Add '-validate-clang-modules-once' and '-clang-build-session-file' corresponding to Clang's '-fmodules-validate-once-per-build-session' and '-fbuild-session-file='. Ensure they are propagated to module interface build sub-invocations.
We require these to be first-class Swift options in order to ensure they are propagated to both: ClangImporter and implicit interface build compiler sub-invocations.
Compiler portion of rdar://105982120
Use the new "grouped diagnostics" feature of the swift-syntax
diagnostic rendering to emit printed diagnostics under the
swift-syntax diagnostic style. This emits macro expansion buffers as
text to the terminal, inset in a box where the macro was expanded, so
that there is more context for understanding how the macro was
expanded and what went wrong inside it.
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.
Previously, combining `-enable-lexical-lifetimes=false` with
`-enable-experimental-move-only` resulted in an error. But this is a
valid combination: the former sets the `LexicalLifetimes` option to
`DiagnosticMarkersOnly`, which is what `-enable-experimental-move-only`
relies on.
Add a compiler option `-load-plugin-executable <path>#<module names>`.
Where '<path>' is a path to a plugin executable, '<module-name>' is a
comma-separated module names the plugin provides.
Nothing is using it at this point. Actual plugin infratructure are
introduced in follow-up commits
In https://github.com/apple/swift/pull/42486 new behavior was introduced to ignore adjacent .swiftmodule files in the SDK. This behavior has caught a few people off guard so it seems like there should be diagnostics clarifying why a rebuild is occurring in this scenario.
Resolves rdar://105477473
Macro expansion buffers, along with other generated source buffers,
need more precise "original source ranges" that can be had with the
token-based `SourceRange`. Switch over to `CharSourceRange` and provide
more thoughtfully-determined original source ranges.
Introduce `-plugin-path <path>` to add a search path where we will look
for compiler plugins. When resolving an external macro definition, look
for libraries in these search paths whose names match the module name
of the macro.
Implements rdar://105095761.
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.
Clang no longer supports `-working-directory<path>`. Split this argument
into `-working-directory <path>` so that clients passing the old
argument still work.