For clang symbols marked with SPI_AVAILABLE, we add SPIAccessControlAttr to them so they will be
considered as SPIs in the AST. To be able to use all these symbols, we also add an implicit SPI import
statement for all clang modules. All clang SPIs belong to the same SPI group named "OBJC_DEFUALT_SPI_GROUP" because clang
currently doesn't support custom SPI group.
rdar://73902734
7856f2d83d only partially skipped writing
out destructors when they were invalid, ie. it skipped writing the decl
itself but not the records common to all decls. This would cause any
records on the destructor to be applied on the next serialized decl.
Make sure to skip serializing anything to do with the destructor when
it's invalid and does not have a class context.
Most invalid attributes are skipped from serialization entirely, but
custom attributes don't have their invalid bit set - the particular
custom attribute (eg. a property wrapper) is requested when needed and
skipped if invalid. Those checks can't set the invalid bit since the
attribute could be a different custom attribute (eg. a result builder).
When allowing errors any attribute could be on any decl, so don't verify
whether an attribute can appear on a decl. Note that these attributes
aren't serialized anyway since they'll be set to invalid during
typechecking and hence skipped.
The `-force-single-frontend-invocation` flag predates WMO and is now an
alias for `-whole-module-optimization`. We should use the latter and let
the former fade into history.
This means that we no longer have the invariant that the extendedType always
contains the generic parameters. So we need to fix the assertions/test cases
for it.
...instead of crashing. Also drop the class if its generic
requirements depend on a type that can't be loaded (instead of
crashing).
rdar://problem/50125674
These can be recreated if needed in a client library. To do this, I've
added a new ConformanceLookupKind::NonInherited, which can also be
used elsewhere in the project where we're already filtering out
inherited conformances some other way.
Note that this doesn't drop inherited conformances from the entire
serialized interface, just from the list that a class explicitly
declares. They still get referenced sometimes.
rdar://problem/50541451 and possibly others
There was only one remaining usage other than in testing tools.
Note that when a declaration mangling was passed in, the old entry
point would (try to) return the type of the declaration.
The new entry point no longer has this behavior. I changed the
bridging-header-first test to run lldb-moduleimport-test with
-decl-from-mangled instead of -type-from-mangled-old to preserve
the behavior of the test.
Also, I removed test/DebugInfo/DumpTypeFromMangledName.swift
completely. This test only covered a handful of cases, and a bunch
of them were declaration manglings rather than type manglings.
The new tests in test/TypeDecoder/ are much more comprehensive.
The -type-from-mangled flag now uses the new API. The -type-from-mangled-old flag
uses the old API, ide::getTypeFromMangledSymbolname().
For now, just change all existing tests to use the -type-from-mangled-old flag;
I'll be adding new tests for the new API shortly.
The SDK directory is now confusing as the Windows target also has a SDK
overlay. In order to make this more uniform, move the SDK directory to
Darwin which covers the fact that this covers the XNU family of OSes.
The Windows directory contains the SDK overlay for the Windows target.
This is enough to let the test case in rdar://problem/40899824 pass,
and any callers of this function already need to be able to handle a
nullptr result. There's a lot more work to do in this area, but it's
nice to get the simple things working again.
Because we've changed how conformances represent associated types
since Swift 4.2 branched, the error described there no longer occurs.
But we still want to make sure we don't regress.
https://bugs.swift.org/browse/SR-7337
This will also preserve @escaping and @autoclosure, which were previously
dropped. This could lead to mismatches between expected and actual
canonical types for serialization cross-references.
https://bugs.swift.org/browse/SR-8045, likely others
Cross-references are identified by their containing module, with the
assumption that two modules will never have the same name. However, an
overlay has the same name as its underlying Clang module, which means
that there can be two declarations with the same name, the same type,
and the same module name. This is the underlying cause of the
'UIEdgeInsetsZero' problem, but it also affects the CloudKit overlay.
By tracking a bit that just says "this came from Clang", we're able
to resolve otherwise ambiguous cross-references.
(Why didn't we do it this way all along? Because if a declaration
moves from Clang to Swift or vice versa, that would break the
cross-reference. But that's only interesting if the swiftmodule format
is meant to be persistent across changing dependencies, and it looks
like we're moving away from that anyway. It's also a little weird for
SerializedModuleLoader to have special cases for Clang, but this isn't
the first.)
Note that I'm not reverting the UIEdgeInsetsZero workaround here; the
end state will have that coming just from UIKit as originally
described.
rdar://problem/40839486
This mirrors how a bridging header is processed when compiling source
files: before any of the imports. This is important for LLDB to
recreate the source environment as closely as possible to how the
compiler does it; in the test case being added involving a non-modular
header file, failure to do so resulted in a deserialization
cross-reference crash.
Note that Serialization still sorts imports, which normal resolution
of imports in source does not do. So we're still not consistent. But
this is less important than handling textual includes (bridging
headers) before modular imports.
rdar://problem/40471329
...instead of relying on the one in the overlay in pre-4.2 versions of
Swift. This caused crashes in deserialization, which (deliberately)
doesn't respect availability.
There are three changes here:
- Remove UIEdgeInsets.zero and UIOffset.zero from the UIKit overlay.
- Always use the 4.2 name for UIEdgeInsetsZero and UIOffsetZero from
the underlying UIKit framework. (This is the nested name.)
- Ignore the unavailability messages for those two constants in
pre-4.2 Swift, since we're now relying on them being present.
The latter two, the compiler changes, can go away once UIKit's API
notes no longer specify different pre-4.2 behavior, but meanwhile we
need to keep compatibility with the SDKs released in Xcode 10b1.
https://bugs.swift.org/browse/SR-7879
This failed in Swift 4.1 because of a problem with Substitutions,
but that whole model's been ripped out and redone in Swift 4.2.
https://bugs.swift.org/browse/SR-7978
This was fixed by 897effe, which I had originally thought would be a
no-functionality-change commit because it just made things lazier.
Turns out requirement signature deserialization can result in
circularity with sufficiently cross-referential conformances.
This isn't exactly a reduced test case because it still depends on
subclassing NSObject, which probably means there are hidden
dependencies on conforming to standard library protocols. But it's
better than nothing.
https://bugs.swift.org/browse/SR-5191
Some cases of using isSuperset can cause crashes, this was caused by improper subclassing callouts; this pr resolves those failures (and provides unit tests for that case)
The cases where the bridge was traversed too much now only causes a single bridge out call (without needing to reallocate or thrash retain/release)
String.components(separatedBy: CharacterSet) should be considerably faster now not only for more apporpriate bridging calls but also no longer needing to bridge arrays back and forth.
Resolves the following issues:
rdar://problem/17281998
rdar://problem/26611771
rdar://problem/29738989
Previously looking up an extension would result in all extensions for
types with the same name (nested or not) being deserialized; this
could even bring in base types that had not been deserialized yet. Add
in a string to distinguish an extension's base type; in the top-level
case this is just a module name, but for nested types it's a full
mangled name.
This is a little heavier than I'd like it to be, since it means we
mangle names and then throw them away, and since it means there's a
whole bunch of extra string data in the module just for uniquely
identifying a declaration. But it's correct, and does less work than
before, and fixes a circularity issue with a nested type A.B.A that
apparently used to work.
https://bugs.swift.org/browse/SR-3915
This "fixes" two issues:
- The name of a non-public typealias would leak into the public
interface if the extension had any public members.
- A common pattern of defining a platform-specific typealias for an
imported class and then extending that type would lead to
circularity when trying to deserialize the typealias. We /shouldn't/
be loading the extension at that point, but fixing that would be
much harder.
The "right" answer is to (a) check that the typealias is public if the
extension has any public members, and (b) somehow ensure there is no
circularity issue (either by not importing the extension as a result
of importing the typealias, or by the extension being able to set its
sugared base type later).
rdar://problem/29694978
Two of them are user-facing, with the following sort of message:
If you're seeing a crash here, check that your SDK and
dependencies match the versions used to build 'SwiftLib'
Prompted by rdar://problem/28282310, which took a while to figure out.
The added test case is a simplified version of the issue. (Obviously
we'd prefer to not crash here, but that's hard---there's an inherited
conformance that's no longer valid, and there may be generic types
depending on that conformance.)
Extend NSNumber bridging to cover not only `Int`, `UInt`, `Double`, and `Bool`, but all of the standard types as well. Extend the `TypePreservingNSNumber` subclass to accommodate all of these types, so that we preserve type identity for `AnyHashable` and dynamic casting of Swift-bridged NSNumbers. If a pure Cocoa NSNumber is cast, just trust that the user knows what they're doing.
This XFAILs a couple of serialization tests that attempt to build the Foundation overlay, but which don't properly handle `gyb` files.
