Add SGX target to rustc
This adds the `x86_64-fortanix-unknown-sgx` target specification to the Rust compiler. See #56066 for more details about this target.
Clarifying documentation for collections::hash_map::Entry::or_insert
Previous version does not show that or_insert does not insert the passed value, as the passed value was the same value as what was already in the map.
Return &T / &mut T in ManuallyDrop Deref(Mut) impl
Without this change the generated documentation looks like this:
fn deref(&self) -> &<ManuallyDrop<T> as Deref>::Target
Returning the actual type directly makes the generated docs more clear:
fn deref(&self) -> &T
Basically, compare how the impl for `Box<T>` and `ManuallyDrop<T>` looks in this screenshot:
This is a follow up to 8aa9267 which changed the driver to use lld
directly rather than invoking it through Clang. This change ensures
we pass all the necessary flags to lld.
Doc total order requirement of sort(_unstable)_by
I took the definition of what a total order is from the Ord trait
docs. I specifically put "elements of the slice" because if you
have a slice of f64s, but know none are NaN, then sorting by
partial ord is total in this case. I'm not sure if I should give
such an example in the docs or not.
r? @GuillaumeGomez
This test ensures that everything in `env::vars()` is inherited but
that's not actually true because other tests may add env vars after we
spawn the process, causing the test to be flaky! This commit moves the
test to a run-pass test where it can execute in isolation.
Along the way this removes a lot of the platform specificity of the
test, using iteslf to print the environment instead of a foreign process.
The BTreeSet and BTreeMap gdb pretty-printers did not take the node
structure into account, and consequently only worked for shallow sets.
This fixes the problem by iterating over child nodes when needed.
This patch avoids the current approach of implementing some of the
value manipulations in debugger-indepdendent code. This was done for
convenience: a type lookup was needed for the first time, and there
currently are no lldb formatters for these types.
Closes#55771
Without this, the `vis` does not wind up in the tree anywhere, and
then we get ICEs because the node-ids it refers to are not present.
The motivation seemed to be documentation, but `ListStem` HIR nodes
are ignored in rustdoc, from what I can tell.
Update books for Rust 2018
This PR:
1. updates all of the books
* I don't know if @Gankro has further plans for the nomicon or not
2. updates the build process because TRPL is only distributing one edition now
3. fixes up the stdlib links
I think that this passes but it's 3:20 am and so I'm sending it in and will fix up anything i missed in the morning.
/cc @alexcrichton for the big beta backport
resolve: Make "empty import canaries" invisible from other crates
Empty imports `use prefix::{};` are desugared into `use prefix::{self as _};` to make sure the prefix is checked for privacy/stability/etc.
This caused issues in cross-crate scenarios because gensyms are lost in crate metadata (the `_` is a gensym).
Fixes https://github.com/rust-lang/rust/issues/55811
rustc: remove {FxHash,Node,DefId,HirId,ItemLocal}{Map,Set} "constructor" fns.
These are cruft left over from a time when `Foo::default()` didn't "just work".
