Don't recompute SymbolExportLevel for upstream crates.
The data collected in #48373 suggests that we can avoid generating up to 30% of the LLVM definitions by only instantiating function monomorphizations once with a given crate graph. Some more data, collected with a [proof-of-concept implementation](https://github.com/michaelwoerister/rust/commits/share-generics) of re-using monomorphizations, which is less efficient than the MIR-only RLIB approach, suggests that it's still around 25% LLVM definitions that we can save.
So far, this PR only cleans up handling of symbol export status. Too early to review still.
Slight modification to the as_ref example of std::option::Option
A user in a reddit thread was confused by the name of the variable
"num_as_int"; they thought the example was trying to convert the
string "10" as if it were binary 2 by calling str::len(). In reality,
the example is simply demonstrating how to take an immutable reference
to the value of an Option. The confusion comes from the coincidence
that the length of the string "10" is also its binary representation,
and the implication from the variable names that a conversion was
occuring ("num_as_str" to "num_as_int").
This PR changes the example number to 12 instead of 10, and changes
the variable name from "num_as_int" to "num_length" to better
communicate what the example is doing.
The reddit thread:
https://www.reddit.com/r/rust/comments/7zpvev/notyetawesome_rust_what_use_cases_would_you_like/dur39xw/
Turn feature-gate table into a query so it is covered by dependency tracking.
Turn access to feature gates into a query so we handle them correctly during incremental compilation.
Features are still available via `Session` through `features_untracked()`. I wish we had a better way of hiding untracked information. It would be great if we could remove the `sess` field from `TyCtxt`.
Fixes#47003.
Have Vec use slice's implementations of Index<I> and IndexMut<I>
This PR simplifies the implementation of Index and IndexMut on Vec, and in the process enables indexing Vec by any user types that implement SliceIndex.
The stability annotations probably need to be changed, but I wasn't sure of the right way to do that. It also wasn't completely clear to me if this change could break any existing code.
rustbuild: Pass `ccache` to build scripts
This is a re-attempt at #48192 hopefully this time with 100% less randomly
[blocking builds for 20 minutes][block]. To work around #48192 the sccache
server is started in the `run.sh` script very early on in the compilation
process.
[block]: https://github.com/rust-lang/rust/issues/48192
rust: Import LLD for linking wasm objects
This commit imports the LLD project from LLVM to serve as the default linker for
the `wasm32-unknown-unknown` target. The `binaryen` submoule is consequently
removed along with "binaryen linker" support in rustc.
Moving to LLD brings with it a number of benefits for wasm code:
* LLD is itself an actual linker, so there's no need to compile all wasm code
with LTO any more. As a result builds should be *much* speedier as LTO is no
longer forcibly enabled for all builds of the wasm target.
* LLD is quickly becoming an "official solution" for linking wasm code together.
This, I believe at least, is intended to be the main supported linker for
native code and wasm moving forward. Picking up support early on should help
ensure that we can help LLD identify bugs and otherwise prove that it works
great for all our use cases!
* Improvements to the wasm toolchain are currently primarily focused around LLVM
and LLD (from what I can tell at least), so it's in general much better to be
on this bandwagon for bugfixes and new features.
* Historical "hacks" like `wasm-gc` will soon no longer be necessary, LLD
will [natively implement][gc] `--gc-sections` (better than `wasm-gc`!) which
means a postprocessor is no longer needed to show off Rust's "small wasm
binary size".
LLD is added in a pretty standard way to rustc right now. A new rustbuild target
was defined for building LLD, and this is executed when a compiler's sysroot is
being assembled. LLD is compiled against the LLVM that we've got in tree, which
means we're currently on the `release_60` branch, but this may get upgraded in
the near future!
LLD is placed into rustc's sysroot in a `bin` directory. This is similar to
where `gcc.exe` can be found on Windows. This directory is automatically added
to `PATH` whenever rustc executes the linker, allowing us to define a `WasmLd`
linker which implements the interface that `wasm-ld`, LLD's frontend, expects.
Like Emscripten the LLD target is currently only enabled for Tier 1 platforms,
notably OSX/Windows/Linux, and will need to be installed manually for compiling
to wasm on other platforms. LLD is by default turned off in rustbuild, and
requires a `config.toml` option to be enabled to turn it on.
Finally the unstable `#![wasm_import_memory]` attribute was also removed as LLD
has a native option for controlling this.
[gc]: https://reviews.llvm.org/D42511
This commit refactors how the path to the linker that we're going to invoke is
selected. Previously all targets listed *both* a `LinkerFlavor` and a `linker`
(path) option, but this meant that whenever you changed one you had to change
the other. The purpose of this commit is to avoid coupling these where possible.
Target specifications now only unconditionally define the *flavor* of the linker
that they're using by default. If not otherwise specified each flavor now
implies a particular default linker to run. As a result, this means that if
you'd like to test out `ld` for example you should be able to do:
rustc -Z linker-flavor=ld foo.rs
whereas previously you had to do
rustc -Z linker-flavor=ld -C linker=ld foo.rs
This will hopefully make it a bit easier to tinker around with variants that
should otherwise be well known to work, for example with LLD, `ld` on OSX, etc.
This commit imports the LLD project from LLVM to serve as the default linker for
the `wasm32-unknown-unknown` target. The `binaryen` submoule is consequently
removed along with "binaryen linker" support in rustc.
Moving to LLD brings with it a number of benefits for wasm code:
* LLD is itself an actual linker, so there's no need to compile all wasm code
with LTO any more. As a result builds should be *much* speedier as LTO is no
longer forcibly enabled for all builds of the wasm target.
* LLD is quickly becoming an "official solution" for linking wasm code together.
This, I believe at least, is intended to be the main supported linker for
native code and wasm moving forward. Picking up support early on should help
ensure that we can help LLD identify bugs and otherwise prove that it works
great for all our use cases!
* Improvements to the wasm toolchain are currently primarily focused around LLVM
and LLD (from what I can tell at least), so it's in general much better to be
on this bandwagon for bugfixes and new features.
* Historical "hacks" like `wasm-gc` will soon no longer be necessary, LLD
will [natively implement][gc] `--gc-sections` (better than `wasm-gc`!) which
means a postprocessor is no longer needed to show off Rust's "small wasm
binary size".
LLD is added in a pretty standard way to rustc right now. A new rustbuild target
was defined for building LLD, and this is executed when a compiler's sysroot is
being assembled. LLD is compiled against the LLVM that we've got in tree, which
means we're currently on the `release_60` branch, but this may get upgraded in
the near future!
LLD is placed into rustc's sysroot in a `bin` directory. This is similar to
where `gcc.exe` can be found on Windows. This directory is automatically added
to `PATH` whenever rustc executes the linker, allowing us to define a `WasmLd`
linker which implements the interface that `wasm-ld`, LLD's frontend, expects.
Like Emscripten the LLD target is currently only enabled for Tier 1 platforms,
notably OSX/Windows/Linux, and will need to be installed manually for compiling
to wasm on other platforms. LLD is by default turned off in rustbuild, and
requires a `config.toml` option to be enabled to turn it on.
Finally the unstable `#![wasm_import_memory]` attribute was also removed as LLD
has a native option for controlling this.
[gc]: https://reviews.llvm.org/D42511
Remove --host and --target arguments to configure in Dockerfiles
These arguments are passed to the relevant x.py invocation in all cases
anyway. As such, there is no need to separately configure them. x.py
will ignore the configuration when they are passed on the command line
anyway.
r? @alexcrichton
std: Add `arch` and `simd` modules
This commit imports the `stdsimd` crate into the standard library,
creating an `arch` and `simd` module inside of both libcore and libstd.
Both of these modules are **unstable** and will continue to be so until
RFC 2335 is stabilized.
As a brief recap, the modules are organized as so:
* `arch` contains all current architectures with intrinsics, for example
`std::arch::x86`, `std::arch::x86_64`, `std::arch::arm`, etc. These
modules contain all of the intrinsics defined for the platform, like
`_mm_set1_epi8`.
* In the standard library, the `arch` module also exports a
`is_target_feature_detected` macro which performs runtime detection to
determine whether a target feature is available at runtime.
* The `simd` module contains experimental versions of strongly-typed
lane-aware SIMD primitives, to be fully fleshed out in a future RFC.
The main purpose of this commit is to start pulling in all these
intrinsics and such into the standard library on nightly and allow
testing and such. This'll help allow users to easily kick the tires and
see if intrinsics work as well as allow us to test out all the
infrastructure for moving the intrinsics into the standard library.
Don't produce TOCs for doc markdown files
Currently, we are producing headers for markdown files,
which is generally not what we want. As such, passing this
flag causes them to render normally.
https://doc.rust-lang.org/nightly/book/ is an example page currently where this is done incorrectly.
Backport LLVM fixes for a JumpThreading / assume intrinsic bug
This fixes the original cause of #48116 and restores the assume intrinsic that was removed as a workaround.
r? @alexcrichton
Restore the download of rust-mingw
The build might otherwise break due to mixing MinGW object files from rust-std and the local MinGW which might be newer/older than the version used to build rust-std.
Fixes#48272
r? @alexcrichton
Improve --help performance for x.py
Since compiling the bootstrap command doesn't require any submodules,
we can skip updating submodules when a --help command is passed in.
On my machine, this saves 1 minute if the submodules are already
downloaded, and 10 minutes if run on a clean repo.
This commit also adds a message before compiling/downloading anything
when a --help command is passed in, to tell the user WHY --help
takes so long to complete. It also points the user to the bootstrap
README.md for faster help.
Finally, this fixes one warning message that still referenced using
make instead of x.py, even though x.py is now the standard way of
building rust.
Closes#37305
Disable NEON on musl ARMv7
`armv7-unknown-linux-musleabihf` target diverged a bit from `armv7-unknown-linux-gnueabihf` target. This PR re-syncs them.
Fix#47765.
