sammccall added inline comments.
================
Comment at: clang/include/clang/Sema/Sema.h:11968
bool AfterAmpersand);
+ void CodeCompleteFunctionDeleteAndDefaultSpecifier();
----------------
nit: "specifier" isn't an accurate term here. I don't know that there is one,
and it seems better to name the method after the situation rather than the
concrete completions it might offer.
Maybe just CodeCompleteAfterFunctionEquals?
================
Comment at: clang/lib/Sema/SemaCodeComplete.cpp:6274
+ // special functions.
+ Results.AddResult("default");
+ // FIXME(liuhui): Ideally, we should only provide `delete` completion for the
----------------
I'm not sure offering default whether it's valid or not is an improvement over
never offering it.
I think the following checks are easy to implement based on the `Declarator`
and "good enough":
- offered for destructors
- offered for operator= (regardless of signature)
- offered for constructors with 0 or 1 arguments (regardless of type)
- optional: offered for operator<, operator>, operator==, operator<=> (if
C++20 is enabled, even if this is not a member function!)
In particular, we shouldn't offer =default for arbitrary methods (consider `=
0` on a virtual method), and we shouldn't offer it for non-members (I think
currently it will be offered outside class context)
Repository:
rG LLVM Github Monorepo
CHANGES SINCE LAST ACTION
https://reviews.llvm.org/D82548/new/
https://reviews.llvm.org/D82548
_______________________________________________
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
