bitv: add larger tests, better benchmarks & remove dead code.
There were no tests for iteration etc. with more than 5 elements,
i.e. not even going beyond a single word. This situation is rectified.
Also, the only benchmarks for `set` were with a constant bit value,
which was not indicative of every situation, due to inlining & branch
removal. This adds a benchmark at the other end of the spectrum: random
input.
closes#16800
r? @nikomatsakis - I'm not 100% sure this is the right approach, it is kind of ad-hoc. The trouble is we don't have any intrinsic notion of which types are sized and which are not, we only have the Sized bound, so I have nothing to validate the Sized bound against.
As outlined in
https://aturon.github.io/style/naming/conversions.html
`to_` functions names should only be used for expensive operations.
Thus `to_option` is better named `as_option`. Also, putting type
names into method names is considered bad style; what the user is
really trying to get is a reference. This `as_ref` is even better.
Also, we are missing a mutable version of this method.
Finally, there is a bug in the signature of `to_option` which has
been around since lifetime elision: originally the returned reference
had 'static lifetime, but since the elision changes this become
the lifetime of the raw pointer (which does not make sense, since
the pointer lifetime and referent lifetime are unrelated). We fix
the bug to return a reference with a fresh lifetime which will be
inferred from the calling context.
[breaking-change]
There were no tests for iteration etc. with more than 5 elements,
i.e. not even going beyond a single word. This situation is rectified.
Also, the only benchmarks for `set` were with a constant bit value,
which was not indicative of every situation, due to inlining & branch
removal. This adds a benchmark at the other end of the spectrum: random
input.
Fixes#16597
I'm not 100% sure this is the correct way to handle this - but I wasn't able to find a better way without doing way more refactoring of the code that I was comfortable with. Comments and criticism are appreciated 😄
For example `let _x: &Trait = &*(box Foo as Box<Trait>);`. There was a bug where no cleanup would be scheduled by the deref.
No test because cleanup-auto-borrow-obj.rs is a test for this once we remove trait cross-borrowing (done on another branch).
The tcp-accept-stress, despite the previous modifications, is still deadlocking
on the osx buildbots. When building/testing/running repeatedly locally, it was
discovered that the test would often fail with TcpStream::connect returning the
error `address not available`.
This test opens up quite a large number of sockets, and it looks like by default
osx isn't the speediest at recycling those sockets for further use.
The test has been modified (and verified) to not deadlock in this error case,
and the test is not just officially ignored on OSX (with no FIXME). I believe
that we'll get good coverage of the relevant code on the linux builders, so this
isn't so much of a loss.
At the same time I turned down the stress parameters to hopefully lighten the
socket load on other platforms.
This assert was likely inherited from some point, but it's not quite valid as a
no-timeout read may enter this loop, but data could be stolen by any other read
after the socket is deemed readable.
I saw this fail in a recent bors run where the assertion was tripped.
Not sure if this is addressing the root cause or just patching up a symptom. Also not sure if I should be adding a diagnostic code for this.
Fixes#16750Fixes#15812
Enables any macros using `MacExpr` to be treated as patterns when
they produce a literal in the form `ExprLit` (e.g. `stringify!` or `line!`).
Fixes#16876.
These are the additions to liblibc required for raw/custom socket support. I've broken this into a separate pull request due to the upcoming I/O overhaul (was originally part of pull #15741).
cc @alexcrichton.
