Previously we imported a Core Foundation type "CCFooRef" as "CCFoo",
but also provided a typealias "CCFooRef". In Swift 3, we decided to
mark "CCFooRef" unavailable to force developers to consistently use
"CCFoo". Now that we have infrastructure to mark /all/ renamed
declarations as unavailable, just use that to track the renaming,
i.e. pretend that "CCFooRef" was the "Swift 2" name for the type.
This doesn't change the conflict resolution behavior: if there's
another name "CCFoo" in the same module, the CF type will be
imported as just "CCFooRef".
Groundwork cleanup for rdar://problem/26347297, which notes that our
import-as-member fix-its use the "Ref" names rather than the short
names.
More specifically, don't include declarations of methods and properties
in the list of "all imported Objective-C members" if said method or
property is in a generated header. (We actually key off of whether the
enclosing class, protocol, or category is marked as coming from Swift,
but since users aren't supposed to modify generated headers themselves
it's much the same thing.)
This previously caused a crash because we tried to import a Clang member
onto a Swift decl in order to provide the particular member on AnyObject.
rdar://problem/25955831 and probably also rdar://problem/25828987, which
deals with the fix-it to migrate to #selector. (We do an AnyObject-like
lookup to find out which class likely implements the specified selector.)
Sema was dutifully tracking conformances that were "used" as part of
type checking, so it could make sure that those conformances got
completed for SILGen to use. However, this information never actually
made it to SILGen, which included its own (more conservative, not
broad enough) heuristics for finding "used" conformances. Teach Sema
to record conformances within the appropriate source file, and have
SILGen reference the conformances when it emits SIL for the source
file.
Special case logic for CF types, which may be coming in as
unmanaged. In this case, we will use audit information if present to
import with the new type, otherwise we have to fall back to
Unmanaged<CF...>.
We still import global variables that must be unmanaged onto the new
type, though they keep their unmanaged types. This helps to
consolidate the definitions, as well as make future migration easier
if they get audited.
Test cases included.
Code cleaup and generalization to also catch typedefs of CF reference
types. Tests will be included in a subsequent commit, as this now
exhibits a latent problem when interacting with Unmanaged.
This protocol lets us identify swift_newtype'd types. More
importantly, use protocol extensions to make it easy to transfer
specific conformances from the underlying type to the wrapper
type. So, for example, if the underlying type is Hashable, make the
wrapper type Hashable as well. Do the same for Equatable, Comparable,
and _ObjectiveCBridgeable. Fixes rdar://problem/26010804.
We now specially import global decls who we identify as fitting the
notification pattern: extern NSStrings whose name ends in
"Notification". When we see them, we import them as a member of
NSNotificationName and, if NSNotificationName is swift_newtype-ed, we
use that new type.
Test cases included.
When swift_newtype is introducing a new Swift type that wraps a
bridged type, the actual storage (called _rawValue) matches the
Objective-C type and "rawValue" itself is a computed
properly. Properly create the getter for that computed property, which
performs a bridging conversion.
When attempting to compile Swift 2 code (or any Swift code using the
Swift 2 names) in Swift 3, the compiler diagnostics are often entirely
useless because the names have changed radically enough that one
generally gets "no member named 'foo'" errors rather than a helpful
"'foo' was renamed to 'bar'" error. This makes for a very poor user
experience when (e.g.) trying to move Swift 2 code forward to Swift 3.
To improve the experience, when the Swift 2 and Swift 3 names of an
API differ, the Clang importer will produce a "stub" declaration that
matches the Swift 2 API. That stub will be marked with a synthesized
attribute
@available(unavailable, renamed: "the-swift-3-name")
that enables better diagnostics (e.g., "'foo' is unavailable: renamed
to 'bar') along with Fix-Its (courtesy of @jrose-apple's recent work)
that fix the Swift 2 code to compile in Swift 3.
This change addresses much of rdar://problem/25309323 (concerning QoI
of Swift 2 code compiled with a Swift 3 compiler), but some cleanup
remains.
...when renaming C functions. This is currently rejected by Clang
itself, but showed up when testing apinotes. Rather than complicate
the apinotes by including parameter placeholders, just do the
obvious thing (and future-proof against us potentially allowing this
in the future).
rdar://problem/26133744
This is a squash of the following commits:
* [SE-0054] Import function pointer arg, return types, typedefs as optional
IUOs are only allowed on function decl arguments and return types, so
don't import typedefs or function pointer args or return types as IUO.
* [SE-0054] Only allow IUOs in function arg and result type.
When validating a TypeRepr, raise a diagnostic if an IUO is found
anywhere other thn the top level or as a function parameter or return
tpye.
* [SE-0054] Disable inference of IUOs by default
When considering a constraint of the form '$T1 is convertible to T!',
generate potential bindings 'T' and 'T?' for $T1, but not 'T!'. This
prevents variables without explicit type information from ending up with
IUO type. It also prevents implicit instantiation of functions and types
with IUO type arguments.
* [SE-0054] Remove the -disable-infer-iuos flag.
* Add nonnull annotations to ObjectiveCTests.h in benchmark suite.
For swift_newtype structs that we create, we sometimes need to provide
a bridged type interface. In these cases, we use the original
non-bridged type as an underlying stored value, and create a computed
rawValue of bridged type. We similarly create an init() taking the
bridged type, and cast it appropriately to/from storage.
Tests updated.
The Clang importer implicitly synthesizes @discardableResult for
nearly all imported functions. Printing this attribute in the
generated interface leads to a lot of noise. Mark it as implicit so we
don't print it.
This is a temporary solution that implements swift_newtype(enum) as
though it were written swift_newtype(struct). This is to work around
to the fact that a String-backed enum does not actually have a String
stored, and a struct is closer to reflecting that storage
properly. Struct provides most of the functionality and appearance for
now, though it does not allow for switching over the values.
Full support for swift_newtype(enum) as a Swift enum is forthcoming.
This introduces support for swift_newtype(struct) attribute (also
known as swift_wrapper). The Clang importer will create a brand new
struct corresponding to the annotated typedefs, which has a backing
raw value. Globals of that type are imported as static members on the
struct.
Additionally, this interacts seamlessly with prior import-as-member
work, meaning that the newly created type can be imported onto. Tests
included.
While inferring get/set, we paired them up even when one of them was
available through a custom objc header (e.g. a private
header). Instead, fail to pair them up. Test case added.
Some of the callers to importDeclContextOf could and should have
passed an effective context, but didn't, so (e.g.) an enum defined in
C couldn't be mapped to a nested type in Swift. Eliminate the
single-argument importDeclContextOf honeypot so we remember to pass
down the effective Clang context. Fixes rdar://problem/25502497.
For the most part this was just "check isInstanceProperty"; the one feature not yet implemented
is the emission of ObjC metadata for class properties.
rdar://problem/16830785
When importing members of an NS_OPTIONS (aka an option set), mark imported
members that have a value of 0 with an unavailable error. This produces an
experience like this:
x = NSRuncingOptions.none // expected-error {{'none' is unavailable: use [] to construct an empty option set}}
This is important to do, because option set members with a value of zero
do not act like members of the option set. For example, they always fail a
"contains" check.
This ensures their bodies are fragile in SIL. While they can
be synthesized again, this change will avoid tripping up the
SIL verifier after some upcoming changes.
Part of https://bugs.swift.org/browse/SR-267.
