Renamed the invert() function in iter.rs to flip().
Also renamed the Invert<T> type to Flip<T>.
Some related code comments changed. Documentation that I could find has
been updated, and all the instances I could locate where the
function/type were called have been updated as well.
`Zero` and `One` have precise definitions in mathematics as the identities of the `Add` and `Mul` operations respectively. As such, types that implement these identities are now also required to implement their respective operator traits. This should reduce their misuse whilst still enabling them to be used in generalized algebraic structures (not just numbers). Existing usages of `#[deriving(Zero)]` in client code could break under these new rules, but this is probably a sign that they should have been using something like `#[deriving(Default)]` in the first place.
For more information regarding the mathematical definitions of the additive and multiplicative identities, see the following Wikipedia articles:
- http://wikipedia.org/wiki/Additive_identity
- http://wikipedia.org/wiki/Multiplicative_identity
Note that for floating point numbers the laws specified in the doc comments of `Zero::zero` and `One::one` may not always hold. This is true however for many other traits currently implemented by floating point numbers. What traits floating point numbers should and should not implement is an open question that is beyond the scope of this pull request.
The implementation of `std::num::pow` has been made more succinct and no longer requires `Clone`. The coverage of the associated unit test has also been increased to test for more combinations of bases, exponents, and expected results.
Ignore all newline characters in Base64 decoder to make it compatible with other Base64 decoders.
Most of the Base64 decoder implementations ignore all newline characters in the input string. There are some examples:
Python:
```python
>>> "
A
Q
=
=
".decode("base64")
'\x01'
```
Ruby:
```ruby
irb(main):001:0> "
A
Q
=
=
".unpack("m")
=> [""]
```
Erlang:
```erlang
1> base64:decode("
A
Q
=
=
").
<<1>>
```
Moreover some Base64 encoders append newline character at the end of the output string by default:
Python:
```python
>>> "".encode("base64")
'AQ==
'
```
Ruby:
```ruby
irb(main):001:0> [""].pack("m")
=> "AQ==
"
```
So I think it's fairly important for Rust Base64 decoder to accept Base64 inputs even with newline characters in the padding.
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)
```
I update the example of json use to the last update of the json.rs code. I delete the old branch.
From my last request, I remove the example3 because it doesn't compile. I don't understand why and I don't have the time now to investigate.
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
This is my first patch so feedback appreciated!
Bug when initialising `bitv:Bitv::new(int,bool)` when `bool=true`. It created a `Bitv` with underlying representation `!0u` rather than the actual desired bit layout ( e.g. `11111111` instead of `00001111`). This works OK because a size attribute is included which keeps access to legal bounds. However when using `BitvSet::from_bitv(Bitv)`, we then find that `bitvset.contains(i)` can return true when `i` should not in fact be in the set.
```
let bs = BitvSet::from_bitv(Bitv::new(100, true));
assert!(!bs.contains(&127)) //fails
```
The fix is to create the correct representation by treating various cases separately and using a bitshift `(1<<nbits) - 1` to generate correct number of `1`s where necessary.
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
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
* 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)
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
The test run summary currently prints the wrong number of tests run. This PR fixes it by adding a newline to the log output, and also adds support for counting bench runs.
Closes#11381
Use a lookup table, SHIFT_MASK_TABLE, that for every possible four
bit prefix holds the number of times the value should be right shifted and what
the right shifted value should be masked with. This way we can get rid of the
branches which in my testing gives approximately a 2x speedup.
Timings on Intel(R) Core(TM) i5-3570K CPU @ 3.40GHz
-- Before --
running 5 tests
test ebml::tests::test_vuint_at ... ok
test ebml::bench::vuint_at_A_aligned ... bench: 494 ns/iter (+/- 3)
test ebml::bench::vuint_at_A_unaligned ... bench: 494 ns/iter (+/- 4)
test ebml::bench::vuint_at_D_aligned ... bench: 467 ns/iter (+/- 5)
test ebml::bench::vuint_at_D_unaligned ... bench: 467 ns/iter (+/- 5)
-- After --
running 5 tests
test ebml::tests::test_vuint_at ... ok
test ebml::bench::vuint_at_A_aligned ... bench: 181 ns/iter (+/- 2)
test ebml::bench::vuint_at_A_unaligned ... bench: 192 ns/iter (+/- 1)
test ebml::bench::vuint_at_D_aligned ... bench: 181 ns/iter (+/- 3)
test ebml::bench::vuint_at_D_unaligned ... bench: 197 ns/iter (+/- 6)
Use a lookup table, SHIFT_MASK_TABLE, that for every possible four
bit prefix holds the number of times the value should be right shifted and what
the right shifted value should be masked with. This way we can get rid of the
branches which in my testing gives approximately a 2x speedup.
Since reader::vuint_at() returns a result of type reader::Res it makes sense
to make it public.
Due to rust's current behavior of externally referenced private structures,
https://github.com/mozilla/rust/issues/10573, you could still use the result and
assign it to a variable if you let the compiler do the type assignment,
but you could not explicitly annotate a variable to hold a reader::Res.
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.
The methods contained in `std::num::{Algebraic, Trigonometric, Exponential, Hyperbolic}` have now been moved into `std::num::Real`. This is part of an ongoing effort to simplify `std::num` (see issue #10387).
`std::num::RealExt` has also been removed from the prelude because it is not a commonly used trait.
r? @alexcrichton
This trait seems to stray too far from the mandate of a standard library as implementations may vary between use cases. Third party libraries should implement their own if they need something like it.
This closes#5316.
r? @alexcrichton, @pcwalton
This is just an unnecessary trait that no one's ever going to parameterize over
and it's more useful to just define the methods directly on the types
themselves. The implementors of this type almost always don't want
inner_mut_ref() but they're forced to define it as well.
The methods contained in `std::num::{Algebraic, Trigonometric, Exponential, Hyperbolic}` have now been moved into `std::num::Real`. This is part of an ongoing effort to simplify `std::num` (see issue #10387).
`std::num::RealExt` has also been removed from the prelude because it is not a commonly used trait.
A typed arena is a type of arena that can only allocate objects of one
type. It is 3x faster than the existing arena and 13x faster than malloc
on Mac.
r? @brson
A typed arena is a type of arena that can only allocate objects of one
type. It is 3x faster than the existing arena and 13x faster than malloc
on Mac.
Apologies for junking up the feed with all of these separate pull requests. I'm still getting the hang of git and will hopefully be doing less of this nonsense soon. I opened up another PR and closed the one from earlier today because the first PR was coming from the wrong branch of my repo.
Anyway, this contains a fleshed-out implementation of TotalEq/TotalOrd/Clone/ToStr for the whole B-tree structure and relevant tests, integrating suggestions and comments from several community members.
r? @catamorphism
- Add `mut_iter`, `mut_lower_bound`, `mut_upper_bound`
- Remove some internal iterators
- Add benchmarks
- Improve performance of `{mut_,}{lower,upper}_bound`
- Minor clean-up of `extra::treemap` after I realised I wasn't exploiting macros to their full DRY potential.
After writing some benchmarks for ebml::reader::vuint_at() I noticed that LLVM doesn't seem to inline the from_be32 function even though it only does a call to the bswap32 intrinsic in the x86_64 case. Marking the functions with #[inline(always)] fixes that and seems to me a reasonable thing to do. I got the following measurements in my vuint_at() benchmarks:
- Before
test ebml::bench::vuint_at_A_aligned ... bench: 1075 ns/iter (+/- 58)
test ebml::bench::vuint_at_A_unaligned ... bench: 1073 ns/iter (+/- 5)
test ebml::bench::vuint_at_D_aligned ... bench: 1150 ns/iter (+/- 5)
test ebml::bench::vuint_at_D_unaligned ... bench: 1151 ns/iter (+/- 6)
- Inline from_be32
test ebml::bench::vuint_at_A_aligned ... bench: 769 ns/iter (+/- 9)
test ebml::bench::vuint_at_A_unaligned ... bench: 795 ns/iter (+/- 6)
test ebml::bench::vuint_at_D_aligned ... bench: 758 ns/iter (+/- 8)
test ebml::bench::vuint_at_D_unaligned ... bench: 759 ns/iter (+/- 8)
- Using vuint_at_slow()
test ebml::bench::vuint_at_A_aligned ... bench: 646 ns/iter (+/- 7)
test ebml::bench::vuint_at_A_unaligned ... bench: 645 ns/iter (+/- 3)
test ebml::bench::vuint_at_D_aligned ... bench: 907 ns/iter (+/- 4)
test ebml::bench::vuint_at_D_unaligned ... bench: 1085 ns/iter (+/- 16)
As expected inlining from_be32() gave a considerable speedup.
I also tried how the "slow" version fared against the optimized version and noticed that it's
actually a bit faster for small A class integers (using only two bytes) but slower for big D class integers (using four bytes)
Right now on linux, an empty executable with LTO still depends on librt becaues
of the clock_gettime function in rust_builtin.o, but this commit moves this
dependency into a rust function which is subject to elimination via LTO.
At the same time, this also drops libstd's dependency on librt on unices that
are not OSX because the library is only used by extra::time (and now the
dependency is listed in that module instead).
Right now on linux, an empty executable with LTO still depends on librt becaues
of the clock_gettime function in rust_builtin.o, but this commit moves this
dependency into a rust function which is subject to elimination via LTO.
At the same time, this also drops libstd's dependency on librt on unices that
are not OSX because the library is only used by extra::time (and now the
dependency is listed in that module instead).
