Rather than looking for a given BraceStmtScope child for a particular
catch node, check whether the given catch node is the parent of the
innermost BraceStmtScope we've found, looking through an intermediate
source file if needed. This ensures it works correctly when the
BraceStmtScope is in a macro expansion.
rdar://149036108
Sema now type-checks the alternate ABI-providing decls inside of @abi attributes.
Making this work—particularly, making redeclaration checking work—required making name lookup aware of ABI decls. Name lookup now evaluates both API-providing and ABI-providing declarations. In most cases, it will filter ABI-only decls out unless a specific flag is passed, in which case it will filter API-only decls out instead. Calls that simply retrieve a list of declarations, like `IterableDeclContext::getMembers()` and friends, typically only return API-providing decls; you have to access the ABI-providing ones through those.
As part of that work, I have also added some basic compiler interfaces for working with the API-providing and ABI-providing variants. `ABIRole` encodes whether a declaration provides only API, only ABI, or both, and `ABIRoleInfo` combines that with a pointer to the counterpart providing the other role (for a declaration that provides both, that’ll just be a pointer to `this`).
Decl checking of behavior specific to @abi will come in a future commit.
Note that this probably doesn’t properly exercise some of the new code (ASTScope::lookupEnclosingABIAttributeScope(), for instance); I expect that to happen only once we can rename types using an @abi attribute, since that will create distinguishable behavior differences when resolving TypeReprs in other @abi attributes.
Introduce a number of fixes to allow us to fully use declarations that
are produced by applying a peer macro to an imported declarations.
These changes include:
* Ensuring that we have the right set of imports in the source file
containing the macro expansion, because it depends only on the module
it comes from
* Ensuring that name lookup looks in that file even when the
DeclContext hierarchy doesn't contain the source file (because it's
based on the Clang module structure)
Expand testing to be sure that we're getting the right calls,
diagnostics, and generated IR symbols.
Switch from promising a DeclContext to a
PatternBindingInitializer.
This has a couple of benefits:
- It eliminates a few places where we were force
`cast`'ing to PatternBindingInitializer.
- It improves the clarity of what's being stored,
it's not whatever the parent context of the
initializer is, it's specifically the
PatternBindingInitializer context if it exists.
There is a small bug fix here in the identification of the catch node,
where the leading `{` of a closure was considered to be "inside" the
closure for code like
{ ... }()
causing us to assume that the call to the closure would catch the error
within the closure.
Other than that, introduce the thrown error type into the type checker's
modeling of `withoutActuallyEscaping(_:do:)`, and mirror that in the
library declaration.
Both `try?` and `try!` are catch nodes, because they catch an error
thrown in their subexpression and handle it. Introduce an ASTScope for
all `try/try?/try1` expressions so we can find them, and model them as
catch nodes.
Fixes rdar://119216455.
The implementation of "lookup the catch node" is scattered throughout
various `ASTScopeImpl` subclasses. Remove that, and instead centralize
the code to find the innermost catch scope.
Replace a number of virtual query functions like `getDeclIfAny`, which had
to be overridden at various points in the hierarchy, with simpler
implementations based on the scope kind and `dyn_cast`.
Macro-metaprogram whether each scope node is associated with a
particular declaration/statement/expression/attribute so these
implementations can be efficiently macro-metaprogrammed.
The ASTScope class hierarchy rooted at ASTScopeImpl is relying entirely
on overridden virtual functions and, in some cases, string comparisons
on the class name, to provide customization for the various classes.
Introduce proper LLVM dynamic casting support to this class hierarchy,
eliminating the string comparisons and the manually-specified
implementations of `getClassName` (which is now used only for printing).
We've been meaning to do this for years, but today I got angry enough.
ASTScope assertions that a new child node inserted into its parent list
is after all other child nodes in the source, which included some bespoke
code for dealing with macro expansions. Replace this bespoke code with
uses of the newer SourceManager operations that check ordering of
source locations in a manner that respects macro expansions.
SourceManager's `isBeforeInBuffer` in buffer takes two source locations
that are assumed to be within the same source buffer and determines
whether the first precedes the second. However, due to macro
expansion, there can be source locations in different source buffers
that are nonetheless conceptually part of the same source file. For
such cases, `isBeforeInBuffer` will silently produce wrong results.
Extend `SourceManager` with a new `isBefore` operation that determines
whether one source location precedes another in the same conceptual
source file, even if they are in different source buffers that (e.g.)
represent macro expansions within that source file. Also introduce
`isAtOrBefore` to cover the case where the source locations might be
equivalent, which was previously handled via a separate equality check
on the source locations that won't work with macro expansions.
Add caching for the list of accessors of a buffer, so we aren't doing
a full walk up the source-buffer chain ever time we do a comparison of
source locations. LCA is still linear-time, but this eliminates
extraneous hash table lookups along the way.
Over time, we should both move more clients over to these new variants
and introduce more assertions into the old (in-buffer) versions to
catch improper uses of them.
The prior Least Common Ancestor (LCA) implementation in ASTScope, which
is used to search child scopes to find a particular location, assumed
that the source range it was given was contained within a single
buffer and was specific to the child-scope search task. Generalize
this to a more fundamental "is before" operation on source locations
that respects macro expansions, simplifying the code in the process.
Introduce a new API to find the AST node that catches or rethrows an
error thrown from the given source location. Use it to determine the
thrown error type to use for type checking a `throw` statement, which
begins as `any Error` within a `do..catch` and is later refined.
Addresses recently discovered performance regression.
`findChildContaining` has to handle macros and for that
it needs a parent module, looking it up every time is
expensive because it involves walking up back to a source
file scope.
Resolves: rdar://114667408
Resolves: rdar://114667496
namespace.
This moves the `isInMacroArgument` predicate and `lookupMacros` into `namelookup`.
ASTScope still encapsulates the scope tree and contains the operation to lookup
the enclosing macro scope, which then invokes a callback to determine whether a
potential macro scope is indeed a macro, because answering this question requires
name lookup.
This patch replaces the stateful generation of SILScope information in
SILGenFunction with data derived from the ASTScope hierarchy, which should be
100% in sync with the scopes needed for local variables. The goal is to
eliminate the surprising effects that the stack of cleanup operations can have
on the current state of SILBuilder leading to a fully deterministic (in the
sense of: predictible by a human) association of SILDebugScopes with
SILInstructions. The patch also eliminates the need to many workarounds. There
are still some accomodations for several Sema transformation passes such as
ResultBuilders, which don't correctly update the source locations when moving
around nodes. If these were implemented as macros, this problem would disappear.
This necessary rewrite of the macro scope handling included in this patch also
adds proper support nested macro expansions.
This fixes
rdar://88274783
and either fixes or at least partially addresses the following:
rdar://89252827
rdar://105186946
rdar://105757810
rdar://105997826
rdar://105102288
Macro expansion buffers, along with other generated source buffers,
need more precise "original source ranges" that can be had with the
token-based `SourceRange`. Switch over to `CharSourceRange` and provide
more thoughtfully-determined original source ranges.
Introduce SingleValueStmtExpr, which allows the
embedding of a statement in an expression context.
This then allows us to parse and type-check `if`
and `switch` statements as expressions, gated
behind the `IfSwitchExpression` experimental
feature for now. In the future,
SingleValueStmtExpr could also be used for e.g
`do` expressions.
For now, only single expression branches are
supported for producing a value from an
`if`/`switch` expression, and each branch is
type-checked independently. A multi-statement
branch may only appear if it ends with a `throw`,
and it may not `break`, `continue`, or `return`.
The placement of `if`/`switch` expressions is also
currently limited by a syntactic use diagnostic.
Currently they're only allowed in bindings,
assignments, throws, and returns. But this could
be lifted in the future if desired.
scopes.
ASTScope lookup can no longer assume that all scopes are within the same source
file. Scope trees can now contain scopes that are in macro expansion buffers, which
live in different source files with independent source ranges from the root of a
given scope tree. To handle this when searching for a scope containing a given source
location, ASTScope lookup needs to walk up the chain of macro expansion buffers to find
the lowest common buffer in which to compare source locations.
This fixes a number of issues with unqualified lookup into and from within macro-expanded
code.
We want to make sure not to cause a circular reference
by fetching the full generic parameter list when we are
creating opaque paramters and result types.
* Drop some unused fields
* const-qualify a consumption method that is logically const - though it
isn't physically const given the mutating use in
ASTScopeDeclConsumerForUnqualifiedLookup::lookInMembers
* Privatize some internal fields
This doesn't really fit the request evaluator model since the
result of evaluating the request is the new insertion point,
but we don't have a way to get the insertion point of an
already-expanded scope.
Instead, let's change the callers of the now-removed
expandAndBeCurrentDetectingRecursion() to instead call
expandAndBeCurrent(), while checking first if the scope was
already expanded.
Also, set the expanded flag before calling expandSpecifically()
from inside expandAndBeCurrent(), to ensure that re-entrant
calls to expandAndBeCurrent() are flagged by the existing
assertion there.
Finally, replace a couple of existing counters, and the
now-gone request counter with a single ASTScopeExpansions
counter to track expansions.
The top-level scope for a conditional clause with a pattern is now
ConditionalClausePatternUseScope, which introduces the pattern's
bindings.
Since the patterns are not visible in their own initializer, a new
ConditionalClauseInitializerScope is used for the initializer.
While it is nested inside of the ConditionalClausePatternUseScope,
it's lookup parent skips one level, giving us the desired behavior.
It wasn't used for anything, and it was always set based on whether
the declaration in question was a GenericTypeParamDecl, a ParamDecl,
or something else.
The old behavior was that ASTScope would introduce all VarDecls
defined in a BraceStmt at the beginning of the BraceStmt.
I recently enabled the use of PatternEntryDeclScopes, which
introduce the binding at its actual source location instead of
at the beginning of the parent statement.
This patch now makes use of the new information by having
UnqualifiedLookupFlags::IncludeOuterResults toggle between
the two behaviors. When searching for outer results, we also
consider all VarDecls in a BraceStmt, not just those in scope.
This is implemented by giving AbstractASTScopeDeclConsumer a
new entry point, consumePossiblyNotInScope(). When looking up
into a BraceStmt, all VarDecls are passed in to this entry
point.
The default implementation does nothing, which means that
ASTScope::lookupSingleLocalDecl() now respects source locations
when searching for bindings, just like parse-time lookup.
However, Sema's preCheckExpression() pass, which sets
Flags::IgnoreOuterResults, will continue to find
forward-referenced VarDecls, just as it did with the old
context-based DeclContext lookup.
This will be used to implement re-declaration checking for local
declarations. Currently this is handled by parse-time lookup.
To make it work with ASTScope, we need to perform lookups that
look into the innermost local scope only; for example, this is
an invalid redeclaration:
do {
let x = 321
let x = 123
}
But the following is fine, even though both VarDecls are in the same
*DeclContext*:
do {
let x = 321
do {
let x = 123
}
}
If we're searching for a declaration with a given name, the name
should be entirely encapsulated inside the DeclConsumer.
Otherwise, there might not be a specific name at all, if we're
performing code completion for example (once LookupVisibleDecls
starts to use ASTScope, anyway).
Bindings are not visible from inside their own initializer, eg this
is invalid:
var (x, y) = (y, x)
The PatternEntryDeclScope which starts after the 'var' introduces x
and y, and it contains the PatternEntryInitializerScope which starts
after the '='.
To model this properly in ASTScope, the PatternEntryInitializerScope
overrides getLookupParent() to skip the PatternEntryDeclScope that
contains it.
I believe this is simpler than having PatternEntryDeclScope begin
at the source location following the initializer, because it is hard
to compute this source location in a way that works with invalid
code, for example the 'var' might be nested inside of a BraceStmt
with the closing '}' missing.
