aaron.ballman added a comment.
In https://reviews.llvm.org/D44231#1031380, @pfultz2 wrote:
> > Can you elaborate a bit more about this?
>
> This catches problems when calling `sizeof(f())` when `f` returns an
> integer(or enum). This is because, most likely, the integer represents the
> type to be chosen, however, `sizeof` is being computed for an integer and not
> the type the integer represents. That is, the user has an enum for the
> `data_type`:
>
> enum data_type {
> float_type,
> double_type
> };
>
>
> At some point the user may call a function to get the data type(perhaps
> `x.GetType()`) and pass it on to `sizeof`, like `sizeof(x.GetType())`, which
> is incorrect.
Can you point to some real world code that would benefit from this check? I
don't see this as being a common issue with code and I am concerned about false
positives. For instance, it's reasonable to write `sizeof(func_call())` in a
context where you don't want to repeat the type name in multiple places. I've
seen this construct used in LLVM's code base (see Prologue::getLength() in
DWARFDebugLine.h for one such instance).
Repository:
rCTE Clang Tools Extra
https://reviews.llvm.org/D44231
_______________________________________________
cfe-commits mailing list
cfe-commits@lists.llvm.org
http://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
