| Age | Commit message (Collapse) | Author |
|---|
|
Instead of avoiding overflow-checking builtins with AK_COMPILER_CLANG,
we can use the preprocessor's __has_builtin() mechanism to check if
they are available.
|
|
This is used by the Jakt runtime.
|
|
This is used in Jakt.
|
|
This patch adds the `USING_AK_GLOBALLY` macro which is enabled by
default, but can be overridden by build flags.
This is a step towards integrating Jakt and AK types.
|
|
Doesn't use them in libc headers so that those don't have to pull in
AK/Platform.h.
AK_COMPILER_GCC is set _only_ for gcc, not for clang too. (__GNUC__ is
defined in clang builds as well.) Using AK_COMPILER_GCC simplifies
things some.
AK_COMPILER_CLANG isn't as much of a win, other than that it's
consistent with AK_COMPILER_GCC.
|
|
|
|
|
|
Either not returning *this, or in the case of Variant, not checking for
self assignment. In AK::Atomic, we can't return *this due to the wrapper
semantics Atomic implements.
|
|
By replacing MakeUnsigned<Source> in this specific specialization with a
simple negativity check this now works for floating point source types.
Previously it would attempt a comparison of the destination type and
void.
|
|
These were already implicitly required to be integral via the usage of
the is_within_range templated function, but making them explicit should
produce nicer error messages when building, and make the IDE highlight
the incorrect usage.
|
|
|
|
Signed integer overflow can occur with division when the RHS is -1,
as the negative values' range is one larger than the positives.
|
|
|
|
|
|
(...and ASSERT_NOT_REACHED => VERIFY_NOT_REACHED)
Since all of these checks are done in release builds as well,
let's rename them to VERIFY to prevent confusion, as everyone is
used to assertions being compiled out in release.
We can introduce a new ASSERT macro that is specifically for debug
checks, but I'm doing this wholesale conversion first since we've
accumulated thousands of these already, and it's not immediately
obvious which ones are suitable for ASSERT.
|
|
The overflow state from both Checkeds is OR'ed in the result.
|
|
Problem:
- `constexpr` functions are decorated with the `inline` specifier
keyword. This is redundant because `constexpr` functions are
implicitly `inline`.
- [dcl.constexpr], §7.1.5/2 in the C++11 standard): "constexpr
functions and constexpr constructors are implicitly inline (7.1.2)".
Solution:
- Remove the redundant `inline` keyword.
|
|
Problem:
- `Checked` is not `constexpr`-aware.
Solution:
- Decorate member functions with `constexpr` keyword.
- Add tests to ensure the functionality where possible.
|
|
Problem:
- Compiler-generated functions are being defined which results in
extra code to maintain.
Solution:
- Switch to compiler-generated default functions for default
construction, copy assignment, move assignment, copy construction
and move construction.
|
|
Specifically:
- post-increment actually implemented pre-increment
- helper-templates that provided operator{+,-,*,/}() couldn't possibly work,
because the interface of add (etc) were incompatible (not taking a Checked<>,
and returning void)
|
|
Once again, we need to hint the compiler that it should inline the function, and
then it is able to eliminate the assertion.
|
|
The two-argument version doesn't need an extra template parameter.
|
|
Apparently Clang does not have __builtin_foo_overflow_p()
Fixes #2044.
|
|
And switch the two-argument version of Checked::multiplication_would_overflow()
to use __builtin_mul_overflow_p(). This helps GCC optimize the code better.
|
|
This allows you to comfortably test if multiply 2 or 3 values would
cause arithmetic overflow.
|
|
A Checked<T> is a boxed integer type that asserts if you try to use its
value after an arithmetic overflow.
|
