Main motivation was to update docs for the removal or "demotion" of certain extension traits. The update to the slice docs was larger, since the text was largely outdated.
Reject specialized Drop impls.
See Issue #8142 for discussion.
This makes it illegal for a Drop impl to be more specialized than the original item.
So for example, all of the following are now rejected (when they would have been blindly accepted before):
```rust
struct S<A> { ... };
impl Drop for S<i8> { ... } // error: specialized to concrete type
struct T<'a> { ... };
impl Drop for T<'static> { ... } // error: specialized to concrete region
struct U<A> { ... };
impl<A:Clone> Drop for U<A> { ... } // error: added extra type requirement
struct V<'a,'b>;
impl<'a,'b:a> Drop for V<'a,'b> { ... } // error: added extra region requirement
```
Due to examples like the above, this is a [breaking-change].
(The fix is to either remove the specialization from the `Drop` impl, or to transcribe the requirements into the struct/enum definition; examples of both are shown in the PR's fixed to `libstd`.)
----
This is likely to be the last thing blocking the removal of the `#[unsafe_destructor]` attribute.
Fix#8142Fix#23584
This commit provides a safe, but unstable interface for the `try` functionality
of running a closure and determining whether it panicked or not.
There are two primary reasons that this function was previously marked `unsafe`:
1. A vanilla version of this function exposes the problem of exception safety by
allowing a bare try/catch in the language. It is not clear whether this
concern should be directly tied to `unsafe` in Rust at the API level. At this
time, however, the bounds on `ffi::try` require the closure to be both
`'static` and `Send` (mirroring those of `thread::spawn`). It may be possible
to relax the bounds in the future, but for now it's the level of safety that
we're willing to commit to.
2. Panicking while panicking will leak resources by not running destructors.
Because panicking is still controlled by the standard library, safeguards
remain in place to prevent this from happening.
The new API is now called `catch_panic` and is marked as `#[unstable]` for now.
This commit alters the behavior of the `Read::read_to_end()` method to zero all
memory instead of passing an uninitialized buffer to `read`. This change is
motivated by the [discussion on the internals forum][discuss] where the
conclusion has been that the standard library will not expose uninitialized
memory.
[discuss]: http://internals.rust-lang.org/t/uninitialized-memory/1652Closes#20314
There have been some recent panics on the bots and this commit is an attempt to
appease them. Previously it was considered invalid to run `rt::at_exit` after
the handlers had already started running. Due to the multithreaded nature of
applications, however, it is not always possible to guarantee this. For example
[this program][ex] will show off the abort.
[ex]: https://gist.github.com/alexcrichton/56300b87af6fa554e52d
The semantics of the `rt::at_exit` function have been modified as such:
* It is now legal to call `rt::at_exit` at any time. The return value now
indicates whether the closure was successfully registered or not. Callers must
now decide what to do with this information.
* The `rt::at_exit` handlers will now be run for a fixed number of iterations.
Common cases (such as the example shown) may end up registering a new handler
while others are running perhaps once or twice, so this common condition is
covered by re-running the handlers a fixed number of times, after which new
registrations are forbidden.
Some usage of `rt::at_exit` was updated to handle these new semantics, but
deprecated or unstable libraries calling `rt::at_exit` were not updated.
See Issue 8142 for discussion.
This makes it illegal for a Drop impl to be more specialized than the
original item.
So for example, all of the following are now rejected (when they would
have been blindly accepted before):
```rust
struct S<A> { ... };
impl Drop for S<i8> { ... } // error: specialized to concrete type
struct T<'a> { ... };
impl Drop for T<'static> { ... } // error: specialized to concrete region
struct U<A> { ... };
impl<A:Clone> Drop for U<A> { ... } // error: added extra type requirement
struct V<'a,'b>;
impl<'a,'b:a> Drop for V<'a,'b> { ... } // error: added extra region requirement
```
Due to examples like the above, this is a [breaking-change].
(The fix is to either remove the specialization from the `Drop` impl,
or to transcribe the requirements into the struct/enum definition;
examples of both are shown in the PR's fixed to `libstd`.)
----
This is likely to be the last thing blocking the removal of the
`#[unsafe_destructor]` attribute.
Includes two new error codes for the new dropck check.
Update run-pass tests to accommodate new dropck pass.
Update tests and docs to reflect new destructor restriction.
----
Implementation notes:
We identify Drop impl specialization by not being as parametric as the
struct/enum definition via unification.
More specifically:
1. Attempt unification of a skolemized instance of the struct/enum
with an instance of the Drop impl's type expression where all of
the impl's generics (i.e. the free variables of the type
expression) have been replaced with unification variables.
2. If unification fails, then reject Drop impl as specialized.
3. If unification succeeds, check if any of the skolemized
variables "leaked" into the constraint set for the inference
context; if so, then reject Drop impl as specialized.
4. Otherwise, unification succeeded without leaking skolemized
variables: accept the Drop impl.
We identify whether a Drop impl is injecting new predicates by simply
looking whether the predicate, after an appropriate substitution,
appears on the struct/enum definition.
This permits all coercions to be performed in casts, but adds lints to warn in those cases.
Part of this patch moves cast checking to a later stage of type checking. We acquire obligations to check casts as part of type checking where we previously checked them. Once we have type checked a function or module, then we check any cast obligations which have been acquired. That means we have more type information available to check casts (this was crucial to making coercions work properly in place of some casts), but it means that casts cannot feed input into type inference.
[breaking change]
* Adds two new lints for trivial casts and trivial numeric casts, these are warn by default, but can cause errors if you build with warnings as errors. Previously, trivial numeric casts and casts to trait objects were allowed.
* The unused casts lint has gone.
* Interactions between casting and type inference have changed in subtle ways. Two ways this might manifest are:
- You may need to 'direct' casts more with extra type information, for example, in some cases where `foo as _ as T` succeeded, you may now need to specify the type for `_`
- Casts do not influence inference of integer types. E.g., the following used to type check:
```
let x = 42;
let y = &x as *const u32;
```
Because the cast would inform inference that `x` must have type `u32`. This no longer applies and the compiler will fallback to `i32` for `x` and thus there will be a type error in the cast. The solution is to add more type information:
```
let x: u32 = 42;
let y = &x as *const u32;
```
This commit alters the behavior of the `Read::read_to_end()` method to zero all
memory instead of passing an uninitialized buffer to `read`. This change is
motivated by the [discussion on the internals forum][discuss] where the
conclusion has been that the standard library will not expose uninitialized
memory.
[discuss]: http://internals.rust-lang.org/t/uninitialized-memory/1652Closes#20314
The method with which backwards compatibility was retained ended up leading to
documentation that rustdoc didn't handle well and largely ended up confusing.
Previously, impls for `[T; n]` were collected in the same place as impls for `[T]` and `&[T]`. This splits them out into their own primitive page in both core and std.
Linking `__pthread_get_minstack`, even weakly, was causing Debian’s `dpkg-shlibdeps` to detect an unnecessarily strict versioned dependency on libc6.
Closes#23628.
This is a [breaking-change]. When indexing a generic map (hashmap, etc) using the `[]` operator, it is now necessary to borrow explicitly, so change `map[key]` to `map[&key]` (consistent with the `get` routine). However, indexing of string-valued maps with constant strings can now be written `map["abc"]`.
r? @japaric
cc @aturon @Gankro
This commit implements [RFC 909](https://github.com/rust-lang/rfcs/pull/909):
The `std::thread_local` module is now deprecated, and its contents are
available directly in `std::thread` as `LocalKey`, `LocalKeyState`, and
`ScopedKey`.
The macros remain exactly as they were, which means little if any code
should break. Nevertheless, this is technically a:
[breaking-change]
Closes#23547
This small commit stabilizes the `Error` trait as-is, except that `Send`
and `Debug` are added as constraints. The `Send` constraint is because
most uses of `Error` will be for trait objects, and by default we would
like these objects to be transferrable between threads. The `Debug`
constraint is to ensure that e.g. `Box<Error>` is `Debug`, and because
types that implement `Display` should certainly implement `Debug` in any case.
In the near future we expect to add `Any`-like downcasting features to
`Error`, but this is waiting on some additional
mechanisms (`Reflect`). It will be added before 1.0 via default methods.
[breaking-change]
r? @alexcrichton
Closes#21790
Fix regression in -C rpath that causes failures with symlinks
The new `relative_from` method no longer supports the case on unix
where both paths are absolute, which `-C rpath` depended on. This
version fixes the problem by copying the old path_relative_from
function into the rpath module.
Fixes#23140
After experimenting with the new `relative_from` function on `Path` I'm not sure what it's use case is. It no longer even figures out that the relative path from `/foo/bar` to `/foo/baz/qux` is `../baz/qux`.
This commit:
* Introduces `std::convert`, providing an implementation of
RFC 529.
* Deprecates the `AsPath`, `AsOsStr`, and `IntoBytes` traits, all
in favor of the corresponding generic conversion traits.
Consequently, various IO APIs now take `AsRef<Path>` rather than
`AsPath`, and so on. Since the types provided by `std` implement both
traits, this should cause relatively little breakage.
* Deprecates many `from_foo` constructors in favor of `from`.
* Changes `PathBuf::new` to take no argument (creating an empty buffer,
as per convention). The previous behavior is now available as
`PathBuf::from`.
* De-stabilizes `IntoCow`. It's not clear whether we need this separate trait.
Closes#22751Closes#14433
[breaking-change]
This commit marks as `#[stable]` the `Entry` types for the maps provided
by `std`. The main reason these had been left unstable previously was
uncertainty about an eventual trait design, but several plausible
designs have been proposed that all work fine with the current type definitions.
r? @Gankro
Impls on `clean::Type::FixedVector` are now collected in the array
primitive page instead of the slice primitive page.
Also add a primitive docs for arrays to `std`.
This commit implements [RFC
909](https://github.com/rust-lang/rfcs/pull/909):
The `std::thread_local` module is now deprecated, and its contents are
available directly in `std::thread` as `LocalKey`, `LocalKeyState`, and
`ScopedKey`.
The macros remain exactly as they were, which means little if any code
should break. Nevertheless, this is technically a:
[breaking-change]
Closes#23547
This small commit stabilizes the `Error` trait as-is, except that `Send`
and `Debug` are added as constraints. The `Send` constraint is because
most uses of `Error` will be for trait objects, and by default we would
like these objects to be transferrable between threads. The `Debug`
constraint is to ensure that e.g. `Box<Error>` is `Debug`, and because
types that implement `Display` should certainly implement `Debug` in any case.
In the near future we expect to add `Any`-like downcasting features to
`Error`, but this is waiting on some additional
mechanisms (`Reflect`). It will be added before 1.0 via default methods.
[breaking-change]
Linking __pthread_get_minstack, even weakly, was causing Debian’s
dpkg-shlibdeps to detect an unnecessarily strict versioned dependency
on libc6.
Closes#23628.
Signed-off-by: Anders Kaseorg <andersk@mit.edu>
This commit removes the `IndexMut` impls on `HashMap` and `BTreeMap`, in
order to future-proof the API against the eventual inclusion of an
`IndexSet` trait.
Ideally, we would eventually be able to support:
```rust
map[owned_key] = val;
map[borrowed_key].mutating_method(arguments);
&mut map[borrowed_key];
```
but to keep the design space as unconstrained as possible, we do not
currently want to support `IndexMut`, in case some other strategy will
eventually be needed.
Code currently using mutating index notation can use `get_mut` instead.
[breaking-change]
Closes#23448
r? @Gankro
There have been some recent panics on the bots and this commit is an attempt to
appease them. Previously it was considered invalid to run `rt::at_exit` after
the handlers had already started running. Due to the multithreaded nature of
applications, however, it is not always possible to guarantee this. For example
[this program][ex] will show off the abort.
[ex]: https://gist.github.com/alexcrichton/56300b87af6fa554e52d
The semantics of the `rt::at_exit` function have been modified as such:
* It is now legal to call `rt::at_exit` at any time. The return value now
indicates whether the closure was successfully registered or not. Callers must
now decide what to do with this information.
* The `rt::at_exit` handlers will now be run for a fixed number of iterations.
Common cases (such as the example shown) may end up registering a new handler
while others are running perhaps once or twice, so this common condition is
covered by re-running the handlers a fixed number of times, after which new
registrations are forbidden.
Some usage of `rt::at_exit` was updated to handle these new semantics, but
deprecated or unstable libraries calling `rt::at_exit` were not updated.
The method with which backwards compatibility was retained ended up leading to
documentation that rustdoc didn't handle well and largely ended up confusing.
This commit removes the reexports of `old_io` traits as well as `old_path` types
and traits from the prelude. This functionality is now all deprecated and needs
to be removed to make way for other functionality like `Seek` in the `std::io`
module (currently reexported as `NewSeek` in the io prelude).
Closes#23377Closes#23378
This commit removes the reexports of `old_io` traits as well as `old_path` types
and traits from the prelude. This functionality is now all deprecated and needs
to be removed to make way for other functionality like `Seek` in the `std::io`
module (currently reexported as `NewSeek` in the io prelude).
Closes#23377Closes#23378
This reverts commit aec67c2.
Closes#20012
This is temporarily rebased on #23245 as it would otherwise conflict, the last commit is the only one relevant to this PR though.
This commit removes the `IndexMut` impls on `HashMap` and `BTreeMap`, in
order to future-proof the API against the eventual inclusion of an
`IndexSet` trait.
Ideally, we would eventually be able to support:
```rust
map[owned_key] = val;
map[borrowed_key].mutating_method(arguments);
&mut map[borrowed_key];
```
but to keep the design space as unconstrained as possible, we do not
currently want to support `IndexMut`, in case some other strategy will
eventually be needed.
Code currently using mutating index notation can use `get_mut` instead.
[breaking-change]
Closes#23448
As @alexcrichton says, this was really a libgreen thing, and isn't
relevant now.
As this removes a technically-public function, this is a
[breaking-change]
The IP and socket address types all had `FromStr` implemented but the
implementations were not marked stable, nor was the error type returned ready to
be properly stabilized.
This commit marks the implementations of `FromStr` as stable and also renamed
the `ParseError` structure to `AddrParseError`. The error is now also an opaque
structure that cannot be constructed outside the standard library.
cc #22949
[breaking-change]
As @alexcrichton says, this was really a libgreen thing, and isn't
relevant now.
As this removes a technically-public function, this is a
[breaking-change]
Conflicts:
src/libtest/lib.rs
This commit stabilizes the `ErrorKind` enumeration which is consumed by and
generated by the `io::Error` type. The purpose of this type is to serve as a
cross-platform namespace to categorize errors into. Two specific issues are
addressed as part of this stablization:
* The naming of each variant was scrutinized and some were tweaked. An example
is how `FileNotFound` was renamed to simply `NotFound`. These names should not
show either a Unix or Windows bias and the set of names is intended to grow
over time. For now the names will likely largely consist of those errors
generated by the I/O APIs in the standard library.
* The mapping of OS error codes onto kinds has been altered. Coalescing no
longer occurs (multiple error codes become one kind). It is intended that each
OS error code, if bound, corresponds to only one `ErrorKind`. The current set
of error kinds was expanded slightly to include some networking errors.
This commit also adds a `raw_os_error` function which returns an `Option<i32>`
to extract the underlying raw error code from the `Error`.
Closes#16666
[breaking-change]
This commit stabilizes the `ErrorKind` enumeration which is consumed by and
generated by the `io::Error` type. The purpose of this type is to serve as a
cross-platform namespace to categorize errors into. Two specific issues are
addressed as part of this stablization:
* The naming of each variant was scrutinized and some were tweaked. An example
is how `FileNotFound` was renamed to simply `NotFound`. These names should not
show either a Unix or Windows bias and the set of names is intended to grow
over time. For now the names will likely largely consist of those errors
generated by the I/O APIs in the standard library.
* The mapping of OS error codes onto kinds has been altered. Coalescing no
longer occurs (multiple error codes become one kind). It is intended that each
OS error code, if bound, corresponds to only one `ErrorKind`. The current set
of error kinds was expanded slightly to include some networking errors.
This commit also adds a `raw_os_error` function which returns an `Option<i32>`
to extract the underlying raw error code from the `Error`.
This commit marks as `#[stable]` the `Entry` types for the maps provided
by `std`. The main reason these had been left unstable previously was
uncertainty about an eventual trait design, but several plausible
designs have been proposed that all work fine with the current type definitions.
unbreak openbsd/bitrig build
- remove `pub` from `struct` (error: visibility has no effect inside functions)
- move `pthread_main_np` into function
r? @alexcrichton
The IP and socket address types all had `FromStr` implemented but the
implementations were not marked stable, nor was the error type returned ready to
be properly stabilized.
This commit marks the implementations of `FromStr` as stable and also renamed
the `ParseError` structure to `AddrParseError`. The error is now also an opaque
structure that cannot be constructed outside the standard library.
cc #22949
[breaking-change]
Nothing inside of the read/write interface itself can panic, so any
poison must have been the result of user code which the lock isn't
protecting.
This seems safe to me, but if we don't want to go this route we should update the docs to indicate that these methods can panic.
r? @alexcrichton
Very minor fix: in `std::net::Ipv6Addr::new`, the documentation had an incomplete representation of the resulting address, missing the last two groups.
This commit clarifies some of the unstable features in the `str` module by
moving them out of the blanket `core` and `collections` features.
The following methods were moved to the `str_char` feature which generally
encompasses decoding specific characters from a `str` and dealing with the
result. It is unclear if any of these methods need to be stabilized for 1.0 and
the most conservative route for now is to continue providing them but to leave
them as unstable under a more specific name.
* `is_char_boundary`
* `char_at`
* `char_range_at`
* `char_at_reverse`
* `char_range_at_reverse`
* `slice_shift_char`
The following methods were moved into the generic `unicode` feature as they are
specifically enabled by the `unicode` crate itself.
* `nfd_chars`
* `nfkd_chars`
* `nfc_chars`
* `graphemes`
* `grapheme_indices`
* `width`
This patch changes the type of byte string literals from `&[u8]` to `&[u8; N]`.
It also implements some necessary traits (`IntoBytes`, `Seek`, `Read`, `BufRead`) for fixed-size arrays (also related to #21725) and adds test for #17233, which seems to be resolved.
Fixes#18465
[breaking-change]
This commit clarifies some of the unstable features in the `str` module by
moving them out of the blanket `core` and `collections` features.
The following methods were moved to the `str_char` feature which generally
encompasses decoding specific characters from a `str` and dealing with the
result. It is unclear if any of these methods need to be stabilized for 1.0 and
the most conservative route for now is to continue providing them but to leave
them as unstable under a more specific name.
* `is_char_boundary`
* `char_at`
* `char_range_at`
* `char_at_reverse`
* `char_range_at_reverse`
* `slice_shift_char`
The following methods were moved into the generic `unicode` feature as they are
specifically enabled by the `unicode` crate itself.
* `nfd_chars`
* `nfkd_chars`
* `nfc_chars`
* `graphemes`
* `grapheme_indices`
* `width`
This function is the current replacement for `std::old_io::timer` which will
soon be deprecated. This function is unstable and has its own feature gate as it
does not yet have an RFC nor has it existed for very long.
This function is the current replacement for `std::old_io::timer` which will
soon be deprecated. This function is unstable and has its own feature gate as it
does not yet have an RFC nor has it existed for very long.
Very minor fix: in `std::net::Ipv6Addr::new`, the documentation had an incomplete representation of the resulting address, missing the last two groups.
The [associated RFC][rfc] for possibly splitting out `flush` has been closed and
as a result there are no more blockers for stabilizing this method, so this
commit marks the method as such.
[rfc]: https://github.com/rust-lang/rfcs/pull/950
- Allow inherent implementations on `char`, `str`, `[T]`, `*const T`, `*mut T` and all the numeric primitives.
- copy `unicode::char::CharExt` methods into `impl char`
- remove `unicode::char::CharExt`, its re-export `std::char::CharExt` and `CharExt` from the prelude
- copy `collections::str::StrExt` methods into `impl str`
- remove `collections::str::StrExt` its re-export `std::str::StrExt`, and `StrExt` from the prelude
- copy `collections::slice::SliceExt` methods into `impl<T> [T]`
- remove `collections::slice::SliceExt` its re-export `std::slice::SliceExt`, and `SliceExt` from the prelude
- copy `core::ptr::PtrExt` methods into `impl<T> *const T`
- remove `core::ptr::PtrExt` its re-export `std::ptr::PtrExt`, and `PtrExt` from the prelude
- copy `core::ptr::PtrExt` and `core::ptr::MutPtrExt` methods into `impl<T> *mut T`
- remove `core::ptr::MutPtrExt` its re-export `std::ptr::MutPtrExt`, and `MutPtrExt` from the prelude
- copy `core::num::Int` and `core::num::SignedInt` methods into `impl i{8,16,32,64,size}`
- copy `core::num::Int` and `core::num::UnsignedInt` methods into `impl u{8,16,32,64,size}`
- remove `core::num::UnsignedInt` and its re-export `std::num::UnsignedInt`
- move `collections` tests into its own crate: `collectionstest`
- copy `core::num::Float` methods into `impl f{32,64}`
Because this PR removes several traits, this is a [breaking-change], however functionality remains unchanged and breakage due to unresolved imports should be minimal. If you encounter an error due to an unresolved import, simply remove the import:
``` diff
fn main() {
- use std::num::UnsignedInt; //~ error: unresolved import `std::num::UnsignedInt`.
-
println!("{}", 8_usize.is_power_of_two());
}
```
---
cc #16862
[preview docs](http://japaric.github.io/inherent/std/index.html)
[unicode::char](http://japaric.github.io/inherent/unicode/primitive.char.html)
[collections::str](http://japaric.github.io/inherent/collections/primitive.str.html)
[std::f32](http://japaric.github.io/inherent/std/primitive.f32.html)
This commit performs a stabilization pass over the std::net module,
incorporating the changes from RFC 923. Specifically, the following actions were
taken:
Stable functionality:
* `net` (the name)
* `Shutdown`
* `Shutdown::{Read, Write, Both}`
* `lookup_host`
* `LookupHost`
* `SocketAddr`
* `SocketAddr::{V4, V6}`
* `SocketAddr::port`
* `SocketAddrV4`
* `SocketAddrV4::{new, ip, port}`
* `SocketAddrV6`
* `SocketAddrV4::{new, ip, port, flowinfo, scope_id}`
* Common trait impls for socket addr structures
* `ToSocketAddrs`
* `ToSocketAddrs::Iter`
* `ToSocketAddrs::to_socket_addrs`
* `ToSocketAddrs for {SocketAddr*, (Ipv*Addr, u16), str, (str, u16)}`
* `Ipv4Addr`
* `Ipv4Addr::{new, octets, to_ipv6_compatible, to_ipv6_mapped}`
* `Ipv6Addr`
* `Ipv6Addr::{new, segments, to_ipv4}`
* `TcpStream`
* `TcpStream::connect`
* `TcpStream::{peer_addr, local_addr, shutdown, try_clone}`
* `{Read,Write} for {TcpStream, &TcpStream}`
* `TcpListener`
* `TcpListener::bind`
* `TcpListener::{local_addr, try_clone, accept, incoming}`
* `Incoming`
* `UdpSocket`
* `UdpSocket::bind`
* `UdpSocket::{recv_from, send_to, local_addr, try_clone}`
Unstable functionality:
* Extra methods on `Ipv{4,6}Addr` for various methods of inspecting the address
and determining qualities of it.
* Extra methods on `TcpStream` to configure various protocol options.
* Extra methods on `UdpSocket` to configure various protocol options.
Deprecated functionality:
* The `socket_addr` method has been renamed to `local_addr`
This commit is a breaking change due to the restructuring of the `SocketAddr`
type as well as the renaming of the `socket_addr` method. Migration should be
fairly straightforward, however, after accounting for the new level of
abstraction in `SocketAddr` (protocol distinction at the socket address level,
not the IP address).
[breaking-change]
The [associated RFC][rfc] for possibly splitting out `flush` has been closed and
as a result there are no more blockers for stabilizing this method, so this
commit marks the method as such.
[rfc]: https://github.com/rust-lang/rfcs/pull/950
This commit deprecates the `count`, `range` and `range_step` functions
in `iter`, in favor of range notation. To recover all existing
functionality, a new `step_by` adapter is provided directly on `ops::Range`
and `ops::RangeFrom`.
[breaking-change]
r? @alexcrichton
