vbvictor wrote: I think this new option should go through RFC process as described in https://llvm.org/docs/RFCProcess.html.
What is the plan for this option to leave "experimental state"? I personally don't like that we will eventually have an option that will speed-up analysis at cost of FP/FN that _we_ as maintainers will have to triage/reason about. --------- > Known affected patterns include checks such as misc-confusable-identifiers, > misc-unused-using-decls, and > readability-inconsistent-declaration-parameter-name, where skipped > declarations can affect the check's internal state. Generated headers should > also be included in the filter when they are semantically part of the focused > module. In the future, these checks could be not the only ones to be affected by this options. How can the user know that some checks are fine with this option and others are not? Does they have to read checkers code to understand it? Or maintainers have to mark each check if it is safe with this option or not? Again I'm against taking more churn on maintaining this since we don't have much capacity even today. https://github.com/llvm/llvm-project/pull/194370 _______________________________________________ cfe-commits mailing list [email protected] https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
