Changes

allow SemanticVersion prerelease field to be int or string (#98)

rdar://147937598

Include try and await expressions which are parents of a freestanding expression macro in lexicalContext

Resolves rdar://109470248

Handle `try` and `await` in unfolded sequences

We can say that any `try`/`await` element also covers all elements to
the right of it in an unfolded sequence. Cases where this isn't true
will be rejected by the compiler, e.g:

```
0 * try foo() + bar()
_ = try foo() ~~~ bar() // Assuming `~~~` has lower precedence than `=`
```

rdar://109470248

Add `unsafe` as a lexical context too

Match the behavior of `try` and `await`.

[cxx-interop] Fix nested structs for non-opaque types

The test for nested constructs used library evolution forcing all types
to be opaque. As a result some code paths for non-opaque types were not
updated to support nested types. This patch updates the rest of the code
making sure we use fully qualified names (so they also work in the
context of the nested classes), and generate correct names for the C
compatibility structs that cannot contain "::".

Fixes #80291

rdar://147882976

Prepare diagnostics docs for docc generation

Update the diagostic docs so that they can be built with docc:
    * Add technologyroot (diagnostics.md) to aggregate all markdown files
    * Fixed warnings in a few diagnostics docs

This allows the creation of docc documentation with a single
invocation of docc. It will generate the landing page from
diagnostics.md and include all of the diagnostics as subarticles.

The subarticles are accessible via the name of the markdown file,
which is being used as the ID for these diagnostics.
E.g. existential-any.

The docc docs for these documents can be generated locally with the
following command:
>docc preview --allow-arbitrary-catalog-directories userdocs/diagnostics

SILGen: Correct ownership forwarding of `Builtin.emplace`.

Forward the owning cleanup for the temporary buffer (if needed) instead of
creating a new cleanup, to avoid a double-free when both the initialization
cleanup and the value cleanup execute. Fixes rdar://147961840.

Sema: Fix regression with 'Failure' type witness inference hack

We have a special hack to infer the 'Failure' type witness in
'AsyncIteratorProtocol' by considering the type of 'next()'
witness.

In eaf06eae0e0a140b240a0b941c8add98dfc424db I added a check to
fix some assertions that could happen if 'next()' was witnessed
by a declaration in a protocol extension or superclass, which
has a different generic signature.

However my check was too narrow, because it also prohibited
this form of inference when 'next()' was in a different
extension of the same nominal type.

Allow this again.

Fixes https://github.com/swiftlang/swift/issues/79367.
Fixes rdar://problem/145341658.

AST: Remove origType parameter from ProtocolConformanceRef::getAssociatedConformance()

AST: Build the right kind of ErrorType inside subst() instead of getTypeWitness()

AST: Remove origType parameter from ProtocolConformanceRef::getTypeWitness()

AST: Remove type parameter from ProtocolConformanceRef::getTypeWitnessByName()

AST: Preserve sugar in getIdentitySubstitutionMap()

We implement getIdentitySubstitutionMap() by passing in
a no-op substitution callback. Remove the unnecessary
getCanonicalType() call, to ensure that the abstract
conformance in the identity substitution map uses the
sugared form of the subject type.

AST: Remove type parameter from ProtocolConformanceRef::getWitnessByName()

AST: Remove origType parameter from ProtocolConformanceRef::subst()

AST: Remove origType parameter from ProtocolConformanceRef::getAssociatedType()

AST: Add new form of SubstitutionMap::getProtocolSubstitutions()

AST: Eliminate GenericEnvironment::mapConformanceRefIntoContext()

AST: Remove origType parameter from substOpaqueTypesWithUnderlyingTypes()

AST: Remove getAssociatedType() call from PackConformance

IRGen: Remove calls to ProtocolConformanceRef::getAssociatedType()

SILGen: Remove a usage of ProtocolConformanceRef::getAssociatedType()

AST: Remove ProtocolConformanceRef::getAssociatedType()

AST: Remove AssociatedType

[AST/Sema] Add `@extensible` attribute on `enum` declarations

This attribute controls whether cross-module access to the declaration
needs `@unknown default:` because it's allowed to gain new cases even
if the module is non-resilient.

[Frontend/Serialization] Remove `ExtensibleEnums` experimental flag

For now the semantics provided by `@extensible` keyword on per-enum
basis. We might return this as an upcoming feature in the future with
a way to opt-out.

Canonicalize different spellings of the same integer generic parameter.

`Foo<256>`, `Foo<2_56>`, and `Foo<0x100>` are all canonically the same type.
Fixes rdar://144736386.

SILLinker: make sure to de-serialize base protocol witness tables in embedded mode

Fixes unresolved symbol linker errors or compiler crashes
rdar://148538336

SILLinker: de-serialize Executor witness tables for the xExecutorRef builtins

Fixes unresolved symbol linker errors in embedded mode.
rdar://148538336

Revert "Add support for inout lifetime dependence"

[clang][modules] Determine if the SDK supports builtin modules independent of the target

Whether the SDK supports builtin modules is a property of the SDK itself, and really has nothing to do with the target. This was already worked around for Mac Catalyst, but there are some other more esoteric non-obvious target-to-sdk mappings that aren't handled. Have the SDK parse its OS out of CanonicalName and use that instead of the target to determine if builtin modules are supported.

[lldb][debugserver] Save and restore the SVE/SME register state (#134184)

debugserver isn't saving and restoring the SVE/SME register state around
inferior function calls.

Making arbitrary function calls while in Streaming SVE mode is generally
a poor idea because a NEON instruction can be hit and crash the
expression execution, which is how I missed this, but they should be
handled correctly if the user knows it is safe to do.

Re-landing this change after fixing an incorrect behavior on systems
without SME support.

rdar://146886210
(cherry picked from commit f1c6612202d88cbde224387621327a31609f2177)