| Age | Commit message (Collapse) | Author | Lines |
|---|
|
very minor cleanups
- add `must_use` to `early_error_no_abort`
this was already being used at its only callsite, but this ensures that new code remembers to use it if it's called in the future. found this while investigating https://github.com/rust-lang/rust/issues/110090.
- remove outdated and incorrect comment in `builder.rs`. `doc_rust_lang_org_channel` doesn't exist in rustdoc, it gets it from an env var instead: https://github.com/rust-lang/rust/blob/b275d2c30b6e88cc48747f349f7137076d450658/src/librustdoc/clean/utils.rs#L569-L573
|
|
... using server-reported timestamp.
This allows us to track changes to the downloaded artifact more easily
and in a more reproducible manner.
Co-authored-by: Zixing Liu <zixing.liu@canonical.com>
|
|
|
|
- add `must_use` to `early_error_no_abort`
this was already being used at its only callsite, but this ensures
that new code remembers to use it if it's called in the future.
- remove outdated and incorrect comment in `builder.rs`.
`doc_rust_lang_org_channel` doesn't exist in rustdoc, it gets it from
an env var instead.
|
|
Fix typo in bootstrap command description
Unless `duild` is some clever name for `dist build`, this is probably a typo.
|
|
Update serde in workspace and non-synced dependencies
The main workspace, bootstrap, cargo, miri, and rust-analyzer all lock serde to different versions. It's preferable to share the same version where possible, so update it.
Rustfmt is synced from another repository and has its own Cargo.lock, but since it's added to the overall workspace it should respect the version here.
Cargo is already at the latest version. Miri and rust-analyzer would require upstream updates.
|
|
|
|
|
|
|
|
|
|
|
|
Generate shell completions for bootstrap with Clap
Now that #110693 has been merged, we can look at generating shell completions for x.py with `clap_complete`. Leaving this as draft for now as I'm not sure of the best way to integration the completion generator. Additionally, the generated completions for zsh are completely broken (will need to be resolved upstream, it doesn't seem to handle subcommands + global arguments well).
I don't have Fish installed and would be interested to know how well completions work there.
Alternative to #107827
|
|
|
|
object -> 0.31.1
thorin-dwp -> 0.6.0
Required to fix watchOS breakage.
|
|
|
|
r=albertlarsan68
support set `rpath` option for each target independently
Currently the `rpath` option is a global config and it's effect on all targets.
But sometimes when developers edit the rustc code and try to release rust toolchains themselves, they may not want to add `rpath` in all targets to avoid dynamically linked shared object library privilege escalation attack.
This PR supports set `rpath` option for each target independently .
Common developers are not aware of the existence of this configuration option and do not affect the existing development process. This configuration option takes effect only after developers explicitly sets .
r? ``@albertlarsan68``
|
|
Start using `windows sys` for Windows FFI bindings in std
Switch to using windows-sys for FFI. In order to avoid some currently contentious issues, this uses windows-bindgen to generate a smaller set of bindings instead of using the full crate.
Unlike the windows-sys crate, the generated bindings uses `*mut c_void` for handle types instead of `isize`. This to sidestep opsem concerns about mixing pointer types and integers between languages. Note that `SOCKET` remains defined as an integer but instead of being a usize, it's changed to fit the [standard library definition](https://github.com/rust-lang/rust/blob/a41fc00eaf352541008965fec0dee811e44373b3/library/std/src/os/windows/raw.rs#L12-L16):
```rust
#[cfg(target_pointer_width = "32")]
pub type SOCKET = u32;
#[cfg(target_pointer_width = "64")]
pub type SOCKET = u64;
```
The generated bindings also customizes the `#[link]` imports. I hope to switch to using raw-dylib but I don't want to tie that too closely with the switch to windows-sys.
---
Changes outside of the bindings are, for the most part, fairly minimal (e.g. some differences in `*mut` vs. `*const` or a few types differ). One issue is that our own bindings sometimes mix in higher level types, like `BorrowedHandle`. This is pretty adhoc though.
|
|
Rollup of 7 pull requests
Successful merges:
- #110297 (Make `(try_)subst_and_normalize_erasing_regions` take `EarlyBinder`)
- #110827 (Fix lifetime suggestion for type aliases with objects in them)
- #111022 (Use smaller ints for bitflags)
- #111056 (Fix some suggestions where a `Box<T>` is expected.)
- #111262 (Further normalize msvc-non-utf8-ouput)
- #111265 (Make generics_of has_self on RPITITs delegate to the opaque)
- #111323 (Give a more helpful error when running the rustc shim directly)
Failed merges:
r? `@ghost`
`@rustbot` modify labels: rollup
|
|
r=Mark-Simulacrum
enable `rust_2018_idioms` lint group for doctests
With this change, `rust_2018_idioms` lint group will be enabled for compiler/libstd doctests.
Resolves #106086
Resolves #99144
Signed-off-by: ozkanonur <work@onurozkan.dev>
|
|
|
|
The main workspace, bootstrap, cargo, miri, and rust-analyzer all lock
serde to different versions. It's preferable to share the same version
where possible, so update it.
Rustfmt is synced from another repository and has its own Cargo.lock,
but since it's added to the overall workspace it should respect the
version here.
Cargo is already at the latest version. Miri and rust-analyzer would
require upstream updates.
|
|
|
|
cc https://rust-lang.zulipchat.com/#narrow/stream/326414-t-infra.2Fbootstrap/topic/Building.20.60coretests.60.20by.20hand
|
|
Migrate bootstrap to Clap-based argument parsing
Supercedes #108083
I chose to re-do the work rather than rebase the onto the large changes since the original PR. If it's preferred I can instead force-push the original PR to this version.
cc `@jyn514` `@albertlarsan68`
|
|
|
|
Signed-off-by: ozkanonur <work@onurozkan.dev>
|
|
|
|
|
|
Use `free-args` consistently in bootstrap
Previously, this was only passed to miri and compiletest. Extended it to all other tests and binaries as well.
cc https://rust-lang.zulipchat.com/#narrow/stream/326414-t-infra.2Fbootstrap/topic/Running.20a.20single.20doctest
r? `@clubby789`
|
|
bootstrap: add llvm-project/runtimes to the sources
This is needed to build libunwind for LLVM 16:
https://discourse.llvm.org/t/runtimes-removed-support-for-llvm-enable-projects-in-libcxx-libcxxabi-and-libunwind/65707
Zulip thread: https://rust-lang.zulipchat.com/#narrow/stream/187780-t-compiler.2Fwg-llvm/topic/missing.20llvm-project.2Fruntimes.20in.20beta.20srcs
Checked by running `x.py dist` and verifying the produced source archive (build/dist/rustc-1.71.0-dev-src.tar.gz) contains the new directory.
|
|
Previously, this was only passed to miri and compiletest. Extended it to
all other tests and binaries as well.
|
|
bootstrap: add .gitmodules to the sources
The bootstrap builder now expects this file to exist: https://github.com/rust-lang/rust/blob/6f8c0557e0b73c73a8a7163a15f4a5a3feca7d5c/src/bootstrap/builder.rs#L494
Zulip thread: https://rust-lang.zulipchat.com/#narrow/stream/187780-t-compiler.2Fwg-llvm/topic/missing.20llvm-project.2Fruntimes.20in.20beta.20srcs
|
|
Add a `sysroot` crate to represent the standard library crates
This adds a dummy crate named `sysroot` to represent the standard library target instead of using the `test` crate. This allows the removal of `proc_macro` as a dependency of `test` allowing these 2 crates to build in parallel saving around 9 seconds locally.
|
|
The bootstrap builder now expects this file to exist:
https://github.com/rust-lang/rust/blob/6f8c0557e0b73c73a8a7163a15f4a5a3feca7d5c/src/bootstrap/builder.rs#L494
|
|
This is needed to build libunwind for LLVM 16:
https://discourse.llvm.org/t/runtimes-removed-support-for-llvm-enable-projects-in-libcxx-libcxxabi-and-libunwind/65707
Doesn't work: with this, running `python3 x.py dist` produces a tar
archive that still doesn't contain the runtimes subdirectory?
|
|
remove pointless `FIXME` in `bootstrap::download`
The suggestion given by `FIXME` to use `CompilerMetadata` for `download_toolchain` in `bootstrap::download` can result in more confusion. This is because `stamp_key` is not always a date; it can also be a commit hash. Additionally, unlike in `download_beta_toolchain`, in the `download_ci_rustc` function, `version` and `commit` values are calculated separately.
|
|
Wrap all of x.py in `if __name__ == '__main__':` to avoid problems with `multiprocessing`
Make the pool sizing better
|
|
The suggestion given by `FIXME` to use `CompilerMetadata` for
`download_toolchain` in `bootstrap::download` can result in more
confusion. This is because `stamp_key` is not always a date; it
can also be a commit hash. Additionally, unlike in `download_beta_toolchain`,
in the `download_ci_rustc` function, `version` and `commit` values
are calculated separately.
Signed-off-by: ozkanonur <work@onurozkan.dev>
|
|
Parallelize initial Rust download in bootstrap
Parallelize the initial download of Rust in `bootstrap.py`
`time ./x.py --help` after `rm -r build`
Before: 33s
After: 27s
|
|
This is quite slow and embarassingly parallel, even in python.
This speeds up the initial bootstrap build by about 5-10s.
|
|
Output some bootstrap messages on stderr
Fixes #110995
|
|
bootstrap: Fix compile error: unused-mut
Compile errors:
```
Compiling bootstrap v0.0.0 (/home/hev/rust/rust/src/bootstrap)
error: variable does not need to be mutable
--> config.rs:1312:17
|
1312 | let mut build_target = config
| ----^^^^^^^^^^^^
| |
| help: remove this `mut`
|
= note: `-D unused-mut` implied by `-D warnings`
error: could not compile `bootstrap` (lib) due to previous error
```
|
|
download-rustc: Give a better error message if artifacts can't be dowloaded
It should be very rare in practice to happen; people would need to both have `download-ci-llvm` disabled and `download-rustc` enabled. I think it may be more common if we start turning this on by default, though.
Helps with https://github.com/rust-lang/rust/issues/81930.
Before:
```
downloading https://ci-artifacts.rust-lang.org/rustc-builds/bf5cad8e775fb326465e5c1b98693e5d259da156/rust-std-nightly-x86_64-unknown-linux-gnu.tar.xz
curl: (22) The requested URL returned error: 404
```
After:
```
downloading https://ci-artifacts.rust-lang.org/rustc-builds/bf5cad8e775fb326465e5c1b98693e5d259da156/rust-std-nightly-x86_64-unknown-linux-gnu.tar.xz
curl: (22) The requested URL returned error: 404
error: failed to download pre-built rustc from CI
note: old builds get deleted after a certain time
help: if trying to compile an old commit of rustc, disable `download-rustc` in config.toml:
[rust]
download-rustc = false
```
|
|
|
|
bootstrap: Unify test argument handling
Fixes #104198. Does *not* help with https://github.com/rust-lang/rust/issues/80124 because I couldn't figure out a reasonable way to omit `--lib` only for `panic_abort` and not other `std` dependencies.
- Remove unnecessary `test_kind` field and `TestKind` struct. These are just subsets of the existing `builder.kind` / `Kind` struct.
- Add a new `run_cargo_test` function which handles passing arguments to cargo based on `builder.config`
- Switch all Steps in `mod test` to `run_cargo_test` where possible
- Combine several steps into one `CrateBootstrap` step. These tests all do the same thing, just with different crate names.
- Fix `x test --no-doc`. This is much simpler after the refactors mentioned earlier, but I'm happy to split it into a separate PR if desired. Before, this would panic a lot because steps forgot to pass `--lib`.
|
|
|
|
- Use `cargo metadata` to determine whether a crate has a library
package or not
- Collect metadata for all workspaces, not just the root workspace and
cargo
- Don't pass `--lib` for crates without a library
- Use `run_cargo_test` for rust-installer
- Don't build documentation in `lint-docs` if `--no-doc` is passed
|
|
|
|
|
|
This fixes the following recurring error on windows:
```
Traceback (most recent call last):
File "C:\Users\jyn\src\rust\x.py", line 29, in <module>
bootstrap.main()
File "C:\Users\jyn\src\rust\src\bootstrap\bootstrap.py", line 963, in main
bootstrap(args)
File "C:\Users\jyn\src\rust\src\bootstrap\bootstrap.py", line 927, in bootstrap
build.download_toolchain()
File "C:\Users\jyn\src\rust\src\bootstrap\bootstrap.py", line 437, in download_toolchain
shutil.rmtree(bin_root)
File "C:\Users\jyn\AppData\Local\Programs\Python\Python311\Lib\shutil.py", line 759, in rmtree
return _rmtree_unsafe(path, onerror)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "C:\Users\jyn\AppData\Local\Programs\Python\Python311\Lib\shutil.py", line 617, in _rmtree_unsafe
_rmtree_unsafe(fullname, onerror)
File "C:\Users\jyn\AppData\Local\Programs\Python\Python311\Lib\shutil.py", line 622, in _rmtree_unsafe
onerror(os.unlink, fullname, sys.exc_info())
File "C:\Users\jyn\AppData\Local\Programs\Python\Python311\Lib\shutil.py", line 620, in _rmtree_unsafe
os.unlink(fullname)
PermissionError: [WinError 5] Access is denied: 'C:\\Users\\jyn\\src\\rust\\build\\x86_64-pc-windows-msvc\\stage0\\bin\\rust-analyzer-proc-macro-srv.exe'
```
|
