This biggest change is:
- LayoutConstraintInfo is now a FoldingSetNode, which allows for proper canonicalization of LayoutConstraints. This is important for the correctness of type comparisons if types contain layout constraints.
No functionality changes from the client's point of view.
New generic environments should be created directly from the generic
signature, without having to explicitly create an archetype
builder. Now, only the canonical archetype builders are ever used to
create a generic environment.
Clean up the representation of PotentialArchetype in a few small ways:
* Eliminate the GenericTypeParamType* at the root, and instead just
store a GenericParamKey. That makes the potential archetypes
independent of a particular set of generic parameters.
* Give potential archetypes a link back to their owning
ArchetypeBuilder, so we can get contextual information (etc.) when
needed. We can remove the "builder" arguments as a separate step.
Also, collapse getName()/getDebugName()/getFullName() into
getNestedName() and getDebugName(). Generic parameters don't have
"names" per se, so they should only show up in debug dumps.
In support of the former, clean up some of the diagnostics emitted by
the archetype builder that were using 'Identifier' or 'StringRef'
where they should have been using a 'Type' (i.e., the type behind the
dependent archetype).
For this we are linking the new re-mangler instead of the old one into the swift runtime library.
Also we are linking the new de-mangling into the swift runtime library.
It also switches to the new mangling for class names of generic swift classes in the metadata.
Note that for non-generic class we still have to use the old mangling, because the ObjC runtime in the OS depends on it (it de-mangles the class names).
But names of generic classes are not handled by the ObjC runtime anyway, so there should be no problem to change the mangling for those.
The reason for this change is that it avoids linking the old re-mangler into the runtime library.
These are no longer necessary now that we have combineSubstitutionMaps(),
and will not make sense once we switch to a more compact representation
for SubstitutionMap.
This method maps interface types to archetypes, which in general
requires a module for performing conformance lookups, if mapping
a member type of a generic parameter which has been made concrete.
However, in practice the types we are mapping here are all canonical
with respect to the generic signature, because they came from
GenericSignature::getAllDependentTypes(), so we actually don't need
to do conformance lookups.
This allows some code to be simplified.
SubstitutionList is going to be a more compact representation of
a SubstitutionMap, suitable for inline allocation inside another
object.
For now, it's just a typedef for ArrayRef<Substitution>.
A new SubstitutionMap::getProtocolSubstitutions() method handles
the case where we construct a trivial SubstitutionMap to replace
the protocol Self type with a concrete type.
When substituting one opened existential archetype for another,
use the form of Type::subst() that takes two callbacks instead of
building a SubstitutionMap. SubstitutionMaps are intended to be
used with keys that either come from a GenericSignature or a
GenericEnvironment, so using them to replace opened archetypes
doesn't fit the conceptual model we're going for.
Just like the recently-added GenericSignature::getSubstitutionMap(),
this takes a type substitution function and conformance lookup
function and produces a new SubstitionMap.
In several places, we construct a SubstitutionMap by taking
generic parameters from one SubstitutionMap up to a certain
depth, with the rest coming from a second SubstitutionMap.
Factor this out into a new utility method, to help with
hiding the internal representation of SubstitutionMap from
clients.
This was meant to be NFC, but it actually fixes a crash in
the devirtualizer, because the old logic there was slightly
wrong, so I added a test for this.
There was only one place where we lowered AST types with a non-zero
uncurry level, in SILGenApply.cpp. Add a new overload of the
getSILFunctionType() method that takes an uncurry level. All the
other methods no longer have to thread it through.
In all cases the DeclCtx field was supposed to be initialized from the
SILLocation of the function, so we can save one pointer per
SILFunction.
There is one test case change where a different (more precise)
diagnostic is being generated after this change.
This is useful to discover when a specific cleanup is being eliminated while
debugging. The implementation is compiled out when assertions are disabled.
rdar://29791263
Another pile of changes to use a side map for types in the constraint
solver and only write them directly into expressions once we have a
known good solution that we want to apply.
Still incomplete, we continue to write the types into expressions along
the way at the moment.
First, add some new utility methods to create SubstitutionMaps:
- GenericSignature::getSubstitutionMap() -- provides a new
way to directly build a SubstitutionMap. It takes a
TypeSubstitutionFn and LookupConformanceFn. This is
equivalent to first calling getSubstitutions() with the two
functions to create an ArrayRef<Substitution>, followed by
the old form of getSubstitutionMap() on the result.
- TypeBase::getContextSubstitutionMap() -- replacement for
getContextSubstitutions(), returning a SubstitutionMap.
- TypeBase::getMemberSubstitutionMap() -- replacement for
getMemberSubstitutions(), returning a SubstitutionMap.
With these in place, almost all existing uses of subst() taking
a ModuleDecl can now use the new form taking a SubstitutionMap
instead. The few remaining cases are explicitly written to use a
TypeSubstitutionFn and LookupConformanceFn.
When lowering closures, we avoid capturing the enclosing generic context when possible. However, the generic context may still be necessary to canonicalize types mentioned inside the closure, such as when an associated type is referred to that is same-typed to a concrete type. Fixes rdar://problem/30254048.
The canonicalization of dependent member types had some
nondeterminism. The root of the problem was that we couldn't
round-trip dependent member types through the archetype
builder---resolving them to a potential archetype lost the specific
associated type that was recorded in the dependent member type, which
affected canonicalization. Maintain that information, make sure that
we always get the right archetype anchor, and tighten up the
canonicalization logic within a generic signature.
Fixes rdar://problem/30274260 and should unblock some other work
that depends on sanity from the archetype builder and generic
signature canonicalization.
Unfortunately, I'm flying blind a bit --- this code is hard to
test outside of an integrated environment on a target using the
ABI in question, and I don't have such an environment set up ---
so I'm left with this rather ugly process of iteratively submitting
patches for trivial bugs and waiting for the bots to deliver judgment.
The environment could be mocked with a test environment that used
a nonstandard ObjC ABI, but that would actually take a lot of work
because of Swift's assumptions about the ObjC ABI: we'd basically
have to configure the test environment as a real, proper target,
possibly all the way through LLVM.
