Dylan DPC [Wed, 14 Oct 2020 00:30:34 +0000 (02:30 +0200)]
Rollup merge of #77765 - amshafer:master, r=petrochenkov
Add LLVM flags to limit DWARF version to 2 on BSD
This has been a thorn in my side for a while, I can finally generate flamegraphs of rust programs on bsd again. This fixes dtrace profiling on freebsd, I think it might help with lldb as well but I can't test that because my current rust-lldb setup is messed up.
I'm limiting the dwarf version to 2 on all bsd's (netbsd/openbsd/freebsd) since it looks like this applies to all of them, but I have only tested on freebsd.
Let me know if there's anything I can improve!
---
Currently on FreeBSD dtrace profiling does not work and shows jumbled/incorrect
symbols in the backtraces. FreeBSD does not support the latest versions of DWARF
in dtrace (and lldb?) yet, and needs to be limited to DWARF2 in the same way as macos.
This adds an is_like_bsd flag since it was missing. NetBSD/OpenBSD/FreeBSD all
match this.
This effectively copies #11864 but targets FreeBSD instead of macos.
Yuki Okushi [Tue, 13 Oct 2020 21:02:35 +0000 (06:02 +0900)]
Rollup merge of #77886 - LingMan:ast_pretty_bool_matches, r=petrochenkov
Replace trivial bool matches with the `matches!` macro
This derives `PartialEq` on one enum (and two structs it contains) to enable the `==` operator for it. If there's some downside to this, I could respin with the `matches!` macro instead.
Yuki Okushi [Tue, 13 Oct 2020 21:02:29 +0000 (06:02 +0900)]
Rollup merge of #77825 - ethanboxx:min_const_generics_diagnostic, r=lcnr
`min_const_generics` diagnostics improvements
As disscussed in [zulip/project-const-generics/non-trivial anonymous constant](https://rust-lang.zulipchat.com/#narrow/stream/260443-project-const-generics/topic/non-trivial.20anonymous.20constants).
Yuki Okushi [Tue, 13 Oct 2020 21:02:21 +0000 (06:02 +0900)]
Rollup merge of #77722 - fusion-engineering-forks:safe-unsupported-locks, r=Mark-Simulacrum
Remove unsafety from sys/unsupported and add deny(unsafe_op_in_unsafe_fn).
Replacing `UnsafeCell`s by a `Cell`s simplifies things and makes the mutex and rwlock implementations safe. Other than that, only unsafety in strlen() contained unsafe code.
Yuki Okushi [Tue, 13 Oct 2020 21:02:19 +0000 (06:02 +0900)]
Rollup merge of #77719 - fusion-engineering-forks:const-new-mutex-attr-cleanup, r=Mark-Simulacrum
Remove unnecessary rustc_const_stable attributes.
These attributes were added in https://github.com/rust-lang/rust/pull/74033#discussion_r450593156 because of [std::io::lazy::Lazy::new](https://github.com/rust-lang/rust/blob/0c03aee8b81185d65b5821518661c30ecdb42de5/src/libstd/io/lazy.rs#L21-L23). But [std::io::lazy::Lazy is gone now](https://github.com/rust-lang/rust/pull/77154), so this can be cleaned up.
bors [Tue, 13 Oct 2020 20:20:27 +0000 (20:20 +0000)]
Auto merge of #75213 - dingxiangfei2009:yield-point-in-match-guard, r=tmandry
[generator] Special cases for match guard when analyzing interior types in generators
Fix #72651
This proposes one of ways to fix the mentioned issue. One cause of #72651 is that the interior type analysis misses out types of match pattern locals. Those locals are manifested as temporary borrows in the scopes of match arm guards. If uses of these locals appear after yield points, the borrows from them were not considered live across the yield points. However, this is not the case since the borrowing always happens at the very beginning of the match guard.
This calls for special treatment to analysis of types appearing in the match guard. Those borrows are recorded as the HIR tree is walked by `InteriorVisitor` and their uses are recorded whenever a yield point is crossed.
Austin Shafer [Fri, 9 Oct 2020 19:08:18 +0000 (15:08 -0400)]
Add a target option for selecting a DWARF version
Certain platforms need to limit the DWARF version emitted (oxs, *bsd). This
change adds a dwarf_version entry to the options that allows a platform to
specify the dwarf version to use. By default this option is none and the default
DWARF version is selected.
Also adds an option for printing Option<u32> json keys
bors [Tue, 13 Oct 2020 12:11:47 +0000 (12:11 +0000)]
Auto merge of #77762 - pietroalbini:dist-build-manifest, r=Mark-Simulacrum
Dist build manifest
This PR makes two changes that should remove a significant chunk of the time spent in our release process: cloning the `rust-lang/rust` monorepo, all its submodules, and building `bootstrap` to then invoke `build-manifest`:
* `build-manifest` doesn't rely on a clone of the monorepo being present anymore. The only remaining bit of information it fetched from it (the Rust version) is instead bundled in the binary.
* A new "component" is added, `build-manifest`. That component includes a prebuilt version of the tool, and it's *not* included in the Rustup manifest. This will allow `promote-release` to directly invoke the tool without interacting with our build system.
* The Linux x86_64 CI is changed to also build the component mentioned above. It's the only CI builder tasked to do so, and to cleanly support this a new `--include-default-paths` flag was added to `./x.py`.
* The `BUILD_MANIFEST_NUM_THREADS` environment variable is added to configure the number of threads at runtime.
This PR is best reviewed commit-by-commit.
r? `@Mark-Simulacrum`
bors [Tue, 13 Oct 2020 08:27:01 +0000 (08:27 +0000)]
Auto merge of #77759 - tblah:fix_riscv_qemu, r=pietroalbini
ci: Fix riscv64gc linux test QEMU fault, plus doc link fix
Newer versions of the `qemu` package (used for riscv64gc-unknown-linux-gnu testing) don't work with the version of the RISC-V bootloader we were using. https://github.com/rust-lang/rust/commit/a4a0342cf59a1bff43ed79586065eb97dba0cddb bumps to a revision which should fix the problem.
bors [Tue, 13 Oct 2020 04:41:08 +0000 (04:41 +0000)]
Auto merge of #76196 - r-52:r-coverage-allow-missing-docs, r=jyn514
rustdoc: skip #[allow(missing docs)] for docs in coverage report
During the document coverage reporting with:
```bash
rustdoc something.rs -Z unstable-options --show-coverage
```
the coverage report counts code that is marked with `#[allow(missing_docs)]` for the calculation, which outputs lower numbers in the coverage report even though these parts should be ignored for the calculation.
Right now I'm not sure how this can be tested (CI)? (I verified it by hand and ran the unit tests)
bors [Mon, 12 Oct 2020 19:09:08 +0000 (19:09 +0000)]
Auto merge of #77867 - JohnTitor:rollup-0odg1n4, r=JohnTitor
Rollup of 10 pull requests
Successful merges:
- #77550 (add shims for WithOptConstParam query calls)
- #77699 (Add word wrap for short descriptions)
- #77724 (Implement `AsRawFd` for `StdinLock` etc. on WASI.)
- #77746 (Fix `x.py setup` sets `changelog-seen`)
- #77784 (Fix intra-docs link in core::ffi::VaList)
- #77811 (rustdoc: Make some functions private that don't need to be public)
- #77818 (Mono collector: replace pair of ints with Range)
- #77831 (Use std methods on char instead of open coding them)
- #77852 (update url in bootstrap README (gcc-rs -> cc-rs))
- #77863 (Remove `mark-i-m` from rustc-dev-guide maintainers)
Yuki Okushi [Mon, 12 Oct 2020 19:08:02 +0000 (04:08 +0900)]
Rollup merge of #77818 - bugadani:range, r=oli-obk
Mono collector: replace pair of ints with Range
I found the initial PR (#33171) that introduced this piece of code but I didn't find any information about why a tuple was preferred over a `Range<usize>`.
I'm hoping there are no technical reasons to not do this.
Yuki Okushi [Mon, 12 Oct 2020 19:07:58 +0000 (04:07 +0900)]
Rollup merge of #77784 - aDotInTheVoid:ffi-sealed_trait-intra-docs, r=jyn514
Fix intra-docs link in core::ffi::VaList
At some point, `VaList` was changes to be a [wrapper](https://github.com/rust-lang/rust/blob/1661f77/library/core/src/ffi.rs#L177-L201) over `VaListImpl`, and now the `Arg` method exists on [`VaListImpl`](https://github.com/rust-lang/rust/blob/1661f77/library/core/src/ffi.rs#L333-L336). This PR fixes the intradoc links so that when `--document-private-items` is ran on std (via [patch](https://gist.github.com/aDotInTheVoid/42c82306210203f9c9093c952b765ab4)), it works
Pietro Albini [Fri, 9 Oct 2020 13:34:57 +0000 (15:34 +0200)]
build-manifest: accept the Rust version instead of the monorepo path
This commit changes the way build-manifest is invoked, to let it accept
the Rust version directly instead of requiring the path of the Rust
monorepo and letting build-manifest figure out the path on its own.
This allows to run build-manifest without a clone of the monorepo.
bors [Mon, 12 Oct 2020 12:12:54 +0000 (12:12 +0000)]
Auto merge of #77821 - tmiasko:discriminant-value-is-safe, r=jonas-schievink
Remove unnecessary unsafe block around calls to discriminant_value
Since 63793 the discriminant_value intrinsic is safe to call. Remove
unnecessary unsafe block around calls to this intrinsic in built-in
derive macros.
Roman [Tue, 1 Sep 2020 12:19:24 +0000 (14:19 +0200)]
rustdoc: skip allow missing doc in cover. report
During the document coverage reporting with
```bash
rustdoc something.rs -Z unstable-options --show-coverage
```
the coverage report also includes parts of the code that are marked
with `#[allow(missing_docs)]`, which outputs lower numbers in the
coverage report even though these parts should be ignored for the
calculation.
Co-authored-by: Joshua Nelson <joshua@yottadb.com>
James Gill [Wed, 7 Oct 2020 01:54:05 +0000 (21:54 -0400)]
Stabilize slice_select_nth_unstable
This stabilizes the functionality in slice_partition_at_index,
but under the names `select_nth_unstable*`. The functions
`partition_at_index*` are left as deprecated, to be removed in
a later release.
bors [Sun, 11 Oct 2020 20:50:02 +0000 (20:50 +0000)]
Auto merge of #76859 - Aaron1011:fix/llvm-cgu-reuse, r=davidtwco,nikic
Use llvm::computeLTOCacheKey to determine post-ThinLTO CGU reuse
During incremental ThinLTO compilation, we attempt to re-use the
optimized (post-ThinLTO) bitcode file for a module if it is 'safe' to do
so.
Up until now, 'safe' has meant that the set of modules that our current
modules imports from/exports to is unchanged from the previous
compilation session. See PR #67020 and PR #71131 for more details.
However, this turns out be insufficient to guarantee that it's safe
to reuse the post-LTO module (i.e. that optimizing the pre-LTO module
would produce the same result). When LLVM optimizes a module during
ThinLTO, it may look at other information from the 'module index', such
as whether a (non-imported!) global variable is used. If this
information changes between compilation runs, we may end up re-using an
optimized module that (for example) had dead-code elimination run on a
function that is now used by another module.
Fortunately, LLVM implements its own ThinLTO module cache, which is used
when ThinLTO is performed by a linker plugin (e.g. when clang is used to
compile a C proect). Using this cache directly would require extensive
refactoring of our code - but fortunately for us, LLVM provides a
function that does exactly what we need.
The function `llvm::computeLTOCacheKey` is used to compute a SHA-1 hash
from all data that might influence the result of ThinLTO on a module.
In addition to the module imports/exports that we manually track, it
also hashes information about global variables (e.g. their liveness)
which might be used during optimization. By using this function, we
shouldn't have to worry about new LLVM passes breaking our module re-use
behavior.
In LLVM, the output of this function forms part of the filename used to
store the post-ThinLTO module. To keep our current filename structure
intact, this PR just writes out the mapping 'CGU name -> Hash' to a
file. To determine if a post-LTO module should be reused, we compare
hashes from the previous session.
This should unblock PR #75199 - by sheer chance, it seems to have hit
this issue due to the particular CGU partitioning and optimization
decisions that end up getting made.