If possible, add imported members to the StructDecl's LookupTable rather than adding them directly as members. This will fix the issues with ordering that #39436 poorly attempted to solve during IRGen.
This also allows us to break out most of the test changes from #39436.
This improves the test pass rate for the IDE tests on Windows. Some
failures remain. Tests which expect the compiler to be built with
libxml2 cause 2 failures. Another set of tests fail due to `stdint.h`
not being accessible to Windows due to include path ordering. Some
other failures seem to stem from incomplete processing of sources.
Technically, these operations belong in the ObjectiveC module, where NSObject
is defined. Keep them there. However, we need to build the mock ObjectiveC
overlay with `-disable-objc-attr-requires-foundation-module` now.
When the swift_bridged_typedef attribute is present on a typedef,
import the underlying type as bridged (e.g., String) rather than as
its unbridged type (e.g., NSString).
Fixes rdar://problem/39497900.
Right now the ClangImporter relies on Sema to synthesize typealiases
in some cases, which doesn't work if a Sema instance is not available,
for example when the ClangImporter is asked to import a declaration
while deserializing SIL during optimization.
Begin addressing this by synthesizing typealiases for the following
conformances:
- ExpressibleByArrayLiteral.ArrayLiteralElement while importing
OptionSets.
- RawRepresentable.RawValue while importing raw-valued enums.
- ObjectiveCBridgeable._ObjectiveCType while importing bridged
newtypes.
Until further changes land, this is mostly NFC, except for a change
in generated interface printing where the new typealiases are now
explicitly shown.
Note that non-bridged newtypes whose raw type is ObjectiveCBridgeable
still rely on associated type inference, because the ClangImporter
cannot safely "copy" the _ObjectiveCType typealias from the raw type
to the newtype, for various reasons mostly having to do with
circularity. Subsequent patches will address this, in a rather novel
fashion that will shock you.
Ever since we stopped associating the top-level struct of an imported
NS_ERROR_ENUM with the Clang enum declaration, we've been unable to
print imported NS_ERROR_ENUMs. The module-printing infrastructure
would drop them thinking they aren't imported declarations.
This also affected NS_ERROR_ENUMs that were imported as members of
another type, as well as other types imported as members.
Fixes rdar://problem/32497693.
Rather than requiring associated type witness inference to go and
figure out the Element type from the other witnesses for imported
OptionSet types, synthesize the typealias directly in the
importer. This is a simplification and a performance optimization.
Rather than requiring associated type witness inference to go and
figure out the RawValue type from the rawValue witnesses for imported,
RawRepresentable types, synthesize the typealias directly in the
importer. This is a simplification and a performance optimization.
We already have an option to skip /all/ unavailable decls. Singling
out the ones that are specifically unavailable-in-Swift just makes
testing harder.
This will not affect interface generation in Xcode, which sets that
option to skip all unavailable decls.
Switch printing off of using Function's ExtInfo for autoclosure and
escaping, and onto the ParameterTypeFlags, which let us do precise and
accurate context-sensitive printing of these parameter type
attributes. This fixes a huge list of issues where we were printing
@escaping for things like optional ObjC completion handlers, among
many others. We now correctly print @escaping in more places, and
don't print it when it's not correct.
Also updates the dumper to be consistent and give a good view of the
AST as represented in memory. Tests updated, more involved testing
coming soon.
* [ClangImporter] Remove importer-based NS stripping.
As Tony puts it, in the end we wound up with more Foundation
declarations imported as members or keeping "NS" than those that
dropped it, and any further decisions will be made on a case-by-case
basis. Move all of the existing cases of prefix-stripping into
Foundation's API notes and drop the logic from the compiler.
Tested by dumping the generated interface for Foundation and its
submodules for both macOS and the iOS simulator, and comparing the
results. A few cases did slip through here because of the interaction
between "SwiftName" and "Availability: nonswift".
The next commit will re-add "NS" to some stragglers that we missed.
rdar://problem/26880017
* APINotes: Add "NS" back to a few types.
NSKeyedUnarchiverDelegate
NSKeyedArchiverDelegate
NSTextCheckingTypes
NSBinarySearchingOptions
NSEnumerationOptions
NSSortOptions
More rdar://problem/26880017
* Remove now-redundant SwiftNames from API notes.
No change observed in the generated interface of Foundation and its
submodules.
Finishes rdar://problem/26880017.
This flips the switch to have @noescape be the default semantics for
function types in argument positions, for everything except property
setters. Property setters are naturally escaping, so they keep their
escaping-by-default behavior.
Adds contentual printing, and updates the test cases.
There is some further (non-source-breaking) work to be done for
SE-0103:
- We need the withoutActuallyEscaping function
- Improve diagnostics and QoI to at least @noescape's standards
- Deprecate / drop @noescape, right now we allow it
- Update internal code completion printing to be contextual
- Add more tests to explore tricky corner cases
- Small regressions in fixits in attr/attr_availability.swift
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.
If a particular typealias is unavailable in Swift, don't form types
that reference that typealias. Use the underlying type instead.
Fixes rdar://problem/25450337.
Similar with @keyword, manifesting @recommended and @recommendedover content in code
completion results can help IDE users to choose the right API in the long candidate list.
This commit extract these two attributes from Clang doc comments and insert/cache them in
code completion results.
rdar://23101030 and rdar://23101029
Conventionally, code completion results are matched with user input solely by
names. However, names are limited in expressiveness. From this comments, we start to
decorate code completion results with @keywords fields extracted from Clang doc comments.
These fields are added by API authors to comment the decl with information that
is not manifested clear enough through names. Code completion users' typing of the
keyword leads to the corresponding code completion results being selected as well.
Keywords can be arbitrarily long and can be multiple.
For instance, a function called "index()" has "@keyword find" in its doc comment.
Users' typing of "find" leads to "index()" being selected in the code completion list.
...and add one extra check for invalid macro info.
This reinstates the tests that were disabled in my previous commit, now that the
Clang issue has been fixed.
Finishes rdar://problem/21480635.
Patch by Jordan Rose.
Swift SVN r29705
This was broken in upstream Clang (LLVM PR23929). When it's fixed, this commit
should be reverted to restore all the disabled tests. See rdar://problem/21480635.
Patch by Jordan Rose.
Swift SVN r29704
This came out of today's language review meeting.
The intent is to match #available with the attribute
that describes availability.
This is a divergence from Objective-C.
Swift SVN r28484
The rule changes are as follows:
* All functions (introduced with the 'func' keyword) have argument
labels for arguments beyond the first, by default. Methods are no
longer special in this regard.
* The presence of a default argument no longer implies an argument
label.
The actual changes to the parser and printer are fairly simple; the
rest of the noise is updating the standard library, overlays, tests,
etc.
With the standard library, this change is intended to be API neutral:
I've added/removed #'s and _'s as appropriate to keep the user
interface the same. If we want to separately consider using argument
labels for more free functions now that the defaults in the language
have shifted, we can tackle that separately.
Fixes rdar://problem/17218256.
Swift SVN r27704
This reverts commits r26369 & r26381.
After upstream clang changes, reverted r25843 which was compensating for a previous upstream clang change.
But keep test/IDE/complete_from_clang_framework.swift disabled for further investigation, because it
seems to have some duplications in a code-completion test.
rdar://20245658&20247922
Swift SVN r26430
Move the availability printing tests from IDE/print_clang_decls.swift to
IDE/print_clang_framework.swift (which is only run on OS X). This commit should unbreak
the iOS build bots, which I broke in r24973.
This commit does not address Jordan's concerns with the tests. I will do that in a latter
commit.
Swift SVN r24985
