Inline trivial (noop) flush calls
At work I noticed that `writer.flush()?` didn't get optimized away in cases where the flush is obviously a no-op, which I had expected (well, desired).
I went through and added `#[inline]` to a bunch of cases that were obviously noops, or delegated to ones that were obviously noops. I omitted platforms I don't have access to (some tier3). I didn't do this very scientifically, in cases where it was non-obvious I left `#[inline]` off.
merge functionality of `io::Sink` into `io::Empty`
Many times, there is a need for a simple dummy `io::Read`er + `io::Write`r, but currently the only options are `io::Empty` and `io::Sink` respectively. Having both of their functionality together requires writing your own boilerplate for something that makes sense to have in the standard library. This PR adds the functionality of `io::Sink` to `io::Empty`, making `io::Empty` be able to perform the tasks of both of the previous structs. (This idea was first mentioned in #24235)
Note: I also updated some doc comments in `io::utils` in this pull request to fix inconsistencies between `io::Sink` and `io::Empty`.
API Change Proposal: https://github.com/rust-lang/libs-team/issues/49
io: soften ‘at most one write attempt’ requirement in io::Write::write
At the moment, documentation of std::io::Write::write indicates that
call to it ‘represents at most one attempt to write to any wrapped
object’. It seems that such wording was put there to contrast it with
pre-1.0 interface which attempted to write all the data (it has since
been changed in [RFC 517]).
However, the requirement puts unnecessary constraints and may
complicate adaptors which perform non-trivial transformations on the
data. For example, they may maintain an internal buffer which needs
to be written out before the write method accepts more data. It might
be natural to code the method such that it flushes the buffer and then
grabs another chunk of user data. With the current wording in the
documentation, the adaptor would be forced to return Ok(0).
This commit softens the wording such that implementations can choose
code structure which makes most sense for their particular use case.
While at it, elaborate on the meaning of `Ok(0)` return pointing out
that the write_all methods interprets it as an error.
[RFC 517]: https://rust-lang.github.io/rfcs/0517-io-os-reform.html
previously it was only able to use BufWriter. This was due to a limitation in the
BufReader generics that prevented specialization. This change works around the issue
by using `where Self: Read` instead of `where I: Read`. This limits our options, e.g.
we can't access BufRead methods, but it happens to work out if we rely on some
implementation details.
Spelling library
Split per https://github.com/rust-lang/rust/pull/110392
I can squash once people are happy w/ the changes. It's really uncommon for large sets of changes to be perfectly acceptable w/o at least some changes.
I probably won't have time to respond until tomorrow or the next day
Limit read size in `File::read_to_end` loop
Fixes#110650.
Windows file reads have perf overhead that's proportional to the buffer size. When we have a reasonable expectation that we know the file size, we can set a reasonable upper bound for the size of the buffer in one read call.
Clarify `Error::last_os_error` can be weird
Fundamentally, querying the OS for error codes is a process that is deeply subject to the whims of chance and fortune. We can account for OS, but not for every combination of platform APIs. A compiled binary may not recognize new errors introduced years later. We should clarify a few especially odd situations, and what they mean: We can effectively promise nothing... if you ask for Rust to decode errors where none have occurred.
This allows removing mention of ErrorKind::Uncategorized.
That error variant is hidden deliberately, so we should not explicitly mention it.
This fixes#106937.
Since you had an opinion also: Does this solution seem acceptable?
r? ``@ChrisDenton``
Implement read_buf for a few more types
Implement read_buf for TcpStream, Stdin, StdinLock, ChildStdout,
ChildStderr (and internally for AnonPipe, Handle, Socket), so
that it skips buffer initialization.
The other provided methods like read_to_string and read_to_end are
implemented in terms of read_buf and so benefit from the optimization
as well.
This commit also implements read_vectored and is_read_vectored where
applicable.
Guarantee that when `read_buf_exact` returns, all bytes read will be
appended to the buffer. Including the case when the operations fails.
The motivating use case are operations on a non-blocking reader. When
`read_buf_exact` fails with `ErrorKind::WouldBlock` error, the operation
can be resumed at a later time.
Implement read_buf for TcpStream, Stdin, StdinLock, ChildStdout,
ChildStderr (and internally for AnonPipe, Handle, Socket), so
that it skips buffer initialization.
The other provided methods like read_to_string and read_to_end are
implemented in terms of read_buf and so benefit from the optimization
as well.
This commit also implements read_vectored and is_read_vectored where
applicable.
At the moment, documentation of std::io::Write::write indicates that
call to it ‘represents at most one attempt to write to any wrapped
object’. It seems that such wording was put there to contrast it
with pre-1.0 interface which attempted to write all the data (it has
since been changed in [RFC 517]).
However, the requirement puts unnecessary constraints and may complicate
adaptors which perform non-trivial transformations on the data. For
example, they may maintain an internal buffer which needs to be written
out before the write method accepts more data. It might be natural to
code the method such that it flushes the buffer and then grabs another
chunk of user data. With the current wording in the documentation, the
adaptor would be forced to return Ok(0).
This commit softens the wording such that implementations can choose
code structure which makes most sense for their particular use case.
While at it, elaborate on the meaning of `Ok(0)` return pointing out
that the write_all methods interprets it as an error.
[RFC 517]: https://rust-lang.github.io/rfcs/0517-io-os-reform.html
Fundamentally, querying the OS for error codes is a process
that is deeply subject to the whims of chance and fortune.
We can account for OS, but not for every combination of platform APIs.
A compiled binary may not recognize new errors introduced years later.
We should clarify a few especially odd situations, and what they mean:
We can effectively promise nothing.
This allows removing mention of ErrorKind::Uncategorized.
That error variant is hidden quite deliberately, so we
should not explicitly mention it.
