Allow subclass!() of superclasses with protected destructors
#1481
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Currently, we only synthesize a default constructor when the superclass destructor is public. This is to comply with 11.4.5.2 of the C++ standard[1], which states that a default constructor is deleted if "any potentially constructed subobject [e.g. the virtual base class] has a type with a destructor that is deleted or inaccessible from the defaulted default constructor". We correctly exclude private base class destructors, which are not accessible from the subclass's constructor. But we also exclude protected base class destructors, which are! Fix that by matching protected superclass destructors too. [1] https://github.com/cplusplus/draft/releases/tag/n4917
It's exactly like what we already have for Enum, Typedef, Function, and Struct. This commit just makes the boring structural changes: the field itself is still empty. A future commit will use it to track superclass destructor visibility.
Currently, we emit as_<superclass>_unique_ptr helpers for all Rust subclasses. But this breaks subclasses derived from classes with private or protected destructors, since unique_ptr needs to be able to destruct the object it references. Fix that by omitting those helpers when we know the superclass destructor is inaccessible. The core changes are in codegen_rs and codegen_cpp; everything else is just plumbing to get information about superclass destructor visibility where it needs to be.
tchebb
changed the title
subclass!() for superclasses with protected destructors
tchebb
changed the title
subclass!() for superclasses with protected destructorssubclass!() of superclasses with protected destructors
tchebb
added a commit
to tchebb/openwv
that referenced
this pull request
Apr 16, 2025
This includes [1] and [2], which are both needed to generate bindings from content_decryption_module.h without errors. If and when upstream takes the fixes, we can pull from crates.io again. [1] google/autocxx#1478 [2] google/autocxx#1481
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
See individual commit messages for details. There were two blockers here:
as_<superclass>_unique_ptrfunctions failed to compile on the C++ side becauseunique_ptrneeds to point to a destructible type. I fixed that by just omitting those functions when the superclass is not destructible.CppPeerConstructorimplementation, were erroneously omitted forprotectedsuperclass destructors rather than only forprivateones.I'm not sure if my approach here, especially on commits 2 and 3, is ideal. I added a new analysis field for
Api::Subclass, which resulted in a ton of new boilerplate. The new code's clean, but there's a lot of it. If you can think of a way to plumb the information through in a less invasive way, I'd be happy to do that instead.