In task-to-thread concurrency mode, `_getMainExecutorAsSerialExecutor`
returns `SerialExecutorRef::generic()`, which is, give or take, NULL.
Unfortunately we'd declared it as returning `any SerialExecutor`,
rather than `(any SerialExecutor)?`, and then the `getMainExecutor()`
function was calling `asUnownedSerialExecutor()` on it, which then
crashes.
rdar://153152063
Without this, llvm would sometimes wrongly assume there's no indirect
accesses and the optimizations can lead to a runtime crash, by
optimizing away initializing options properly.
Resolves rdar://152548190
When using the Swift static build on Windows, projects like Foundation
use the clang headers from the toolchain build `lib` directory. No
target in the toolchain build depends on the static clang headers, so
they do not get installed. As a workaround, this makes the static clang
headers symlink target a dependency of the regular dynamic version of
the headers, ensuring they are installed as part of the toolchainb
build.
This only modifies the runtime function `swift_task_enqueueGlobalWithDeadline` to take new clock primitive to
interoperate with existing dispatch wall clock values.
The _Builtin_float and Synchronization modules are getting the SDK_NAME encoded in their embedded modules, preventing them from being used outside of macOS in embedded mode.
Making it transparent evidently induces new retain/release traffic in Array’s subscript, even though I can find no indication that `ClosedRange.contains` is ever called by that code path. Oh well.
These are really popular shorthands for implementing bounds checks, but currently they aren’t even marked `@inline(__always)`, so in larger functions they tend to remain outlined, defeating optimizations that would otherwise happen.
The corresponding variants on the partial range types are `@_transparent`; that helps unoptimized performance, so let’s apply the same attribute here.
rdar://151177326
`line-directive` was expanding response files for the swiftc invocation,
which was causing issues on CI when the file list got too long. This
changes the script to stop expanding response files and pass them as-is
to the swiftc invocation instead. The format for the response files was
also incorrect, with each individual file being quoted, so this fixes
the CMake script generating those.
Finally, this changes `line-directive` to dump the command line on
failure.
