Regardless of consumes of copies, if the original lexical value is not
consumed on a dead-end path, it must remain live to that dead-end.
rdar://145226757
At one point, OpenedArchetypeType did not exist as a separate subclass
of ArchetypeType, so this method did something. Now, it's just
equivalent to calling is<> or isa<>.
I also removed a couple of asserts that were obvious no-ops as a result.
If we have a self value that does not dominate loop preheader,
and the array semantics call does not consume the self value,
that means there will be instructions that consume the self value
with the loop.
In ossa, we cannot hoist such semantic calls because there is no
support for creating destroys for them in the preheader.
Add a bailout to avoid the ownership error.
rdar://145673368
To handle borrowing operands that produce a dependence value but do not create a
nested borrow scope. This includes non-reborrow borrowed-from and guaranteed
mark_dependence [nonescaping].
Only return false if the visitor returns false. Clients were ignoring the
result.
If the BorrowingOperand does not create a borrow scope, call visitUnknownUse
instead.
Until we have complete lifetimes, to avoid breaking code that cannot handle dead
defs, consider a dead borrow scope to be an unknown use.
This also fixes a case where we would have inferred the wrong isolation (which
the assert caught). I think we missed this since it only comes up with final
classes.
rdar://142568522
In these cases, we want to lookthrough so we propagate through
nonisolated(unsafe) and make it easier to discover that we are processing
keypaths (the reason I am making this change).
Currently we don't support hoisting ownership instructions.
But the check was missing a store of Optional.none because such a value has no ownership even if the optional is not trivial.
Fixes a SIL verifier crash.
https://github.com/swiftlang/swift/issues/79491
From talking with @dgregor, it became clear that this comment was easily
interpreted as saying that AssignFresh always introduced a disconnected value...
which is not the case. Instead, AssignFresh just introduces a new value that
could have any form of isolation. The actual isolation of the value is assigned
via tryToTrackValue and eventually SILIsolationInfo::get().
If the default argument generator (and, consequently, the function taking this default argument) has public visibility, it's OK to have a closure (which always has private visibility) as the default value of the argument.
CSE uses OSSA rauw which creates copies and copies that are created to optimize
across borrow scopes are unoptimizable. This PR avoids this situation for now.
An indirect argument (except `@inout_aliasable`) must not alias with another indirect argument.
Now, if we would replace tempAddr with destAddr in
```
apply %f(%tempAddr, %destAddr) : (@in T) -> @out T
```
we would invalidate this rule.
This is even true if the called function does not read from destAddr.
Fixes a SIL verification error.
rdar://145090659
Inside fragile functions, we expect function derivatives to be public, which could be achieved by either explicitly marking the functions as differentiable or having a public explicit derivative defined for them. This is obviously not
possible for single and double curry thunks which are a special case of `AutoClosureExpr`.
Instead of looking at the thunk itself, we unwrap it and look at the function being wrapped. While the thunk itself and its differentiability witness will not have public visibility, it's not an issue for the case where the function being wrapped (and its witness) have public visibility.
Fixes#54819Fixes#75776
If a pass forgot to call invalidateAnalysis but deleted some instructions, the pass-manager can fix this.
Currently following passes do not invalidate analysis when they change the SIL:
* LowerTupleAddrConstructor
* DestroyAddrHoisting
* MoveOnlyChecker
* PredictableDeadAllocationElimination
Ideally we should fix those passes. But with this addition in the pass-manager it's not strictly necessary.
Fixes a compiler crash.
Although it's not used anymore we still have to support it to be able to read old Swift.interface files which still contain the builtin.
rdar://144781646
Introduce a new experimental feature StrictSendableMetatypes that stops
treating all metatypes as `Sendable`. Instead, metatypes of generic
parameters and existentials are only considered Sendable if their
corresponding instance types are guaranteed to be Sendable.
Start with enforcing this property within region isolation. Track
metatype creation instructions and put them in the task's isolation
domain, so that transferring them into another isolation domain
produces a diagnostic. As an example:
func f<T: P>(_: T.Type) {
let x: P.Type = T.self
Task.detached {
x.someStaticMethod() // oops, T.Type is not Sendable
}
}
