Travis CI has new infrastructure using the Google Compute Engine which has both
faster CPUs and more memory, and we've been encouraged to switch as it should
help our build times! The only downside currently, however, is that IPv6 is
disabled, causing a number of standard library tests to fail.
Consequently this commit tweaks our travis config in a few ways:
* ccache is disabled as it's not working on GCE just yet
* Docker is used to run tests inside which reportedly will get IPv6 working
* A system LLVM installation is used instead of building LLVM itself. This is
primarily done to reduce build times, but we want automation for this sort of
behavior anyway and we can extend this in the future with building from source
as well if needed.
* gcc-specific logic is removed as the docker image for Ubuntu gives us a
recent-enough gcc by default.
it makes rustc compatible with gcc installation that are using
`--program-transform-name' configure flag (on OpenBSD for example).
- detects at configure the name of stdc++ library on the system
- use the detected name in llvm makefile (with enable-static-stdcpp),
and pass it to mklldeps.py
- generate mklldeps.rs using this detected name
note that CFG_STDCPP_NAME is about stdc++ name, not about libc++. If
using libc++, the default name will be `stdc++', but it won't be used
when linking.
Commit 9104a902c0 fixed the generated
files, but that change would be lost (or require additional manual
intervention) if they are re-generated of if new architectures are
added.
cc #28273
This also involved adding `[TYPE;N]` syntax and aggregate indexing
support to the generator script: it's the only way to be able to have a
parameterised intrinsic that returns an aggregate, since one can't refer
to previous elements of the current aggregate (and that was harder to
implement).
I believe everything that doesn't take a constant integer up to SSE4.2
should now be correct (I don't have any reason to believe that those
that do take constant integers are wrong; they're just more complicated
and I just haven't tested them in detail).
This python script will consume an appropriately formatted JSON file and
output either a Rust file for use in librustc_platform_intrinsics, or an
extern block for importing the intrinsics in an external library.
The --help flag has details.
This commit removes all unstable and deprecated functions in the standard
library. A release was recently cut (1.3) which makes this a good time for some
spring cleaning of the deprecated functions.
Completely rewrite the conversion of decimal strings to `f64` and `f32`. The code is intended to be absolutely positively completely 100% accurate (when it doesn't give up). To the best of my knowledge, it achieves that goal. Any input that is not rejected is converted to the floating point number that is closest to the true value of the input. This includes overflow, subnormal numbers, and underflow to zero. In other words, the rounding error is less than or equal to 0.5 units in the last place. Half-way cases (exactly 0.5 ULP error) are handled with half-to-even rounding, also known as banker's rounding.
This code implements the algorithms from the paper [How to Read Floating Point Numbers Accurately][paper] by William D. Clinger, with extensions to handle underflow, overflow and subnormals, as well as some algorithmic optimizations.
# Correctness
With such a large amount of tricky code, many bugs are to be expected. Indeed tracking down the obscure causes of various rounding errors accounts for the bulk of the development time. Extensive tests (taking in the order of hours to run through to completion) are included in `src/etc/test-float-parse`: Though exhaustively testing all possible inputs is impossible, I've had good success with generating millions of instances from various "classes" of inputs. These tests take far too long to be run by @bors so contributors who touch this code need the discipline to run them. There are `#[test]`s, but they don't even cover every stupid mistake I made in course of writing this.
Another aspect is *integer* overflow. Extreme (or malicious) inputs could cause overflow both in the machine-sized integers used for bookkeeping throughout the algorithms (e.g., the decimal exponent) as well as the arbitrary-precision arithmetic. There is input validation to reject all such cases I know of, and I am quite sure nobody will *accidentally* cause this code to go out of range. Still, no guarantees.
# Limitations
Noticed the weasel words "(when it doesn't give up)" at the beginning? Some otherwise well-formed decimal strings are rejected because spelling out the value of the input requires too many digits, i.e., `digits * 10^abs(exp)` can't be stored in a bignum. This only applies if the value is not "obviously" zero or infinite, i.e., if you take a near-infinity or near-zero value and add many pointless fractional digits. At least with the algorithm used here, computing the precise value would require computing the full value as a fraction, which would overflow. The precise limit is `number_of_digits + abs(exp) > 375` but could be raised almost arbitrarily. In the future, another algorithm might lift this restriction entirely.
This should not be an issue for any realistic inputs. Still, the code does reject inputs that would result in a finite float when evaluated with unlimited precision. Some of these inputs are even regressions that the old code (mostly) handled, such as `0.333...333` with 400+ `3`s. Thus this might qualify as [breaking-change].
# Performance
Benchmarks results are... tolerable. Short numbers that hit the fast paths (`f64` multiplication or shortcuts to zero/inf) have performance in the same order of magnitude as the old code tens of nanoseconds. Numbers that are delegated to Algorithm Bellerophon (using floats with 64 bit significand, implemented in software) are slower, but not drastically so (couple hundred nanoseconds).
Numbers that need the AlgorithmM fallback (for `f64`, roughly everything below 1e-305 and above 1e305) take far, far longer, hundreds of microseconds. Note that my implementation is not quite as naive as the expository version in the paper (it needs one to four division instead of ~1000), but division is fundamentally pretty expensive and my implementation of it is extremely simple and slow.
All benchmarks run on a mediocre laptop with a i5-4200U CPU under light load.
# Binary size
Unfortunately the implementation needs to duplicate almost all code: Once for `f32` and once for `f64`. Before you ask, no, this cannot be avoided, at least not completely (but see the Future Work section). There's also a precomputed table of powers of ten, weighing in at about six kilobytes.
Running a stage1 `rustc` over a stand-alone program that simply parses pi to `f32` and `f64` and outputs both results reveals that the overhead vs. the old parsing code is about 44 KiB normally and about 28 KiB with LTO. It's presumably half of that + 3 KiB when only one of the two code paths is exercised.
| rustc options | old | new | delta |
|--------------------------- |--------- |--------- |----------- |
| [nothing] | 2588375 | 2633828 | 44.39 KiB |
| -O | 2585211 | 2630688 | 44.41 KiB |
| -O -C lto | 1026353 | 1054981 | 27.96 KiB |
| -O -C lto -C link-args=-s | 414208 | 442368 | 27.5 KiB |
# Future Work
## Directory layout
The `dec2flt` code uses some types embedded deeply in the `flt2dec` module hierarchy, even though nothing about them it formatting-specific. They should be moved to a more conversion-direction-agnostic location at some point.
## Performance
It could be much better, especially for large inputs. Some low-hanging fruit has been picked but much more work could be done. Some specific ideas are jotted down in `FIXME`s all over the code.
## Binary size
One could try to compress the table further, though I am skeptical. Another avenue would be reducing the code duplication from basically everything being generic over `T: RawFloat`. Perhaps one can reduce the magnitude of the duplication by pushing the parts that don't need to know the target type into separate functions, but this is finicky and probably makes some code read less naturally.
## Other bases
This PR leaves `f{32,64}::from_str_radix` alone. It only replaces `FromStr` (and thus `.parse()`). I am convinced that `from_str_radix` should not exist, and have proposed its [deprecation and speedy removal][deprecate-radix]. Whatever the outcome of that discussion, it is independent from, and out of scope for, this PR.
Fixes#24557Fixes#14353
r? @pnkfelix
cc @lifthrasiir @huonw
[paper]: http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.45.4152
[deprecate-radix]: https://internals.rust-lang.org/t/deprecate-f-32-64-from-str-radix/2405
This commit removes all unstable and deprecated functions in the standard
library. A release was recently cut (1.3) which makes this a good time for some
spring cleaning of the deprecated functions.
This commit leverages the runtime support for DWARF exception info added
in #27210 to enable unwinding by default on 64-bit MSVC. This also additionally
adds a few minor fixes here and there in the test harness and such to get
`make check` entirely passing on 64-bit MSVC:
* The invocation of `maketest.py` now works with spaces/quotes in CC
* debuginfo tests are disabled on MSVC
* A link error for librustc was hacked around (see #27438)
As there’s no C++ runtime any more there’s really no point in having
anything but Rust tags being made.
I’ve also taken the liberty of excluding the compiler parts of this in
the `librust%,,` pattern substitution. Whether or not this is “correct”
will depend on whether you want tags for the compiler or for general
use. For myself, I want it for general use.
I’m not sure how much people use the tags files anyway. I definitely do,
but with Racer existing the tags files aren’t quite so necessary.
I added it because it was easy (same a `char::to_lowercase`,
just a different table), but it doesn’t make sense to have this
in std but not str::to_titlecase, which would require
https://github.com/unicode-rs/unicode-segmentation
At some point in the future this feature will be available
(both on char and str) in a crates.io crate.
GDB and LLDB pretty printers have some common functionality and also access some common information, such as the layout of standard library types. So far, this information has been duplicated in the two pretty printing python modules. This PR introduces a common module used by both debuggers.
This PR also implements proper rendering of `String` and `&str` values in LLDB.
GDB and LLDB pretty printers have some common functionality
and also access some common information, such as the layout of
standard library types. So far, this information has been
duplicated in the two pretty printing python modules. This
commit introduces a common module used by both debuggers.
