kallehuttunen added a comment. In D59103#1423307 <https://reviews.llvm.org/D59103#1423307>, @aaron.ballman wrote:
> In D59103#1422775 <https://reviews.llvm.org/D59103#1422775>, @kallehuttunen > wrote: > > > Another idea that came to my mind would be to enable this check only for > > annotated types. So warning for missing field access would be only given > > for types that have for example `[[clang::annotate("value type")]]` > > annotation. Possibly other kinds of checks could be also developed for > > types that have the given annotation. > > > I don't think `annotate` would be a good choice because the primary purpose > of that one is to pass that attribute information down to the backend and > using it for this purpose feels a bit hackish. However, we could always add a > new attribute if needed, but I'm not convinced an attribute is the right > approach either (but then again, I'm also lacking information I'm sure). > > What other kinds of checks do you have in mind and what are the semantics of > the attribute you're thinking of? I was thinking about something along the lines of having an attribute to mark types to be a "value type" as defined in Herb's metaclasses proposal (chapter 3.5 in https://herbsutter.files.wordpress.com/2017/07/p0707r1.pdf). There could be then checks for having default ctor, no virtual functions etc in addition to this check for those types. But then again now that I think about it, a "string" class is an counterexample for this: it can be a value (regular) type, but when comparing instances you don't want to access the "capacity" member. Repository: rCTE Clang Tools Extra CHANGES SINCE LAST ACTION https://reviews.llvm.org/D59103/new/ https://reviews.llvm.org/D59103 _______________________________________________ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
