It's not okay to filter to only ErrorType results, since we may be
trying to chain to an error type result foo.bar.getError(). And the
existing logic had no way to handle results from other modules, so we
were missing key results like 'NSError'.
Eventually we'll want to bring back something like this that handles all
modules, but as a way to bump the priority of ErrorType results rather
than to filter out everything else.
rdar://problem/20985515
Swift SVN r28716
protocol P { typealias T; func foo() -> T }
func invalid(x: P) { x.#^COMPLETE^#
We were trying to access a null Type created because the associated type
doesn't make sense in the protocol type P (we can only use P as a
generic constraint, but it shouldn't crash code completion if we use it
incorrectly).
For rdar://problem/20305938
Swift SVN r28588
We were asserting (and doing the wrong thing) when trying to code
complete
nil #^HERE^#
The issue is that we tried to apply a solution to the expression that
contained free type variables (converted to generic parameters). This
trips us up when we expect the new type to conform to protocols. In code
completion we generally only need the type of the expression, so this
commit switches to getting that more explicitly. That said, this did
cause us to drop non-API parameter names in call patterns after an
opening '(' in some cases (covered by rdar://20962472).
Thanks to Doug for suggesting this solution!
rdar://problem/20891867
Swift SVN r28584
This came out of today's language review meeting.
The intent is to match #available with the attribute
that describes availability.
This is a divergence from Objective-C.
Swift SVN r28484
This will let us implement caching in the client (e.g. SourceKit) at
some point and simplifies adding more levels of caching. Requires a
corresponding SourceKit change.
Swift SVN r28365
Modules occupy a weird space in the AST now: they can be treated like
types (Swift.Int), which is captured by ModuleType. They can be
treated like values for disambiguation (Swift.print), which is
captured by ModuleExpr. And we jump through hoops in various places to
store "either a module or a decl".
Start cleaning this up by transforming Module into ModuleDecl, a
TypeDecl that's implicitly created to describe a module. Subsequent
changes will start folding away the special cases (ModuleExpr ->
DeclRefExpr, name lookup results stop having a separate Module case,
etc.).
Note that the Module -> ModuleDecl typedef is there to limit the
changes needed. Much of this patch is actually dealing with the fact
that Module used to have Ctx and Name public members that now need to
be accessed via getASTContext() and getName(), respectively.
Swift SVN r28284
Now get() and set() manage determining whether the results are stale,
and getResults() can just rely on that.
Also drive-by fix a data race where we were inserting our results sink
into the cache before it was finished being modified.
Swift SVN r28175
We want to be able to synthesize new results inside SourceKit. At this
point, the simplest way to do that is to expose the constructors for
CodeCompletionResult and a create() function for CodeCompletionString.
The expectation that any strings are stored properly inside a
CodeCompletionResultSink is documented.
Swift SVN r27822
Inference of type witnesses for associated types was previously
implemented as part of value witness matching in the constraint
solver. This led to a number of serious problems, including:
- Recursion problems with the solver hunting for a type witness,
which triggers more attemts to match value witnesses...
- Arbitrarily crummy attempts to break the recursion causing
type-check failures in fun places.
- Ordering dependencies abound: different results depending on which
value witnesses were satisfied first, failures because of the order
in which we attempted to infer type witnesses, etc.
This new implementation of type witness inference uses a separate pass
that occurs whenever we're looking for any type witness, and solves
all of the type witnesses within a given conformance
simultaneously. We still look at potential value witnesses to infer
type witnesses, but we match them structurally, without invoking the
constraint solver.
There are a few caveats to this implementation:
* We're not currently able to infer type witnesses from value
witnesses that are global operators, so some tricks involving global
operators (*cough* ~> *cough*) might require some manually-specified
type witnesses. Note that the standard library doesn't include any
such cases.
* Yes, it's another kind of solver. At simple one, fortunately.
On the other hand, this implementation should be a big step forward:
* It's far more predictable, order-invariant, and non-recursive.
* The diagnostics for failures to infer type witnesses have
improved.
Fixes rdar://problem/20598513.
Swift SVN r27616
Previously, the only way to get initializers was completing after the
name of the type:
Foo#^complete_here^#
Foo(#^or_here^#
And now it will also work in unadorned expressions:
#^a_top_level_completion^#
bar(a, #^walked_into_a_bar^#
Unfortunately, not all our clients handle this well yet, so it's
protected by a language option.
-code-complete-inits-in-postfix-expr
Swift SVN r27275
When a call (func, method, initializer) has default arguments, produce
both the *with* and *without* default argument versions as if they were
overloaded.
rdar://problem/18573874
Swift SVN r27118
Previously some parts of the compiler referred to them as "fields",
and most referred to them as "elements". Use the more generic 'elements'
nomenclature because that's what we refer to other things in the compiler
(e.g. the elements of a bracestmt).
At the same time, make the API better by providing "getElement" consistently
and using it, instead of getElements()[i].
NFC.
Swift SVN r26894
This improves run time of this particular function by up to 8x in my
testing. For *really* large sets of code completions (I tested top-level
completion with Cocoa imported) this gives a 30-40% improvement across
the request, not including time spent in the client. The tradeoff is a
slight increase in peak memory usage (I saw 0.7% or 1.2 MB for the
top-level Cocoa case).
<rdar://problem/20355626>
Swift SVN r26748
getImportedModules is the canonical way to get imports, whether private,
public, or both. This is especially true now that we have more flags
for SourceFile imports that really shouldn't be consumed by anyone
outside of SourceFile.
In this same vein, provide addImports instead of setImports, since imports
are always additive.
No visible functionality change.
Swift SVN r26634
Previously, when code completing Optional<T>., we only show
the member functions of T; This fix adds the member functions
of Optional to the code completion list.
rdar://20316534
Swift SVN r26622
When the code completion token appears after throw keyword,
a set of visible decls and instances conforming _ErrorType are
recommended as completion.
Swift SVN r26565
Currently a no-op, but effective access for entities within the current
module will soon need to take testability into account. This declaration:
internal func foo() {}
has a formal access of 'internal', but an effective access of 'public' if
we're in a testable mode.
Part of rdar://problem/17732115 (testability)
Swift SVN r26472
This patch also introduces some SILGen infrastructure for
dividing the function into "ordinary" and "postmatter"
sections, with error-handling-like stuff going into the
final section. Currently, this is largely undermined by
SILBuilder, but I'm going to fix that in a follow-up.
Swift SVN r26422
We explicitly whitelist these "stdlib private" decls in interface
generation, because they may contain methods that users are required to
implement. But in code-completion, there's no good reason to show them.
We still show completions for the methods themselves if you complete on
a public protocol that inherits from the private protocol. So,
<complete> => doesn't show _CollectionType
let a: CollectionType = ...
a.<complete> => *does* show startIndex, which comes from _CollectionType
rdar://problem/20086106
Swift SVN r26355
For decls, we get the most specific submodule from clang modules. For
macros, we don't yet have a way to get this information, since the
mapping from module ID to submodule is burried in the ClangImporter.
Having submodule information for macros would also help reduce the space
cost of storing the module names, since we would hit the single-element
module name cache more often.
There is no special handling of imported headers, which currently come
through with their internal clang module name '__ObjC'. Possibly these
should be treated as part of the current module.
A future commit will add the module name to swift-ide-test's output and update
the tests.
Swift SVN r26260
When a developer auto-completes an override function,
we collect the keywords that have already been specified
by the developer, so that the code completion strings do
not duplicate them.
rdar://16738036
Swift SVN r26230
When code completing a literal expr, it is likely that code completion engine only collects the expr
that is not fully type checked. Therefore, no members of the literal can be suggested. To address this,
we try to climb up expr hierarchy in AST to find an expr with a nominal type, and use the nominal type
to finish code completion.
rdar://20059173
Swift SVN r26116
auto-completing @attributes. By delaying the handling of code completion token after the entire decl being parsed, we know
what are the targets of the attribute to finishe, thus, only suggesting those applicable attributes.
Swift SVN r25938
context-sensitive. The first step is to recommend parameter-applicable
attributes only when the code completion token is found inside a
param decl.
Swift SVN r25810
