Instead of the code querying the compiler's built-in Clang instance, refactor the
dependency scanner to explicitly keep track of module output path. It is still
set according to '-module-cache-path' as it has been prior to this change, but
now the scanner can use a different module cache for scanning PCMs, as specified
with '-clang-scanner-module-cache-path', without affecting module output path.
Resolves rdar://113222853
Unlike `swift-frontend -scan-dependencies` option, when dependency
scanner is used as a library by swift driver, the SwiftScanningService
is shared for multiple driver invocations. It can't keep states (like
common file dependencies) that can change from one invocation to
another.
Instead, the clang/SDK file dependencies are computed from each driver
invocations to avoid out-of-date information when scanning service is
reused.
The test case for a shared Service will be added to swift-driver repo
since there is no tool to test it within swift compiler.
Reformatting everything now that we have `llvm` namespaces. I've
separated this from the main commit to help manage merge-conflicts and
for making it a bit easier to read the mega-patch.
This is phase-1 of switching from llvm::Optional to std::optional in the
next rebranch. llvm::Optional was removed from upstream LLVM, so we need
to migrate off rather soon. On Darwin, std::optional, and llvm::Optional
have the same layout, so we don't need to be as concerned about ABI
beyond the name mangling. `llvm::Optional` is only returned from one
function in
```
getStandardTypeSubst(StringRef TypeName,
bool allowConcurrencyManglings);
```
It's the return value, so it should not impact the mangling of the
function, and the layout is the same as `std::optional`, so it should be
mostly okay. This function doesn't appear to have users, and the ABI was
already broken 2 years ago for concurrency and no one seemed to notice
so this should be "okay".
I'm doing the migration incrementally so that folks working on main can
cherry-pick back to the release/5.9 branch. Once 5.9 is done and locked
away, then we can go through and finish the replacement. Since `None`
and `Optional` show up in contexts where they are not `llvm::None` and
`llvm::Optional`, I'm preparing the work now by going through and
removing the namespace unwrapping and making the `llvm` namespace
explicit. This should make it fairly mechanical to go through and
replace llvm::Optional with std::optional, and llvm::None with
std::nullopt. It's also a change that can be brought onto the
release/5.9 with minimal impact. This should be an NFC change.
Teach swift dependency scanner to use CAS to capture the full dependencies for a build and construct build commands with immutable inputs from CAS.
This allows swift compilation caching using CAS.
Avoid parsing the syntax tree up-front, and instead
only parse it when required, which happens when either:
1. ASTGen parsing is enabled (currently disabled
by default)
2. Round trip checking is enabled for a primary
file (enabled by default in a debug build,
except when dep scanning or doing an IDE
operation)
3. We need to evaluate a macro in that file
This change therefore means that we now no longer
need to parse the syntax tree for secondary files
by default unless we specifically need to evaluate
a macro in them (e.g if we need to lookup a member
on a decl with an attached macro). And the same
for primaries in release builds.
rdar://109283847
They may be a super-set of the ones that appear in the textual interface - e.g. 'internal' imports will be contained in the adjacent binary module, but not the textual interface
For a `@Testable` import in program source, if a Swift interface dependency is discovered, and has an adjacent binary `.swiftmodule`, open up the module, and pull in its optional dependencies. If an optional dependency cannot be resolved on the filesystem, fail silently without raising a diagnostic.
On both input moduel source-files and interface files.
This currently yields dramatic scanning performance improvements at no cost - we do not require an AST during scan.
Since https://github.com/apple/swift/pull/63178 added support for Clang modules in the explicit module map, it is possible for there to be multiple modules with the same name: a Swift module and a Clang module. The current parsing logic just overwrites the corresponding entry module in a hashmap so we always only preserved the module that comes last, with the same name.
This change separates the parsing of the modulemap JSON file to produce a separate Swift module map and Clang module map. The Swift one is used by the 'ExplicitSwiftModuleLoader', as before, and the Clang one is only used to populate the ClangArgs with the requried -fmodule-... flags.
If a module was first read using the adjacent swiftmodule and then
reloaded using the swiftinterface, we would do an up to date check on
the adjacent module but write out the unit using the swiftinterface.
This would cause the same modules to be indexed repeatedly for the first
invocation using a new SDK. On the next run we would instead raad the
swiftmodule from the cache and thus the out of date check would match
up.
The impact of this varies depending on the size of the module graph in
the initial compilation and the number of jobs started at the same time.
Each SDK dependency is re-indexed *and* reloaded, which is a drain on
both CPU and memory. Thus, if many jobs are initially started and
they're all going down this path, it can cause the system to run out of
memory very quickly.
Resolves rdar://103119964.
- '-o <output_path>'
- '-disable-implicit-swift-modules'
- '-Xcc -fno-implicit-modules' and '-Xcc -fno-implicit-module-maps'
- '-candidate-module-file'
These were previously supplied by the driver. Instead, they will now be ready to be run directly from the dependency scanner's output.
This changes the scanner's behavior to "resolve" a discovered module's dependencies to a set of Module IDs: module name + module kind (swift textual, swift binary, clang, etc.).
The 'ModuleDependencyInfo' objects that are stored in the dependency scanner's cache now carry a set of kind-qualified ModuleIDs for their dependencies, in addition to unqualified imported module names of their dependencies.
Previously, the scanner's internal state would cache a module dependnecy as having its own set of dependencies which were stored as names of imported modules. This led to a design where any time we needed to process the dependency downstream from its discovery (e.g. cycle detection, graph construction), we had to query the ASTContext to resolve this dependency's imports, which shouldn't be necessary. Now, upon discovery, we "resolve" a discovered dependency by executing a lookup for each of its imported module names (this operation happens regardless of this patch) and store a fully-resolved set of dependencies in the dependency module info.
Moreover, looking up a given module dependency by name (via `ASTContext`'s `getModuleDependencies`) would result in iterating over the scanner's module "loaders" and querying each for the module name. The corresponding modules would then check the scanner's cache for a respective discovered module, and if no such module is found the "loader" would search the filesystem.
This meant that in practice, we searched the filesystem on many occasions where we actually had cached the required dependency, as follows:
Suppose we had previously discovered a Clang module "foo" and cached its dependency info.
-> ASTContext.getModuleDependencies("foo")
--> (1) Swift Module "Loader" checks caches for a Swift module "foo" and doesn't find one, so it searches the filesystem for "foo" and fails to find one.
--> (2) Clang Module "Loader" checks caches for a Clang module "foo", finds one and returns it to the client.
This means that we were always searching the filesystem in (1) even if we knew that to be futile.
With this change, queries to `ASTContext`'s `getModuleDependencies` will always check all the caches first, and only delegate to the scanner "loaders" if no cached dependency is found. The loaders are then no longer in the business of checking the cached contents.
To handle cases in the scanner where we must only lookup either a Swift-only module or a Clang-only module, this patch splits 'getModuleDependencies' into an alrady-existing 'getSwiftModuleDependencies' and a newly-added 'getClangModuleDependencies'.
Introduce a new source file kind to describe source files for macro
expansions, and include the macro expression that they expand. This
establishes a "parent" relationship
Also track every kind of auxiliary source file---whether for macro
expansions or other reasons---that is introduced into a module, adding
an operation that allows us to find the source file that contains a
given source location.
Previously, when evaluating a `#if canImport(Module, _version: 42)` directive the compiler could diagnose and ignore the directive under the following conditions:
- The associated binary module is corrupt/bogus.
- The .tbd for an underlying Clang module is missing a current-version field.
This behavior is surprising when there is a valid `.swiftinterface` available and it only becomes apparent when building against an SDK with an old enough version of the module that the version in the `.swiftinterface` is too low, making this failure easy to miss. Some modules have different versioning systems for their Swift and Clang modules and it can also be intentional for a distributed binary `.swiftmodule` to contain bogus data (to force the compiler to recompile the `.swiftinterface`) so we need to handle both of these cases gracefully and predictably.
Now the compiler will enumerate all module loaders, ask each of them to attempt to parse the module version and then consistently use the parsed version from a single source. The `.swiftinterface` is preferred if present, then the binary module if present, and then finally the `.tbd`. The `.tbd` is still always used exclusively for the `_underlyingVersion` variant of `canImport()`.
Resolves rdar://88723492
When looking for a Swift module on disk, we were scanning all module search paths if they contain the module we are searching for. In a setup where each module is contained in its own framework search path, this scaled quadratically with the number of modules being imported. E.g. a setup with 100 modules being imported form 100 module search paths could cause on the order of 10,000 checks of `FileSystem::exists`. While these checks are fairly fast (~10µs), they add up to ~100ms.
To improve this, perform a first scan of all module search paths and list the files they contain. From this, create a lookup map that maps filenames to the search paths they can be found in. E.g. for
```
searchPath1/
Module1.framework
searchPath2/
Module1.framework
Module2.swiftmodule
```
we create the following lookup table
```
Module1.framework -> [searchPath1, searchPath2]
Module2.swiftmodule -> [searchPath2]
```
These kinds of modules differ from `SwiftTextual` modules in that they do not have an interface and have source-files.
It is cleaner to enforce this distinction with types, instead of checking for interface optionality everywhere.
We should hold off actually building the binary module file until it is imported.
`canImport` queries can happen, for example, during dependency scanning, when we do not wish to have the scanner tool execute any module builds.
Resolves rdar://82603098
The dependency scanner's cache persists across different queries and answering a subsequent query's module lookup with a module not in the query's search path is not correct.
For example, suppose we are looking for a Swift module `Foo` with a set of search paths `SP`.
And dependency scanner cache already contains a module `Foo`, for which we found an interface file at location `L`. If `L`∉`SP`, then we cannot re-use the cached entry because we’d be resolving the scanning query to a filesystem location that the current scanning context is not aware of.
Resolves rdar://81175942
This mechanism allows the compiler to use a backup interface file to build into a binary module when
a corresponding interface file from the SDK is failing for whatever reasons. This mechansim should be entirely opaque
to end users except several diagnostic messages communicating backup interfaces are used.
Part of rdar://77676064
canImport should be able to take an additional parameter labeled by either version or
underlyingVersion. We need underlyingVersion for clang modules with Swift overlays because they
have separate version numbers. The library users are usually interested in checking the importability
of the underlying clang module instead of its Swift overlay.
Part of rdar://73992299
This matches the behavior of the current client (`swift-driver`) and reduces ambiguity in how the nodes in the graph are to be treated. Swift dependencies with a textual interface, for example, must be built into a binary module by clients. Swift dependencies without a textual interface, with only a binary module, are to be used directly, without any up-to-date checks.
Note, this is distinct from Swift dependencies that have a textual interface, for which we also detect potential pre-build binary module candidates. Those are still reported in the `details` field of textual Swift dependencies as `prebuiltModuleCandidates`.
This refactoring allows us to drop ModuleInterfaceLoader when explicit modules
are enabled. Before this change, the dependencies scanner needs the loader to be
present to access functionalities like collecting prebuilt module candidates.
This is meant to address a problem that arises on incremental package builds:
A re-scan on an already-built package instead of a placeholder dependency produces a graph that contains a dependency consisting solely of the previously-built swiftmodule.
This is not the behaviour we would like. Instead, we should respect that this is a placeholder dependency and ensure that the dependency graph for that dependency itself captures the fact that a previously-built module exists using compiledModuleCandidates field of the dependency graph.
In the fast dependency scanner, depending on whether a module intrface was found via the import search path or framework search path, encode into the dependency graph Swift module details, whether a given module is a framework.
Some implicitly imported modules aren't printed in the textual interface file as explicit import,
e.g. SwiftOnoneSupport. We should check implicit imports and add them to the dependency graph.
