The code inside this conditional will not work on FreeBSD 10+ because
those versions use clang and libc++ rather than libstdc++.
Since FreeBSD comes with libc++ in the base, presumably all 10+ systems
will have it present.
Searching for libstdc++.a will not work if it is not present. As a
result, this would previously have set `LLVM_STATIC_STDCPP=libstdc++.a`,
which isn't a valid path and caused problems later on when building
`librustc_llvm`.
This could possibly be updated in the future to look for `libc++.a` on
FreeBSD, by expanding the code inside the conditional. In one attempt
to run this on x86_64-freebsd, I found that libc++ was not compiled with
PIC, so it failed anyway.
The main goal here is to use FreeBSD's normal libc++, instead of
statically linking the libstdc++ packaged with GCC, because that
libstdc++ has bugs that cause rustc to deadlock inside LLVM.
But the easiest way to use libc++ is to switch the build from GCC to
Clang, and the Clang package in the Ubuntu image already knows how to
cross-compile (given a sysroot and preferably cross-binutils), so the
toolchain script now uses that instead of building a custom compiler.
This also de-duplicates the `build-toolchain.sh` script.
Move collector to librustc_mir::monomorphize
cc https://github.com/rust-lang/rust/issues/44334 and https://github.com/rust-lang/rust/issues/45276
* I moved the collector to rustc_mir
* I renamed `TransItem` to `MonoItem`. _(I still need to fix up comments and variable names)_
* I got rid of `common.rs` and `monomorphize.rs` from `librustc_trans_utils`. I moved most of the functionality into `TyCtxt`. I realized that the `librustc_trans_utils::common.rs` was just copy pasted from `librustc_trans::common.rs`.
Should I also get rid of the `librustc_trans::common.rs` in this PR? Most of the functionality seems a bit useless, I decided to put some of it into `TyCtxt` but maybe that is not the correct action here.
Should I also get rid of `librustc_trans_utils` completely here? Or should I do it in a separate PR?
rustc: ensure optimized enums have a properly aligned size.
Fixes#46769 by padding the optimized enums wrapping packed data as necessary.
Note that this is not the only way to solve this - on nightly, #46436 makes it easier to fix without adding new padding because of the replacement of `packed` flags with a non-redundant scheme.
But because it can't be backported, the optimal fix will be in a separate nightly-only PR (#46809).
Add lossless debug implementation for unix OsStrs
Fixes#22766
Invalid utf8 byte sequences are replaced with `\xFF` style escape codes, while valid utf8 goes through the normal `Debug` implementation.
This is necessarily different from the windows Debug implementation, which uses `\u{xxxx}` style escape sequences for unpaired surrogates, but both implementations are consistent in that they are both lossless, and display invalid sequences in the way most similar to existing language syntax.
r? @dtolnay
Add sync module to rustc_data_structures
This PR is split out from https://github.com/rust-lang/rust/pull/45912, since github apparently can't handle such large PRs.
r? @arielb1
Improve error messages on LLVM bitcode parsing failure
The LLVM error causing the parse failure is now printed, in the style
of the other thin LTO error messages. This prevents a flood of
assertion failure messages if the bitcode can’t be parsed.
