We previously computed cross-imports by comparing N transitive imports against N transitive imports. This is wasteful, because at least one of the two modules in a pair has to actually declare a cross-import overlay for us to discover one, and the vast majority of modules don’t declare any.
This commit makes us instead compare N transitive imports against M transitive imports which are known to declare at least one cross-import overlay. Since N is potentailly in the thousands while M is perhaps in the double digits, this should be good for a substantial time savings.
However, this optimization has made a test of another cross-import performance optimization fail—not because we have regressed on that, but because it skips work the test case expects us to perform. I have XFAILed that test for now.
Fixes <rdar://problem/59538458>.
When a “separately imported overlay” is added to a SourceFile, two things happen:
1. The direct import of the underlying module is removed from getImports*() by default. It is only visible if the caller passes ImportFilterKind:: ShadowedBySeparateOverlay. This means that non-module-scoped lookups will search _OverlayModule before searching its re-export UnderlyingModule, allowing it to shadow underlying declarations.
2. When you ask for lookupInModule() to look in the underlying module in that source file, it looks in the overlays instead. This means that UnderlyingModule.foo() can find declarations in _OverlayModule.
VS2015 had an issue with the deletion of an operator. Since VS2017 is
the minimum version that LLVM uses, we can assume that VS2017+ is in use
(_MSC_VER >= 1910). Clean up the now defunct workaround.
Add an alternative to getTopLevelDecls and getDeclChecked to limit which
decls are deserialized by first looking at their attributes. If the
attributes are accepted by a function passed as argument the decl is
fully deserialized, otherwise it is ignored.
The filter is included in the signature of existing functions in the
Serilalization services, but I’ve added new methods for it in FileUnit
and its subclasses to leave existing implementations untouched.
ProtocolConformanceRef already has an invalid state. Drop all of the
uses of Optional<ProtocolConformanceRef> and just use
ProtocolConformanceRef::forInvalid() to represent it. Mechanically
translate all of the callers and callsites to use this new
representation.
Most of AST, Parse, and Sema deal with FileUnits regularly, but SIL
and IRGen certainly don't. Split FileUnit out into its own header to
cut down on recompilation times when something changes.
No functionality change.
Removes duplicated logic from the implementations of
FileUnit::lookupValue, and simplifies the interface to
ModuleDecl::lookupValue, where everyone was passing an empty
(non-filtering) access path anyway /except/ during actual lookup from
source code. No functionality change.
We already do this for other ASTContext-allocated types (see
Decl.cpp). This will prevent the sort of mistakes in the previous two
commits.
Note that if any particular subclass of FileUnit wants to have its
destructor run, it can opt into that manually using
ASTContext::addDestructorCleanup. SourceFile and BuiltinUnit both do
this. But we generally don't /want/ to do this if we can avoid it
because it adds to compiler teardown time.
Implementing it in LoadedFile is nice in theory, but causes a leak in
practice because that type is ASTContext-allocated and usually never
destroyed.
https://bugs.swift.org/browse/SR-11366
...rather than the buffer, for a compiled module that came from a
module interface.
This was already happening at a higher level
(ModuleDecl::getModuleFilename) so pushing it down to the low-level
ModuleFile::getModuleFilename doesn't really change things much. The
important fix that goes with this is that SerializedASTFile no longer
leaks this name by storing it outside of ModuleFile.
https://bugs.swift.org/browse/SR-11365
These operations are called frequently, and would get rather
expensive when a ModuleDecl contains a large number of
SourceFiles:
- lookupValue()
- lookupVisibleDecls()
- lookupClassMember()
- lookupClassMembers()
Add fast paths to handle the case where all file units in the
module are SourceFiles, by refactoring SourceFile::LookupCache
from a per-file to a per-module cache.
There was an error when including Module.h in an external project with
std::default_delete being instantiated when making a std::unique_ptr.
This change declares destructor in header file and sets it to default in
cpp file.
Otherwise, there's no guarantee of binary compatibility, and whoever
turned on library evolution support shouldn't be lulled into a false
sense of security.
This is just a warning for now, but will be promoted to an error later
once clients have shaken out any places where they're doing this.
Note that the still-experimental '@_implementationOnly' opts out of
this check, because that enforces that the import doesn't make its way
into the current module's public source or binary interface.
rdar://50261171
Previously 'isSystemModule()' returns true only if the module is:
- Standard library
- Clang module and that is `IsSystem`
- Swift overlay for clang `IsSystem` module
Now:
- Clang module and that is `IsSystem`; or
- Swift overlay for clang `IsSystem` module
- Swift module found in either of these directories:
- Runtime library directoris (including stdlib)
- Frameworks in `-Fsystem` directories
- Frameworks in `$SDKROOT/System/Library/Frameworks/` (Darwin)
- Frameworks in `$SDKROOT/Library/Frameworks/` (Darwin)
rdar://problem/50516314
- In Sema, don't traverse nested declarations while deducing the opaque return type. This would
cause returns inside nested functions to clobber the return type of the outer function.
- In IRGen, walk the list of opaque return types we keep in the SourceFile already for type
reconstruction, instead of trying to visit them ad-hoc as part of walking the AST, since
IRGen doesn't normally walk the bodies of function decls directly.
Fixes rdar://problem/50459091
Way back in Swift 1 I was trying to draw a distinction between
"overlays", separate libraries that added Swift content to an existing
Objective-C framework, and "the Swift part of a mixed-source
framework", even though they're implemented in almost exactly the same
way. "Adapter module" was the term that covered both of those. In
practice, however, no one knew what "adapter" meant. Bring an end to
this confusion by just using "overlay" within the compiler even for
the mixed-source framework case.
No intended functionality change.
Sema no longer adds conformances to a per-SourceFile list that it thinks
are going to be "used" by SILGen, IRGen and the runtime. Instead, previous
commits already ensure that SILGen determines the set of conformances to be
emitted, triggering conformance checking as needed.
When printing a swiftinterface, represent opaque result types using an attribute that refers to
the mangled name of the defining decl for the opaque type. To turn this back into a reference
to the right decl's implicit OpaqueTypeDecl, use type reconstruction. Since type reconstruction
doesn't normally concern itself with non-type decls, set up a lookup table in SourceFiles and
ModuleFiles to let us handle the mapping from mangled name to opaque type decl in type
reconstruction.
(Since we're invoking type reconstruction during type checking, when the module hasn't yet been
fully validated, we need to plumb a LazyResolver into the ASTBuilder in an unsightly way. Maybe
there's a better way to do this... Longer term, at least, this surface design gives space for
doing things more the right way--a more request-ified decl validator ought to be able to naturally
lazily service this request without the LazyResolver reference, and if type reconstruction in
the future learns how to reconstruct non-type decls, then the lookup tables can go away.)
When a Swift module built with debug info imports a library without
debug info from a textual interface, the textual interface is
necessary to reconstruct types defined in the library's interface. By
recording the Swift interface files in DWARF dsymutil can collect them
and LLDB can find them.
rdar://problem/49751363
This is an attribute that gets put on an import in library FooKit to
keep it from being a requirement to import FooKit. It's not checked at
all, meaning that in this form it is up to the author of FooKit to
make sure nothing in its API or ABI depends on the implementation-only
dependency. There's also no debugging support here (debugging FooKit
/should/ import the implementation-only dependency if it's present).
The goal is to get to a point where it /can/ be checked, i.e. FooKit
developers are prevented from writing code that would rely on FooKit's
implementation-only dependency being present when compiling clients of
FooKit. But right now it's not.
rdar://problem/48985979
...in preparation for me adding a third kind of import, making the
existing "All" kind a problem. NFC, except that I did rewrite the
ClangModuleUnit implementation of getImportedModules to be simpler!
It's a pretty obscure feature (and one we wish we didn't need), but
sometimes API is initially exposed through one module in order to
build another one, and we want the canonical presented name to be
something else. Push this concept into Swift's AST properly so that
other parts of the compiler stop having to know that this is a
Clang-specific special case.
No functionality change in this commit; will be used in the next
commit.
This is only currently exercised by swift-remoteast-test, so do the
minimum to ensure that we’re getting cached mangled names, but don’t
fret over the linear-time search.
