[lldb] Improve handling of artificial subclasses #7423
Conversation
|
@swift-ci test |
augusto2112
force-pushed
the
artificial-class
branch
from
e92daeb to
981dd9b
Compare
| @@ -1080,6 +1080,10 @@ CompilerType SwiftLanguageRuntimeImpl::GetChildCompilerTypeAtIndex( | |||
| LLDBTypeInfoProvider tip(*this, *instance_ts); | |||
| reflection_ctx->ForEachSuperClassType( | |||
| &tip, pointer, [&](SuperClassType sc) -> bool { | |||
| // If the typeref is invalid, we don't want to process it. | |||
| if (!sc.get_typeref()) | |||
| return false; | |||
There was a problem hiding this comment.
Is this an error, or expected? We might want to log it if it's an error.
There was a problem hiding this comment.
If it's expected, because this is normal for artificial base classes, maybe the comment could be more explicit about that?
There was a problem hiding this comment.
It's expected if the current class is artificial, I'll add a comment
There was a problem hiding this comment.
is there a way we can know that the current class is artificial? In the case where the current class is not artificial, then we'd want a log or better.
There was a problem hiding this comment.
Remote Mirrors doesn't really tell us why getting the typeref failed (it could be an artificial class or conceivably something else). It'd be pretty straight forward to add an API asking if the class is artificial or not, if we want that.
An artificial subclass is a subclass which was registered at runtime, which is possible to do when Obj-C interop is enabled. Fix variables of these types being printed incorrectly. rdar://101519300
augusto2112
force-pushed
the
artificial-class
branch
from
981dd9b to
58f6d82
Compare
|
@swift-ci test |
|
@augusto2112 Can you rebase this on swift/release/5.10? |
An artificial subclass is a subclass which was registered at runtime, which is possible to do when Obj-C interop is enabled. Fix variables of these types being printed incorrectly.
rdar://101519300