martong added a comment. Okay, thanks for the explanation.
> The in-class initializer expression is not always stored in the AST, in the > ToTU it is missing initially. The field has the flag set that it contains > in-class initializer but the actual expression is not set yet (probably > because the code parser works this way). So, the parser can really create such a node when the expression is not set? That seems like a bug in the parser. Anyway, would be nice to see a test case for it and perhaps fix in the parser. But I have doubts that it would be a parser issue. I wonder whether it is rather this case: first we do an import that fails to set the expression and then the subsequent import crashes? Repository: rG LLVM Github Monorepo CHANGES SINCE LAST ACTION https://reviews.llvm.org/D120824/new/ https://reviews.llvm.org/D120824 _______________________________________________ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits