Fixes#6593
Currently, Rust provides no way to print very large or very small floating point values which come up routinely in scientific and modeling work. The classical solution to this is to use the scientific/exponential notation, which not-coincidentally, corresponds to how floating point values are encoded in memory. Given this, there are two solutions to the problem. One is what, as far as I understand it, Python does. I.e. for floating point numbers in a certain range it does what we do today with the `'f'` formatting flag, otherwise it switches to exponential notation. The other way is to provide a set of formatting flags to explicitly choose the exponential notation, like it is done in C. I've chosen the second way as I think its important to provide that kind of control to the user.
This pull request changes the `std::num::strconv::float_to_str_common` function to optionally format floating point numbers using the exponential (scientific) notation. The base of the significant can be varied between 2 and 25, while the base of the exponent can be 2 or 10.
Additionally this adds two new formatting specifiers to `format!` and friends: `'e'` and `'E'` which switch between outputs like `1.0e5` and `1.0E5`. Mostly parroting C stdlib in this sense, although I wasn't going for an exact output match.
Native timers are a much hairier thing to deal with than green timers due to the
interface that we would like to expose (both a blocking sleep() and a
channel-based interface). I ended up implementing timers in three different ways
for the various platforms that we supports.
In all three of the implementations, there is a worker thread which does send()s
on channels for timers. This worker thread is initialized once and then
communicated to in a platform-specific manner, but there's always a shared
channel available for sending messages to the worker thread.
* Windows - I decided to use windows kernel timer objects via
CreateWaitableTimer and SetWaitableTimer in order to provide sleeping
capabilities. The worker thread blocks via WaitForMultipleObjects where one of
the objects is an event that is used to wake up the helper thread (which then
drains the incoming message channel for requests).
* Linux/(Android?) - These have the ideal interface for implementing timers,
timerfd_create. Each timer corresponds to a timerfd, and the helper thread
uses epoll to wait for all active timers and then send() for the next one that
wakes up. The tricky part in this implementation is updating a timerfd, but
see the implementation for the fun details
* OSX/FreeBSD - These obviously don't have the windows APIs, and sadly don't
have the timerfd api available to them, so I have thrown together a solution
which uses select() plus a timeout in order to ad-hoc-ly implement a timer
solution for threads. The implementation is backed by a sorted array of timers
which need to fire. As I said, this is an ad-hoc solution which is certainly
not accurate timing-wise. I have done this implementation due to the lack of
other primitives to provide an implementation, and I've done it the best that
I could, but I'm sure that there's room for improvement.
I'm pretty happy with how these implementations turned out. In theory we could
drop the timerfd implementation and have linux use the select() + timeout
implementation, but it's so inaccurate that I would much rather continue to use
timerfd rather than my ad-hoc select() implementation.
The only change that I would make to the API in general is to have a generic
sleep() method on an IoFactory which doesn't require allocating a Timer object.
For everything but windows it's super-cheap to request a blocking sleep for a
set amount of time, and it's probably worth it to provide a sleep() which
doesn't do something like allocate a file descriptor on linux.
This routine is currently only used to clean up the timer helper thread in the
libnative implementation, but there are possibly other uses for this.
The documentation is clear that the procedures are *not* run with any task
context and hence have very little available to them. I also opted to disallow
at_exit inside of at_exit and just abort the process at that point.
* Stop using hardcoded numbers that have to all get updated when something changes (inevitable errors and rebase conflicts) as well as removes some unneeded -Z options (obsoleted over time).
* Remove `std::rt::borrowck`
The implementation has been made more succinct and no longer requires Clone. The coverage of the associated unit test has also been increased to check more combinations of bases, exponents, and expected results.
There was an old and barely used implementation of pow, which expected
both parameters to be uint and required more traits to be implemented.
Since a new implementation for `pow` landed, I'm proposing to remove
this old impl in favor of the new one.
The benchmark shows that the new implementation is faster than the one being removed:
```
test num::bench::bench_pow_function ..bench: 9429 ns/iter (+/- 2055)
test num::bench::bench_pow_with_uint_function ...bench: 28476 ns/iter (+/- 2202)
```
The `malloc` family of functions may return a null pointer for a
zero-size allocation, which should not be interpreted as an
out-of-memory error.
If the implementation does not return a null pointer, then handling
this will result in memory savings for zero-size types.
This also switches some code to `malloc_raw` in order to maintain a
centralized point for handling out-of-memory in `rt::global_heap`.
Closes#11634
There was an old and barely used implementation of pow, which expected
both parameters to be uint and required more traits to be implemented.
Since a new implementation for `pow` landed, I'm proposing to remove
this old impl in favor of the new one.
The benchmark shows that the new implementation is faster than the one
being removed:
test num::bench::bench_pow_function ..bench: 9429 ns/iter (+/- 2055)
test num::bench::bench_pow_with_uint_function ...bench: 28476 ns/iter (+/- 2202)
As part of #10387, this removes the `Primitive::{bits, bytes, is_signed}` methods and removes the trait's operator trait constraints for the reasons outlined below:
- The `Primitive::{bits, bytes}` associated functions were originally added to reflect the existing `BITS` and `BYTES`statics included in the numeric modules. These statics are only exist as a workaround for Rust's lack of CTFE, and should be deprecated in the future in favor of using the `std::mem::size_of` function (see #11621).
- `Primitive::is_signed` seems to be of little utility and does not seem to be used anywhere in the Rust compiler or libraries. It is also rather ugly to call due to the `Option<Self>` workaround for #8888.
- The operator trait constraints are already covered by the `Num` trait.
The `malloc` family of functions may return a null pointer for a
zero-size allocation, which should not be interpreted as an
out-of-memory error.
If the implementation does not return a null pointer, then handling
this will result in memory savings for zero-size types.
This also switches some code to `malloc_raw` in order to maintain a
centralized point for handling out-of-memory in `rt::global_heap`.
Closes#11634
The patch adds the missing pow method for all the implementations of the
Integer trait. This is a small addition that will most likely be
improved by the work happening in #10387.
Fixes#11499
This stores the stack of iterators inline (we have a maximum depth with
`uint` keys), and then uses direct pointer offsetting to manipulate it,
in a blazing fast way:
Before:
bench_iter_large ... bench: 43187 ns/iter (+/- 3082)
bench_iter_small ... bench: 618 ns/iter (+/- 288)
After:
bench_iter_large ... bench: 13497 ns/iter (+/- 1575)
bench_iter_small ... bench: 220 ns/iter (+/- 91)
This removes the `Primitive::{bits, bytes, is_signed}` methods and removes the operator trait constraints, for the reasons outlined below:
- The `Primitive::{bits, bytes}` associated functions were originally added to reflect the existing `BITS` and `BYTES` statics included in the numeric modules. These statics are only exist as a workaround for Rust's lack of CTFE, and should probably be deprecated in the future in favor of using the `std::mem::size_of` function (see #11621).
- `Primitive::is_signed` seems to be of little utility and does not seem to be used anywhere in the Rust compiler or libraries. It is also rather ugly to call due to the `Option<Self>` workaround for #8888.
- The operator trait constraints are already covered by the `Num` trait.
* Reexport io::mem and io::buffered structs directly under io, make mem/buffered
private modules
* Remove with_mem_writer
* Remove DEFAULT_CAPACITY and use DEFAULT_BUF_SIZE (in io::buffered)
cc #11119
* Reexport io::mem and io::buffered structs directly under io, make mem/buffered
private modules
* Remove with_mem_writer
* Remove DEFAULT_CAPACITY and use DEFAULT_BUF_SIZE (in io::buffered)
Major changes:
- Define temporary scopes in a syntax-based way that basically defaults
to the innermost statement or conditional block, except for in
a `let` initializer, where we default to the innermost block. Rules
are documented in the code, but not in the manual (yet).
See new test run-pass/cleanup-value-scopes.rs for examples.
- Refactors Datum to better define cleanup roles.
- Refactor cleanup scopes to not be tied to basic blocks, permitting
us to have a very large number of scopes (one per AST node).
- Introduce nascent documentation in trans/doc.rs covering datums and
cleanup in a more comprehensive way.
r? @pcwalton
The patch adds a `pow` function for types implementing `One`, `Mul` and
`Clone` trait.
The patch also renames f32 and f64 pow into powf in order to still have
a way to easily have float powers. It uses llvms intrinsics.
The pow implementation for all num types uses the exponentiation by
square.
Fixes bug #11499
Major changes:
- Define temporary scopes in a syntax-based way that basically defaults
to the innermost statement or conditional block, except for in
a `let` initializer, where we default to the innermost block. Rules
are documented in the code, but not in the manual (yet).
See new test run-pass/cleanup-value-scopes.rs for examples.
- Refactors Datum to better define cleanup roles.
- Refactor cleanup scopes to not be tied to basic blocks, permitting
us to have a very large number of scopes (one per AST node).
- Introduce nascent documentation in trans/doc.rs covering datums and
cleanup in a more comprehensive way.
The failure functions are generic, meaning they're candidates for getting
inlined across crates. This has been happening, leading to monstrosities like
that found in #11549. I have verified that the codegen is *much* better now that
we're not inlining the failure path (the slow path).
Unique pointers and vectors currently contain a reference counting
header when containing a managed pointer.
This `{ ref_count, type_desc, prev, next }` header is not necessary and
not a sensible foundation for tracing. It adds needless complexity to
library code and is responsible for breakage in places where the branch
has been left out.
The `borrow_offset` field can now be removed from `TyDesc` along with
the associated handling in the compiler.
Closes#9510Closes#11533
The failure functions are generic, meaning they're candidates for getting
inlined across crates. This has been happening, leading to monstrosities like
that found in #11549. I have verified that the codegen is *much* better now that
we're not inlining the failure path (the slow path).
Unique pointers and vectors currently contain a reference counting
header when containing a managed pointer.
This `{ ref_count, type_desc, prev, next }` header is not necessary and
not a sensible foundation for tracing. It adds needless complexity to
library code and is responsible for breakage in places where the branch
has been left out.
The `borrow_offset` field can now be removed from `TyDesc` along with
the associated handling in the compiler.
Closes#9510Closes#11533
Currently, we have c_void defined to be represented as an empty struct,
but LLVM expects C's void* to be represented as i8*. That means we
currently generate code in which LLVM doesn't recognize malloc() and
free() and can't apply certain optimization that would remove calls to
those functions.
These functions are of little utility outside a small subset of use cases. If people need them for their own projects then they can use their own bindings for libm (which aren't hard to make).
This fixes#11336
I guess the type sizes are correct for both OS X and iOS, but i am not certain.
In any case, i'd rather have any iOS build at all, so that we have something to improve upon.
The `print!` and `println!` macros are now the preferred method of printing, and so there is no reason to export the `stdio` functions in the prelude. The functions have also been replaced by their macro counterparts in the tutorial and other documentation so that newcomers don't get confused about what they should be using.
