Alex Crichton [Wed, 29 Jul 2015 04:13:02 +0000 (21:13 -0700)]
std: Fix sub-second Condvar::wait_timeout_ms
The API we're calling requires us to pass an absolute point in time as an
argument (`pthread_cond_timedwait`) so we call `gettimeofday` ahead of time to
then add the specified duration to. Unfortuantely the current "add the duration"
logic forgot to take into account the current time's sub-second precision (e.g.
the `tv_usec` field was ignored), causing sub-second duration waits to return
spuriously.
Auto merge of #27260 - alexcrichton:cap-lints, r=nrc
This commit is an implementation of [RFC 1193][rfc] which adds the ability to
the compiler to cap the lint level for the entire compilation session. This flag
will ensure that no lints will go above this level, and Cargo will primarily use
this flag passing `--cap-lints allow` to all upstream dependencies.
Alex Crichton [Fri, 24 Jul 2015 05:19:12 +0000 (22:19 -0700)]
rustc: Add a --cap-lints flag to the compiler
This commit is an implementation of [RFC 1193][rfc] which adds the ability to
the compiler to cap the lint level for the entire compilation session. This flag
will ensure that no lints will go above this level, and Cargo will primarily use
this flag passing `--cap-lints allow` to all upstream dependencies.
Auto merge of #26846 - P1start:print-maybe-styled-macro, r=pnkfelix
`EmitterWriter::print_maybe_styled` was basically always used with `format!`, so this macro makes some code cleaner. It should also remove some unnecessary allocations (most `print_maybe_styled` invocations allocated a `String` previously, whereas the new macro uses `write_fmt` to write the formatted string directly to the terminal).
This probably could have been part of #26838, but it’s too late now. It’s also rebased on #26838’s branch because otherwise pretty much all of the changes in this PR would conflict with the other PR’s changes.
Auto merge of #27173 - mark-buer:split-android-ndks, r=alexcrichton
Allows a multi-Android-target Rust compiler to be built.
Without these (or similar) changes, only a single-Android-target Rust compiler is possible.
Please see https://internals.rust-lang.org/t/dual-target-android-rust-compiler/2382/3 for additional context.
Auto merge of #26173 - pnkfelix:fsk-trans-nzmove-take3, r=nikomatsakis
Add dropflag hints (stack-local booleans) for unfragmented paths in trans. Part of #5016.
Added code to maintain these hints at runtime, and to conditionalize drop-filling and calls to destructors.
In this early stage of my incremental implementation strategy, we are using hints, so we are always free to leave out a flag for a path -- then we just pass `None` as the dropflag hint in the corresponding schedule cleanup call. But, once a path has a hint, we must at least maintain it: i.e. if the hint exists, we must ensure it is never set to "moved" if the data in question might actually have been initialized. It remains sound to conservatively set the hint to "initialized" as long as the true drop-flag embedded in the value itself is up-to-date.
I hope the commit series has been broken up to be readable; most of the commits in the series should build (though I did not always check this).
----
Oh, I think this technically qualifies as a:
[breaking-change]
because it removes drop-filling in some cases where one could previously observe it. That should only affect `unsafe` code; no safe code should be able to inspect whether the drop-fill was present or not. For an example of code that needed to change to account for this, see commit a81c24ae0216ab47df59acd724f8a33124fb6d97 (a unit test of the `move_val_init` intrinsic). I have not encountered actual code that needed to be updated to account for the change, since this should only be skipping the drop-filling on *moved* values, not on dropped one, and so even types that use `unsafe_no_drop_flag` should be unchanged by this particular PR. (Their time will come later.)
Reduced the Lvalue constructors to a kernel of three constructors.
Updated all call sites that used the other contructors to uniformly
call `Lvalue::new_with_hint`, passing along the appropriate kind
of hint for each context.
Add dropflag hints (stack-local booleans) for unfragmented paths in trans.
Added code to maintain these hints at runtime, and to conditionalize
drop-filling and calls to destructors.
In this early stage, we are using hints, so we are always free to
leave out a flag for a path -- then we just pass `None` as the
dropflag hint in the corresponding schedule cleanup call. But, once a
path has a hint, we must at least maintain it: i.e. if the hint
exists, we must ensure it is never set to "moved" if the data in
question might actually have been initialized. It remains sound to
conservatively set the hint to "initialized" as long as the true
drop-flag embedded in the value itself is up-to-date.
----
Here are some high-level details I want to point out:
* We maintain the hint in Lvalue::post_store, marking the lvalue as
moved. (But also continue drop-filling if necessary.)
* We update the hint on ExprAssign.
* We pass along the hint in once closures that capture-by-move.
* You only call `drop_ty` for state that does not have an associated hint.
If you have a hint, you must call `drop_ty_core` instead.
(Originally I passed the hint into `drop_ty` as well, to make the
connection to a hint more apparent, but the vast majority of
current calls to `drop_ty` are in contexts where no hint is
available, so it just seemed like noise in the resulting diff.)
Extend `trans::datum::Lvalue` so that it carrys an optional dropflag hint.
Instrumented calls sites that construct Lvalues to ease tracking down
cases that we might need to change whether or not they carry a hint.
Note that this commit does not do anything to actually *construct*
the `lldropflag_hints` map, nor does it change anything about codegen
itself. Those parts are in follow-on commits.
Prep for dropflag hints: refactor `trans:_match` to pass around `MatchInput` rather than `ValueRef`.
(already thumbs-upped pre-rebase by nikomatsakis)
The refactoring here is trivial because `trans::datum::Lvalue`
currently carries no payload. However, future commits will start
adding a payload to `Lvalue`, and thus will force us either
1. to thread the payload through the `_match` code (a long term goal), or
2. to ensure the payload has some reasonable default value.
Auto merge of #27319 - diaphore:pr_debug_osstr_escape, r=alexcrichton
I had to modify some tests : since `wtf8buf_show` and `wtf8_show` were doing the exact same thing, I repurposed `wtf8_show` to `wtf8buf_show_str` which ensures `Wtf8Buf` `Debug`-formats the same as `str`.
`write_str_escaped` might also be shared amongst other `fmt` but I just left it there within `Wtf8::fmt` for review.
Auto merge of #27280 - bluss:siphash-perf, r=alexcrichton
Improve siphash performance for longer data
Use `ptr::copy_nonoverlapping` (aka memcpy) to load an u64 from the
byte stream. This is correct for any alignment, and the compiler will
use the appropriate instruction to load the data.
Also contains small tweaks that should benefit hashing short data too,
both the commit that removes a variable and the autovectorization of
the hash state initialization (in SipHash::reset).
Benchmarks show that hashing longer data benefits for the improved word loading.
Before (using benchmarks from the first commit in the PR):
Auto merge of #26914 - alexcrichton:deprecate-easy, r=aturon
Many of these have long since reached their stage of being obsolete, so this
commit starts the removal process for all of them. The unstable features that
were deprecated are:
Alex Crichton [Wed, 8 Jul 2015 15:33:13 +0000 (08:33 -0700)]
std: Deprecate a number of unstable features
Many of these have long since reached their stage of being obsolete, so this
commit starts the removal process for all of them. The unstable features that
were deprecated are:
Visual Studio 2015, recently released, includes the Universal CRT, a different
flavor than was provided before. The binaries and header files for this library
are included in new locations not previously known about by gcc-rs, and this
commit adds support for the necessary probing to find these.
Unfortunately there are no prior examples of this probing to be found in
frameworks like CMake or clang, so this is done is a bit of a sketchy method
today. It assumes that the installation is in a relatively standard format and
then blindly looks for the location of the UCRT. I'd love to switch this over to
using registry keys for probing, but I was currently unable to find such keys.
This should enable the compiler to work outside VS 2015 dev tools prompts.
Auto merge of #26216 - azerupi:doc-experiments, r=steveklabnik
So I have tried to improve the rustbook engine:
- The sidebar now looks a lot more like gitbook (I thinks it cleaner)
- Added the Open Sans font, in my opinion more readable for prolonged periods of time
- Changed the style for code blocks a little
I encountered 1 problem. In `build.rs` I added this google font url (I commented out the non-relevant parts for clarity)
```rust
let rustdoc_args: &[String] = &[
//"".to_string(),
//preprocessed_path.display().to_string(),
//format!("-o{}", out_path.display()),
//format!("--html-before-content={}", prelude.display()),
//format!("--html-after-content={}", postlude.display()),
//format!("--markdown-playground-url=http://play.rust-lang.org"),
//format!("--markdown-css={}", item.path_to_root.join("rust-book.css").display()),
format!("--markdown-css=http://fonts.googleapis.com/css?family=Open+Sans:400italic,700italic,400,700"),
//"--markdown-no-toc".to_string(),
];
```
As you can see, I had to escape `=` with `=` because the string would get truncated if I didn't. Is that normal behaviour? Is that for security measures? If it is, isn't it a little weak if you can circumvent it by escaped characters? I don't know the reason behind, but I thought it was at least worth mentioning :)
Take your time for this PR, I still want to add multiple improvements:
- Like gitbook, possibility to change font by user
- Put `css` and `js` in their respective files (not hardcoded in rust)
- button to hide sidebar
- ...
So I'm not in a hurry to get this merged ;) But if you think it's good enough to be merged, go ahead. I will make another PR when I have other improvements.
In the image below is a screen of the improvements
This commit caused a lot of unintended breakage for many Cargo builds. The problem is that Cargo compiles build scripts with `-C prefer-dynamic`, so the standard library is always dynamically linked and hence any imports need to be marked with `dllimport`. Dependencies of build scripts, however, were compiled as rlibs and did not have their imports tagged with `dllimport`, so build scripts would fail to link.
While known that this situation would break, it was unknown that it was a common scenario in the wild. As a result I'm just reverting these heuristics for now.
Auto merge of #27284 - lastorset:default-methods, r=Gankro
Instead of bar/baz, use valid/invalid as default methods. This
illustrates why you might want default methods, and shows that you can
call other trait methods from a default method.
Auto merge of #27311 - kballard:thread-mod-desc-remove-scoped, r=huonw
It's deprecated and unsafe, so we shouldn't be encouraging people to use
it. Move it to `std::thread::scoped` instead, since it's still useful
information to anyone who is using the API.
Auto merge of #27294 - eddyb:deep-unsize-hinting, r=nrc
`Rc::new(RefCell::new(x)): Rc<RefCell<Trait>>` should not mean `RefCell::new(x): RefCell<Trait>`.
The latter is impossible, as an rvalue can't have an unsized type.
We were already handling unsized argument hints, but not when dealing with unsized structures.
Kevin Ballard [Sun, 26 Jul 2015 22:26:47 +0000 (15:26 -0700)]
Remove the module-level documentation for thread::scoped
It's deprecated and unsafe, so we shouldn't be encouraging people to use
it. Move it to `std::thread::scoped` instead, since it's still useful
information to anyone who is using the API.
This PR tries to clarify uses of "character" where it means "code point" or "UTF-8 sequence", which are almost, but not quite the same. Edge cases added to some examples to demonstrate this.
However, I've kept use of the term "code point" instead of "Unicode scalar value", because in UTF-8 they're the same, and "code point" is more widely known.
Auto merge of #27297 - mitaa:cleanup_E0005, r=alexcrichton
This does two things:
* removes ast::LocalSource, where only one variant was used because for-loop expansion has changed. One reason that this slipped into here is because the code in `check_local` which checks for `LocalSource::LocalFor` would report the same error as in `check_exhaustive` while using the wrong error code (E0005 instead of E0297).
* silences the warning about already used diagnostic code E0005 (fixes #27279)
Auto merge of #26870 - jroesch:default-typaram-fallback, r=nikomatsakis
This PR completes [RFC 213](https://github.com/rust-lang/rfcs/blob/master/text/0213-defaulted-type-params.md) by allowing default type parameters to influence inference. This is almost certainly a breaking change due to interactions between default type parameters and the old fallback algorithm used for integral and floating point literals.
The error messages still require polish but I wanted to get early review and feedback from others on the the changes, error messages, and test cases. I also imagine we will want to run anywhere from 1-3 versions of this on crater and evaluate the impact, and it would be best to get that ball rolling.
The only outstanding issue I'm aware of is that type alias defaults don't work. It seems this may require significant restructuring, since during inference type aliases have already been expanded. @nikomatsakis might be able to provide some clarity here.
Remove `ast::LocalSource` with only one used variant
`LocalSource` indicated wether a let binding originated from for-loop desugaring to enable specialized error messages, but for-loop expansion has changed and this is now achieved through `MatchSource::ForLoopDesugar`.
Jared Roesch [Mon, 13 Jul 2015 03:33:17 +0000 (20:33 -0700)]
Refactor the default type parameter algorithm
The algorithm was not correctly detecting conflicts after moving
defaults into TypeVariableValue. The updated algorithm
correctly detects and reports conflicts with information about
where the conflict occured and which items the defaults were
introduced by. The span's for said items are not being correctly
attached and still need to be patched.
Jared Roesch [Tue, 7 Jul 2015 22:50:02 +0000 (15:50 -0700)]
Implement Default TyParam fallback
This patch allows type parameter defaults to influence type inference. This is a possible breaking change since it effects the way type inference works and will have different behavior when mixing defaults and literal fallback.
Instead of bar/baz, use valid/invalid as default methods. This
illustrates why you might want default methods, and shows that you can
call other trait methods from a default method.
Auto merge of #27258 - nikomatsakis:issue-26952, r=eddyb
Correct regression in type-inference caused by failing to reconfirm that
the object trait matches the required trait during trait selection. The
existing code was checking that the object trait WOULD match (in a
probe), but never executing the match outside of a probe.
This corrects various regressions observed in the wild, including
issue #26952. Fixes #26952.
Auto merge of #26630 - eefriedman:recursive-static, r=pnkfelix
***Edit: Fixed now.*** I'm pretty sure the way I'm using LLVMReplaceAllUsesWith here is
unsafe... but before I figure out how to fix that, I'd like a
reality-check: is this actually useful?
Auto merge of #27253 - bossmc:unbalanced-delimiters-cause-ice, r=nikomatsakis
This introduces a test for #23389 and improves the error behaviour to treat the malformed LHS as an error, not a compiler bug.
The parse phase that precedes the call to `check_lhs_nt_follows` could possibly be enhanced to police the format itself (which the old code suggests was the original intention), but I'm not sure that's any nicer than just parsing the matcher as generic rust code and then policing the specific requirements for being a macro matcher afterwards (as this does).
siphash: Use ptr::copy_nonoverlapping for efficient data loading
Use `ptr::copy_nonoverlapping` (aka memcpy) to load an u64 from the
byte stream. This is correct for any alignment, and the compiler will
use the appropriate instruction to load the data.
Use unchecked indexing.
This results in a large improvement of throughput (hashed bytes
/ second) for long data. Maximum improvement benches at a 70% increase
in throughput for large values (> 256 bytes) but already values of 16
bytes or larger improve.
Introducing unchecked indexing is motivated to reach as good throughput
as possible. Using ptr::copy_nonoverlapping without unchecked indexing
would land the improvement some 20-30 pct units lower.
We use a debug assertion so that the test suite checks our use of
unchecked indexing.
projects / rust.git / log