We'd like to kill this enum off eventually, since the runtime inevitably needs to be able to handle arbitrary checked casts in opaque contexts, and SILGen and IRGen can deal with picking more optimal runtime entry points for specific casts. Only the container bridging kinds are still depended on anymore, and even those ought to eventually be handlable by the runtime in 'x as T' situations. NFC yet.
Swift SVN r23127
This change includes a number of simplifications that allow us to
eliminate the type checker hack that specifically tries
AssertString. Doing so provides a 25% speedup in the
test/stdlib/ArrayNew.swift test (which is type-checker bound).
The specific simplifications here:
- User-level
assert/precondition/preconditionalFailure/assertionFailer/fatalError
always take an autoclosure producing a String, eliminating the need
for the StaticString/AssertString dance.
- Standard-library internal _precondition/_sanityCheck/etc. always
take a StaticString. When we want to improve the diagnostics in the
standard library, we can provide a separate overload or
differently-named function.
- Remove AssertString, AssertStringType, StaticStringType, which are
no longer used or needed
- Remove the AssertString hack from the compiler
- Remove the "BooleanType" overloads of these functions, because
their usefuless left when we stopped making optional types conform
to BooleanType (sorry, should have been a separate patch).
Swift SVN r22139
When the solver has attempted to produce a solution using the default
literal type tha has failed, dont' try every type that conforms to
that literal type. Instead, try the bridged class type (which deals
with the common AnyObject case) or one of two other options:
- For integer literals, try the default floating point type (Double)
- For string literals, try the standard library's AssertString (this
is a temporary hack)
This limits exponential blow-up in cases where the literal's type
cannot be determined from context. Addresses rdar://problem/18307267.
Swift SVN r22131
Trying a collection literal early often means that we can determine
the element type from context, which saves us the work of trying to
guess at the element type firsthand.
Doing this seems to help some cases significantly:
- test/stdlib/ArrayNew.swift got about 20% faster in a release build
- I had to drop the threshold for the "expression too complex" test
case by 20x to still trigger the issue.
Swift SVN r22097
That, and _BuiltinCharacterLiteralConvertible, are not actually useful
without special hidden switches being passed to the compiler. We don't
want to have to explain them to users.
Swift SVN r22036
Conforming to BooleanLiteralConvertible now requires
init(booleanLiteral: Bool)
rather than
static func convertFromBooleanLiteral(value: Bool) -> Self
This posed a problem for NSNumber's conformance to
BooleanLiteralConvertible. A class needs a required initializer to
satisfy an initializer requirement, but one cannot add a required
initializer via an extension. To that end, we hack the Clang importer
to import NSNumber's initWithBool with the name
init(booleanLiteral:)
and add back the expected init(bool:) initializer in the
overlay. These tricks make NSNumber even harder to subclass, but we
don't really care: it's nearly impossible to do well anyway, and is
generally a Bad Idea.
Part of rdar://problem/18154091.
Swift SVN r21961
t2.swift:3:1: error: argument for generic parameter 'U' could not be
inferred
f(i)
^
t2.swift:2:6: note: in call to function 'f'
func f<T, U>(t: T) -> U? { return nil }
^
Our lack of decent locator information means that we don't get notes
in all of the cases we want them. I'll look at that separately.
Swift SVN r21921
Eliminate the implicit conversion from NSNumber to one of its bridged
value types (Int, Float, Double, Bool, etc.). One can use "as" to
perform the conversion instead. This eliminates a class of
accepts-dubious that involve lossy conversions from NSNumber to Int.
Our eventual arc is to eliminate all of these conversions. This is a
small, high-value step along that path rdar://problem/18269449.
Swift SVN r21879
You'll notice that emitting this diagnostic will make some already noisy closure-related errors slightly more so. This is unfortunate, but for the time-being it's better than crashing.
Swift SVN r21817
This already can't happen in most circumstances because of trailing closures, but we didn't explicitly disallow it at the beginning of a BraceStmt or following a statement production. Fixes the parser part of rdar://problem/17850752 (though there's a type checker bug there too).
Swift SVN r21663
A tuple of lvalues behaves semantically like an lvalue, so for consistency with scalar lvalues, we should diagnose ignored lvalue tuples in offline code, or coerce them to rvalues in REPL or playground contexts, because these contexts really expect rvalues for presentation purposes. Fixes rdar://problem/17057039.
Swift SVN r21549
This allows UnicodeScalars to be constructed from an integer, rather
then from a string. Not only this avoids an unnecessary memory
allocation (!) when creating a UnicodeScalar, this also allows the
compiler to statically check that the string contains a single scalar
value (in the same way the compiler checks that Character contains only
a single extended grapheme cluster).
rdar://17966622
Swift SVN r21198
reserve ? itself as a special token that cannot be defined (protecting ternary, postfix ?,
etc) but add some defensive code to prevent people from defining those operators.
<rdar://problem/17923322> allow ? as a general operator character
Swift SVN r21051
A checked cast such as "x as String" or "x as? [String]", where x is of
class or Objective-C existential type, is now handled as a normal
checked cast rather than a Sema-generated call to the corresponding
witness. This eliminates a pile of hairy code from constraint
application and takes a step toward <rdar://problem/17408934>.
The part of the switch_objc.swift test I removed wasn't testing
anything useful; that's what <rdar://problem/17408934> is about.
Swift SVN r20970
The _forceBridgeFromObjectiveC and _conditionallyBridgeFromObjectiveC
requirements of the _ObjectiveCBridgeable protocol previously returned
Self and Self?, respectively, where 'Self' is the value type that is
bridged. This use of returns is fairly hostile to the idea of calling
the witnesses for these requirements from the C++ part of the runtime,
leading to "interesting" tricks with OpaqueExistentialContainer that
made it hard to use these witnesses within the dynamic casting
infrastructure.
Replace the returns with inout Self? parameters, which are far easier
to deal with in the C++ part of the runtime. Despite the churn because
we're changing the _ObjectiveCBridgeable protocol, this is NFC.
Swift SVN r20934
This is essentially NFC, but the protocols we were calling into are
changing in a way that makes it hard to use the witnesses directly
from the type checker.
Swift SVN r20933
Start capitalizing on some of the new diagnostic machinery in a few different ways:
- When mining constraints for type information, utilize constraints "favored" by the overload resolution process.
- When printing type variables, if the variable was created by opening a literal expression, utilize the literal
default type or conformance if possible.
- Utilize syntactic information when crafting diagnostics:
- If the constraint miner can produce a better diagnostic than the recorded failure, diagnose via constraints.
- Factor in the expression kind when choosing which types to include in a diagnostic message.
- Start customizing diagnostics based on the amount of type data available.
What does all this mean?
- Fewer type variables leaking into diagnostic messages.
- Far better diagnostics for overload resolution failures. Specifically, we now print proper argument type data
for failed function calls.
- No more "'Foo' is not convertible to 'Foo'" error messages
- A greater emphasis on type data means less dependence on the ordering of failed constraints. This means fewer
inscrutable diagnostics complaining about 'UInt8' when all the constituent expressions are of type Float.
So we still have a ways to go, but these changes should greatly improve the number of head-scratchers served up
by the type checker.
These changes address the following radars:
rdar://problem/17618403
rdar://problem/17559042
rdar://problem/17007456
rdar://problem/17559042
rdar://problem/17590992
rdar://problem/17646988
rdar://problem/16979859
rdar://problem/16922560
rdar://problem/17144902
rdar://problem/16616948
rdar://problem/16756363
rdar://problem/16338509
Swift SVN r20927
Squash _[Conditionally]BridgedToObjectiveC into one protocol. This
change results in simpler bridging code with fewer dynamic protocol
conformance checks, and solves the nasty naming/semantics problem that
resulted from having _ConditionallyBridgedToObjectiveC refining
_BridgedToObjectiveC.
Also, rename things so they're more symmetrical and less confusing.
Swift SVN r20664
To limit user confusion when using conditional expressions of type Bool?, we've decided to remove the BooleanType (aka "LogicValue") conformance from optional types. (If users would like to use an expression of type Bool? as a conditional, they'll need to check against nil.)
Note: This change effectively regresses the "case is" pattern over types, since it currently demands a BooleanType conformance. I've filed rdar://problem/17791533 to track reinstating it if necessary.
Swift SVN r20637
Change the lexing of '?' to be similar to '!', where we special-case the postfix case for the intrinsic postfix optional operator, but fall back to lexing as an operator when it isn't left-bound. For now, only accept '??' as an operator name--we could easily generalize this, but that warrants discussion first.
Swift SVN r20591
Modify TypeBase::getRValueType to structurally convert lvalues embedded in tuple and paren types. Inside the constraint solver, coerce types to rvalues based on the structural 'isLValueType' test rather than shallow 'is<LValueType>' checking. Fixes <rdar://problem/17507421>, but exposes an issue with call argument matching and lvalues <rdar://problem/17786730>.
Swift SVN r20442
enforce its own little constraints. The type checker isn't using it for
anything, and it is just clutter.
This resolves <rdar://problem/16656024> Remove @assignment from operator implementations
Swift SVN r19960
modifiers and with the func implementations of the operators. This resolves the rest of:
<rdar://problem/17527000> change operator declarations from "operator prefix" to "prefix operator" & make operator a keyword
Swift SVN r19931
eliminating the @'s from them when used on func's. This is progress towards
<rdar://problem/17527000> change operator declarations from "operator prefix" to "prefix operator" & make operator a keyword
This also consolidates rejection of custom operator definitions into one
place and makes it consistent, and adds postfix "?" to the list of rejected
operators.
This also changes the demangler to demangle weak/inout/postfix and related things
without the @.
Swift SVN r19929
This only tackles the protocol case (<rdar://problem/17510790>); it
does not yet generalize to an arbitrary "class" requirement on either
existentials or generics.
Swift SVN r19896
