[PATCH] D45702: [clang-tidy] Add a new check, readability-redundant-data-call, that finds and removes redundant calls to .data().

[Shuai Wang via Phabricator via cfe-commits]

shuaiwang added inline comments.


================
Comment at: clang-tidy/readability/RedundantDataCallCheck.cpp:45
+                  anyOf(TypesMatcher, pointerType(pointee(TypesMatcher)))))),
+              callee(namedDecl(hasName("data"))))
+              .bind("call")))),
----------------
aaron.ballman wrote:
> shuaiwang wrote:
> > aaron.ballman wrote:
> > > Eugene.Zelenko wrote:
> > > > aaron.ballman wrote:
> > > > > Should this check apply equally to `std::string::c_str()` as well as `std::string::data()`?
> > > > readability-redundant-string-cstr do both.
> > > Yup! But that makes me wonder if the name "redundant-data-call" is an issue. Perhaps the check name should focus more on the array subscript in the presence of an operator[]()?
> > How about "readability-circumlocutionary-subscript"?
> > "readability-circumlocutionary-element-access"?
> > "circumlocutionary" -> "verbose"?
> hah, I think circumlocutionary might be a bit too much. ;-) I think `readability-simplify-array-subscript-expr` might be reasonable, however. Right now, the simplification is just for `foo.data()[0]` but it seems plausible that there are other array subscript simplifications that could be added in the future, like `a[1 + 1]` being converted to `a[2]` or `x ? a[200] : a[400]` going to `a[x ? 200 : 400]` (etc).
Just `readability-simplify-subscript-expr`?
Since after simplification the subscript operation is done by calling overloaded `operator[]` on an object instead of built-in subscript operator on an array.

Let me know if this name looks good to you and I'll do the actual renaming (together with addressing other comments) after your confirmation.


Repository:
  rCTE Clang Tools Extra

https://reviews.llvm.org/D45702