Make the Keccak permutation public, as it's useful for more than
SHA-3 (kMAC, SHAKE, TurboSHAKE, TupleHash, etc).
Our Keccak implementation was accepting f as a comptime parameter,
but always used 64-bit words and 200 byte states, so it actually
didn't work with anything besides f=1600.
That has been fixed. The ability to use reduced-round versions
was also added in order to support M14 and K12.
The state was constantly converted back and forth between bytes
and words, even though only a part of the state is actually used
for absorbing and squeezing bytes. It was changed to something
similar to the other permutations we have, so we can avoid extra
copies, and eventually add vectorized implementations.
In addition, the SHAKE extendable output function (XOF) was
added (SHAKE128, SHAKE256). It is required by newer schemes,
such as the Kyber post-quantum key exchange mechanism, whose
implementation is currently blocked by SHAKE missing from our
standard library.
Breaking change: `Keccak_256` and `Keccak_512` were renamed to
`Keccak256` and `Keccak512` for consistency with all other
hash functions.
When there is a diamond dependency, reuse a *Module instead of creating
a redundant one using the same build.zig file. Otherwise, the compile
error "file exists in multiple modules" would occur.
Partial revert of 1e963053d0ff67361b587b046a917375e963d5e9. Now that
lldb has a new method for accessing this data that does not require
manual updating, there is no longer any reason to share this data.
And make it not do any installation, only objcopying. We already have
install steps for doing installation.
This commit also makes ObjCopyStep properly integrate with caching.
Many `Type`s can correspond to the same `CType`, so this reduces the
number of used locals by 27760 when compiling only-c.
Also, disabled some tests that were only passing by accident and
shouldn't really be considered working.
- Combine mulXi3 routines for follow-up cleanup.
- DRY up Dwords and Twords
- rename both to HalveInt and use instance
* Justification: Not all processors have word size 32 bit.
* remove test file from CMakeLists
* DRY things.
After this, the last MSVC warnings are in behavior/bugs/529.zig:
behavior.c(37971): warning C4133: 'function': incompatible types - from 'A__8479 *' to 'A__8474 *'
behavior.c(37974): warning C4133: 'function': incompatible types - from 'A__8480 *' to 'A__8474 *'
* Support always_tail and never_tail/never_inline with a comptime callee using clang
* Support never_inline using gcc
* Support never_inline using msvc
Unfortunately, can't enable behavior tests because of the conditional support.
And additionally support writing files to source files. This means a
custom build step in zig's own build.zig is no longer needed for copying
zig.h because it is handled by WriteFileStep.
* Implement @fieldParentPtr on a union
* Refactor field access to ensure that it is handled consistently
* Remove `renderTypecast` as it is now behaves the same as `renderType`
Forcing the key to be of the same type as the sorted items used during
the search is a valid use case.
There, however, exists some cases where the key and the items are of
heterogeneous types, like searching for a code point in ordered ranges
of code points:
```zig
const CodePoint = u21;
const CodePointRange = [2]CodePoint;
const valid_ranges = &[_]CodePointRange{
// an ordered array of ranges
};
fn orderCodePointAndRange(
context: void,
code_point: CodePoint,
range: CodePointRange
) std.math.Order {
_ = context;
if (code_point < range[0]) {
return .lt;
}
if (code_point > range[1]) {
return .gt;
}
return .eq;
}
fn isValidCodePoint(code_point: CodePoint) bool {
return std.sort.binarySearch(
CodePointRange,
code_point,
valid_ranges,
void,
orderCodePointAndRange
) != null;
}
```
It is so expected that `std.sort.binarySearch` should therefore support
both homogeneous and heterogeneous keys.
