- 'SwiftModuleScanner' will now be owned directly by the 'ModuleDependencyScanningWorker' and will contain all the necessary custom logic, instead of being instantiated by the module interface loader for each query
- Moves ownership over module output path and sdk module output path directly into the scanning worker, instead of the cache
Initially, the compiler rejected building dependencies that contained OS
versions in an invalid range. However, this happens to be quite
disruptive, so instead allow it and request that these versions be
implicitly bumped based on what `llvm::Triple::getCanonicalVersionForOS`
computes.
resolves: rdar://153205856
Adds an access control field for each imported module identified. When multiple imports of the same module are found, this keeps track of the most "open" access specifier.
Unlike with implicitly-built modules (prior to Swift 6 mode), explicitly-built modules require that all search paths be specified explicitly and no longer inherit search paths serialized into discovered Swift binary modules. This behavior was never intentional and is considered a bug. This change adds a diagnostic note to a scan failure: for each binary Swift module dependency, the scanner will attempt to execute a dependency scanning query for each serialized search path inside that module. If such diagnostic query returns a result, a diagnostic will be emitted to inform the user that the dependency may be found in the search path configuration of another Swift binary module dependency, specifying which search path contains the "missing" module, and stating that such search paths are not automatically inherited by the current compilation.
-nostdimport and -nostdlibimport only remove the toolchain and usr/lib/swift search paths, and they leave the framework search paths intact. That makes it impossible to get a fully custom SDK environment. Make their behavior match clang's -nostdinc/-nostdlibinc behavior: treat framework and non-framework paths the same. In other words, -nostdinc removes *all* compiler provided search paths, and -nostdlibinc removes *all* SDK search paths.
Rename SkipRuntimeLibraryImportPaths to SkipAllImportPaths, and ExcludeSDKPathsFromRuntimeLibraryImportPaths to SkipSDKImportPaths to reflect their updated behavior.
Move the DarwinImplicitFrameworkSearchPaths handling from SearchPathOptions to CompilerInvocation, where RuntimeLibraryImportPaths is managed. Rename it to just ImplicitFrameworkSearchPaths, and filter for Darwin when it's set up so that all of the clients don't have to do Darwin filtering themselves later.
rdar://150557632
For now the semantics provided by `@extensible` keyword on per-enum
basis. We might return this as an upcoming feature in the future with
a way to opt-out.
As the old comment indicates, it was not expected that build systems
would register binary swift modules produced from textual interfaces
with -add_ast_path, but we have found examples of the in the wild and
it leads to the most unexpected side effects. Fixing this is
straightforward, we can just check if a binary module was compiled
from a textual interface and then not bypass resilience.
rdar://145226754
When `ExtensibleEnums` flag is set, it's going to be reflected in
the module file produced by the compiler to make sure that consumers
know that non-`@frozen` enumerations can gain new cases in the
future and switching cannot be exhaustive.
Add ability to automatically chaining the bridging headers discovered from all
dependencies module when doing swift caching build. This will eliminate all
implicit bridging header imports from the build and make the bridging header
importing behavior much more reliable, while keep the compatibility at maximum.
For example, if the current module A depends on module B and C, and both B and
C are binary modules that uses bridging header, when building module A,
dependency scanner will construct a new header that chains three bridging
headers together with the option to build a PCH from it. This will make all
importing errors more obvious while improving the performance.
This failure will most-likely result in the dependency query failure which will fail the scan. It will be helpful if the scanner emitted diagnostic for each such module it rejected to explain the reason why.
Resolves rdar://142906530
In https://github.com/swiftlang/swift/pull/77156, normalization was introduced
for -target-variant triples. That PR also caused -target-variant arguments to
be inherited from the main compilation options whenever building dependency
modules from their interfaces, which is incorrect. The -target-variant option
must only be specified when compiling a "zippered" module, but the dependencies
of zippered modules are not necessarily zippered themselves and
indiscriminantly propagating the option can cause miscompilation.
The new, more targeted approach to normalizing arm64e triples simply uses the
arch and subarch of the -target argument of the main compile to decide whether
the subarch of both the -target and -target-variant arguments of a dependency
need adjustment.
Resolves rdar://135322077 and rdar://141640919.
Extend the module trace format with a field indicating whether a given
module, or any module it depends on, was compiled with strict memory
safety enabled. This separate output from the compiler can be used as
part of an audit to determine what parts of Swift programs are built
with strict memory safety checking enabled.
When Swift passes search paths to clang, it does so directly into the HeaderSearch. That means that those paths get ordered inconsistently compared to the equivalent clang flag, and causes inconsistencies when building clang modules with clang and with Swift. Instead of touching the HeaderSearch directly, pass Swift search paths as driver flags, just do them after the -Xcc ones.
Swift doesn't have a way to pass a search path to clang as -isystem, only as -I which usually isn't the right flag. Add an -Isystem Swift flag so that those paths can be passed to clang as -isystem.
rdar://93951328
Sema now type-checks the alternate ABI-providing decls inside of @abi attributes.
Making this work—particularly, making redeclaration checking work—required making name lookup aware of ABI decls. Name lookup now evaluates both API-providing and ABI-providing declarations. In most cases, it will filter ABI-only decls out unless a specific flag is passed, in which case it will filter API-only decls out instead. Calls that simply retrieve a list of declarations, like `IterableDeclContext::getMembers()` and friends, typically only return API-providing decls; you have to access the ABI-providing ones through those.
As part of that work, I have also added some basic compiler interfaces for working with the API-providing and ABI-providing variants. `ABIRole` encodes whether a declaration provides only API, only ABI, or both, and `ABIRoleInfo` combines that with a pointer to the counterpart providing the other role (for a declaration that provides both, that’ll just be a pointer to `this`).
Decl checking of behavior specific to @abi will come in a future commit.
Note that this probably doesn’t properly exercise some of the new code (ASTScope::lookupEnclosingABIAttributeScope(), for instance); I expect that to happen only once we can rename types using an @abi attribute, since that will create distinguishable behavior differences when resolving TypeReprs in other @abi attributes.
This change ensures that when loading some module dependency 'Bar' which has a package-only dependency on 'Foo', only the following clients attempt to resolve/load 'Foo':
- Source compilation with package-name equal to that of 'Bar'.
- Textual interface compilation of a *'package'* interface with package-name equal to that of 'Bar'.
Ensuring that the following kinds of clients do not attempt to resolve/load 'Foo':
- Source compilation with package-name different to that of 'Bar'
- Textual interface compilation of a public or private interface, regardless of package name.
This fixes the behavior where previously compilation of a Swift textual interface dependency 'X' from its public or private interface, with an interface-specified package-name, from a client without a matching package-name, resulted in a lookup of package-only dependencies of modules loaded into 'X'. This behavior is invalid if we are not building from the package textual interface, becuase the module dependency graph is defined by the package name of the source client, not individual module dependency package name. i.e. In-package module dependencies are resolved/loaded only if the parent source compile matches the package name.
Resolves rdar://139979180
Rather than exposing an `addFile` member on
ModuleDecl, have the `create` members take a
lambda that populates the files for the module.
Once module construction has finished, the files
are immutable.
Improve the version/flags extract from interface file by moving away
from using Regex and limiting the search to the beginning of the file.
Switch away from Regex will give 5-10% improvement in time and
instruction counts, and limiting the search lines can save a lot of time
if the swiftinterface is large. For example, the extract time for Swift
stdlib is 10x faster after the patch.
Current strategey for limiting the line to search is by only parsing the
first comment block.
It might be unexpected to future users that `-swift-compiler-version`
would produce a version aligned to .swiftinterface instead of one used
to build the .swiftmodule file. To avoid this possible confusion, let's
scope down the version to `-interface-compiler-version` flag and
`SWIFT_INTERFACE_COMPILER_VERSION` option in the module.
Based on preliminary work from @rmaz.
The compilation arguments for a swiftinterface file are preprocessed to
modify the `-target` argument to match the preferred target (which comes
from the command line) in cases in which the sub-architecture differs,
but it is compatible (for example using `arm64e` when `arm64` is being
compiled), but this was not done for the target variant, which ended up
with mismatches on the sub-architecture used by the target and target
variant, which fails an assert in assert toolchains.
Add function to handle all macro dependencies kinds in the scanner,
including taking care of the macro definitions in the module interface
for its client to use. The change involves:
* Encode the macro definition inside the binary module
* Resolve macro modules in the dependencies scanners, including those
declared inside the dependency modules.
* Propagate the macro defined from the direct dependencies to track
all the potentially available modules inside a module compilation.
When '.package.swiftinterface' loading ('-experimental-package-interface-load') is disabled and when '-scanner-module-validation' is disabled, the scanner defaults to locating the non-package textual interface and may specify its adjacent binary module as a valid candidate binary module to use. If said candidate is up-to-date and ends up getting used, and belongs to the same package as the loading Swift source, then the source compilation may attempt to load its package-only dependencies. Since the scanner only parsed the non-package textual interface, those dependencies are not located and specified as inputs to compilation. This change causes the scanner, in such cases, to also lookup package-only dependencies in adjacent binary Swift modules of textual Swift module dependencies, if such dependency belongs to the same package as the source target being scanned.
Resolves rdar://135215789
This assert was correctly catching the fact that `-target-variant` is not being
normalized at the same time as `-target` when building arm64e modules from
swiftinterface. That should be fixed, but at the moment it isn't causing any
concrete harm and the assertion fails when building against the SDKs included
with the latest Xcode 16 betas.
Resolves rdar://133020098.
This makes sure that Swift respects `-Xcc -stdlib=libc++` flags.
Clang already has existing logic to discover the system-wide libc++ installation on Linux. We rely on that logic here.
Importing a Swift module that was built with a different C++ stdlib is not supported and emits an error.
The Cxx module can be imported when compiling with any C++ stdlib. The synthesized conformances, e.g. to CxxRandomAccessCollection also work. However, CxxStdlib currently cannot be imported when compiling with libc++, since on Linux it refers to symbols from libstdc++ which have different mangled names in libc++.
rdar://118357548 / https://github.com/swiftlang/swift/issues/69825
This patch allows controlling the automatic import of private dependencies
separately from the DebuggerSupport option, which currently also triggers this
behavior. With explicit modules + precise compiler invocations LLDB is moving
towards no longer needing this behavior.
rdar://133088201
(cherry picked from commit a1ba7159e3)
As-is, this default interferes with the incremental build machinery which conservatively assumes that binary module dependencies must cause dependents to be re-built.
Add a check for the client side flag which explcitly opts in to loading the package interface,
besides whether package-name is empty or in the same package.
rdar://131393508
When the dependency scanner picks a pre-built binary module candidate for a given dependency, it needs to be able to attempt to resolve its cross-import overlays relative to the textual interface that the binary module was built from. For example, if a collection of binary modules are located in, and resolved as dependencies from, a pre-built module directory, the scanner must lookup their corresponding cross-import overlays relative to the defining interface as read out from the binary module's MODULE_INTERFACE_PATH. https://github.com/swiftlang/swift/pull/70817 ensures that binary modules serialize the path to their defining textual interface.
Resolves rdar://130778577
Fix the problem that when the only module can be found is an
invalid/out-of-date swift binary module, canImport and import statement
can have different view for if the module can be imported or not.
Now canImport will evaluate to false if the only module can be found for
name is an invalid swiftmodule, with a warning with the path to the
module so users will not be surprised by such behavior.
rdar://128876895
