trans: always use a memcpy for ABI argument/return casts.
When storing incoming arguments or values returned by call/invoke, always do a `memcpy` from a temporary of the cast type, if there is an ABI cast.
While Clang has gotten smarter ([store](https://godbolt.org/g/EphFuK) vs [memcpy](https://godbolt.org/g/5dikH9)), a `memcpy` will always work.
This is what @dotdash has wanted to do all along, and it fixes#32049.
Have included an example of a Cargo.toml file containing the libc dependency.
The file has been reformatted to use sections. The note on the default features warning is part of the _Using libc_ section but is in bold.
Remove old -Z options that do nothing
Technically, this is a [breaking-change], but I'm not sure what the policy for -Z flags is (especially unused ones).
Update rc.rs
The original description suggests that the original `Rc<T>` itself is downgraded, which doesn't seem to be what the code does. At the same time, `Rc` is one of those types that can do weird things with only a shared reference, so I thought it would be good to be clear.
Minor changes to variable bindings chapter
* In "*... name as another binding, that's currently in scope, will ...*", *"
that's currently in scope"* is not a parenthetical element, and the commas
can be omitted.
* Other minor changes.
add documentation on howto build just rustc without libstd to the build system
I searched for days until I found this, maybe this helps other poor souls 😂
for reference #33990
[MIR] Fix MIR trans edge cases that showed up on crater.
These fixes cover all of the [regressions found by crater](https://gist.github.com/nikomatsakis/88ce89ed06ef7f7f19bfd1e221d7f7ec) (for #34096).
Two of them were `Pair` edge cases (ZSTs and constants) causing LLVM assertions, the other one was causing stack overflows in debug scripts compiled in debug mode, due to the `fn_ret_cast` `alloca` ending up in a loop.
(Includes a regression test, which needed to go into `compile-fail/`
due to weaknesses when combining `#[deny(warnings)]` with
`tcx.sess.span_warn(..)`)
(updated with review feedback from arielb1.)
rustc: add ReErased to be used by trait selection, MIR and trans.
`ReErased` replaces `ReStatic` (i.e. `'static`) for erasing regions.
Using a distinct lifetime helps prevent accidental mix-ups between the two.
It also allows cleaner type printing (see test changes), including in symbol names:
```rust
str..pattern..CharSearcher$LT$$u27$static$GT$::drop.30560::h840c2f2afc03bbea // before
str..pattern..CharSearcher::drop.30561::h6bd31d2af614377a // after
```
Not that we should be producing symbols this way, but it's still better.
* In "... name as another binding, that's currently in scope, will ...", ",
that's currently in scope, " is not a parenthetical element, and the commas
can be omitted.
* Other minor changes.
The book's "Error handling with Box<Error>" section talks about Box<Error>.
In the actual example Box<Error + Send + Sync> is used instead so that the
corresponding From impls could be used to convert a plain string to an error
type. Rust 1.7 added support for conversion from &str/String to
Box<Error>, so this gotcha and later references to it can now be removed.
[MIR] Fix double-rounding of float constants and ignore NaN sign in tests.
Fixes#32805 by handling f32 and f64 separately in rustc_const_eval.
Also removes `#[rustc_no_mir]` from a couple libstd tests by ignoring NaN sign.
Turns out that runtime evaluation of `0.0 / 0.0` produces a NaN with the sign bit set,
whereas LLVM constant folds it to a NaN with the sign bit unset, which we were testing for.
Add error description for E0174
Reference for issue: #32777
r? @GuillaumeGomez
Hey Guillaume, sorry for taking too long to do it. I got some unexpected work during the week.
Waiting for your review :)
Unsupport wget
wget support was removed in #32942 (search for wget in diff), but configure wasn't updated. wget support was introduced in #7498 for Windows, but we now use PowerShell on Windows.
