[Sema] Disallow conditional conformances on objective-c generics. #14628
Conversation
huonw
force-pushed
the
no-conditional-objc-generics
branch
from
3f3cbd3 to
151e6d8
Compare
|
@swift-ci please smoke test. |
| TypeOrExtensionDecl(ExtensionDecl *D); | ||
|
|
||
| /// \brief Return the contained *Decl as the Decl superclass. | ||
| class Decl *getAsDecl() const; |
There was a problem hiding this comment.
The field is called Decl: clang specifically said I should write class Decl to disambiguate.
There was a problem hiding this comment.
Ah, I guess I would have just renamed the field.
|
|
||
| bool operator==(TypeOrExtensionDecl rhs) { return Decl == rhs.Decl; } | ||
| bool operator!=(TypeOrExtensionDecl rhs) { return Decl != rhs.Decl; } | ||
| bool operator<(TypeOrExtensionDecl rhs) { return Decl < rhs.Decl; } |
There was a problem hiding this comment.
I'm not sure this one makes sense to provide.
There was a problem hiding this comment.
Just mindlessly duplicating the PointerUnion API.
| //===----------------------------------------------------------------------===// | ||
| // | ||
| // This file defines the TypeOrExtensionDecl struct, separately to Decl.h so | ||
| // that this can be included in files that Decl.h includes. |
There was a problem hiding this comment.
This probably doesn't actually work, because PointerUnion needs to know the alignment of the types in it to know where to stick spare bits.
There was a problem hiding this comment.
It compiles as long as you also include Decl.h before the template actually has to be instantiated, I think.
There was a problem hiding this comment.
How do I address this then?
There was a problem hiding this comment.
Untangle the headers? Or…ah, it looks like we have swift/AST/TypeAlignments.h for this. ASTNode.h uses it in a similar way.
There was a problem hiding this comment.
Neat; changed it to use that header.
It's used in |
|
Sorry, I know the refactoring is used where it was used before. It doesn't seem to be used in any new places. |
|
Ah yeah, it's a follow up to #14554 (comment) but including it here rather than a separate PR since CI is apparently getting hammered... |
|
Ugh, is there a weird corner case where we are trying to make a nested type of an ObjC generic class conditionally conform to a protocol, but this code wouldn't notice? |
huonw
force-pushed
the
no-conditional-objc-generics
branch
from
151e6d8 to
da6feab
Compare
|
Updated to handle nested types. @swift-ci please smoke test. |
There's no way to look up information about objective-c generic parameters, meaning the runtime cannot check same-type constraints or conformance requirements, so dynamic casts cannot work. We want to keep the static and dynamic systems the same, so we have to disable this functionality from the start (i.e. no conditional conformances on objective-c types :( ). Fixes rdar://problem/37524969.
huonw
force-pushed
the
no-conditional-objc-generics
branch
from
da6feab to
dd845c4
Compare
|
@swift-ci please smoke test and merge. |
3 similar comments
|
@swift-ci please smoke test and merge. |
|
@swift-ci please smoke test and merge. |
|
@swift-ci please smoke test and merge. |
|
@swift-ci please smoke test |
|
@swift-ci please smoke test Linux platform |
There's no way to look up information about objective-c generic
parameters, meaning the runtime cannot check same-type constraints or
conformance requirements, so dynamic casts cannot work. We want to keep
the static and dynamic systems the same, so we have to disable this
functionality from the start (i.e. no conditional conformances on
objective-c types :( ).
Fixes rdar://problem/37524969.
(Plus moving
TypeOrExtensionDecl.)