Add cg_clif as optional codegen backend
Rustc_codegen_cranelift is an alternative codegen backend for rustc based on Cranelift. It has the potential to improve compilation times in debug mode. In my experience the compile time improvements over debug mode LLVM for a clean build are about 20-30% in most cases.
This PR adds cg_clif as optional codegen backend. By default it is only enabled for `./x.py check`. It can be enabled for `./x.py build` too by adding `cranelift` to the `rust.codegen-backends` array in `config.toml`.
MCP: https://github.com/rust-lang/compiler-team/issues/270
r? `@Mark-Simulacrum`
This commit changes the way build-manifest is invoked, to let it accept
the Rust version directly instead of requiring the path of the Rust
monorepo and letting build-manifest figure out the path on its own.
This allows to run build-manifest without a clone of the monorepo.
The format of the tarballs produced by CI is roughly the following:
{component}-{release}-{target}.{ext}
While on the beta and nightly channels `{release}` is just the channel
name, on the stable channel is either the Rust version or the version of
the component we're shipping:
cargo-0.47.0-x86_64-unknown-linux-gnu.tar.xz
clippy-0.0.212-x86_64-unknown-linux-gnu.tar.xz
llvm-tools-1.46.0-x86_64-unknown-linux-gnu.tar.xz
miri-0.1.0-x86_64-unknown-linux-gnu.tar.xz
rls-1.41.0-x86_64-unknown-linux-gnu.tar.xz
rust-1.46.0-x86_64-unknown-linux-gnu.tar.xz
...
This makes it really hard to get the package URL without having access
to the manifest (and there is no manifest on ci-artifacts.rlo), as there
is no consistent version number to use.
This commit addresses the problem by always using the Rust version
number as `{release}` for the stable channel, regardless of the version
number of the component we're shipping. I chose that instead of "stable"
to avoid breaking the URL scheme *that* much.
Rustup should not be affected by this change, as it fetches the URLs
from the manifest. Unfortunately we don't have a way to test other
clients before making a stable release, as this change only affects the
stable channel.
Refactor versions detection in build-manifest
This PR refactors how `build-manifest` handles versions, making the following changes:
* `build-manifest` now detects the "package releases" on its own, without relying on rustbuild providing them through CLI arguments. This drastically simplifies calling the tool outside of `x.py`, and will allow to ship the prebuilt tool in a tarball in the future, with the goal of stopping to invoke `x.py` during `promote-release`.
* The `tar` command is not used to extract the version and the git hash from tarballs anymore. The `flate2` and `tar` crates are used instead. This makes detecting those pieces of data way faster, as the archive is decompressed just once and we stop parsing the archive once all the information is retrieved.
* The code to extract the version and the git hash now stores all the collected data dynamically, without requiring to add new fields to the `Builder` struct every time.
I tested the changes locally and it should behave the same as before.
r? `@Mark-Simulacrum`
Don't dynamically link LLVM tools unless rustc is too
This PR initially tried to support link-shared on all of our target platforms (other than Windows), but ran into a number of difficulties:
* LLVM doesn't really support a shared link on macOS (llvm-config runs into problems with the version suffix)
* LLVM doesn't seem to support a shared link when cross-compiling (the libLLVM.so ends up empty and symbols are not found)
So, this PR has now been revised such that we don't attempt to dynamically link LLVM tools (even if that would, otherwise, be supported) on targets where LLVM is statically linked to rustc. Currently that's basically everything except for x86_64-unknown-linux-gnu (where we dynamically link to avoid rerunning ThinLTO in each stage).
Follow-up to #76708.
Fixes#76698.
Move the version number to a plaintext file
The Rust version number is currently embedded in bootstrap's source code, which makes it hard to update it automatically or access it outside of ./x.py (as you'd have to parse the source code).
This PR moves the version number to a standalone plaintext file, which makes accessing or updating it trivial.
r? @Mark-Simulacrum
Previously we would have some platforms where LLVM was linked to rustc
statically, but to the LLVM tools dynamically. That meant we were distributing
two copies of LLVM: one as a separate dylib and one statically linked in to
librustc_driver.
The Rust version number is currently embedded in bootstrap's source
code, which makes it hard to update it automatically or access it
outside of ./x.py (as you'd have to parse the source code).
This commit moves the version number to a standalone plaintext file,
which makes accessing or updating it trivial.
rustbuild: avoid trying to inversely cross-compile for build triple from host triples
This changes rustbuild's cross compilation logic to better match what users expect,
particularly, avoiding trying to inverse cross-compile for the build triple from host triples.
That is, if build=A, host=B, target=B, we do not want to try and compile for A from B.
Indeed, the only "known to run" triple when cross-compiling is the build triple A.
When testing for a particular target we need to be able to run binaries compiled for
that target though.
The last commit also modifies the default set of host/target triples to avoid producing
needless artifacts for the build triple:
The new behavior is to respect --host and --target when passed as the *only*
configured triples (no triples are implicitly added). The default for --host is
the build triple, and the default for --target is the host triple(s), either
configured or the default build triple.
Fixes#76333
r? `@alexcrichton` if possible, otherwise we'll need to hunt down a reviewer
rustc is a natively cross-compiling compiler, and generally none of our steps
should care whether they are using a compiler built of triple A or B, just the
--target directive being passed to the running compiler. e.g., when building for
some target C, you don't generally want to build two stds: one with a host A
compiler and the other with a host B compiler. Just one std is sufficient.
This currently includes libLLVM, llvm-config, and FileCheck, but will perhaps
expand to more tooling overtime. It should be considered entirely unstable and
may change at any time.
Fix windows-gnu host cross-compilation
Fixes https://github.com/rust-lang/rust/issues/64218
Also turns out it's faster to run Linux virtual machine on Windows and cross-compile `./x.py dist` than doing it on Windows directly...
This commit is a proof-of-concept for switching the standard library's
backtrace symbolication mechanism on most platforms from libbacktrace to
gimli. The standard library's support for `RUST_BACKTRACE=1` requires
in-process parsing of object files and DWARF debug information to
interpret it and print the filename/line number of stack frames as part
of a backtrace.
Historically this support in the standard library has come from a
library called "libbacktrace". The libbacktrace library seems to have
been extracted from gcc at some point and is written in C. We've had a
lot of issues with libbacktrace over time, unfortunately, though. The
library does not appear to be actively maintained since we've had
patches sit for months-to-years without comments. We have discovered a
good number of soundness issues with the library itself, both when
parsing valid DWARF as well as invalid DWARF. This is enough of an issue
that the libs team has previously decided that we cannot feed untrusted
inputs to libbacktrace. This also doesn't take into account the
portability of libbacktrace which has been difficult to manage and
maintain over time. While possible there are lots of exceptions and it's
the main C dependency of the standard library right now.
For years it's been the desire to switch over to a Rust-based solution
for symbolicating backtraces. It's been assumed that we'll be using the
Gimli family of crates for this purpose, which are targeted at safely
and efficiently parsing DWARF debug information. I've been working
recently to shore up the Gimli support in the `backtrace` crate. As of a
few weeks ago the `backtrace` crate, by default, uses Gimli when loaded
from crates.io. This transition has gone well enough that I figured it
was time to start talking seriously about this change to the standard
library.
This commit is a preview of what's probably the best way to integrate
the `backtrace` crate into the standard library with the Gimli feature
turned on. While today it's used as a crates.io dependency, this commit
switches the `backtrace` crate to a submodule of this repository which
will need to be updated manually. This is not done lightly, but is
thought to be the best solution. The primary reason for this is that the
`backtrace` crate needs to do some pretty nontrivial filesystem
interactions to locate debug information. Working without `std::fs` is
not an option, and while it might be possible to do some sort of
trait-based solution when prototyped it was found to be too unergonomic.
Using a submodule allows the `backtrace` crate to build as a submodule
of the `std` crate itself, enabling it to use `std::fs` and such.
Otherwise this adds new dependencies to the standard library. This step
requires extra attention because this means that these crates are now
going to be included with all Rust programs by default. It's important
to note, however, that we're already shipping libbacktrace with all Rust
programs by default and it has a bunch of C code implementing all of
this internally anyway, so we're basically already switching
already-shipping functionality to Rust from C.
* `object` - this crate is used to parse object file headers and
contents. Very low-level support is used from this crate and almost
all of it is disabled. Largely we're just using struct definitions as
well as convenience methods internally to read bytes and such.
* `addr2line` - this is the main meat of the implementation for
symbolication. This crate depends on `gimli` for DWARF parsing and
then provides interfaces needed by the `backtrace` crate to turn an
address into a filename / line number. This crate is actually pretty
small (fits in a single file almost!) and mirrors most of what
`dwarf.c` does for libbacktrace.
* `miniz_oxide` - the libbacktrace crate transparently handles
compressed debug information which is compressed with zlib. This crate
is used to decompress compressed debug sections.
* `gimli` - not actually used directly, but a dependency of `addr2line`.
* `adler32`- not used directly either, but a dependency of
`miniz_oxide`.
The goal of this change is to improve the safety of backtrace
symbolication in the standard library, especially in the face of
possibly malformed DWARF debug information. Even to this day we're still
seeing segfaults in libbacktrace which could possibly become security
vulnerabilities. This change should almost entirely eliminate this
possibility whilc also paving the way forward to adding more features
like split debug information.
Some references for those interested are:
* Original addition of libbacktrace - #12602
* OOM with libbacktrace - #24231
* Backtrace failure due to use of uninitialized value - #28447
* Possibility to feed untrusted data to libbacktrace - #21889
* Soundness fix for libbacktrace - #33729
* Crash in libbacktrace - #39468
* Support for macOS, never merged - ianlancetaylor/libbacktrace#2
* Performance issues with libbacktrace - #29293, #37477
* Update procedure is quite complicated due to how many patches we
need to carry - #50955
* Libbacktrace doesn't work on MinGW with dynamic libs - #71060
* Segfault in libbacktrace on macOS - #71397
Switching to Rust will not make us immune to all of these issues. The
crashes are expected to go away, but correctness and performance may
still have bugs arise. The gimli and `backtrace` crates, however, are
actively maintained unlike libbacktrace, so this should enable us to at
least efficiently apply fixes as situations come up.
With this, we can now use a target JSON file to build a
cross-compiler:
```
x.py install --host ../aarch64-apple-darwin.json --target aarch64-apple-darwin
```
`rustc` allows passing in predefined target triples as well as JSON
target specification files. This change allows bootstrap to have the
first inkling about those differences. This allows building a
cross-compiler for an out-of-tree architecture (even though that
compiler won't work for other reasons).
Even if no one ever uses this functionality, I think the newtype
around the `Interned<String>` improves the readability of the code.
The current plan is that submodule tracks the `release` branch of
rust-analyzer, which is updated once a week.
rust-analyzer is a workspace (with a virtual manifest), the actual
binary is provide by `crates/rust-analyzer` package.
Note that we intentionally don't add rust-analyzer to `Kind::Test`,
for two reasons.
*First*, at the moment rust-analyzer's test suite does a couple of
things which might not work in the context of rust repository. For
example, it shells out directly to `rustup` and `rustfmt`. So, making
this work requires non-trivial efforts.
*Second*, it seems unlikely that running tests in rust-lang/rust repo
would provide any additional guarantees. rust-analyzer builds with
stable and does not depend on the specifics of the compiler, so
changes to compiler can't break ra, unless they break stability
guarantee. Additionally, rust-analyzer itself is gated on bors, so we
are pretty confident that test suite passes.
Remove legacy InnoSetup GUI installer
On Windows the InnoSetup `.exe` installer was superseded by the MSI installer long ago. It's no longer needed.
The `.exe` installer hasn't been linked from the [other installation methods](https://forge.rust-lang.org/infra/other-installation-methods.html#standalone) page in many years. As far as I can tell the intent was always to remove this installer once the MSI proved itself. Though admittedly both installers feel very "legacy" at this point.
Removing this would mean we only maintain one Windows GUI installer and would speed up the distribution phase.
As a result of removing InnoSetup, this closes#24397
Create self-contained directory and move there some of external binaries/libs
One of the steps to reach design described in https://github.com/rust-lang/rust/issues/68887#issuecomment-633048380
This PR moves things around and allows link code to handle the new directory structure.
Replace old GDB and LLDB pretty-printers with new ones
which were originally written for IntelliJ Rust.
New LLDB pretty-printers support synthetic children.
New GDB/LLDB pretty-printers support all Rust types
supported by old pretty-printers, and also support:
Rc, Arc, Cell, Ref, RefCell, RefMut, HashMap, HashSet.
For running the compiler, we usually only need LLVM from `$sysroot/lib`,
which rustup will make available with `LD_LIBRARY_PATH`. We've also been
shipping LLVM in the `$target/lib` directory, which bloats the download
and installed size. The only times we do need the latter are for the
RPATH of `llvm-tools-preview` binaries, and for linking `rustc-dev`
libraries. We'll move it to the `llvm-tools-preview` component directly,
and `rustc-dev` will have an implicit dependency on it.
Here are the dist sizes that I got before and after this change:
llvm-tools-1.45.0-dev-x86_64-unknown-linux-gnu.tar.gz 1.3M 24M
llvm-tools-1.45.0-dev-x86_64-unknown-linux-gnu.tar.xz 748K 17M
rustc-1.45.0-dev-x86_64-unknown-linux-gnu.tar.gz 83M 61M
rustc-1.45.0-dev-x86_64-unknown-linux-gnu.tar.xz 56M 41M
The installed size should reduce by exactly one `libLLVM.so` (~70-80M),
unless you also install `llvm-tools`, and then it should be identical.
