[lldb] SwiftLanguageRuntime: Support Clang typedef-to-enums. #10392
Conversation
This also relaxes the check for the name of the member in GetIndexOfChildMemberWithName for Clang enums, since the result is always 1 anyway. rdar://147905814
adrian-prantl
requested review from
augusto2112 and
kastiglione
and removed request for
a team
|
@swift-ci test |
| @@ -695,6 +695,15 @@ void LogUnimplementedTypeKind(const char *function, CompilerType type) { | |||
| #endif | |||
| } | |||
|
|
|||
| CompilerType GetCanonicalClangType(CompilerType type) { | |||
| for (unsigned i = 0; i < 32; ++i) { | |||
There was a problem hiding this comment.
Why 32? Is this just an arbitrary limit? If so I'd raise it to 1024 or something like that...
There was a problem hiding this comment.
It's an arbitrary limit to guard against cycles.
It's for chains of typedefs, so practically I wouldn't expect more than two to ever appear, but I'm fine with dialing up the limit a bit.
| @@ -695,6 +695,15 @@ void LogUnimplementedTypeKind(const char *function, CompilerType type) { | |||
| #endif | |||
| } | |||
|
|
|||
| CompilerType GetCanonicalClangType(CompilerType type) { | |||
There was a problem hiding this comment.
Why is this called CanonicalClangType? There doesn't seem to be anything specific about clang in this function.
There was a problem hiding this comment.
I guess this would only be a canonical type if it's a clang type, since untypedefing wouldn't necessarily result in a canonical Swift type, right? Maybe add an assert or early return if this isn't a clang type?
There was a problem hiding this comment.
You're right, the function itself is generic, but I wanted to differentiate it from TypeSystemSwiftTypeRef::GetCanonicalNode, which also resolves arbitrarily deep nested Swift type aliases.
I like the idea of asserting that it's a clang type.
|
@swift-ci test macos |
|
I'll address the feedback in a follow-up commit. |
This also relaxes the check for the name of the member in
GetIndexOfChildMemberWithName for Clang enums, since the result is
always 1 anyway.
rdar://147905814