Currently symbol graphs are always written in files that contain 1 to 2
module names. It's possible for Swift module names to be very long, so
combining 2 of them in file name like `module1@module2...` in the same
path component means the name can be too long for some file systems. The
new option `-symbol-graph-shorten-output-names` changes the symbol graph
output files to use a MD5 hash of the module name(s) as the filename and
outputs an additional JSON file with the original names mapped to the
real filename. The module names JSON can be used to construct a VFS
overlay with the original naming scheme.
fix#83723
I considered using vfsoverlay, which seems like a viable solution, but
the vfsoverlay options don't seem to apply to any of the outputs from
the compiler. When I set an overlay to remap the symbol graph file
outputs, the remapped external paths aren't used so the root problem of
too long file names remains.
Replace the one-off compiler flag for Library Evolution with an
optional language feature. This makes the
`hasFeature(LibraryEvolution)` check work in an `#if`, and is
otherwise just cleanup.
Tracked by rdar://161125572.
Textual interfaces for 'Darwin' built with recent compilers specify that it is built witout C++ interop enabled. However, to ensure compatibility with versions of the 'Darwin' module built with older compilers, we hard-code this fact. This is required to break the module cycle that occurs when building the 'Darwin' module with C++ interop enabled, where the underlying 'Darwin' clang module depends on C++ standard library for which the compiler brings in the 'CxxStdlib' Swift overlay, which depends on 'Darwin'.
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
Updates swift-symbolgraph-extract to parse "-cxx-interoperability-mode"
flags and update the underlying compiler invocation. This fixes a bug
where were are unable to extract the symbol graph from swiftmodules with
transitive cxx modules because we parsed cxx headers as c headers.
rdar://128888548 (Add support for parsing cxx headers)
`getValue` -> `value`
`getValueOr` -> `value_or`
`hasValue` -> `has_value`
`map` -> `transform`
The old API will be deprecated in the rebranch.
To avoid merge conflicts, use the new API already in the main branch.
rdar://102362022
This includes:
- bumping the SWIFT_SYMBOLGRAPH_FORMAT_MINOR version
- introduction of the "swift.extension" symbol and "extensionTo" relationship
- adding support for ExtensionDecl to the Symbol class
- adding a "typeKind" field to the symbol's extension mixin which indicates what kind
of symbol was extended
- intoduction of the -emit-extension-block-symbols flag, which enables the behavior
outlined below
- adaptions to SymbolGraphASTWalker that ensure a swift.extension symbol is emitted
for each extension to a type that does not exist in the local symbol graph
- adaptions to SymbolGraph and SymbolGraphASTWalker that ensure member and conformance
relationships are correctly associated with the swift.extension symbol instead of
the original type declaration's (extended nominal's) symbol where applicable
- adaptions to SymbolGraphASTWalker that ensure swift.extension symbols are connected
to their respective extended nominal's symbol using an extensionTo relationship
Testing:
- adds SymbolGraph tests that test behavior only relevant in
-emit-extension-block-symbols mode
- adapts some SymbolGraph tests to additionally test similar behavior for
extensions to external types in -emit-extension-block-symbols mode
- adapts some SymbolGraph tests to (additionally or exclusively) test the
behavior with -emit-extension-block-symbols mode enabled
Bugfixes:
- fixes a bug where some conformsTo relationships implicated by the conformances
declared on an extension to an external type were not emitted
(see test/SymbolGraph/Relationships/ConformsTo/Indirect.swift)
Further changes:
- documents the strategy for naming and associating children declared in extensions
to typealiases (see test/SymbolGraph/Relationships/MemberOf/Typealias.swift,
test/SymbolGraph/Symbols/Names.swift)
Basic should not be allowed to link Parse, yet it was doing so
to allow Version to provide a constructor that would conveniently
parse a StringRef. This entrypoint also emitted diagnostics, so it
pulled in libAST.
Sink the version parser entrypoint down into Parse where it belongs
and point all the clients to the right place.
With the change to include `SmallVector.h` directly in `LLVM.h` rather
than forward declaring in the only case it matters (ie. Clang <= 5),
these fixes are no longer needed. Since defaulted version is preferred
when there's no better choice (which is presumably the case if that's
how they were originally added), use it instead. Some uses were instead
changed to add `llvm::` so remove that too.
This was added recently but not initialized in symbolgraph-extract.
We never output Clang imported symbols from this tool, so always passing
false is fine for now.
Instead of checking that the stdlib can be loaded in a variety of places, check it when setting up the compiler instance. This required a couple more checks to avoid loading the stdlib in cases where it’s not needed.
To be able to differentiate stdlib loading failures from other setup errors, make `CompilerInstance::setup` return an error message on failure via an inout parameter. Consume that error on the call side, replacing a previous, more generic error message, adding error handling where appropriate or ignoring the error message, depending on the context.
This patch adds a number of stack elements to the number of overlays in
the swift_symbolgraph_extract tool. We don't get the defaults from the
llvm type on Linux so it needs to be manually specified.
I never saw the number of overlays go above zero when dumping while
running the test suite, so I don't know what a "reasonable" number looks
like here. 8 seemed like a good number at the time.
