Add explicit "Derivable" and "How can I implement `Default`" sections.
Copied relevant sections from the module-level documentation, but also
linked to there-- it has a more comprehensive narrative with examples
that show implementation AND use. Decided to just put implementation
example in the trait documentation.
The new order puts all the "when" questions together and puts the "how"
question with the "derivable" section. So you have to scroll past (and
hopefully read) the can/cannot/should caveats and guidelines to get to
the information about how to actually go about doing it once you've
determined that you can and should, with derivable information first so
that you can just use the derived implementation if that applies.
Previous order:
* General explanation
* When can my type be `Copy`?
* How can I implement `Copy`?
* When can my type _not_ be `Copy`?
* When should my type be `Copy`?
* Derivable
New order:
* General explanation
* When can my type be `Copy`?
* When can my type _not_ be `Copy`?
* When should my type be `Copy`?
* Derivable
* How can I implement `Copy`?
I think these just got out of sync, but both use a lexicographic
ordering.
Relevant commits in the history of these explanations:
* 8b81f76 on 2015-06-30
* e22770b on 2016-02-09
Efficient trie lookup for boolean Unicode properties
Replace binary search of ranges with trie lookup using leaves of
64-bit bitmap chunks. Benchmarks suggest this is approximately 10x
faster than the bsearch approach.
Use libc::abort, not intrinsics::abort, in rtabort!
intrinsics::abort compiles down to an illegal instruction, which on
Unix-like platforms causes the process to be killed with SIGILL. A more
appropriate way to kill the process would be SIGABRT; this indicates
better that the runtime has explicitly aborted, rather than some kind of
compiler bug or architecture mismatch that SIGILL might indicate.
For rtassert!, replace this with libc::abort. libc::abort raises
SIGABRT, but is defined to do so in such a way that it will terminate
the process even if SIGABRT is currently masked or caught by a signal
handler that returns.
On non-Unix platforms, retain the existing behavior. On Windows we
prefer to avoid depending on the C runtime, and we need a fallback for
any other platforms that may be defined. An alternative on Windows
would be to call TerminateProcess, but this seems less essential than
switching to using SIGABRT on Unix-like platforms, where it is common
for the process-killing signal to be printed out or logged.
This is a [breaking-change] for any code that depends on the exact
signal raised to abort a process via rtabort!
cc #31273
cc #31333
intrinsics::abort compiles down to an illegal instruction, which on
Unix-like platforms causes the process to be killed with SIGILL. A more
appropriate way to kill the process would be SIGABRT; this indicates
better that the runtime has explicitly aborted, rather than some kind of
compiler bug or architecture mismatch that SIGILL might indicate.
For rtassert!, replace this with libc::abort. libc::abort raises
SIGABRT, but is defined to do so in such a way that it will terminate
the process even if SIGABRT is currently masked or caught by a signal
handler that returns.
On non-Unix platforms, retain the existing behavior. On Windows we
prefer to avoid depending on the C runtime, and we need a fallback for
any other platforms that may be defined. An alternative on Windows
would be to call TerminateProcess, but this seems less essential than
switching to using SIGABRT on Unix-like platforms, where it is common
for the process-killing signal to be printed out or logged.
This is a [breaking-change] for any code that depends on the exact
signal raised to abort a process via rtabort!
cc #31273
cc #31333
Added a `rustdoc` shortcut for collapse/expand all
Now when the user presses the "+" key all sections will collapse/expand.
Also added a note to the help screen which describes this behavior.
This required increasing the height of the help screen.
Clarify the English translation of `?Sized`
* It wasn't clear whether `?Sized` meant "not `Sized`" or "`Sized` or not `Sized`". According to #rust IRC, it does indeed mean "`Sized` or not `Sized`".
* Use the same language as [Trait std::marker::Sized](https://doc.rust-lang.org/std/marker/trait.Sized.html) about how `Sized` is implicitly bound.
* Refer to the syntax as `?Sized`, since it's currently the only allowed trait that can follow `?`.
Book: small improvement to a table to make it clearer
This table is used as an example of four heap values where two of them got deallocated leaving a gap.
It also has stack variables.
Instead of four stack variables I propose three, that way there's no misleading connection between the number of stack variables and heap variables.
But more importantly: three of the four stack variables had the same name (**y**) which could be confusing to a beginner, I changed this as well.
rustdoc: Add doc snippets for trait impls, with a read more link
The read more link only appears if the documentation is more than one line long.
It currently does not appear on non-defaulted methods, since you can document them directly. I could make it so that default documentation gets forwarded if regular docs don't exist.
Fixes#33672
r? @alexcrichton
cc @steveklabnik
rustbuild: Touch up some test suites
This adds in some missing test suites, primarily a few pretty suites. It also starts optimizing tests by default as the current test suite does, but also recognizes `--disable-optimize-tests`.
Currently the optimization of tests isn't recognized by crate tests because Cargo doesn't support the ability to compile an unoptimized test suite against an optimized library. Perhaps a feature to add, though!
* Use "special bound syntax" instead of "special syntax". `?Sized` is technically a "bound", but `?Sized` is specialized syntax that _only_ works with `Sized`, and no other Trait.
* Replace "constant size" with "sized".
Originally fixed in #29961 the bug was unfortunately still present in the face
of crates using `#[macro_use]`. This commit refactors for the two code paths to
share common logic to ensure that they both pick up the same bug fix.
Closes#33762
std: Cache HashMap keys in TLS
This is a rebase and extension of #31356 where we not only cache the keys in
thread local storage but we also bump each key every time a new `HashMap` is
created. This should give us a nice speed bost in creating hash maps along with
retaining the property that all maps have a nondeterministic iteration order.
Closes#27243
