This modifies the builder to download and use the LLVM tools from the
last known good build on the WebAssembly buildbot waterfall, since these
tools are built with the WebAssembly LLVM backend enabled.
This is used in wasm32-experimental-emscripten to ensure that emscripten
links against the libc bitcode files produced by the wasm LLVM backend,
instead of using fastcomp.
This adds the experimental targets option to configure so it can be used
by the builders and changes the wasm32 Dockerfile accordingly. Instead
of using LLVM from the emsdk, the builder's emscripten tools now uses
the Rust in-tree LLVM, since this is the one built with wasm support.
Change the for-loop desugar so the `break` does not affect type inference. Fixes#42618
Rewrite the `for` loop desugaring to avoid contaminating the inference results. Under the older desugaring, `for x in vec![] { .. }` would erroneously type-check, even though the type of `vec![]` is unconstrained. (written by @nikomatsakis)
Pass path to python from bootstrap.py to bootstrap.rs
When bootstrap is executed with python not in `$PATH`, (e. g.
`c:\Python27\python.exe x.py test`) bootstrap cannot find python
and crashes.
This commit passes path to python in `BOOTSTRAP_PYTHON` env var.
Make rustc errors colorful.
Rustbuild passes --message-format=json to Cargo to learn about the
dependencies for a given build, which then makes Cargo steal the
stderr/stdout for the compiler process, leading to non colorful output.
To avoid this, detection of stderr being a tty is added to rustbuild,
and an environment variable is used to communicate with the rustc shim.
Fixes https://github.com/rust-lang/rust/issues/42801.
r? @alexcrichton
Impl Clone for DefaultHasher
It's useful for a hasher to be `Clone`. It's also strange for any type to not be `Clone`. `DefaultHasher` is not meant to be used directly, but being in std it can be useful as a placeholder. I don't see any forward compatibility hazard if the hasher is changed since it's very rare for something to not be `Clone`.
Fixes bootstrapping with custom cargo/rustc.
config.mk is now always read when parsing the configuration to prevent
this from reoccurring in the future, hopefully.
Fixes https://github.com/rust-lang/rust/issues/42543.
r? @alexcrichton
cc @infinity0 @kyrias
Integrate jobserver support to parallel codegen
This commit integrates the `jobserver` crate into the compiler. The crate was
previously integrated in to Cargo as part of rust-lang/cargo#4110. The purpose
here is to two-fold:
* Primarily the compiler can cooperate with Cargo on parallelism. When you run
`cargo build -j4` then this'll make sure that the entire build process between
Cargo/rustc won't use more than 4 cores, whereas today you'd get 4 rustc
instances which may all try to spawn lots of threads.
* Secondarily rustc/Cargo can now integrate with a foreign GNU `make` jobserver.
This means that if you call cargo/rustc from `make` or another
jobserver-compatible implementation it'll use foreign parallelism settings
instead of creating new ones locally.
As the number of parallel codegen instances in the compiler continues to grow
over time with the advent of incremental compilation it's expected that this'll
become more of a problem, so this is intended to nip concurrent concerns in the
bud by having all the tools to cooperate!
Note that while rustc has support for itself creating a jobserver it's far more
likely that rustc will always use the jobserver configured by Cargo. Cargo today
will now set a jobserver unconditionally for rustc to use.
mark calls in the unwind path as !noinline
The unwind path is always cold, so that should not have bad performance
implications. This avoids catastrophic exponential inlining, and also
decreases the size of librustc.so by 1.5% (OTOH, the size of `libstd.so`
increased by 0.5% for some reason).
Fixes#41696.
r? @nagisa
Rustbuild passes --message-format=json to Cargo to learn about the
dependencies for a given build, which then makes Cargo steal the
stderr/stdout for the compiler process, leading to non colorful output.
To avoid this, detection of stderr being a tty is added to rustbuild,
and an environment variable is used to communicate with the rustc shim.
When bootstrap is executed with python not in `$PATH`, (e. g.
`c:\Python27\python.exe x.py test`) bootstrap cannot find python
and crashes.
This commit passes path to python in `BOOTSTRAP_PYTHON` env var.
This commit integrates the `jobserver` crate into the compiler. The crate was
previously integrated in to Cargo as part of rust-lang/cargo#4110. The purpose
here is to two-fold:
* Primarily the compiler can cooperate with Cargo on parallelism. When you run
`cargo build -j4` then this'll make sure that the entire build process between
Cargo/rustc won't use more than 4 cores, whereas today you'd get 4 rustc
instances which may all try to spawn lots of threads.
* Secondarily rustc/Cargo can now integrate with a foreign GNU `make` jobserver.
This means that if you call cargo/rustc from `make` or another
jobserver-compatible implementation it'll use foreign parallelism settings
instead of creating new ones locally.
As the number of parallel codegen instances in the compiler continues to grow
over time with the advent of incremental compilation it's expected that this'll
become more of a problem, so this is intended to nip concurrent concerns in the
bud by having all the tools to cooperate!
Note that while rustc has support for itself creating a jobserver it's far more
likely that rustc will always use the jobserver configured by Cargo. Cargo today
will now set a jobserver unconditionally for rustc to use.
Memoize types in `is_representable` to avoid exponential worst-case
I could have made representability a cached query, but that would have
been added complexity for not much benefit - outside of the exponential
worst-case, this pass is fast enough already.
Fixes#42747.
r? @eddyb
Remove in-tree flate/getopts crates
Remove `src/libflate` in favor of `flate2` on crates.io and `src/libgetopts` in favor of `getopts` on crates.io. The replacements have slightly different APIs and the usage in the compiler has been updated to reflect this.
This uncovered an unfortunate limitation of the compiler today to deal with linking everything correctly, and the workaround can be found documented in `src/librustc/Cargo.toml`.
Add a Read::initializer method
This is an API that allows types to indicate that they can be passed
buffers of uninitialized memory which can improve performance.
cc @SimonSapin
r? @alexcrichton
Clearer Error Message for Duplicate Definition
Clearer use of the error message and span labels to communicate duplication definitions/imports.
fixes#42061
Convert `Into<Box<[T]>> for Vec<T>` into `From<Vec<T>> for Box<[T]>`
As the `collections` crate has been merged into `alloc` in #42648 this impl is now possible. This is the final part of #42129 missing from #42227.
add `FromStr` Impl for `char`
fixes#24939.
is it possible to use pub(restricted) instead of using a stability attribute for the internal error representation? is it needed at all?
This commit deletes the in-tree `getopts` crate in favor of the crates.io-based
`getopts` crate. The main difference here is with a new builder-style API, but
otherwise everything else remains relatively standard.
The unwind path is always cold, so that should not have bad performance
implications. This avoids catastrophic exponential inlining, and also
decreases the size of librustc.so by 1.5% (OTOH, the size of `libstd.so`
increased by 0.5% for some reason).
Fixes#41696.
