i.e.
turning on more features should never break compatibility), but it
doesn't matter for raw cfg tokens which are not intended to be
"constraint solved" by Cargo or anything else.
To support this use-case we create a new feature, "global-oom-handling",
on by default, and put the global OOM handler infra and everything else
it that depends on it behind it. By default, nothing is changed, but
users concerned about global handling can make sure it is disabled, and
be confident that all OOM handling is local and explicit.
For this first iteration, non-flat collections are outright disabled.
`Vec` and `String` don't yet have `try_*` allocation methods, but are
kept anyways since they can be oom-safely created "from parts", and we
hope to add those `try_` methods in the future.
bors [Wed, 5 May 2021 17:45:41 +0000 (17:45 +0000)]
Auto merge of #84956 - RalfJung:rollup-m70mx2n, r=RalfJung
Rollup of 11 pull requests
Successful merges:
- #83553 (Update `ptr` docs with regards to `ptr::addr_of!`)
- #84183 (Update RELEASES.md for 1.52.0)
- #84709 (Add doc alias for `chdir` to `std::env::set_current_dir`)
- #84803 (Reduce duplication in `impl_dep_tracking_hash` macros)
- #84808 (Account for unsatisfied bounds in E0599)
- #84843 (use else if in std library )
- #84865 (rustbuild: Pass a `threads` flag that works to windows-gnu lld)
- #84878 (Clarify documentation for `[T]::contains`)
- #84882 (platform-support: Center the contents of the `std` and `host` columns)
- #84903 (Remove `rustc_middle::mir::interpret::CheckInAllocMsg::NullPointerTest`)
- #84913 (Do not ICE on invalid const param)
Ralf Jung [Wed, 5 May 2021 15:52:29 +0000 (17:52 +0200)]
Rollup merge of #84913 - estebank:issue-84831, r=varkor
Do not ICE on invalid const param
When encountering a path that can't have generics, do not call
`generics_of`. This would happen when writing something like
`path::this_is_a_mod<const_val>`.
Removing it per https://github.com/rust-lang/rust/pull/84842#discussion_r625589674: it's a dead enum variant.
Note that `PointerArithmeticTest` also seems dead:
```
$ rg -F PointerArithmeticTest -C5
compiler/rustc_middle/src/mir/interpret/error.rs
169-
170-/// Details of why a pointer had to be in-bounds.
171-#[derive(Debug, Copy, Clone, TyEncodable, TyDecodable, HashStable)]
172-pub enum CheckInAllocMsg {
173- MemoryAccessTest,
174: PointerArithmeticTest,
175- InboundsTest,
176-}
177-
178-impl fmt::Display for CheckInAllocMsg {
179- /// When this is printed as an error the context looks like this
--
182- write!(
183- f,
184- "{}",
185- match *self {
186- CheckInAllocMsg::MemoryAccessTest => "memory access",
187: CheckInAllocMsg::PointerArithmeticTest => "pointer arithmetic",
188- CheckInAllocMsg::InboundsTest => "inbounds test",
189- }
190- )
191- }
192-}
```
Not sure if that is also desirable to be removed, however.
Ralf Jung [Wed, 5 May 2021 15:52:25 +0000 (17:52 +0200)]
Rollup merge of #84865 - petrochenkov:llthread, r=Mark-Simulacrum
rustbuild: Pass a `threads` flag that works to windows-gnu lld
MinGW driver for COFF LLD doesn't currently translate GNU-style `--threads=N` to native `/threads:N`, so we have to pass the option in its native form to avoid an error.
Also pass the `threads` flag to lld-link (windows-msvc lld) as well.
Ralf Jung [Wed, 5 May 2021 15:52:20 +0000 (17:52 +0200)]
Rollup merge of #84709 - joshtriplett:doc-alias-chdir, r=dtolnay
Add doc alias for `chdir` to `std::env::set_current_dir`
Searching for `chdir` in the Rust documentation produces no useful
results.
I wrote some code recently that called `libc::chdir` and manually
handled errors, because I didn't realize that the safe
`std::env::set_current_dir` existed. I searched for `chdir` and
`change_dir` and `change_directory` (the latter two based on the
precedent of unabbreviating set by `create_dir`), and I also read
through `std::fs` expecting to potentially find it there. Given that
none of those led to `std::env::set_current_dir`, I think that provides
sufficient justification to add this specific alias.
Ralf Jung [Wed, 5 May 2021 15:52:18 +0000 (17:52 +0200)]
Rollup merge of #83553 - jfrimmel:addr-of, r=m-ou-se
Update `ptr` docs with regards to `ptr::addr_of!`
This updates the documentation since `ptr::addr_of!` and `ptr::addr_of_mut!` are now stable. One might remove the distinction between the sections `# On packed structs` and `# Examples`, as the old section on packed structs was primarily to prevent users of doing undefined behavior, which is not necessary anymore.
Technically there is now wrong/outdated documentation on stable, but I don't think this is worth a point release 😉
Rich Kadel [Mon, 3 May 2021 15:47:43 +0000 (08:47 -0700)]
Disallows `#![feature(no_coverage)]` on stable and beta
using allow_internal_unstable (as recommended)
Fixes: #84836
```shell
$ ./build/x86_64-unknown-linux-gnu/stage1/bin/rustc src/test/run-make-fulldeps/coverage/no_cov_crate.rs
error[E0554]: `#![feature]` may not be used on the dev release channel
--> src/test/run-make-fulldeps/coverage/no_cov_crate.rs:2:1
|
2 | #![feature(no_coverage)]
| ^^^^^^^^^^^^^^^^^^^^^^^^
error: aborting due to previous error
For more information about this error, try `rustc --explain E0554`.
```
bors [Wed, 5 May 2021 14:15:19 +0000 (14:15 +0000)]
Auto merge of #84200 - CDirkx:os, r=m-ou-se
Move all `sys::ext` modules to `os`
This PR moves all `sys::ext` modules to `os`, centralizing the location of all `os` code and simplifying the dependencies between `os` and `sys`.
Because this also removes all uses `cfg_if!` on publicly exported items, where after #81969 there were still a few left, this should properly work around https://github.com/rust-analyzer/rust-analyzer/issues/6038.
Pietro Albini [Wed, 5 May 2021 09:57:10 +0000 (11:57 +0200)]
Revert PR 83866
That PR caused multiple test failures when Rust's channel is changed
from nightly to anything else. The commit will have to be landed again
after the test suite is fixed.
bors [Wed, 5 May 2021 11:05:27 +0000 (11:05 +0000)]
Auto merge of #84886 - Xanewok:update-rls-and-rustfmt, r=Xanewok
Update RLS and Rustfmt
Closes #84537. Closes #84538.
I know there's https://github.com/rust-lang/rust/pull/82208 in progress but I'm not sure which we want to land first.
Also, I'm getting Rustfmt test failures due to inner attributes not permitted:
<details>
```
error: an inner attribute is not permitted in this context
--> tests/target/issue-3592.rs:4:13
|
4 | #![cfg(unix)]
| ^^^^^^^^^^^^^
|
= note: inner attributes, like `#![no_std]`, annotate the item enclosing them, and are usually found at the beginning of source files. Outer attributes, like `#[test]`, annotate the item following them.
error: an inner attribute is not permitted in this context
--> tests/target/issue-3592.rs:8:13
|
8 | #![cfg(not(unix))]
| ^^^^^^^^^^^^^^^^^^
|
= note: inner attributes, like `#![no_std]`, annotate the item enclosing them, and are usually found at the beginning of source files. Outer attributes, like `#[test]`, annotate the item following them.
error: an inner attribute is not permitted in this context
--> tests/source/match.rs:413:9
|
413 | #![allow(simple_match)]
| ^^^^^^^^^^^^^^^^^^^^^^^
|
= note: inner attributes, like `#![no_std]`, annotate the item enclosing them, and are usually found at the beginning of source files. Outer attributes, like `#[test]`, annotate the item following them.
error: an inner attribute is not permitted in this context
--> tests/target/match.rs:444:9
|
444 | #![allow(simple_match)]
| ^^^^^^^^^^^^^^^^^^^^^^^
|
= note: inner attributes, like `#![no_std]`, annotate the item enclosing them, and are usually found at the beginning of source files. Outer attributes, like `#[test]`, annotate the item following them.
test test::system_tests ... FAILED
test test::idempotence_tests ... FAILED
```
bors [Wed, 5 May 2021 03:56:26 +0000 (03:56 +0000)]
Auto merge of #84794 - ChrisDenton:dedup-native-libs, r=petrochenkov
Deduplicate native libs before they are passed to the linker
Stop spamming the linker with the same native library over and over again, if they directly follow from each other. This would help prevent [this situation](https://github.com/MSxDOS/ntapi/issues/2).
Issue #38460 has been open since 2016 so I think it's worth making an incomplete fix that at least addresses the most common symptom and without otherwise changing how Rust handles native libs. This PR is intended to be easy to revert (if necessary) when a more permanent fix is implemented.
Esteban Küber [Tue, 4 May 2021 15:41:40 +0000 (08:41 -0700)]
Do not ICE on invalid const param
When encountering a path that can't have generics, do not call
`generics_of`. This would happen when writing something like
`path::this_is_a_mod<const_val>`.
bors [Tue, 4 May 2021 10:42:13 +0000 (10:42 +0000)]
Auto merge of #84017 - Smittyvb:int-literal-underscores, r=jyn514
Valid underscores in hex/octal/binary literal docs
Currently hex/octal/binary literals with computed values are displayed like `0_xff_fff_fffu32`, which is invalid since underscores can't be in the middle of integer prefixes. This properly formats prefixed integers.
This causes [`std::u32::MAX`](https://doc.rust-lang.org/std/u32/constant.MAX.html) to be displayed as
```rust
pub const MAX: u32 = u32::MAX; // 0_xff_fff_fffu32
```
This PR changes it to be displayed as:
```rust
pub const MAX: u32 = u32::MAX; // 0xffff_ffffu32
```
bors [Tue, 4 May 2021 05:40:24 +0000 (05:40 +0000)]
Auto merge of #84874 - joshtriplett:ci-extract-llvm-win64-installer, r=Mark-Simulacrum
CI: Extract LLVM win64 installer directly, using 7z
Currently, we have LLVM tarballs for win64, generated by someone running
the installer via wine and tarring up the result.
7z knows how to extract NSIS installers directly, and the result is
identical to our tarball, except that it doesn't include `Uninstall.exe`
(which we don't care about) and it includes the NSIS plugin directory
(which we also don't care about).
This simplifies the process of upgrading CI, and allows us to just
mirror the upstream release .exe directly. This also improves our
supply chain.