Alex Crichton [Tue, 20 Dec 2016 19:16:46 +0000 (11:16 -0800)]
Rollup merge of #38472 - alexcrichton:fix-cargotest, r=brson
travis: Fix the cargotest bot
Recently the Cargo revision running on cargotest was updated in #38051, and the
updated version of Cargo pulls in the openssl-sys 0.9 crate instead of the old
openssl-sys 0.7 crate. The 0.9 release of openssl-sys has slightly different
requirements and logic for detecting OpenSSL, namely it requires `pkg-config` to
be present on Linux typically.
This commit fixes this problem by installing pkg-config in the cargotest
container that's running on the bots. This in turn should hopefully fix the
build script and allow it to find the already-installed local OpenSSL libraries.
Alex Crichton [Tue, 20 Dec 2016 19:16:45 +0000 (11:16 -0800)]
Rollup merge of #38471 - alexcrichton:run-debuginfo-tests, r=brson
rustbuild: Run debuginfo tests by default
This fixes an accidental regression in rustbuild which stopped running debuginfo
tests by default. Here we flag the test suites as `default(true)` to ensure that
they're run on bots, for example.
Alex Crichton [Tue, 20 Dec 2016 19:16:44 +0000 (11:16 -0800)]
Rollup merge of #38470 - alexcrichton:really-fix-osx, r=brson
rustbuild: Update Cargo download location
I updated the beta compiler used to bootstrap the master branch in #38438 with
the intention of fixing Travis OSX linkage issues but I mistakenly forgot that
the PR only updated rustc, not Cargo itself. Cargo has a new release process
with downloads in a different location, so this commit updates rustbuild to
download from this new location by tracking revisions instead of Cargo nightly
dates.
Alex Crichton [Tue, 20 Dec 2016 19:16:42 +0000 (11:16 -0800)]
Rollup merge of #38468 - xen0n:tarball-wrangling, r=alexcrichton
rustbuild: Eliminate duplication of dist tarballs
Fixes #38365 by not constructing the duplicate steps in the first place, as suggested. The source package step is lacking the check as in other steps, so it is added as well.
Tested locally with the `alexcrichton/rust-slave-linux-cross:2016-11-11` container (with the build slave init replaced with no-op, of course).
Alex Crichton [Tue, 20 Dec 2016 19:16:41 +0000 (11:16 -0800)]
Rollup merge of #38463 - japaric:asm-args, r=alexcrichton
target spec: add an asm-args field to pass arguments to the external ..
assembler
The main use case is the (still out of tree) msp430 target. For that target we use an external assembler, `mps430-elf-gcc`, to produce object files from the assembly rustc/llvm outputs. The problem is that by default `msp430-elf-gcc` produces object files for the MSP430**X** ABI but we want to use produce objects using the MSP430 (note: no X) ABI. To do that we have to pass `-mcpu=msp430` to the assembler and that's what this flag is for.
Alex Crichton [Tue, 20 Dec 2016 19:16:40 +0000 (11:16 -0800)]
Rollup merge of #38451 - semarie:openbsd-rustbuild, r=alexcrichton
adaptation to rustbuild for openbsd
Since the switch to rustbuild, the build for openbsd is broken:
- [X] `ar` inference based on compiler name is wrong (OpenBSD usually use `egcc`, but `ear` doesn't exist)
- [X] `make` isn't GNU-make under OpenBSD (and others BSD platforms)
- [x] `stdc++` isn't the right stdc++ library to link with (it should be `estdc++`)
- [x] corrects tests that don't pass anymore (problems related to rustbuild)
Alex Crichton [Tue, 20 Dec 2016 19:16:38 +0000 (11:16 -0800)]
Rollup merge of #38418 - michaelwoerister:def_path_cleanup, r=eddyb
Cleanup refactoring around DefPath handling
This PR makes two big changes:
* All DefPaths of a crate are now stored in metadata in their own table (as opposed to `DefKey`s as part of metadata `Entry`s.
* The compiler will no longer allocate a pseudo-local DefId for inlined HIR nodes (because those are gross). Inlined HIR nodes will have a NodeId but they don't have there own DefId anymore. Turns out they were not needed anymore either. Hopefully HIR inlining will be gone completely one day but if until then we start needing to be able to map inlined NodeIds to original DefIds, we can add an additional table to metadata that allows for reconstructing this.
Overall this makes for some nice simplifications and removal of special cases.
Alex Crichton [Tue, 20 Dec 2016 19:16:35 +0000 (11:16 -0800)]
Rollup merge of #38388 - redox-os:config_toml_prefix, r=alexcrichton
Add prefix to config.toml
This allows `rustbuild` to be used to install to a prefix.
```toml
[build]
prefix = "/path/to/install"
```
For example, the following `config.toml` will cause `x.py dist --install` to install to `/path/to/install`
Alex Crichton [Tue, 20 Dec 2016 19:16:32 +0000 (11:16 -0800)]
Rollup merge of #38315 - jonhoo:better-E0309, r=pnkfelix
Expand E0309 explanation with motivating example
I recently started reading @Gankro's "[Learning Rust With Entirely Too Many Linked Lists](http://cglab.ca/~abeinges/blah/too-many-lists/book/README.html)", and came across [a part](http://cglab.ca/~abeinges/blah/too-many-lists/book/second-iter.html) where he comes across `E0309`, and after showing `rustc --explain E0309` prompty says
> This is dumb. I think it's dumb. You have to do it.
Humor aside, I think this says something about the current explanation being somewhat lacking.
This patch introduces a motivating example saying why `T: 'a` is a necessary restriction. Hopefully, this will help new Rustaceans understand why leaving out the `'a` bound on `T` might lead to broken code.
Alex Crichton [Tue, 20 Dec 2016 19:16:29 +0000 (11:16 -0800)]
Rollup merge of #38245 - estebank:cast-deref-hint-2, r=nikomatsakis
When cast needs a dereference point at full cast
After the fix of #37453 in PR #37369, instead of pointing at only the cast type, point at the full cast span when a cast needs a dereference, as well as assign the error label to the correct span for proper coloring:
<img width="471" alt="error span pointing at the entire cast" src="https://cloud.githubusercontent.com/assets/1606434/21024245/8797fc2e-bd38-11e6-82c1-66c281c656c1.png">
instead of
<img width="471" alt="error span pointing at the type of the cast" src="https://cloud.githubusercontent.com/assets/1606434/21023777/d4814aa6-bd36-11e6-9fc3-b2a0ea5ee15d.png">
Alex Crichton [Tue, 20 Dec 2016 19:16:20 +0000 (11:16 -0800)]
Rollup merge of #38150 - estebank:fix-23286, r=nikomatsakis
Point out the known type when field doesn't satisfy bound
For file
```rust
use std::path::Path;
fn f(p: Path) { }
```
provide the following error
```nocode
error[E0277]: the trait bound `[u8]: std::marker::Sized` is not satisfied in `std::path::Path`
--> file.rs:3:6
|
3 | fn f(p: Path) { }
| ^ within `std::path::Path`, the trait `std::marker::Sized` is not implemented for `[u8]`
|
= note: `[u8]` does not have a constant size known at compile-time
= note: required because it appears within the type `std::path::Path`
= note: all local variables must have a statically known size
```
Alex Crichton [Mon, 19 Dec 2016 18:42:04 +0000 (10:42 -0800)]
rustbuild: Run debuginfo tests by default
This fixes an accidental regression in rustbuild which stopped running debuginfo
tests by default. Here we flag the test suites as `default(true)` to ensure that
they're run on bots, for example.
Wang Xuerui [Mon, 19 Dec 2016 07:30:14 +0000 (15:30 +0800)]
rustbuild: only plan from build triple for dist
We only want to package each host/target once for `dist`. The obvious
solution takes the form of step dependency, which is implemented at
least for the `dist-rustc` step. Unfortunately since the steps are
created from `hosts x targets` during planning and *not* de-duplicated
afterwards, the problem still persists.
We therefore move the check inside `plan()` instead, to avoid creating
the duplicate steps in the first place.
Alex Crichton [Mon, 19 Dec 2016 17:45:42 +0000 (09:45 -0800)]
rustbuild: Update Cargo download location
I updated the beta compiler used to bootstrap the master branch in #38438 with
the intention of fixing Travis OSX linkage issues but I mistakenly forgot that
the PR only updated rustc, not Cargo itself. Cargo has a new release process
with downloads in a different location, so this commit updates rustbuild to
download from this new location by tracking revisions instead of Cargo nightly
dates.
bors [Mon, 19 Dec 2016 20:07:49 +0000 (20:07 +0000)]
Auto merge of #38072 - nikomatsakis:bootstrap-incremental, r=acrichto
add preliminary support for incremental compilation to rustbuild.py
This implements the integration described in #37929. It requires the use of a local nightly as your bootstrap compiler. The setup is described in `src/bootstrap/README.md`.
This does NOT implement the "copy stage0 libs to stage1" optimization described in #37929, just because that seems orthogonal to me.
In local testing, I do not yet see any incremental re-use when building rustc. I'm not sure why that is, more investigation needed.
(For these reasons, this is not marked as fixing the relevant issue.)
r? @alexcrichton -- I included one random cleanup (`Step::noop()`) that turned out to not be especially relevant. Feel free to tell me you liked it better the old way.
Alex Crichton [Mon, 19 Dec 2016 19:40:36 +0000 (11:40 -0800)]
travis: Fix the cargotest bot
Recently the Cargo revision running on cargotest was updated in #38051, and the
updated version of Cargo pulls in the openssl-sys 0.9 crate instead of the old
openssl-sys 0.7 crate. The 0.9 release of openssl-sys has slightly different
requirements and logic for detecting OpenSSL, namely it requires `pkg-config` to
be present on Linux typically.
This commit fixes this problem by installing pkg-config in the cargotest
container that's running on the bots. This in turn should hopefully fix the
build script and allow it to find the already-installed local OpenSSL libraries.
Seo Sanghyeon [Mon, 19 Dec 2016 07:59:44 +0000 (16:59 +0900)]
Rollup merge of #38459 - nagisa:rustbuild-exec, r=alexcrichton
Use exec for the wrapper on UNIXes
This not only avoids the small – and unnecessary – constant overhead for each compiler invocation,
but also helps somewhat by only having “correct” rustc processes to look for in `/proc/`.
This also makes the wrapper behave effectively as a regular exec wrapper its intended to be.
I also took liberty to change the fallback error code from `1` to `0xfe` (now only relevant on windows) so that when people complain about “compiler exited with code 254”, its obvious where the issue lies (wrapper losing the exit code somehow).
Seo Sanghyeon [Mon, 19 Dec 2016 07:59:43 +0000 (16:59 +0900)]
Rollup merge of #38438 - alexcrichton:update-beta, r=brson
Update beta bootstrap compiler
The current beta that rustc is bootstrapping from contains a bug in Cargo that
erroneously links to OpenSSL in /usr/local, but this is fixed in the most recent
1.14 beta, so let's use that.
Seo Sanghyeon [Mon, 19 Dec 2016 07:59:40 +0000 (16:59 +0900)]
Rollup merge of #38422 - zackmdavis:enjoy_tidy_path_error_macro_nicety, r=bluss
tidy features: use 2-parameter form of internal try macro for open err
This tiny patch merely applies @bluss's suggestion for how to get a more
informative error message when the feature check can't open a file, a
matter that had briefly annoyed the present author, leading to the
filing of #38417.
Seo Sanghyeon [Mon, 19 Dec 2016 07:59:37 +0000 (16:59 +0900)]
Rollup merge of #38413 - rillian:armv7-neon, r=alexcrichton
rustc: Disable NEON on armv7 android.
We thought Google's ABI for arvm7 required neon, but it is
currently optional, perhaps because there is a significant
population of Tegra 2 devices still in use.
This turns off neon code generation outside #[target-feature]
blocks just like we do on armv7-unknown-linux-gnu, but unlike
most other armv7 targets. LLVM defaults to +neon for this target,
so an explicit disable is necessary.
See https://developer.android.com/ndk/guides/abis.html#v7a
for instruction set extension requirements.
This not only avoids the small – and unnecessary – constant overhead for each compiler invocation,
but also helps somewhat by only having “correct” rustc processes to look for in `/proc/`.
This also makes the wrapper behave effectively as a regular exec wrapper its intended to be.
bors [Sun, 18 Dec 2016 07:59:32 +0000 (07:59 +0000)]
Auto merge of #37429 - camlorn:univariant_layout_optimization, r=eddyb
struct field reordering and optimization
This is work in progress. The goal is to divorce the order of fields in source code from the order of fields in the LLVM IR, then optimize structs (and tuples/enum variants)by always ordering fields from least to most aligned. It does not work yet. I intend to check compiler memory usage as a benchmark, and a crater run will probably be required.
I don't know enough of the compiler to complete this work unaided. If you see places that still need updating, please mention them. The only one I know of currently is debuginfo, which I'm putting off intentionally until a bit later.
bors [Sun, 18 Dec 2016 04:38:27 +0000 (04:38 +0000)]
Auto merge of #38440 - alexcrichton:fix-osx-nightlies, r=brson
rustbuild: Fix LC_ID_DYLIB directives on OSX
Currently libraries installed by rustbuild on OSX have an incorrect
`LC_ID_DYLIB` directive located in the dynamic libraries that are
installed. The directive we expect looks like:
@rpath/libstd.dylib
Which means that if you want to find that dynamic library you should
look at the dylib's other `@rpath` directives. Typically our `@rpath`
directives look like `@loader_path/../lib` for the compiler as that's
where the installed libraries will be located. Currently, though,
rustbuild produces dylibs with the directive that looks like:
In other words, the build directory is encoded erroneously. The compiler
already [knows how] to change this directive, but it only passes that
argument when `-C rpath` is also passed. The rustbuild system, however,
explicitly [does not pass] this option explicitly and instead bakes its
own. This logic then also erroneously didn't pass `-Wl,-install_name`
like the compiler.
To fix this regression this patch introduces a new `-Z` flag, `-Z
osx-rpath-install-name` which basically just forces the compiler to take
the previous `-install_name` branch when creating a dynamic library.
Hopefully we can sort out a better rpath story in the future, but for
now this "hack" should suffice in getting our nightly builds back to the
same state as before.
Alex Crichton [Sat, 17 Dec 2016 22:11:02 +0000 (14:11 -0800)]
rustbuild: Fix LC_ID_DYLIB directives on OSX
Currently libraries installed by rustbuild on OSX have an incorrect
`LC_ID_DYLIB` directive located in the dynamic libraries that are
installed. The directive we expect looks like:
@rpath/libstd.dylib
Which means that if you want to find that dynamic library you should
look at the dylib's other `@rpath` directives. Typically our `@rpath`
directives look like `@loader_path/../lib` for the compiler as that's
where the installed libraries will be located. Currently, though,
rustbuild produces dylibs with the directive that looks like:
In other words, the build directory is encoded erroneously. The compiler
already [knows how] to change this directive, but it only passes that
argument when `-C rpath` is also passed. The rustbuild system, however,
explicitly [does not pass] this option explicitly and instead bakes its
own. This logic then also erroneously didn't pass `-Wl,-install_name`
like the compiler.
To fix this regression this patch introduces a new `-Z` flag, `-Z
osx-rpath-install-name` which basically just forces the compiler to take
the previous `-install_name` branch when creating a dynamic library.
Hopefully we can sort out a better rpath story in the future, but for
now this "hack" should suffice in getting our nightly builds back to the
same state as before.
Alex Crichton [Sat, 17 Dec 2016 21:47:31 +0000 (13:47 -0800)]
Update beta bootstrap compiler
The current beta that rustc is bootstrapping from contains a bug in Cargo that
erroneously links to OpenSSL in /usr/local, but this is fixed in the most recent
1.14 beta, so let's use that.
Zack M. Davis [Sat, 17 Dec 2016 01:45:59 +0000 (17:45 -0800)]
tidy features: use 2-parameter form of internal try macro for open err
This tiny patch merely applies @bluss's suggestion for how to get a more
informative error message when the feature check can't open a file, a
matter that had briefly annoyed the present author, leading to the
filing of #38417.