Via preprocessor goop, Serialization.cpp generates a set of static
functions that can be used to verify if a set of attributes
can be serialized for a particular declaration.
This design forces the author of the attribute to specify up front
whether or not an attribute is supported on a given declaration
kind. We can possibly hoist this into semantic analysis as well.
These N separate functions do conceptually replace a simple
variadic template implementation. I'm fine with alternatives,
but the goal was to provide a way for the author of new
attributes to describe the requirements in one place: Attr.def.
Swift SVN r15470
This patch adds in the necessary infrastructure for lazily deserializing
witness tables. This is done by following the same approach as the
deserialization/serialization of SILFunction.
Now if one calls SILModule::lookUpWitnessTable and the given witness table is a
definition, the SILModule will attempt to deserialize it from one of the other
modules.
Swift SVN r15403
This representation is inspired by Clang's internal representation.
The current attribute representation, which is basically a union
of "stuff" in DeclAttributes, is not amendable to richer
attributes, such as @availability, that need to be implemented.
In Clang, attributes are modeled with actual objects that
encode both semantic and syntactic information (e.g., source ranges)
that facilitate richer checking, better diagnostics, and better tools.
This change is foundational for implementing @availability, but
also is a better long-term representation. As a migratory path,
it creates some duplications, with AttrKind and DeclAttrKind, the
two which should eventually become the same thing.
As part of this patch, there is some additional parser recovery
(for the new attribute representation) for duplicate attributes.
The parser now parses the entire duplicate attribute, which could
be quite complex, and then issues a diagnostic that the attribute
is a duplicate (and discarding it). This delayed diagnostic
also allows us to present ranges for the duplicate attribute, which
provides a better user experience.
Swift SVN r15365
This fixes the following two bugs:
1. We sometimes would create new conformances when deserializing a
witness method which would not be mapped in the SILModule to the
appropriate witness table. This would cause us to be unable to perform
devirtualization of this witness method. This is tested via a new
verifier check.
2. Different conformances would be created for an instance of a base
protocol and the original protocol. This would cause IRGen to try to
emit witness table global variables with differing types, hitting an
assertion. This is tested via a traditional test.
Swift SVN r15362
Parse function declarations with the form
func murder inRoom(room: Int) weapon(Int) {}
where the function name ("murder") is separated from the parameter
names. This is the same style used in initializers, i.e.,
init withCString(cstr: CString) encoding(Encoding)
Swift SVN r15140
In a framework containing both Clang headers and a Swift module, the Swift
module gets picked up first, but then automatically imports (and re-exports)
the Clang module as well.
One interesting case here is that it's possible for the Clang side to add
categories to a class declared in Swift. In order to support this, the
Clang importer can now find extensions for Swift classes marked @objc.
(I couldn't think of a way to test this separately; the previous commit
was supposed to do that, but worked without this change.)
<rdar://problem/16295937>
Swift SVN r15084
...due to a code signing issue. See <rdar://problem/16328065> for details.
Also, fix an error where the wrong file name was being used for framework
docs.
No release note because no one is using Swift frameworks yet (besides LLDB).
Swift SVN r15074
Instead of referencing them by name, serialize a reference to their
storage decl instead, then note which accessor to retrieve.
Uncovered by Chris's override work in r15051...thanks, Chris!
Swift SVN r15063
Previously, we were cloning the default arguments completely, which
meant code duplication (when inheriting within a module) or simply a
failure (when inheriting across modules). Now, we reference the
default arguments where we inherited them, eliminating the
duplication. Part of <rdar://problem/16318855>.
Swift SVN r15062
This will help with ensuring that we do not create multiple witness
table "definitions" one of which is null. That situtation yields an
IRGen assertion to be hit since the external declaration (in the guise
of a definition) has a different type from the actual deserialized
definition.
Swift SVN r14999
If a subclass defines no subobject initializers and all of its stored
properties have initial values, "inherit" all of the subobject
initializers of its superclass by creating a new initializer with the
same signature that overrides (and chains to) the corresponding
subobject initializer of its parent. Do this instead of blindly
creating a default initializer.
Note that we aren't yet doing this for generic initializers. That will
be a separate step.
Swift SVN r14995
The driver infers the filename from the module file by replacing the extension,
and passes the explicit path to the swiftdoc file to the frontend. But there
is no option in the driver to control emission of swiftdoc (it is always
emitted, and name is always inferred from the swiftmodule name).
The swiftdoc file consists of a single table that maps USRs to {brief comment,
raw comment}. In order to look up a comment for decl we generate the USR
first. We hope that the performance hit will not be that bad, because most
declarations come from Clang. The advantage of this design is that the
swiftdoc file is not locked to the swiftmodule file, and can be updated,
replaced, and even localized.
Swift SVN r14914
This won't ever happen naturally, but by either messing with the module file
or asking the compiler to load something that isn't a module at all we could
end up hitting the assert.
<rdar://problem/16274875>
Swift SVN r14902
In the short term, we need to be able to emit shared symbols for SILWitnessTables corresponding to Clang-imported modules, and soon, the generic specializer will need to be able to reference *_external witness tables deserialized from library modules.
Swift SVN r14887
When a subclass does not implement a designated initializer of its
superclass, introduce a stub initializer that simply traps. Such stubs
cannot be invoked directly using Swift syntax, but can be invoked
through the Objective-C runtime and from Objective-C code. Catch such
errors rather than allowing them to violate the memory safety of the
language.
Note that we're currently using cond_fail to trap; this will be
improved in the future,
Swift SVN r14839
Let ArchetypeType nested types and PotentialArchetypes be bound to concrete types in addition to archetypes. Constraints to outer context archetypes still suffer type-checker issues, but constraints to true concrete types should work now.
Swift SVN r14832
Currently this consists of private functions and shared functions. Even
though shared functions are technically serializable, we only want to
serialize them if their serialization is requested.
Swift SVN r14825
Previously, serialization of a single source file only includes decls that
are within the SourceFile context. However, new top-level decls can be
added in order to derive protocol conformances; these decls need to be
serialized /somewhere/. Add the concept of decls "forced" to be serialized
along with the decls within the primary SourceFile context, and then mark
all derived top-level decls as "forced" if they come from a decl in the
primary source file.
Possibly a fix for <rdar://problem/16254101>, which crashes nearby.
Swift SVN r14817
Before, we would hit one of the error conditions before updating the path
trace, which was misleading at best.
No functionality change.
Swift SVN r14815
The standard library likes to have default definitions for associated types,
which is good. Often the /choice/ of default type, however, is a type that
(indirectly) conforms to the very protocol containing the associated type.
Rather than try to make sure everything is present all at once, just delay
the deserialization of the default definition until it's actually requested.
This does swell the size of AssociatedTypeDecl by two words. I've filed
<rdar://problem/16266669> to remind myself to try to reduce this.
Part of <rdar://problem/16257259>
Swift SVN r14809
Make the name lookup interfaces all take DeclNames instead of identifiers, and update the lookup caches of the various file units to index their members by both compound name and simple name. Serialized modules are keyed by identifiers, so as a transitional hack, do simple name lookup then filter the results by compound name.
Swift SVN r14768
Add __FUNCTION__ to the repertoire of magic source-location-identifying tokens. Inside a function, it gives the function name; inside a property accessor, it gives the property name; inside special members like 'init', 'subscript', and 'deinit', it gives the keyword name, and at top level, it gives the module name. As a bit of future-proofing, stringify the full DeclName, even though we only ever give declarations simple names currently.
Swift SVN r14710
to the archetypes of their generic parameter declarations.
This was a major inconsistency that was causing a lot of
problems with deserialized generic functions in the presence
of chained module files. In particular, what would happen is
that all the deserialized uses of the archetype would be
mapped to a freshly-created archetype, but the GenericParamList
would refer to GenericTypeParamDecls deserialized from
(potentially) a different ModuleFile and therefore using a
different archetype (since the deserializer assumes that
different modules never share archetypes).
The fix relies on processing a generic parameter list before
any references to the archetypes. Unfortunately, we have
bogus references to archetypes in substitution lists scattered
everywhere in SIL and the AST. I've introduced a really
gross hack where we allow archetypes to be created in that
way but just drop them (in favor of the parameter's archetype)
after processing the generic parameter list; this should
work as long as all the decontextualized references are
basically pointless, but like I said, it's gross, and I've
filed rdar://16240384 to clean it up.
Swift SVN r14694
If we got just the wrong chain of dependencies, we'd create two VarDecls
for the same variable. I'm not sure if any of the other decls have the
same problem, but better safe than sorry.
I couldn't think of a small test case for this, but it does fix the one
in the Radar.
<rdar://problem/16205936>
Swift SVN r14625
