While it's currently not possible for `isDeclAsSpecializedAs` to compare
decls that differ by having a curried self or having a parameter list,
tweak the logic slightly so it could handle that case if it needed to.
If one decl has a parameter list and the other doesn't, then just
compare their self types. If one decl is curried and the other isn't,
then add a curried self to the other and compare as normal.
Use this function to replace various places where the logic is
duplicated.
In addition, isolate the logic where subscripts are treated as having
curried self parameters to CalleeCandidateInfo, as their interface types
don't have a curried self, but get curried with self by
CalleeCandidateInfo. Ideally we'd fix this by having a subscript's
interface type be curried with self, but given that most of this CSDiag
logic should be going away, this may not be necessary.
Introduce an attribute @_disfavoredOverload that can be used to state
that a particular declaration should be avoided if there is a
successful type-check for a non-@_disfavoredOverload. It's a way to
nudge overload resolution away from particular solutions.
Instead of storing information about expression depths in the
solver state (which gets recomputed for salvage) let's track
it directly in constraint system, which also gives solver
access to it when needed e.g. for fixes.
Solving Bind is a little easier than Equal. The only remaining uses of Equal
are in the .member syntax and keypaths; if we can refactor those, we might be
able to simplify LValue handling in the type checker in general.
We've been running doxygen with the autobrief option for a couple of
years now. This makes the \brief markers into our comments
redundant. Since they are a visual distraction and we don't want to
encourage more \brief markers in new code either, this patch removes
them all.
Patch produced by
for i in $(git grep -l '\\brief'); do perl -pi -e 's/\\brief //g' $i & done
Make sure that presence of `@autoclosure` attribute handled
in one place - `matchCallArguments`, which makes it possible
to remove the rest of (now redundant) autoclosure related
logic scattered throughout solver.
As discussed in #18951, we don't want to be comparing two different protocol decls here, as that could lead to spurious 'is more specialized' results. I'm not aware of any case that currently trips this logic up though, so no tests to accompany.
Changes in shadowing behaviour by #15412 caused a property on a concrete type to no longer shadow a protocol property member, which created unintentional ambiguities in 4.2. This commit ensures we at least keep these cases unambiguous in Swift 5 under Swift 4 compatibility mode. This is intentionally narrow in order to best preserve source compatibility under Swift 4 mode by ensuring we don't introduce any new ambiguities.
Resolves SR-7425, SR-7940 & SR-8343.
This rule caused us to lose ambiguities in places where we really want ambiguity for `AnyObject` lookup, so only apply it when not comparing such overloads. This whole situation is a bit of a hack – really we shouldn't be applying any type-based or context-based overload ranking rules to overloads found through `AnyObject` lookup, but unfortunately we don't have syntax to precisely disambiguate overloads.
This commit can be reverted if/when we ever remove `AnyObject` lookup.
This commit adds a new rule to `isDeclAsSpecializedAs` in order to favour a member on a concrete type over a protocol member. This rule is effectively an extension of the existing rule that prefers concrete type members over protocol extension members.
- getAsDeclOrDeclExtensionContext -> getAsDecl
This is basically the same as a dyn_cast, so it should use a 'getAs'
name like TypeBase does.
- getAsNominalTypeOrNominalTypeExtensionContext -> getSelfNominalTypeDecl
- getAsClassOrClassExtensionContext -> getSelfClassDecl
- getAsEnumOrEnumExtensionContext -> getSelfEnumDecl
- getAsStructOrStructExtensionContext -> getSelfStructDecl
- getAsProtocolOrProtocolExtensionContext -> getSelfProtocolDecl
- getAsTypeOrTypeExtensionContext -> getSelfTypeDecl (private)
These do /not/ return some form of 'this'; instead, they get the
extended types when 'this' is an extension. They started off life with
'is' names, which makes sense, but changed to this at some point. The
names I went with match up with getSelfInterfaceType and
getSelfTypeInContext, even though strictly speaking they're closer to
what getDeclaredInterfaceType does. But it didn't seem right to claim
that an extension "declares" the ClassDecl here.
- getAsProtocolExtensionContext -> getExtendedProtocolDecl
Like the above, this didn't return the ExtensionDecl; it returned its
extended type.
This entire commit is a mechanical change: find-and-replace, followed
by manual reformatted but no code changes.
That is, don't look through InOutType anymore, and update callers to
call getInOutObjectType() as well (or not, where it was obvious to me
that InOutType could not appear).
This surfaces more remaining uses of getInOutObjectType() directly.
Since constraint solver has been improved to diagnose more problems
via "fixes", sometimes applying fixes might lead to producing solutions
which are completely ambiguous when compared to each other, and/or are
incomparable, which leads to `findBestSolutions` erasing all of them
while trying to compute best "partial" solution, which is incorrect.
Resolves: rdar://problem/42678836
This is helpful in experimenting with constraint solver changes that
might help us remove some of these unsound options. It's not ever mean
to be enabled, but if we're able to remove the things guarded by the
option we can eventually remove the option.
...unless the argument is an `Any?`, in which case we prefer `f(_: Any?)`.
This change also results in our selecting f<T>(_: T) over f(_:
Any). Coercing with 'as Any' makes it possible to explicitly select
the Any overload. Previously there was no way to select the generic
overload.
Treat non-optional generic parameters as being more specialized than
optional generic parameters, and penalize any solutions that involve
generic arguments that are themselves Optional.
By doing these things, we can remove the special-cased code for the
two overloads of '??' in the stdlib, instead treating the (T?, T)
overload as better than the (T?, T?) overload except where a user
actually passes an optionally-typed value as the second parameter.
Fixes: rdar://problem/19748710
