bors [Sun, 4 Jun 2017 05:44:39 +0000 (05:44 +0000)]
Auto merge of #42265 - Zoxc:for-sugar, r=eddyb
Change for-loop desugar to not borrow the iterator during the loop
This is enables the use of suspend points inside for-loops in movable generators. This is illegal in the current desugaring as `iter` is borrowed across the body.
bors [Sun, 4 Jun 2017 03:23:39 +0000 (03:23 +0000)]
Auto merge of #41968 - kennytm:fix-unreadable-json-test-output-36516, r=nikomatsakis
Introduce 'run-pass' header to 'ui' tests in compiletest. Fix issue #36516.
<del>`ui-run` test is a combination of `ui` test and `run-pass` test. It is used to test lint output.</del>
Added support of `// run-pass` header to `ui` tests.
The compiler message of each test must match the corresponding `*.stderr` file like the traditional `ui` tests. Additionally, the compiled output must be executed successfully like the `run-pass` test.
12 `run-pass`/`run-pass-fulldeps` tests are moved to `ui`/`ui-fulldeps` plus the headers. After this move, no `run-pass`/`run-pass-fulldeps` tests should rely on the compiler's JSON message. This allows us to stop passing `--error-format json` in run-pass tests, thus fixing #36516.
bors [Sat, 3 Jun 2017 18:02:57 +0000 (18:02 +0000)]
Auto merge of #42369 - RalfJung:drop-glue, r=eddyb
array drop glue: avoid using out-of-bounds index lvalues
Avoid using out-of-bounds index lvalues. Miri doesn't like them, and they seem fishy in general.
Instead, use a pointer cast to compute the address of the beginning, and Offset to compute the end.
bors [Sat, 3 Jun 2017 10:56:42 +0000 (10:56 +0000)]
Auto merge of #42332 - michaelwoerister:no-more-retracing, r=nikomatsakis
incr.comp.: Use DefPathHash-based DepNodes in the serialized DepGraph and remove obsolete DefIdDirectory
With this PR we don't store the dep-graph as a set of `DepNode<IndexIntoDefIdDirectory>` anymore but instead as a set of `DepNode<DefPathHash>`. Since a `DefPathHash` is a global identifier that is valid across compilation sessions, we don't need the `DefIdDirectory` anymore.
Since a `DepNode<DefPathHash>` is bigger than a `DepNode<IndexIntoDefIdDirectory>` and our on-disk encoding of the dep-graph is inefficient, this PR will probably increase the amount of space the dep-graph takes up on disk. I'm in the process of gathering some performance data.
The changes in here are a step towards implementing ICH-based `DepNodes` (#42294).
Corey Farwell [Sat, 3 Jun 2017 05:00:55 +0000 (01:00 -0400)]
Rollup merge of #42382 - alexcrichton:remove-rustflags, r=Mark-Simulacrum
rustbuild: Remove RUSTFLAGS logic in rustc shim
This was added in #38072 but I can't recall why and AFAIK Cargo already handles
this. This was discovered through #42146 where passing duplicate flags was
causing problems.
Corey Farwell [Sat, 3 Jun 2017 05:00:54 +0000 (01:00 -0400)]
Rollup merge of #42368 - estebank:call-site, r=nikomatsakis
Use callsite's span for macro calls on suggestion
When suggesting an appropriate mutability for a macro call, use the call
span instead of the expanded macro's span.
```
error[E0308]: mismatched types
--> $DIR/coerce-suggestions.rs:48:9
|
48 | s = format!("foo");
| ^^^^^^^^^^^^^^ expected mutable reference, found struct `std::string::String`
|
= note: expected type `&mut std::string::String`
found type `std::string::String`
= help: try with `&mut format!("foo")`
= note: this error originates in a macro outside of the current crate
```
Fix #41858.
Corey Farwell [Sat, 3 Jun 2017 05:00:53 +0000 (01:00 -0400)]
Rollup merge of #42363 - cuviper:no-fail-fast, r=alexcrichton
rustbuild: Add `./x.py test --no-fail-fast`
This option forwards to each `cargo test` invocation, and applies the
same logic across all test steps to keep going after failures. At the
end, a brief summary line reports how many commands failed, if any.
Note that if a test program fails to even start at all, or if an
auxiliary build command related to testing fails, these are still left
to stop everything right away.
Corey Farwell [Sat, 3 Jun 2017 05:00:52 +0000 (01:00 -0400)]
Rollup merge of #42354 - Mark-Simulacrum:reduce-verbosity, r=alexcrichton
Reduce verbosity of build logs
This does two separate things.
- Sets sccache logging to warn instead of info.
- Makes tests when running for a given PR (not on auto branch) quiet. (cc @eddyb)
rustdoc: Hide `self: Box<Self>` in list of deref methods
These methods can never be called through deref so there is no point
including them. For example you can't call `into_boxed_bytes` or
`into_string` on `String`.
bors [Fri, 2 Jun 2017 19:23:14 +0000 (19:23 +0000)]
Auto merge of #42189 - nikomatsakis:chalk-trait-env-param-env, r=eddyb
querify layout and move param env out of the infcx
The main goal of this PR is to move the parameter environment *out* of the inference context. This is because the inference environment will soon be changing over the course of inference --- for example, when we enter into a `for<'a> fn(...)` type, we will push a new environment with an increasing universe index, rather than skolemizing the `'a` references. Similarly, each obligation will soon be able to have a distinct parameter environment, and therefore the `Obligation` struct is extended to carry a `ParamEnv<'tcx>`. (I debated about putting it into the cause; seems plausible, but also weird.)
Along the way, I also reworked how layout works, moving the layout cache into a proper query along the lines of needs-drop and friends.
Finally, tweaks the inference context API. It seemed to be accumulating parameters at an alarming rate. The main way to e.g. make a subtype or equality relationship is to do the following:
infcx.at(cause, param_env).sub(a, b)
infcx.at(cause, param_env).eq(a, b)
In both cases, `a` is considered the "expected" type (this used to be specified by a boolean). I tried hard to preserve the existing notion of what was "expected", although in some cases I'm not convinced it was being set on purpose one way or the other. This is why in some cases you will see me do `sup(b, a)`, which is otherwise equivalent to `sub(a, b)`, but sets the "expected type" differently.
Alex Crichton [Fri, 2 Jun 2017 16:57:20 +0000 (09:57 -0700)]
rustbuild: Remove RUSTFLAGS logic in rustc shim
This was added in #38072 but I can't recall why and AFAIK Cargo already handles
this. This was discovered through #42146 where passing duplicate flags was
causing problems.
Josh Stone [Fri, 2 Jun 2017 16:27:44 +0000 (09:27 -0700)]
rustbuild: Add `./x.py test --no-fail-fast`
This option forwards to each `cargo test` invocation, and applies the
same logic across all test steps to keep going after failures. At the
end, a brief summary line reports how many commands failed, if any.
Note that if a test program fails to even start at all, or if an
auxiliary build command related to testing fails, these are still left
to stop everything right away.
kennytm [Sat, 13 May 2017 09:54:50 +0000 (17:54 +0800)]
Introduce 'run-pass' header to 'ui' tests in compiletest. Fix issue #36516.
The 'run-pass' header cause a 'ui' test to execute the result. It is used
to test the lint output, at the same time ensure those lints won't cause
the source code to become compile-fail.
12 run-pass/run-pass-fulldeps tests gained the header and are moved to
ui/ui-fulldeps. After this move, no run-pass/run-pass-fulldeps tests should
rely on the compiler's JSON message. This allows us to stop passing
`--error-format json` in run-pass tests, thus fixing #36516.
Mark Simulacrum [Fri, 2 Jun 2017 15:10:45 +0000 (09:10 -0600)]
Rollup merge of #42343 - cuviper:install-executables, r=alexcrichton
Update rust-installer for Windows executable mode
It now marks a few whitelisted extensions as executable in the tarball,
so Windows packages can be extracted on other platforms and directly
execute install.sh.
It also includes a fix for the chmod on bulk dirs, so now the html docs
won't be marked executable en masse.
Mark Simulacrum [Fri, 2 Jun 2017 15:10:44 +0000 (09:10 -0600)]
Rollup merge of #42335 - jcowgill:fingerprint-be, r=michaelwoerister
Don't byteswap Fingerprints when encoding
Byteswapping Fingerprints when encoding is unnessesary and breaks if the Fingerprint is later decoded on a machine with different endianness to the one it was encoded on.
Fixes #42239
This PR fixes a regression caused by #42082. @michaelwoerister
* Bring back colors on Travis, which was disabled since #39036. Append `--color=always` to cargo when running in CI environment.
* Removed `set -x` from the shell scripts. The `retry` function already prints which command it is running, adding `-x` just adds noise to the output. Interesting information can be manually `echo`ed.
* Support `travis_fold`/`travis_time`. Matching pairs of these allow Travis CI to collapse the output in between. This greatly cut down the unnecessary "successful" output one need to scroll through before finding the failed statement.
* Passed `--quiet` to all tests, so tests not failing will not occupy a line of log, reducing bloat in the report.
Also include some minor changes, like changing the `script` of `.travis.yml` to execute a single-line command, so the log won't write the extremely long multi-line
`The command "if [ "$ALLOW_PR" = "" ] && [ "$TRAVIS_BRANCH" != "auto" ]; then … " exited with 0` at the end.
bors [Fri, 2 Jun 2017 07:51:20 +0000 (07:51 +0000)]
Auto merge of #41670 - scottmcm:slice-rotate, r=alexcrichton
Add an in-place rotate method for slices to libcore
A helpful primitive for moving chunks of data around inside a slice.
For example, if you have a range selected and are drag-and-dropping it somewhere else (Example from [Sean Parent's talk](https://youtu.be/qH6sSOr-yk8?t=560)).
(If this should be an RFC instead of a PR, please let me know.)
kennytm [Sun, 21 May 2017 20:27:47 +0000 (04:27 +0800)]
ci: Further tone down the test verbosity.
When `--quiet` is passed to rustbuild, suppress rustdoc test output unless
failure.
Added a `--quiet` flag to `tidy`, which suppresses the features table.
The actual `--quiet` flag is enabled in #42354.
Since details of failed tests will still be printed, and the name of slow
tests taking >60 to runtime will also be printed, the debugging difficulty
caused by information loss should be minimal; but it is very worthwhile to
keep the log under 10000 lines on Travis CI so that common errors can be
spotted without reading the raw log.
kennytm [Sun, 21 May 2017 11:29:42 +0000 (19:29 +0800)]
Change .travis.yml's `language: minimal` to `generic`.
There is no `minimal` language. Due to travis-ci/travis-ci#4895, it will
fallback to `ruby`, which certainly isn't what we want. `generic` is an
undocumented (travis-ci/docs-travis-ci-com#910) language that serves the
desired purpose.
kennytm [Wed, 17 May 2017 16:33:20 +0000 (00:33 +0800)]
ci: Improve log output (mainly Travis).
* Bring back colors on Travis, which was disabled since #39036.
Append --color=always to cargo when running in CI environment.
* Removed `set -x` in the shell scripts. The `retry` function already
prints which command it is running, add `-x` just add noise to the
output.
* Support travis_fold/travis_time. Matching pairs of these allow Travis CI
to collapse the output in between. This greatly cut down the unnecessary
"successful" output one need to scroll through before finding the failed
statement.
bors [Thu, 1 Jun 2017 11:34:13 +0000 (11:34 +0000)]
Auto merge of #42281 - eddyb:well-adjusted, r=nikomatsakis
Decompose Adjustment into smaller steps and remove the method map.
The method map held method callee information for:
* actual method calls (`x.f(...)`)
* overloaded unary, binary, indexing and call operators
* *every overloaded deref adjustment* (many can exist for each expression)
That last one was a historical ~~accident~~ hack, and part of the motivation for this PR, along with:
* a desire to compose adjustments more freely
* containing the autoderef logic better to avoid mutation within an inference snapshot
* not creating `TyFnDef` types which are incompatible with the original one
* i.e. we used to take a`TyFnDef`'s `for<'a> &'a T -> &'a U` signature and instantiate `'a` using a region inference variable, *then* package the resulting `&'b T -> &'b U` signature in another `TyFnDef`, while keeping *the same* `DefId` and `Substs`
* to fix #3548 by explicitly writing autorefs for the RHS of comparison operators
Individual commits tell their own story, of "atomic" changes avoiding breaking semantics.
Future work based on this PR could include:
* removing the signature from `TyFnDef`, now that it's always "canonical"
* some questions of variance remain, as subtyping *still* treats the signature differently
* moving part of the typeck logic for methods, autoderef and coercion into `rustc::traits`
* allowing LUB coercions (joining multiple expressions) to "stack up" many adjustments
* transitive coercions (e.g. reify or unsize after multiple steps of autoderef)
James Cowgill [Wed, 31 May 2017 14:53:35 +0000 (15:53 +0100)]
Don't byteswap Fingerprints when encoding
Byteswapping Fingerprints when encoding is unnessesary and breaks if
the Fingerprint is later decoded on a machine with different endianness
to the one it was encoded on. Fix by removing the Encodable and
Decodable implementations and use the ones derived from RustcEncodable
and RustcDecodable.
bors [Thu, 1 Jun 2017 07:44:01 +0000 (07:44 +0000)]
Auto merge of #42263 - alexcrichton:fix-copies, r=Mark-Simulacrum
rustbuild: Fix copying duplicate crates into the sysroot
After compiling a project (e.g. libstd, libtest, or librustc) rustbuild needs to
copy over all artifacts into the sysroot of the compiler it's assembling.
Unfortunately rustbuild doesn't know precisely what files to copy! Today it has
a heuristic where it just looks at the most recent version of all files that
look like rlibs/dylibs and copies those over. This unfortunately leads to bugs
with different versions of the same crate as seen in #42261.
This commit updates rustbuild's strategy of copying artifacts to work off the
list of artifacts produced by `cargo build --message-format=json`. The build
system will now parse json messages coming out of Cargo to watch for files being
generated, and then it'll only copy over those precise files.
Note that there's still a bit of weird logic where Cargo prints that it's
creating `libstd.rlib` where we actually want `libstd-xxxxx.rlib`, so we still
do a bit of "most recent file" probing for those. This commit should take care
of the crates.io dependency issues, however, as they're all copied over
precisely.