Using IncludeTree::FileList to concat the include tree file systems that
are passed on the command-line. This significantly reduce the
command-line size, and also makes the cache key computation a lot
faster.
rdar://148752988
(cherry picked from commit 201e4faea7)
With `ARCMigrate` and `arcmt-test` removed from clang in
https://github.com/llvm/llvm-project/pull/119269 and the new code
migration experience under way (see
https://github.com/swiftlang/swift-evolution/pull/2673), these options
are no longer relevant nor known to be in use. They were introduced
long ago to support fix-it application in Xcode.
For now, turn them into a no-op and emit a obsoletion warning.
(cherry picked from commit 46c394788a84d5932289c71274dd32ea2d61d9dc)
In expectation, this should never happen. Such a situation means that within the same scanning action, Clang Dependency Scanner has produced two different variants of the same module. This is not supposed to happen, but we are currently hunting down the rare cases where it does, seemingly due to differences in Clang Scanner direct by-name queries and transitive header lookup queries.
Improve diagnostics message for swift caching build by trying to emit
the diagnostics early when there is more context to differentiate the
different kind of problems.
After the improvement, CAS Error should be more closer to when there is
functional problem with the CAS, rather than mixing in other kinds of
problem (like scanning dependency failures) when operating with a CAS.
rdar://145676736
(cherry picked from commit 226552bf23)
It is possible for a module interface (e.g., ModuleA) to be generated
with C++ interop disabled, and then rebuilt with C++ interop enabled
(e.g., because ModuleB, which imports ModuleA, has C++ interop enabled).
This circumstance can lead to various issues when name lookup behaves
differently depending on whether C++ interop is enabled, e.g., when
a module name is shadowed by a namespace of the same name---this only
happens in C++ because namespaces do not exist in C. Unfortunately,
naming namespaces the same as a module is a common C++ convention,
leading to many textual interfaces whose fully-qualified identifiers
(e.g., c_module.c_member) cannot be correctly resolved when C++ interop
is enabled (because c_module is shadowed by a namespace of the same
name).
This patch does two things. First, it introduces a new frontend flag,
-formal-cxx-interoperability-mode, which records the C++ interop mode
a module interface was originally compiled with. Doing so allows
subsequent consumers of that interface to interpret it according to the
formal C++ interop mode. Note that the actual "versioning" used by this
flag is very crude: "off" means disabled, and "swift-6" means enabled.
This is done to be compatible with C++ interop compat versioning scheme,
which seems to produce some invalid (but unused) version numbers. The
versioning scheme for both the formal and actual C++ interop modes
should be clarified and fixed in a subsequent patch.
The second thing this patch does is fix the module/namespace collision
issue in module interface files. It uses the formal C++ interop mode to
determine whether it should resolve C++-only decls during name lookup.
For now, the fix is very minimal and conservative: it only filters out
C++ namespaces during unqualified name lookup in an interface that was
originally generated without C++ interop. Doing so should fix the issue
while minimizing the chance for collateral breakge. More cases other
than C++ namespaces should be added in subsequent patches, with
sufficient testing and careful consideration.
rdar://144566922
Most SDKs use only swiftinterfaces under usr/lib/swift. Let's make sure
we standardize this behavior and use only swiftinterface when they are
present, even if there are also binary swiftmodule files available.
Apply the same logic to SubFrameworks as well while we're at it.
rdar://145316821
Batch dependency scanning was added as a mechanism to support multiple compilation contexts within a single module dependency graph.
The Swift compiler and the Explicitly-built modules model has long since abandoned this approach and this code has long been stale. It is time to remove it and its associated C API.
Specifically, this means `-dump-ast-format json` is
incompatible with `-dump-parse`. This is because the JSON
format is meant to export more details about the AST that
require type checking to have been performed.
I'm open to lifting this restriction in the future.
This only takes the existing AST information and writes it as JSON
instead of S-expressions. Since many of these fields are stringified,
they're not ideal for the kind of analysis clients of the JSON format
would want to do. A future commit will update these values to use a
more structured representation.
Parsing for `-enable-upcoming-feature` and `-enable-experimental-feature` is
lenient by default because some projects need to be compatible with multiple
language versions and compiler toolchains simultaneously, and strict
diagnostics would be a nuisance. On the other hand, though, it would be useful
to get feedback from the compiler when you attempt to enable a feature that
doesn't exist. This change splits the difference by introducing new diagnostics
for potential feature enablement misconfigurations but leaves those diagnostics
ignored by default. Projects that wish to use them can specify `-Wwarning
StrictLanguageFeatures`.
If the output loading failed after cache key lookup, treat that as a
warning and resume as if that is a cache miss. This is not a valid
configuration for builtin CAS but can happen for a remote CAS service
that failed to serve the output. Instead of failing, we should continue
to compile to avoid disruptive failures.
rdar://140822432
Add flag `-load-resolved-plugin` to load macro plugin, which provides a
pre-resolved entry into PluginLoader so the plugins can be loaded based
on module name without searching the file system. The option is mainly
intended to be used by explicitly module build and the flag is supplied
by dependency scanner.
Its functionality has been superseded by `@_spiOnly import`. There are no
longer any known clients and the flag was already unsupported in Swift 6, so
the functionality is now removed (but the flag is only deprecated for Swift 5).
Resolves rdar://136867210.
This commit adds new compiler options -no-warning-as-error/-warning-as-error which allows users to specify behavior for exact warnings and warning groups.
This PR ensures library-evolution is enabled for Package CMO; without it,
it previously fell back to regular CMO, which caused mismatching serialization
attributes if importing another module that had Package CMO enbaled, causing
an assert fail for loadable types.
Resolves rdar://135308288
Foundation needs to be loaded early in the process for Swift's runtime
to properly initialize bridging support; otherwise it may cause issues
like unrecognized selectors. When scripting, load Foundation early in
performFrontend before any swift code runs.
rdar://129528115
Now that API descriptions are emitted during module build jobs when
`-emit-api-descriptor-path` is specified and the build system has been updated
to pass that flag when the output is needed, the `swift-api-extract` frontend
alias is no longer used. Delete it and the tests that were specific to invoking
`swift-api-extract`.
Resolves rdar://116537394.
Teach dependency scanner to report all the module canImport check result
to swift-frontend, so swift-frontend doesn't need to parse swiftmodule
or parse TBD file to determine the versions. This ensures dependency
scanner and swift-frontend will have the same resolution for all
canImport checks.
This also fixes two related issues:
* Previously, in order to get consistant results between scanner and
frontend, scanner will request building the module in canImport check
even it is not imported later. This slightly alters the definition of
the canImport to only succeed when the module can be found AND be
built. This also can affect the auto-link in such cases.
* For caching build, the location of the clang module is abstracted away
so swift-frontend cannot locate the TBD file to resolve
underlyingVersion.
rdar://128067152
Enabling `@_spiOnly` also enables stronger type-checking of SPI decls.
As this could be source breaking, it has always been opt-in. Turn it on
by default in Swift 6 mode where the stronger type-checking will also
become expected.
At the same time, disable the old alternative to `@_spiOnly` which was
designed to be compatible with old compilers. Any user of that feature
should move to `@_spiOnly` or `package import` instead.
we only check if the loaded module is built from a package interface. This is
not enough as a binary module could just contain exportable decls if built with
experimental-skip-non-exportable-decls, essentially resulting in content equivalent
to interface content. This might be made a default behavior so this PR requires
a module to opt in to allow non-resilient access by a participating client in the
same package.
Since it affects module format, SWIFTMODULE_VERSION_MINOR is updated.
rdar://123651270
The Swift compiler can load either the binary swiftmodule file or the
textual swiftinterface file when importing a module. It currently picks
the swiftmodule over the swiftinterface, unless there’s an exception. We
should flip the default for distributed modules, prefer the
swiftinterface over the swiftmodule unless there’s an exception.
rdar://122955640
The `-experimental-lazy-typecheck` and `-experimental-skip-non-exportable-decls`
flags are not safe to use when emitting a non-resilient module because the
clients of non-resilient modules expect to have access to all the members of a
type in order to e.g. compute the size the type. The
`-experimental-skip-non-exportable-decls` flag skips serialization of
non-public members and would therefore cause mis-compilation. The
`-experimental-lazy-typecheck` is theoretically safe for non-resilient modules
but more requestification work is needed before it can be used successfully.
Resolves rdar://122272758
When loading input from CAS, `swift-frontend` relies on the input file
name to determine the type to look from CAS entry. In the case where
file extension is `.private.swiftinterface`, swift mis-identify that as
`.swiftinterface` file and look up the wrong input file. Add a new
file type lookup function that can figure out the type from the full
filename.
Also add few diagnostics during the CAS lookup for the input file to
error out immediately, rather than rely on the lookup failure later.
Diagnose rather than exit when using experimental features that cannot
be enabled in production. Also diagnose if using a compiler without
`SwiftCompilerSources` when enabling embedded.
This is especially important for long-running services like sourcekitd,
which shouldn't exit just because an invalid argument was passed.
Resolves rdar://119724074.
