Two new symbols:
`void std::__detail::__variant::__erased_ctor<std::monostate&, std::monostate&&>(void*, void*)`
`void std::__detail::__variant::__erased_assign<std::monostate&, std::monostate&&>(void*, void*)`
Likely added from #71627.
`swift_willThrow` is called with an error right before it is thrown.
This existing entrypoint requires an already-boxed error existential;
with typed errors, we don't have the error existential on hand, so we
would need to allocate the box to throw a typed error. That's not okay.
Introduce a new `swift_willThrowTypedImpl` entry point into the runtime
that will first check for the presence of an error handler and, if one
is present, box the error to provide to the error handler. This
maintains the no-allocations path for typed errors while still
allowing existing error handlers to work.
This new entrypoint isn't available on older Swift runtimes, so create
a back-deployable shim called by the compiler. On new-enough platforms,
this will call through to `swift_willThrowTypedImpl`. On older
platforms, we drop the error and don't call the registered will-throw
handler at all. This is a compromise that avoids boxing when throwing
typed errors, at the cost of a slightly different experience for this
new feature on older runtimes.
Fixes rdar://119828459.
- Changed 128 integer literal to Int(Int8.max)+1 in unit tests (Int8.Stride is Int).
Co-authored-by: Guillaume Lessard <glessard@users.noreply.github.com>
The core definitions of CGPoint and other types were pushed down from
CoreGraphics to CoreFoundation. This test indirectly depends on the conformance
of those types to CustomReflectable, but nothing in the test was using those
conformances explicitly and so libswiftCoreGraphics would not be auto-linked
when the test is compiled at a high enough deployment target. Use the
conformances directly to force linkage.
Resolves rdar://121343931
Passing in the declared interface type to checkConformance() here
masked a silent failure where if the type declaration was generic
and the conformance conditional, the conditional requirement check
would fail. As a result, we did not diagnose the absence of
@retroactive, nor the unnecessary presence of it.
Since we only care about the existence of some conformance, we
can use lookupConformance() instead.
Make `init(catching:)` and `get()` use typed throws. The former infers
the `Failure` type from the closure provided (once full type inference
is in place) and the latter only throws errors of the `Failure` type.
original source location of a macro expansion when walking up the ancestor
list.
The original source range of a macro expansion buffer contains the entire
range of the code that is replaced by the expansion. In some cases, the
original code is type checked, so ASTScope needs to represent both the
original code and the macro-expanded code in the scope tree.
For accessor macros, the original code -- the initializer expression of
the property -- is type checked, but not before macro expansion, so
unqualified lookup must deal with the scope tree for the initializer and
the expanded accessors. The accessors are inserted in the tree after the
initializer scope, but the `isBeforeInSource` was using the start location
of the original source range for macro expansions, so the accessor scopes
were incorrectly considered to be ordered before a source location in the
initializer expression, which caused unqualified lookup of local variables
within the initializer to fail.
These tests are testing changes that aren't present in older runtimes:
test/stdlib/SwiftValueNSObject.swift
test/stdlib/SwiftObjectNSObject.swift
test/stdlib/BridgeEquatableToObjC.swift
And this test looks for some wording that's different on older runtimes. This one already does availability checks, so we change the requirements to SwiftStdlib 5.11:
test/Casting/CastTraps.swift.gyb
* Add concrete overload for on all stdlib integer types.
* Add very basic tests for type inference of comparands.
* Add a comment explaining why these overloads are present.
