bors [Sat, 22 Jun 2019 15:12:15 +0000 (15:12 +0000)]
Auto merge of #61020 - HeroicKatora:master, r=matthewjasper
librustc_data_structures: Speedup union of sparse and dense hybrid set
This optimization speeds up the union of a hybrid bitset when that
switches it from a sparse representation to a dense bitset. It now
clones the dense bitset and integrate only the spare elements instead of
densifying the sparse bitset, initializing all elements, and then a
union on two dense bitset, touching all words a second time.
It's not completely certain if the added complexity is worth it but I would
like to hear some feedback in any case. Benchmark results from my machine:
This being the second iteration of trying to improve the speed, since I missed the return value in the first, and forgot to run the relevant tests. Oops.
bors [Sat, 22 Jun 2019 05:12:11 +0000 (05:12 +0000)]
Auto merge of #62010 - ecstatic-morse:kill-borrows-of-proj, r=pnkfelix
Kill conflicting borrows of places with projections.
Resolves #62007.
Due to a bug, the previous version of this check did not actually kill all conflicting borrows unless the borrowed place had no projections. Specifically, `sets.on_entry` will always be empty when `statement_effect` is called. It does not contain the set of borrows which are live at this point in the program.
@pnkfelix describes why this was not caught before in #62007, and created an example where the current borrow checker failed unnecessarily. This PR adds their example as a test, but they will likely want to add some additional ones.
Rollup merge of #60971 - rbtcollins:docs-perf, r=rbtcollins,GuillaumeGomez
Add DocFS layer to rustdoc
* Move fs::create_dir_all calls into DocFS to provide a clean
extension point if async extension there is needed.
* Convert callsites of create_dir_all to ensure_dir to reduce syscalls.
* Convert fs::write usage to DocFS.write
(which also removes a lot of try_err! usage for easier reading)
* Convert File::create calls to use Vec buffers and then DocFS.write
in order to both consistently reduce syscalls as well as make
deferring to threads cleaner.
* Convert OpenOptions usage similarly - I could find no discussion on
the use of create_new for that one output file vs all the other
files render creates, if link redirection attacks are a concern
DocFS will provide a good central point to introduce systematic
create_new usage.
* DocFS::write defers to rayon for IO on Windows producing a modest
speedup: before this patch on my development workstation:
$ time cargo +mystg1 doc -p winapi:0.3.7
Documenting winapi v0.3.7
Finished dev [unoptimized + debuginfo] target(s) in 6m 11s
real 6m11.734s
user 0m0.015s
sys 0m0.000s
Afterwards:
$ time cargo +mystg1 doc -p winapi:0.3.7
Compiling winapi v0.3.7
Documenting winapi v0.3.7
Finished dev [unoptimized + debuginfo] target(s) in 49.53s
real 0m49.643s
user 0m0.000s
sys 0m0.015s
I haven't measured how much time is in the compilation logic vs in the
IO and outputting etc, but this takes it from frustating to tolerable
for me, at least for now.
Robert Collins [Mon, 20 May 2019 02:04:04 +0000 (14:04 +1200)]
Add DocFS layer to rustdoc
* Move fs::create_dir_all calls into DocFS to provide a clean
extension point if async extension there is needed.
* Convert callsites of create_dir_all to ensure_dir to reduce syscalls.
* Convert fs::write usage to DocFS.write
(which also removes a lot of try_err! usage for easier reading)
* Convert File::create calls to use Vec buffers and then DocFS.write
in order to consistently reduce syscalls as well, make
deferring to threads cleaner and avoid leaving dangling content if
writing to existing files....
* Convert OpenOptions usage similarly - I could find no discussion on
the use of create_new for that one output file vs all the other
files render creates, if link redirection attacks are a concern
DocFS will provide a good central point to introduce systematic
create_new usage. (fs::write/File::create is vulnerable to link
redirection attacks).
* DocFS::write defers to rayon for IO on Windows producing a modest
speedup: before this patch on my development workstation:
$ time cargo +mystg1 doc -p winapi:0.3.7
Documenting winapi v0.3.7
Finished dev [unoptimized + debuginfo] target(s) in 6m 11s
real 6m11.734s
Afterwards:
$ time cargo +mystg1 doc -p winapi:0.3.7
Compiling winapi v0.3.7
Documenting winapi v0.3.7
Finished dev [unoptimized + debuginfo] target(s) in 49.53s
real 0m49.643s
I haven't measured how much time is in the compilation logic vs in the
IO and outputting etc, but this takes it from frustating to tolerable
for me, at least for now.
bors [Fri, 21 Jun 2019 03:24:56 +0000 (03:24 +0000)]
Auto merge of #61948 - ehuss:update-mdbook, r=steveklabnik
Update mdbook
This updates to mdbook 0.3.0. This release includes the following changes:
https://github.com/rust-lang-nursery/mdBook/blob/master/CHANGELOG.md#mdbook-030
One of the biggest changes is the update to pulldown-cmark.
This also bumps the submodules (except the nomicon which has some broken links).
## reference
6 commits in f8ae436d936f6f4891d3c1bbb1af5865eb8aeadb..08ae27a4921ca53967656a7391c82f6c0ddd1ccc
2019-05-31 14:59:12 +0200 to 2019-06-17 11:24:13 -0700
- Document `underscore_const_names` (rust-lang-nursery/reference#620)
- Describe `#[repr(align(x))]` semantics on `enum`s (rust-lang-nursery/reference#619)
- Fix typo in lazy_static example. (rust-lang-nursery/reference#622)
- Rust 2015 supports Kleene ? macro operator (rust-lang-nursery/reference#617)
- Bitrig support has been removed (rust-lang-nursery/reference#615)
- Fix spell error in attributes/diagnostics.md (rust-lang-nursery/reference#614)
## book
8 commits in 62a8c6f25fbd981c80a046f3b04be9684749af3b..9aacfcc4c5b102c8cda195932addefd32fe955d2
2019-05-28 15:48:23 -0400 to 2019-06-16 21:27:26 -0400
- Add a script to help diff rendered books
- Update mdbook links (rust-lang/book#1989)
- Update for markdown changes. (rust-lang/book#1984)
- Make all mentions of crates.io consistent
- Propagate corrections made in layout
- Add link to Swedish translation
- Propagate small edits made in layout
- Remove snapshot files that have been checked with layout
## rust-by-example
3 commits in d8eec1dd65470b9a68e80ac1cba8fad0daac4916..b27472962986e85c94f4183b1a6d2207660d3ed6
2019-06-08 10:15:22 -0300 to 2019-06-17 15:52:07 -0300
- Update `macro_rules!` formatting. (rust-lang/rust-by-example#1205)
- Add Cargo comment to the `try!` macro example. (rust-lang/rust-by-example#1203)
- Fix typo (rust-lang/rust-by-example#1204)
## rustc-guide
15 commits in 3ac9cfc9c9ab2e366feebf18718112737f572352..f55e97c145cf37fd664db2e0e2f2d05df328bf4f
2019-06-02 19:36:58 -0500 to 2019-06-15 17:29:12 -0500
- fix typos
- Update for TyCtxt<'a, 'gcx, 'tcx> -> TyCtxt<'tcx>.
- fix ci failures, typos, broken links
- Some edits to address review comments
- Notes about closure de-sugaring
- add note about rebuilding llvm
- Changes to config.toml require a clean
- Fix a few typos in type inference chapter
- Add triagebot (rust-lang/rustc-guide#337)
- Update how-to-build-and-run.md
- Update how-to-build-and-run.md
- Update compiler-debugging.md
- use debug instead of debuginfo-level
- Replaced tabs with spaces
- correct indentation
## edition-guide
4 commits in 581c6cccfaf995394ea9dcac362dc8e731c18558..f8072acde5ce29c7570d7986180bbded2d22e287
2019-05-06 12:47:44 -0700 to 2019-06-14 23:27:05 +0200
- Small fixes (rust-lang-nursery/edition-guide#179)
- Cleanup unused NLL link (rust-lang-nursery/edition-guide#167)
- Part 2: Rust 2015 now supports ? macro rep op (rust-lang-nursery/edition-guide#163)
- Part 1: Rust 2015 now supports ? macro rep op (rust-lang-nursery/edition-guide#162)
bors [Thu, 20 Jun 2019 23:44:52 +0000 (23:44 +0000)]
Auto merge of #60293 - nagisa:rustdoc-all-the-auto-traits, r=eddyb
rustdoc: generate implementors for all auto traits
Previously we would only generate a list of synthetic implementations
for two well known traits – Send and Sync. With this patch all the auto
traits known to rustc are considered. This includes such traits like
Unpin and user’s own traits.
Sadly the implementation still iterates through the list of crate items
and checks them against the traits, which for non-std crates containing
their own auto-traits will still not include types defined in std/core.
Dylan MacKenzie [Thu, 20 Jun 2019 22:31:58 +0000 (15:31 -0700)]
Kill conflicting borrows of places with projections.
Resolves #62007.
Due to a bug, the previous version of this check did not actually kill
any conflicting borrows unless the borrowed place had no projections.
Specifically, `entry_set` will always be empty when `statement_effect`
is called. It does not contain the set of borrows which are live at this
point in the program.
bors [Thu, 20 Jun 2019 20:55:02 +0000 (20:55 +0000)]
Auto merge of #62006 - Centril:rollup-4my59er, r=Centril
Rollup of 5 pull requests
Successful merges:
- #61900 (implement Error::source for Box<T: Error>)
- #61979 (Implement Debug for PlaceBase)
- #61981 (Closures implement Copy and Clone, generators don't)
- #61996 (Add unit tests for unescaping raw (byte) strings)
- #62000 (Add test for issue-54189)
bors [Thu, 20 Jun 2019 17:57:04 +0000 (17:57 +0000)]
Auto merge of #61827 - golddranks:lldb_fix, r=nikomatsakis
Fix rust-lldb wrapper scripts.
Currently the `rust-lldb` wrapper provided by Rust project is broken. The error messages it produces on launch are as follows:
```
warning: ignoring unknown option: --one-line-before-file=command script import "/Users/kon/.rustup/toolchains/nightly-2019-05-02-x86_64-apple-darwin/lib/rustlib/etc/lldb_rust_formatters.py"
warning: ignoring unknown option: --one-line-before-file=type summary add --no-value --python-function lldb_rust_formatters.print_val -x ".*" --category Rust
warning: ignoring unknown option: --one-line-before-file=type category enable Rust
(lldb) target create "target/debug/nagare"
Traceback (most recent call last):
File "<string>", line 1, in <module>
File "/Users/kon/.rustup/toolchains/nightly-2019-05-02-x86_64-apple-darwin/lib/rustlib/x86_64-apple-darwin/lib/python2.7/site-packages/lldb/__init__.py", line 1481, in <module>
class SBAddress(object):
File "/Users/kon/.rustup/toolchains/nightly-2019-05-02-x86_64-apple-darwin/lib/rustlib/x86_64-apple-darwin/lib/python2.7/site-packages/lldb/__init__.py", line 1647, in SBAddress
__swig_getmethods__["module"] = GetModule
NameError: name '__swig_getmethods__' is not defined
Traceback (most recent call last):
File "<string>", line 1, in <module>
NameError: name 'run_one_line' is not defined
Traceback (most recent call last):
File "<string>", line 1, in <module>
NameError: name 'run_one_line' is not defined
Traceback (most recent call last):
File "<string>", line 1, in <module>
NameError: name 'run_one_line' is not defined
...etc.
```
The errors stem from two regressions: one caused by an LLVM upgrade and one caused by unintended upgrade to SWIG 4.0 (SWIG is a wrapper generator that is used to generate Python bindings for LLVM and LLDB.)
(Edit: found the exact dates) The SWIG breakage happened because of a Homebrew version upgrade on `nightly-2019-05-01-x86_64-apple-darwin` and the LLVM breakage happened on `nightly-2019-01-27-x86_64-apple-darwin` (likely to have been caused by https://github.com/rust-lang/rust/pull/57675 ).
The fix is to update the LLVM parameter syntax and to "downgrade" to SWIG 3.0.x. SWIG 3.0.x is not going to be supported by Homebrew forever, but should be good for now, until LLDB upgrades to support SWIG 4.0.0. Here's some more info about Homebrew support: https://github.com/Homebrew/homebrew-core/pull/39929 & https://github.com/Homebrew/homebrew-core/pull/40882 I'm going to send a bug & fix to LLDB about SWIG 4.0.0 to get the situation fixed in the future.
It would be good to also backport this to beta, since it's such a small change, and will fix an obvious regression.
rustdoc: generate implementors for all auto traits
Previously we would only generate a list of synthetic implementations
for two well known traits – Send and Sync. With this patch all the auto
traits known to rustc are considered. This includes such traits like
Unpin and user’s own traits.
Sadly the implementation still iterates through the list of crate items
and checks them against the traits, which for non-std crates containing
their own auto-traits will still not include types defined in std/core.
Rollup merge of #61968 - eddyb:hir-noclone, r=petrochenkov
rustc: disallow cloning HIR nodes.
Besides being inefficient, cloning also risks creating broken HIR (without properly recreating all the IDs and whatnot, in which case you might as well reconstruct the entire node without ever `Clone`-ing anything).
We detect *some* detrimental situations (based on the occurrence of `HirId`s, I believe?), but it's better to statically disallow it, IMO.
One of the examples that is fixed by this PR is `tcx.hir().fn_decl{,_by_hir_id}`, which was cloning an entire `hir::FnDecl` *every single time it was called*.
Rollup merge of #60772 - timvermeulen:slice_iter_nth_back, r=scottmcm
Implement nth_back for slice::{Iter, IterMut}
Part of #54054.
I implemented `nth_back` as straightforwardly as I could, and then slightly changed `nth` to match `nth_back`. I believe I did so correctly, but please double-check 🙂
I also added the helper methods `zst_shrink`, `next_unchecked`, and `next_back_unchecked` to get rid of some duplicated code. These changes hopefully make this code easier to understand for new contributors like me.
I noticed the `is_empty!` and `len!` macros which sole purpose seems to be inlining, according to the comment right above them, but the `is_empty` and `len` methods are already marked with `#[inline(always)]`. Does that mean we could replace these macros with method calls, without affecting anything? I'd love to get rid of them.
bors [Thu, 20 Jun 2019 02:36:49 +0000 (02:36 +0000)]
Auto merge of #60341 - mtak-:macos-tlv-workaround, r=alexcrichton
macos tlv workaround
fixes: #60141
Includes:
* remove dead code: `requires_move_before_drop`. This hasn't been needed for a while now (oops I should have removed it in #57655)
* redox had a copy of `fast::Key` (not sure why?). That has been removed.
* Perform a `read_volatile` on OSX to reduce `tlv_get_addr` calls per `__getit` from (4-2 depending on context) to 1.
`tlv_get_addr` is relatively expensive (~1.5ns on my machine).
Previously, in contexts where `__getit` was inlined, 4 calls to `tlv_get_addr` were performed per lookup. For some reason when `__getit` is not inlined this is reduced to 2x - and performance improves to match.
After this PR, I have only ever seen 1x call to `tlv_get_addr` per `__getit`, and macos now benefits from situations where `__getit` is inlined.
I'm not sure if the `read_volatile(&&__KEY)` trick is working around an LLVM bug, or a rustc bug, or neither.
projects / rust.git / log