When emitting SIL for `if #available(Swift ..., *)` queries, call the new
`_isSwiftRuntimeVersionAtLeast()` function in the stdlib to check the
condition. To support back deployment, the implementation of
`_isSwiftRuntimeVersionAtLeast()` is `@_alwaysEmitIntoClient` and performs its
comparison against the result of `_SwiftStdlibVersion.current`, which is
pre-existing ABI that the stdlib exposes for querying the Swift runtime
version.
Resolves rdar://162726037.
This helps avoid producing more downstream errors. This changes
`GenericSignature::forInvalid` to produce the same signature as e.g
`<T where T == Undefined>`. This subsumes the need to introduce
conformance requirements for invertible protocols.
The `_Concurrency` and `_StringProcessing` modules are implementation details of the standard library; to developers, their contents should behave as though they are declared directly within module `Swift`. This is the exact same behavior we expect of cross-import overlays, so treat these modules as though they are cross-import overlays with no bystanding module.
Because these modules don’t re-export the standard library, it’s also necessary to treat `Swift` as a separately imported overlay of itself; do so and make that actually work.
Now that ErrorType prints as `_`, we can use that instead of UnresolvedType
here since the original type is only really used for type printing and
debugging.
While the intent behind this functor was noble, it has grown in complexity
considerably over the years, and it seems to be nothing but a source of
crashes in practice. I don't want to deal with it anymore, so I've decided
to just subsume all usages with LookUpConformanceInModule instead.
When importing custom availability domains with dynamic predicates from Clang
modules, synthesize predicate functions for `if #available` queries and call
them when generating SIL.
Resolves rdar://138441312.
Turns out we can also get solver-allocated original ErrorTypes through
type resolution. Given the original type is only used for
printing/debugging, let's just fold away any type variables and
placeholders into UnresolvedType (which print as placeholders). This
matches what `Solution::simplifyType` does.
I don't believe we ever form these types, and if we did we aren't
correctly handling them in `Solution::simplifyType`. Let's just
enforce we don't get them.
Convert a bunch of places where we're dumping to stderr and calling
`abort` over to using `ABORT` such that the message gets printed to
the pretty stack trace. This ensures it gets picked up by
CrashReporter.
-nostdimport and -nostdlibimport only remove the toolchain and usr/lib/swift search paths, and they leave the framework search paths intact. That makes it impossible to get a fully custom SDK environment. Make their behavior match clang's -nostdinc/-nostdlibinc behavior: treat framework and non-framework paths the same. In other words, -nostdinc removes *all* compiler provided search paths, and -nostdlibinc removes *all* SDK search paths.
Rename SkipRuntimeLibraryImportPaths to SkipAllImportPaths, and ExcludeSDKPathsFromRuntimeLibraryImportPaths to SkipSDKImportPaths to reflect their updated behavior.
Move the DarwinImplicitFrameworkSearchPaths handling from SearchPathOptions to CompilerInvocation, where RuntimeLibraryImportPaths is managed. Rename it to just ImplicitFrameworkSearchPaths, and filter for Darwin when it's set up so that all of the clients don't have to do Darwin filtering themselves later.
rdar://150557632
Previously we would insert StringRefs that reference the keys in the
map, but that breaks if the invocation ever gets copied. Switch to
`std::string`.
rdar://148130166
This hash is also used for the dependency scanning hash. In both cases, PCH contents may differ based on whether a certain module they depend on is found in a system or non-system search path. In dependency scanning, systemness should cause a full change of scanning context requiring a from-scratch scan.
Resolves rdar://150334077
This Boolean flag is used in ClangTypeConverter to indicate whether
a type is being converted in the context of a template type parameter.
This parameter can be made a template parameter because it is always
a compile-time constant.
Swift started to explicitly forbid the instantiation of C++ function
templates with arbitrary types in #77430, because many types cause the
Swift compiler to crash. However, those checks prevented them from being
instantiated with Swift closures (which were previously fine), causing
a regression.
This patch relaxes the convertTemplateArgument() function to also allow
converting Swift function types, and adds some tests to make sure doing
so is fine.
This patch also does some cleanup of existing tests checking the
instantiation of various types, and adds testing for C function pointers
and Obj-C blocks.
rdar://148124104
