This is one more step towards completing #13231
This series of commits add support for default trait implementations. The changes in this PR don't break existing code and they are expected to preserve the existing behavior in the compiler as far as built-in bounds checks go.
The PR adds negative implementations of `Send`/`Sync` for some types and it removes the special cases for `Send`/`Sync` during the trait obligations checks. That is, it now fully relies on the traits check rather than lang items.
Once this patch lands and a new snapshot is created, it'll be possible to add default impls for `Send` and `Sync` and remove entirely the use of `BuiltinBound::{BoundSend,BoundSync}` for positive implementations as well.
This PR also removes the restriction on negative implementations. That is, it is now possible to add negative implementations for traits other than `Send`/`Sync`
We already do this for the function arguments, but miss it for the
retslot pointer, which can lead to LLVM assertions because the retslot
has the wrong type.
Fixes#22663
e. g.
```
let b: Box<Foo> = Box::from_raw(p);
```
instead of
```
let b: Box<Foo> = mem::transmute(p);
```
Patch also changes closure release code in `src/libstd/sys/unix/thread.rs`
when `pthread_create` failed. Raw pointer was transmuted to box of
`FnOnce()` instead of `Thunk`. This code was probably never executed,
because `pthread_create` rarely fails.
(And there are two more patches in PR: fix typo in doc and mark `from_raw` and `into_raw` functions inline.)
- Various grammatical changes.
- Use triple-backtick syntax and `sh` highlighting for code blocks.
- Fix indentation of code block in step 2 of \"Building on Windows\".
- Use title case for \"Getting Help\" subheading.
The Rust Reference should include the tuple indexing (using a number
as a field) notation; currently it is only available on
http://doc.rust-lang.org/std/primitive.tuple.html and not easily
searchable.
Fixed example threaded code in intro doc never printing results. Threads were created with Thread::spawn instead of Thread::scoped. Also added correct thread handling like in the first example of the document.
This is a patch for #17829.
In `compiletest` there are multiple layers which capture the output. The first layer is `run_tests_console` which is used to execute all tests.
Then there are some tests that contain unit tests, which by default also captures output. Therefore `compiletest` adds `RUST_TEST_NOCAPTURE` (and `RUST_TEST_TASKS` for completeness) to the run environment of the task.
Finally, the task used to execute a test redirects stdout and stdin. At the moment, the `VERBOSE=1` prints all captured output of the task (but has to print stdout and stderr separately). So at the moment using `RUST_TEST_NOCAPTURE=1` only makes sense when also using `VERBOSE=1` which seems a little bit cumbersome.
Should I update the patch to only print the output of the tasks that actually execute the test (`VERBOSE=1` includes other stuff, like the output of the task used to compile the test)? This will probably involve adding an extra flag to some functions in `src/compiletest/runtest.rs` to distinguish compilation runs from runs that execute the actual tests.
... to convert between Box and raw pointers. E. g. use
```
let b: Box<Foo> = Box::from_raw(p);
```
instead of
```
let b: Box<Foo> = mem::transmute(p);
```
Patch also changes closure release code in `src/libstd/sys/unix/thread.rs`
when `pthread_create` failed. Raw pointer was transmuted to box of
`FnOnce()` instead of `Thunk`. This code was probably never executed,
because `pthread_create` rarely fails in practice.
This is not a complete implementation of the RFC:
- only existing methods got updated, no new ones added
- doc comments are not extensive enough yet
- optimizations got lost and need to be reimplemented
See https://github.com/rust-lang/rfcs/pull/528
Technically a
[breaking-change]
