================ Comment at: clang-tidy/readability/RedundantVoidArgCheck.cpp:244 @@ +243,3 @@ + removeVoidArgumentTokens( + Result, shrinkRangeByOne(SourceRange(CStyleCast->getLParenLoc(), + CStyleCast->getRParenLoc())), ---------------- alexfh wrote: > Character-based offsets can be brittle in case escaped newlines and > alternative tokens are used. I'd prefer getting the type range directly when > it's possible. Here and for other casts something like this should work: > `CStyleCast->getTypeInfoAsWritten()->getTypeLoc().getSourceRange()`. > > Also, do we need separate functions for `ExplicitCastExpr` and two of its > children classes? I don't think the character based offsets can be avoided here. The nodes don't provide getter functions for all the locations I'm interested in and I have to construct the appropriate source location myself. I will try what you suggest and see how it goes.
I tried doing things with just `ExplicitCastExpr` initially and when I added test cases for the functional and C-style casts, I needed to do different replacement ranges. http://reviews.llvm.org/D7639 EMAIL PREFERENCES http://reviews.llvm.org/settings/panel/emailpreferences/ _______________________________________________ cfe-commits mailing list cfe-commits@cs.uiuc.edu http://lists.cs.uiuc.edu/mailman/listinfo/cfe-commits
