An environment is always associated with a location with a signature, so
having them separate is pointless duplication. This patch also updates
the serialization to round-trip the signature data.
The witnesses in a NormalProtocolConformance have never been
completely serialized, because their substitutions involved a weird
mix of archetypes that blew up the deserialization code. So, only the
witness declarations themselves got serialized. Many clients (the type
checker, SourceKit, etc.) didn't need the extra information, but some
clients (e.g., the SIL optimizers) would end up recomputing this
information. Ick.
Now, serialize the complete Witness structure along with the AST,
including information about the synthetic environment, complete
substitutions, etc. This should obsolete some redundant code paths in
the SIL optimization infrastructure.
This (de-)serialization code takes a new-ish approach to serializing
the synthetic environment in that it avoids serializing any
archetypes. Rather, it maps everything back to interface types during
serialization, and deserialization forms a new generic environment
(with new archetypes!) on-the-fly, mapping deserialized types back
into that environment (and to those archetypes). This way, we don't
have to maintain identity of archetypes in the deserialization code,
and might get some better re-use of the archetypes.
More of rdar://problem/24079818.
Reimplement the witness matching logic used for generic requirements
so that it properly models the expectations required of the witness,
then captures the results in the AST. The new approach has a number of
advantages over the existing hacks:
* The constraint solver no longer requires hacks to try to tangle
together the innermost archetypes from the requirement with the
outer archetypes of the context of the protocol
conformance. Instead, we create a synthetic set of archetypes that
describes the requirement as it should be matched against
witnesses. This eliminates the infamous 'SelfTypeVar' hack.
* The type checker no longer records substitutions involving a weird
mix of archetypes from different contexts (see above), so it's
actually plausible to reason about the substitutions of a witness. A
new `Witness` class contains the declaration, substitutions, and all
other information required to interpret the witness.
* SILGen now uses the substitution information for witnesses when
building witness thunks, rather than computing all of it from
scratch. ``substSelfTypeIntoProtocolRequirementType()` is now gone
(absorbed into the type checker, and improved from there), and the
witness-thunk emission code is simpler. A few other bits of SILGen
got simpler because the substitutions can now be trusted.
* Witness matching and thunk generation involving generic requirements
and nested generics now works, based on some work @slavapestov was
already doing in this area.
* The AST verifier can now verify the archetypes that occur in witness substitutions.
* Although it's not in this commit, the `Witness` structure is
suitable for complete (de-)serialization, unlike the weird mix of
archetypes previously present.
Fixes rdar://problem/24079818 and cleans up an area that's been messy
and poorly understood for a very, very long time.
There is no reason to make a declaration have qualified or unqualified
ownership. Incidently, this also fixes a bug where we were only setting the
unqualified bit for functions created in the deserializer. If we created a
declaration and linked it in later, we would not get the proper setting for
ownership qualification.
rdar://28851920
The reason we are using the parsing heuristic is to ensure that we do
not need to update a ton of test cases. This makes sense since in
general, when parsing we are creating new code that is running for the
first time through the compiler. On the other hand, in
serialization/deserialization we expect to get back exactly the
SILFunction that we serialized. So it makes sense to explicitly
preserve whether we have ownership qualification or not.
rdar://28851920
Over the past day or so I have been thinking about how we are going to need to
manage verification of semantic ARC semantics in the pass pipeline. Specifically
the Eliminator pass really needs to be a function pass to ensure that we can
transparently put it at any stage of the optimization pipeline. This means that
just having a flag on the SILVerifier that states whether or not ownership is
enabled is not sufficient for our purposes. Instead, while staging in the SIL
ownership model, we need a bit on all SILFunctions to state whether the function
has been run through the ownership model eliminator so that the verifier can
ensure that we are in a world with "SIL ownership" or in a world without "SIL
ownership", never in a world with only some "SIL ownership" instructions. We
embed this distinction in SIL by creating the concept of a function with
"qualified ownership" and a function with "unqualified ownership".
Define a function with "qualified ownership" as a function that contains no
instructions with "unqualified ownership" (i.e. unqualified load) and a function
with "unqualified ownership" as a function containing such no "ownership
qualified" instructions (i.e. load [copy]) and at least 1 unqualified ownership
instruction.
This commit embeds this distinction into SILFunction in a manner that is
transparently ignored when compiling with SIL ownership disabled. This is done
by representing qualified or unqualified ownership via an optional Boolean on
SILFunction. If the Boolean is None, then SILOwnership is not enabled and the
verifier/passes can work as appropriate. If the Boolean is not None, then it
states whether or not the function has been run through the Ownership Model
Eliminator and thus what invariants the verifier should enforce.
How does this concept flow through the compilation pipeline for functions in a
given module? When SIL Ownership is enabled, all SILFunctions that are produced
in a given module start with "qualified ownership" allowing them to contain SIL
ownership instructions. After the Ownership Model eliminator has run, the
Ownership Model sets the "unqualified" ownership flag on the SILFunction stating
that no more ownership qualified instructions are allowed to be seen in the
given function.
But what about functions that are parsed or are deserialized from another
module? Luckily, given the manner in which we have categories our functions, we
can categorize functions directly without needing to add anything to the parser
or to the deserializer. This is done by enforcing that it is illegal to have a
function with qualified ownership and unqualified ownership instructions and
asserting that functions without either are considered qualified.
rdar://28685236
As I was preparing a semantic arc commit that worked in this area, I realized I
needed ResultVal to be a SILInstruction, not a ValueBase. As I prepared to
perform a dyn_cast, I realized that ResulVal is always a SILInstruction
anyways... so why not just use the right type in the first place... Thus this
commit.
*sigh*
rdar://28685236
We don't want the machine calling conventions for closure invocation functions to necessarily be tied to the convention for normal thin functions or methods. NFC yet; for now, 'closure' follows the same behavior as the 'method' convention, but as part of partial_apply simplification it will be a requirement that partial_apply takes a @convention(closure) function and a box and produces a @convention(thick) function from them.
llvm r283043 and possibly other recent changes switch to use StringRef
instead of char* pointers. Update Swift to match. In some cases, this is
a clear improvement. It would be good to assess the impact on memory use,
particularly for the Filename component of source locations.
Note that the change to SILLocation::isNull fixes an apparent bug where
the location was treated as null when the filename was *not* null.
Sugared GenericTypeParamTypes point to GenericTypeParamDecls,
allowing the name of the parameter as written by the user to be
recovered. Canonical GenericTypeParamTypes on the other hand
only store a depth and index, without referencing the original
declaration.
When printing SIL, we wish to output the original generic parameter
names, even though SIL only uses canonical types. Previously,
we used to accomplish this by mapping the generic parameter to an
archetype and printing the name of the archetype. This was not
adequate if multiple generic parameters mapped to the same
archetype, or if a generic parameter was mapped to a concrete type.
The new approach preserves the original sugared types in the
GenericEnvironment, adding a new GenericEnvironment::getSugaredType()
method.
There are also some other assorted simplifications made possible
by this.
Unfortunately this makes GenericEnvironments use a bit more memory,
however I have more improvements coming that will offset the gains,
in addition to making substitution lists smaller also.
UnconditionalAvailabilityKind => PlatformAgnosticAvailabilityKind
::UnavailableInCurrentSwift => ::SwiftVersionSpecific
Plus a couple related method renamings. Prep work for SR-2709.
The behavior of std::minmax with rvalue arguments is undefined after the
end of the expression that contains the call to minmax. This code to
special-case the comparisons for ARM vs. Thumb and macOS vs. Darwin was a
little overly clever anyway. Rewrite it to use straightforward comparisons.
This fixes a test failure in Serialization/target-incompatible.swift when
building with a recent version of clang, because these checks were completely
optimized away due to the undefined behavior. rdar://problem/28700005
This is another problem (like 5356cc3) exposed by building with a newer
version of Clang. Construct a temporary SmallVector when needed to create
an ArrayRef.
This lets us get to the goal of +0 guaranteed closure contexts. NFC yet, just add the under-the-hood ability for partial_apply instructions producing callee-guaranteed closures to be parsed, printed, and serialized.
This would have allowed us to triage rdar://problem/28305755 much
sooner. The actual problem is pretty bad: if you have too many methods
with the same selector, serialization just falls over. "Too many" is
in the thousands, which seems unlikely, but 'dealloc' can actually get
there if there are a lot of little classes, and 'init' might as well,
so we really should do better here.
It's the same thing as for alloc_ref: the optional [tail_elems ...] attribute specify the tail elements to allocate.
For details see docs/SIL.rst
This feature is needed so that we can allocate a MangedBuffer with alloc_ref_dynamic.
The ManagedBuffer.create() function uses the dynamic self type to create the buffer instance.
Two of them are user-facing, with the following sort of message:
If you're seeing a crash here, check that your SDK and
dependencies match the versions used to build 'SwiftLib'
Prompted by rdar://problem/28282310, which took a while to figure out.
The added test case is a simplified version of the issue. (Obviously
we'd prefer to not crash here, but that's hard---there's an inherited
conformance that's no longer valid, and there may be generic types
depending on that conformance.)
RequirementReprs stored serialized references to archetypes,
which do not have enough information to reconstruct same-type
requirements.
For this reason, we would serialize the 'as written' requirement
string as well as the actual types, which is a horrible hack.
Now that the ASTPrinter and SourceKit use GenericSignatures,
none of this is needed anymore.
There's a bit of a hack to deal with generic typealiases, but
overall this makes things more logical.
This is the last big refactoring before we can allow constrained
extensions to make generic parameters concrete. All that remains
is a small set of changes to SIL type lowering, and retooling
some diagnostics in Sema.
Long term, we want to refactor the AST to reflect the current
programming model in Swift. This would include refactoring
FunctionType to take a list of ParameterTypeElt, or something with a
better name, that can contain both the type and flags/bits that are
only specific to types in parameter position, such as @autoclosure and
@escaping. At the same time, noescape-by-default has severely hurt our
ability to print types without significant context, as we either have
to choose to too aggressively print @escaping or not print it in every
situation it occurs, or both.
As a gentle step towards the final solution, without uprooting our
overall AST structure, and as a way towards fixing the @escaping
printing ails, put these bits on the TupleTypeElt and ParenType, which
will serve as a model for what ParameterTypeElt will be like in the
future. Re-use these flags on CallArgParam, to leverage shared
knowledge in the type system. It is a little painful to tack onto
these types, but it's minor and will be overhauled soon, which will
eventually result in size savings and less complexity overall.
This includes all the constraint system adjustments to make these
types work and influence type equality and overload resolution as
desired. They are encoded in the module format. Additional tests
added.
This flag switches the "effective language version" of the compiler,
at least to any version supported (as of this change: "3" or "3.0").
At the moment nothing uses it except the language version build
configuration statements (#if swift(...)) and various other places
that report, encode, or otherwise check version numbers.
In the future, it's intended as scaffolding for backwards compatibility.
Fixes SR-2582
The new instructions are: ref_tail_addr, tail_addr and a new attribute [ tail_elems ] for alloc_ref.
For details see docs/SIL.rst
As these new instructions are not generated so far, this is a NFC.
We didn't serialize the their var-info so after de-serializing they were useless anyway.
By not serializing them we get rid of useless debug_value instructions coming from the stdlib.
In future we might change this and fully serialize them.
The new instructions are: ref_tail_addr, tail_addr and a new attribute [ tail_elems ] for alloc_ref.
For details see docs/SIL.rst
As these new instructions are not generated so far, this is a NFC.
We were optimizing away unused pattern binding initializer contexts in
both the parser and in semantic analysis, which led to a
somewhat-unpredictable set of DeclContexts in the AST. Normalize
everything by always creating these contexts.
