There are two moments when a BufRead object needs to empty it's internal
buffer:
- In a seek call;
- In a read call when all data in the internal buffer had been already
consumed and the output buffer has a greater or equal size than the
internal buffer.
In both cases, the buffer was not being properly emptied, but only
marked as consumed (self.pos = self.cap). That should be no problem if
the inner reader is only Read, but if it is Seek as well, then it's
possible to access the data in the buffer by using the seek_relative
method. In order to prevent this from happening, both self.pos and
self.cap should be set to 0.
Two test cases were added to detect that failure:
- test_buffered_reader_invalidated_after_read
- test_buffered_reader_invalidated_after_seek
Both tests are very similar to each other. The inner reader contains the
following data: [5, 6, 7, 0, 1, 2, 3, 4]. The buffer capacity is 3
bytes.
- First, we call fill_buffer, which loads [5, 6, 7] into the internal
buffer, and then consume those 3 bytes.
- Then we either read the 5 remaining bytes in a single read call or we
move to the end of the stream by calling seek. In both cases the
buffer should be emptied to prevent the previous data [5, 6, 7] from
being read.
- We now call seek_relative(-2) and read two bytes, which should give us
the last 2 bytes of the stream: [3, 4].
Before this commit, the the seek_relative method would consider that
we're still in the range of the internal buffer, so instead of fetching
data from the inner reader, it would return the two last bytes that were
incorrectly still in the buffer: [6, 7]. Therefore, the test would fail.
Now, when seek_relative is called the buffer is empty. So the expected
data [3, 4] is fetched from the inner reader and the test passes.
There's lots of comments in the code, but the main gist of this commit
is that the acquisition of the global malloc lock on the
`wasm32-unknown-unknown` target when threads are enabled will not spin
on contention rather than block.
The example had a potential race condition that would still pass the test.
If the thread which was supposed to modify it's own thread local was slower than the instruction to
modify in the main thread, then the test would pass even in case of a failure.
This is would be minor if the child thread was waited for since it check using an `assert_eq` for
the same thing, but vice versa.
However, if the `assert_eq` failed this would trigger a panic, which is not at all caught by the
example since the thread is not waited on.
Signed-off-by: benaryorg <binary@benary.org>
Use the correct stderr when testing libstd
When compiling the unit tests for libstd, there are two copies of `std` in existence, see [lib.rs](919cf42/src/libstd/lib.rs (L335-L341)). This means there are two copies of everything, including thread local variable definitions. Before this PR, it's possible that libtest would configure a stderr sink in one of those copies, whereas the panic logic would inspect the sink in the other copy, resulting in libtest missing the relevant panic message. This PR makes sure that when testing, the panic logic always accesses the stderr sink from “realstd”, using the same logic that libtest uses.
Set secure flags when opening a named pipe on Windows
Fixes https://github.com/rust-lang/rust/issues/42036, see also the previous attempt in https://github.com/rust-lang/rust/pull/44556.
Whether this is correct depends on if it is somehow possible to create a symlink to a named pipe, outside the named pipe filesystem (NPFS). But as far as I can tell that should be impossible.
Also fixes that `security_qos_flags(SECURITY_ANONYMOUS)` does not set the `SECURITY_SQOS_PRESENT` flag, and the incorrect documentation about the default value of `security_qos_flags`.
std: docs: Disable running several Stdio doctests
* A number of `Stdio` related doc examples include running the "rev"
command to illustrate piping commands. The majority of these tests are
marked as `no_run` except for two tests which were not
* Not running these tests is unlikely to cause any negative impact, and
doing so also allows the test suite to pass in environments where the
"rev" command is unavailable
* A number of `Stdio` related doc examples include running the "rev"
command to illustrate piping commands. The majority of these tests are
marked as `no_run` except for two tests which were not
* Not running these tests is unlikely to cause any negative impact, and
doing so also allows the test suite to pass in environments where the
"rev" command is unavailable
Add vectored read and write support
This functionality has lived for a while in the tokio ecosystem, where
it can improve performance by minimizing copies.
r? @alexcrichton
Refactor Windows stdio and remove stdin double buffering
I was looking for something nice and small to work on, tried to tackle a few FIXME's in Windows stdio, and things grew from there.
This part of the standard library contains some tricky code, and has changed over the years to handle more corner cases. It could use some refactoring and extra comments.
Changes/fixes:
- Made `StderrRaw` `pub(crate)`, to remove the `Write` implementations on `sys::Stderr` (used unsynchronised for panic output).
- Remove the unused `Read` implementation on `sys::windows::stdin`
- The `windows::stdio::Output` enum made sense when we cached the handles, but we can use simple functions like `is_console` now that we get the handle on every read/write
- `write` can now calculate the number of written bytes as UTF-8 when we can't write all `u16`s.
- If `write` could only write one half of a surrogate pair, attempt another write for the other because user code can't reslice in any way that would allow us to write it otherwise.
- Removed the double buffering on stdin. Documentation on the unexposed `StdinRaw` says: 'This handle is not synchronized or buffered in any fashion'; which is now true.
- `sys::windows::Stdin` now always only partially fills its buffer, so we can guarantee any arbitrary UTF-16 can be re-encoded without losing any data.
- `sys::windows::STDIN_BUF_SIZE` is slightly larger to compensate. There should be no real change in the number of syscalls the buffered `Stdin` does. This buffer is a little larger, while the extra buffer on Stdin is gone.
- `sys::windows::Stdin` now attempts to handle unpaired surrogates at its buffer boundary.
- `sys::windows::Stdin` no langer allocates for its buffer, but the UTF-16 decoding still does.
### Testing
I did some manual testing of reading and writing to console. The console does support UTF-16 in some sense, but doesn't supporting displaying characters outside the BMP.
- compile stage 1 stdlib with a tiny value for `MAX_BUFFER_SIZE` to make it easier to catch corner cases
- run a simple test program that reads on stdin, and echo's to stdout
- write some lines with plenty of ASCII and emoji in a text editor
- copy and paste in console to stdin
- return with `\r\n\` or CTRL-Z
- copy and paste in text editor
- check it round-trips
-----
Fixes https://github.com/rust-lang/rust/issues/23344. All but one of the suggestions in that issue are now implemented. the missing one is:
> * When reading data, we require the entire set of input to be valid UTF-16. We should instead attempt to read as much of the input as possible as valid UTF-16, only returning an error for the actual invalid elements. For example if we read 10 elements, 5 of which are valid UTF-16, the 6th is bad, and then the remaining are all valid UTF-16, we should probably return the first 5 on a call to `read`, then return an error, then return the remaining on the next call to `read`.
Stdin in Console mode is dealing with text directly input by a user. In my opinion getting an unpaired surrogate is quite unlikely in that case, and a valid reason to error on the entire line of input (which is probably short). Dealing with it is incompatible with an unbuffered stdin, which seems the more interesting guarantee to me.
Simplify the unix `Weak` functionality
- We can avoid allocation by adding a NUL to the function name.
- We can get `Option<F>` directly, rather than aliasing the inner `AtomicUsize`.
Turn duration consts into associated consts
As suggested in https://github.com/rust-lang/rust/issues/57391#issuecomment-459658236, I'm moving `Duration` constants (`SECOND`, `MILLISECOND` and so on; currently behind unstable `duration_constants` feature) into the `impl Duration` block.
cc @frewsxcv @SimonSapin
