An environment is always associated with a location with a signature, so
having them separate is pointless duplication. This patch also updates
the serialization to round-trip the signature data.
Reimplement the witness matching logic used for generic requirements
so that it properly models the expectations required of the witness,
then captures the results in the AST. The new approach has a number of
advantages over the existing hacks:
* The constraint solver no longer requires hacks to try to tangle
together the innermost archetypes from the requirement with the
outer archetypes of the context of the protocol
conformance. Instead, we create a synthetic set of archetypes that
describes the requirement as it should be matched against
witnesses. This eliminates the infamous 'SelfTypeVar' hack.
* The type checker no longer records substitutions involving a weird
mix of archetypes from different contexts (see above), so it's
actually plausible to reason about the substitutions of a witness. A
new `Witness` class contains the declaration, substitutions, and all
other information required to interpret the witness.
* SILGen now uses the substitution information for witnesses when
building witness thunks, rather than computing all of it from
scratch. ``substSelfTypeIntoProtocolRequirementType()` is now gone
(absorbed into the type checker, and improved from there), and the
witness-thunk emission code is simpler. A few other bits of SILGen
got simpler because the substitutions can now be trusted.
* Witness matching and thunk generation involving generic requirements
and nested generics now works, based on some work @slavapestov was
already doing in this area.
* The AST verifier can now verify the archetypes that occur in witness substitutions.
* Although it's not in this commit, the `Witness` structure is
suitable for complete (de-)serialization, unlike the weird mix of
archetypes previously present.
Fixes rdar://problem/24079818 and cleans up an area that's been messy
and poorly understood for a very, very long time.
Sugared GenericTypeParamTypes point to GenericTypeParamDecls,
allowing the name of the parameter as written by the user to be
recovered. Canonical GenericTypeParamTypes on the other hand
only store a depth and index, without referencing the original
declaration.
When printing SIL, we wish to output the original generic parameter
names, even though SIL only uses canonical types. Previously,
we used to accomplish this by mapping the generic parameter to an
archetype and printing the name of the archetype. This was not
adequate if multiple generic parameters mapped to the same
archetype, or if a generic parameter was mapped to a concrete type.
The new approach preserves the original sugared types in the
GenericEnvironment, adding a new GenericEnvironment::getSugaredType()
method.
There are also some other assorted simplifications made possible
by this.
Unfortunately this makes GenericEnvironments use a bit more memory,
however I have more improvements coming that will offset the gains,
in addition to making substitution lists smaller also.
UnconditionalAvailabilityKind => PlatformAgnosticAvailabilityKind
::UnavailableInCurrentSwift => ::SwiftVersionSpecific
Plus a couple related method renamings. Prep work for SR-2709.
Banish the abomination that is clangSemaOverride, a previously
necessary evil. When building the module caches, different Clang
instances will be spawned than the one used by the normal
importer. Since we want to reuse code and get the same name both ways,
this meant threading through alternative clang Semas and preprocessors
throughtout, some of the time. This broke the abstraction and
encapsulation of the Impl, complicated the programming model, and
otherwise made effective caching hard.
Now that we’ve done enough ImportName refactoring, we can create a
NameImporter per Clang instance, and encapsulate naming therein. We
can now remove the sema overrides, as we have already done to the
preprocessor overrides.
This shifts the 2-phase initialization problem to the Impl and the
Clang module writers.
NFC
SwiftNameLookupExtension and ClangImporter::Implementation were
friends, but as time goes on they have drifted apart. As part of the
ImportName refactoring, these are being decoupled to facilitate
multiple-name importing, and fight the existing false encapsulation
present in the Impl.
SwiftNameLookupExtension is now spun off into its own entity, and can
evolve to have and use its own de-coupled NameImporter.
There's a bit of a hack to deal with generic typealiases, but
overall this makes things more logical.
This is the last big refactoring before we can allow constrained
extensions to make generic parameters concrete. All that remains
is a small set of changes to SIL type lowering, and retooling
some diagnostics in Sema.
By refactoring out PlatformAvailability from the ClangImporter, we can
more easily refactor out isUnavailableInSwift from the impl, which
will free us up to do more flexible import naming.
Introduces new files ClangAdapter.h/cpp, which will serve as a
convenient place to put code reasoning about Clang details. Refactors
out most Clang-related is*, has*, and get* methods from the
ImporterImpl. In the future, an adapter class could help serve to
seperate the concerns of the importer from the details of how to
correctly use Clang APIs.
Refactor some 3k lines of code from the class definition (nested 3
indentation levels) to static functions or out-of-line
definitions. Reduces some of code browsing burden of the class, though
I hope to do more cleanup to the file as a whole.
NFC
Refactors out some definitions and types from the
ClangImporter::Implementation into a new component ImportName. Future
work will include more separation and finally some redesigning of name
determination components.
Now that SILFunctions no longer reference a GenericParamList, we
don't need to de-serialize cross-module references to archetypes
anymore.
This was the last remaining usage of AllArchetypes, so we can
finally rip it out.
A GenericEnvironment stores the mapping between GenericTypeParamTypes
and context archetypes (or eventually, concrete types, once we allow
extensions to constrain a generic parameter to a concrete type).
The goals here are two-fold:
- Eliminate the GenericTypeParamDecl::getArchetype() method, and
always use mapTypeIntoContext() instead
- Replace SILFunction::ContextGenericParams with a GenericEnvironment
This patch adds the new data type as well as serializer and AST
verifier support. but nothing else uses it yet.
Note that GenericSignature::get() now asserts if there are no
generic parameters, instead of returning null. This requires a
few tweaks here and there.
Simplify e.g., ASTContext::getBridgedToObjC(), which no longer needs
the optional return.
Eliminate the now-unused constraint kind for checking bridging to
Objective-C.
Like NSObject, CFType has primitive operations CFEqual and CFHash,
so Swift should allow those types to show up in Hashable positions
(like dictionaries). The most general way to do this was to
introduce a new protocol, _CFObject, and then have the importer
automatically make all CF types conform to it.
This did require one additional change: the == implementation that
calls through to CFEqual is in a new CoreFoundation overlay, but the
conformance is in the underlying Clang module. Therefore, operator
lookup for conformances has been changed to look in the overlay for
an imported declaration (if there is one).
https://bugs.swift.org/browse/SR-2388
...because otherwise option sets that get imported as members using
NS_SWIFT_NAME are printed with an 'enum' tag, and the definition of
NS_OPTIONS only declares the typedef under C++.
We should come back and figure out something more principled for this
later, but for now this solves an issue with generated headers
imported into C++ translation units.
rdar://problem/27130343
One minor revision: this lifts the proposed restriction against
overriding a non-open method with an open one. On reflection,
that was inconsistent with the existing rule permitting non-public
methods to be overridden with public ones. The restriction on
subclassing a non-open class with an open class remains, and is
in fact consistent with the existing access rule.
Introduce several new factory methods to create CallExprs, and hide
the constructor. The primary reason for this refactor is to start
moving clients over to the factory method that takes the call
arguments separately from the argument labels. Internally, it
repackages those arguments into a TupleExpr or ParenExpr (as
appropriate) so the result ASTs are the same. However, this will make
it easier for us to tease out the arguments themselves in the
implementation of SE-0111.
The code added here to introduce an unavailable typealias accidentally returned the typealias, causing us to import all 'id's as Any even in unbridgeable positions, using the 'id' alias as their sugar name instead of 'Any'.
- Any is made into a keyword which is always resolved into a TypeExpr,
allowing the removal of the type system code to find TheAnyType before
an unconstrained lookup.
- Types called `Any` can be declared, they are looked up as any other
identifier is
- Renaming/redefining behaviour of source loc methods on
ProtocolCompositionTypeRepr. Added a createEmptyComposition static
method too.
- Code highlighting treats Any as a type
- simplifyTypeExpr also does not rely on source to get operator name.
- Any is now handled properly in canParseType() which was causing
generic param lists containing ‘Any’ to fail
- The import objc id as Any work has been relying on getting a decl for
the Any type. I fix up the clang importer to use Context.TheAnyType
(instead of getAnyDecl()->getDeclaredType()). When importing the id
typedef, we create a typealias to Any and declare it unavaliable.
A given Objective-C error enum, which is effectively an NS_ENUM that
specifies its corresponding error domain, will now be mapped to an
ErrorProtocol-conforming struct that wraps an NSError, much like
NSCocoaError does. The actual enum is mapped to a nested "Code"
enum. For example, CoreLocation's CLError becomes:
struct CLError : ErrorProtocol {
let _nsError: NSError
// ...
@objc enum Code : Int {
case ...
}
}
This implements bullet (2) in the proposed solution of SE-0112, so
that Cocoa error types are mapped into structures that maintain the
underlying NSError to allow more information to be extracted from it.
Changes how we import from just protocol<> to the proper Any
type. This means that uses of AnyObject will either implicitly cast up
to Any, or must be explicitly cast through a temporary if used as an
lvalue. id as Any is still predicated on the id-as-any flags.
The Clang attribute allows one to state that a particular enumeration
type describes an error, and associates it with a particular domain
constant. However, due to lack of API notes support, this attribute
wasn't actually getting used. Instead, we had a number of explicit
extensions to enum types to make them conform to the _BridgedNSError
protocol explicitly.
Now that we have API notes, use them to make these enums into error
enums with the appropriate domain, so that the Clang importer will
synthesize the _BridgedNSError conformances. Then, remove all of the
explicit conformances---and with them, the overlays for 12 frameworks.
There is a small fix to more eagerly consider these conformances as
"used" if an expression is formed with the error enum as a value
type. This better ensures that the conformances will be available at
runtime when needed.
This cleanup is needed to implement SE-0112 (NSError bridging),
although it is useful by itself.
Previously getInterfaceType() would punt to getType() if no
interface type was set. This patch changes getInterfaceType()
to assert if no interface type is set, and updates various
places to set the interface type explicitly.
This brings us a step closer to removing PolymorphicFunctionType.
Before we would construct types containing a mix of interface and
contextual types, and then map them in and out. Straighten this out.
Note that I've also had to start untangling the issue where
synthesized ParamDecls do not have an interface type.
First, enforce that the superclass of a class is an interface type.
Previously, Swift classes used interface types but imported
Objective-C generics used archetypes.
When the superclass type is always an interface type, we
can use the recently-added gatherAllSubstitutions() instead of
rolling our own parent type walk.
Also, this exposed an issue in name lookup where we would call
getSuperclass() on a type whose parent was an unbound generic.
This doesn't make sense, so generalize the existing check there.
