UnsafeContinuations can be stored in variables or properties,
so it's important for RemoteMirror to be able to at least minimally
recognize them.
This just treats an UnsafeContinuation as a refcounted pointer.
Which might be "good enough" for now.
Working towards rdar://110351406
The "generic depth" is used to match up generic type variables.
For example:
```
struct Foo<T> { // `T` at generic depth 0
struct Bar {
struct Baz<U> { // 'U' at generic depth 1
...
}}}
```
Note in the above that `Bar` is not counted in the
generic depth. The previous logic did count `Bar` in
the generic depth calculation, leading to mismatches
when trying to associate references to generic variables.
This adds a new test with cases like the above and of course
corrects the calculation.
Resolves rdar://127450037
This fills in a number of missing cases:
* MPEs with closure payloads
* MPEs with many non-payload cases
* MPEs with class existential payloads
* MPEs with existential metatype payloads
Resolves rdar://132270733
Resolves rdar://128705332
MPE layout code to exclusively use the new code. The key observation: existing
reflection metadata seems to already provide enough information in all cases, so
we can abandon an earlier effort to add spare bitmask data.
Resolves rdar://129281368
This adds a `getSpareBits` method to all the TypeInfo classes
that returns a suitable bitmask indicating the spare bits available
in values of this type. This gives us a way to recursively explore
the type tree and build up a full spare bit mask for an arbitrary type.
Happily, it appears we actually do have enough information to do this
calculation entirely from first principles, without requiring additional
reflection information. So once this is stable, we should remove my
earlier incomplete effort to publish spare bit mask info in the reflection
data, as that adds unnecessary metadata to every binary.
This doubtless still has plenty of holes, but seems sufficient
to handle a few basic enum types, including the stdlib DecodingError
which was used as an example to work out some key issues.
When a no-payload enum is stored inside a multi-payload enum,
the outer enum may be using some of the extra high-order bits.
So when we examine the inner enum, we should just strip those bits.
These tests had Darwin-specific code in them, mostly unnecessary,
that caused them to break on non-Darwin platforms.
Only one (reflect_Enum_SingleCaseCFPayload.swift) was actually
specific to Darwin. That one has been disabled. For the rest,
just delete the unnecessary `import Darwin` and similar.
A "trivial" enum is one that carries no information.
The previous logic asserted that a trivial enum had only one
case and that case had no payload.
But this is also a trivial payload:
```
// Trivial, so zero-sized
enum MyFormatVersion {
case v4
}
// Also trivial, since the payload carries no information
enum Format {
case MyFormat(MyFormatVersion)
}
```
This commit adds a test case similar to the above
and corrects the assertions for trivial enums to
assert that there is either no payload, or that the
payload is zero-sized.
Resolves rdar://116406504
Disable the assert:
// At least one payload is non-empty (otherwise this // would get laid out as a non-payload enum) assert(getNumNonEmptyPayloadCases() > 0);
in TaggedMultiPayloadEnumTypeInfo because it fails when you have
generic but zero-sized payload enum cases.
Also remove unnecessary "REQUIRES: objc_interop" lines from the enum
reflection tests.
I earlier overhauled the enum layout logic to correctly consider
enums with generic cases and cases that have zero size.
This updates the enum projection logic to use that information
as well.
In particular, this fixes a bug where an MPE with zero-sized cases
would be incorrectly projected by RemoteMirror (with consequences
for the `leaks` tool and lldb).
Resolves rdar://111705059
